Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Evaluating the impact of prenatal behavior modification on maternal and infant outcomes: British Columbia’s… Martin, Cheryl Lynn 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1995-0630.pdf [ 7.91MB ]
Metadata
JSON: 831-1.0098977.json
JSON-LD: 831-1.0098977-ld.json
RDF/XML (Pretty): 831-1.0098977-rdf.xml
RDF/JSON: 831-1.0098977-rdf.json
Turtle: 831-1.0098977-turtle.txt
N-Triples: 831-1.0098977-rdf-ntriples.txt
Original Record: 831-1.0098977-source.json
Full Text
831-1.0098977-fulltext.txt
Citation
831-1.0098977.ris

Full Text

E V A L U A T I N G T H E IMPACT OF PRENATAL BEHAVIOR MODIFICATION O N M A T E R N A L A N D INFANT OUTCOMES: BRITISH C O L U M B I A ' S P R E G N A N C Y O U T R E A C H PROJECTS By CHERYL L Y N N MARTIN  B.S.N. The University of Victoria, 1986  A THESIS S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E REQUIREMENTS FOR T H E DEGREE O F MASTER O F SCIENCE (Health Services Planning and Administration) in T H E F A C U L T Y O F G R A D U A T E STUDIES Department of Health Care and Epidemiology  We accept this thesis as conforming to the required standard  T H E U N I V E R S I T Y O F BRITISH C O L U M B I A November  1995  © Cheryl L y n n Martin, 1995  In  presenting  degree freely  at  this  the  available  copying  of  department publication  of  in  partial  fulfilment  University  of  British  Columbia,  for  this or  thesis  reference  thesis by  this  for  his thesis  and  scholarly  or for  her  of  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  OcAdOX 1 1 ^ 5 "  I  I  further  purposes  gain  the  shall  requirements  agree  that  agree  may  representatives.  financial  permission.  Department  study.  of  be  It not  that  the  be  Library  an  by  understood allowed  advanced  shall  permission for  granted  is  for  the that  without  make  it  extensive  head  of  copying my  my or  written  ABSTRACT  Since the introduction of enhanced prenatal care programs for h i g h risk women in the late seventies, their effectiveness has been questioned.  Recent  studies i n Canada and the U n i t e d States have documented mixed results regarding the impact of comprehensive prenatal care programs o n infant outcomes.  This study was undertaken to explore the impact of British  Columbia's Pregnancy Outreach Projects (POP).  Data from the U B C Perinatal Study for the years 1990 to 1991, was used to obtain maternal, delivery and birth outcomes data for the 106 P O P clients and 318 matched controls included in this study.  P O P clients obtained on average significantly more prenatal visits (7.9) than their matched controls (7.2).  They had slightly better results for initiation of  prenatal care, adequacy of prenatal care and maternal morbidity. Although both groups of infants were similar with regard to measures of growth, measures of morbidity were mixed. birth  15.1% versus  P O P infants had significantly higher rates of preterm  8.8%  and congenital  anomalies  14.2% versus  4.7%.  However, they had a significantly lower rate of small for gestational age, 8.5% versus 12.3%.  W i t h the exception of maternal morbidity, early entry P O P clients (less than 20 weeks gestation) had significantly better maternal outcomes than their controls. Infant outcomes i n the early entry subanalysis mirrored the overall analysis,  with significantly less P O P infants (6.7%) b o r n small for gestational  age  compared to their control infants (10.6%).  There were no differences i n maternal outcomes between P O P clients w h o entered late (between 21 and 28 weeks gestation) and their matched controls. Fewer late entry P O P infants were born small for gestational age, though not significantly so.  The P O P exerted its effect through reductions in the rate of small for gestational age for program infants.  This trend was seen for the overall group and in the  subanalysis based on entry into the program.  Results of the subanalysis based o n risk group, showed aboriginal smokers, in addition to single and adolescent Caucasians were the subgroups of P O P clients who received the most benefit from the program. Their infants had lower rates of both small for gestational age and preterm birth.  TABLE OF CONTENTS  ABSTRACT  ii  TABLE OF CONTENTS  iv  LIST O F T A B L E S  viii  LIST O F FIGURES  ix  ACKNOWLEDGEMENTS  x  OVERVIEW  1  CHAPTER 1  INTRODUCTION  REFERENCES CHAPTER 2  4 10  THE IMPACT OF COMPREHENSIVE PRENATAL C A R E PROGRAMS FOR SOCIALLY DISADVANTAGED PREGNANT W O M E N  INTRODUCTION Issues in Studying the Effects of Prenatal Care THE EMERGENCE OF COMPREHENSIVE PRENATAL CARE  13 13 16 19  ANALYSIS O F T H E IMPACT OF COMPREHENSIVE P R E N A T A L C A R E PROGRAMS O N M A T E R N A L A N D INFANT OUTCOMES Multidisciplinary Provision - Large Programs Multidisciplinary Provision - Small Programs Case Management H o m e Visitation  21 23 39 46 55  CONCLUSION  68  REFERENCES  73  C H A P T E R 3 P R E G N A N C Y O U T R E A C H PROJECTS Background Objectives Service Model  79 79 80 81  V  Program Components Assessment Tools Individual Prenatal Risk Identification Tool T - A C E Questionnaire , Vancouver Island Pregnancy Outreach Projects REFERENCES CHAPTER 4  82 84 84 85 86 88  RATIONALE A N D METHODS  89  RATIONALE Questions  89 90  METHODS Study Design Sample Size and Power Data Sources Study Subjects Pregnancy Outreach Project Clients Comparison group Identification of P O P Clients Independent Variable Matching Variables Age at Delivery Race Parity Family Income Data Analysis Maternal Outcomes of Interest Initiation of Prenatal Care Number of Prenatal Visits Adequacy of Prenatal Care Maternal Morbidity Infant Outcomes of Interest Gestational Age Birth Weight H e a d Circumference Length Preterm Delivery Low Birth Weight Small for Gestational Age Large for Gestational Age Perinatal Conditions , Congenital Anomalies  91 91 91 92 93 93 93 94 95 95 96 96 97 97 99 101 101 102 102 102 103 103 103 103 104 104 104 104 104 104 105  REFERENCES  106  vi  CHAPTER 5  RESULTS  110  P O P D A T A BASE D E V E L O P M E N T Pregnancy Outreach Projects Intake Client Retention Lost to Followup Study Eligible P O P Clients Power Calculation Development of Comparison Group Missing Data  110 110 Ill 113 114 116 116 116  ANALYSIS OF M A T E R N A L A N D INFANT O U T C O M E S Maternal Characteristics Primary Analyses Maternal Outcomes Infant Outcomes Subanalysis-Program Entry Early Entry Maternal Outcomes Early Entry Infant Outcomes Late Entry Maternal Outcomes Late Entry Infant Outcomes Subanalysis-Groups at Risk  117 117 121 121 124 128 129 132 136 138 141  SUMMARY  145  REFERENCES CHAPTER 6  .150 DISCUSSION O F RESULTS A N D I M P L I C A T I O N O F STUDY  151  DISCUSSION Maternal Outcomes... Infant Outcomes Impact of Program Entry on Maternal and Infant Outcomes Impact of Program on Subgroups at Risk Other Issues  151 151 153 155 157 159  LIMITATIONS  160  IMPLICATIONS.... Pregnancy Outreach Project Implications Research Implications  164 164 167  REFERENCES  169  APPENDIX A..  170  A P P E N D I X B.  177  APPENDIX C  179  viii  LIST O F T A B L E S  3.1 3.2 4.1 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10 5.11 5.12 5.13 5.14 5.15 5.16 5.17 5.18  Pregnancy Outreach Projects Objectives Program Components Maternal Matching Variables Sources of Referral P O P Clients-Completed and Lost to Followup Study Clients-Maternal Profile Maternal Outcomes-Initiation and Visits Maternal Outcomes-Adequacy and Morbidity Infant Growth Infant Morbidity P O P Clients-Early and Late Entry Early Entry Maternal Outcomes-Initiation and Visits Early Entry Maternal Outcomes-Adequacy and Morbidity Early Entry Analysis-Infant Growth Early Entry Analysis-Infant Morbidity Late Entry Maternal Outcomes-Initiation and Visits Late Entry Maternal Outcomes-Adequacy and Morbidity Late Entry Analysis-Infant Growth Late Entry Analysis-Infant Morbidity Subanalysis-Aboriginals Subanalysis-Caucasians  81 83 95 110 113 120 123 123 125 127 129 131 131 133 135 137 137 139 141 143 144  ix  LIST O F FIGURES  3.1 Pregnancy Outreach Projects Service Model 5.1 Client Retention 5.2 Study Eligible P O P Clients  81 112 115  ACKNOWLEDGEMENTS I would like to acknowledge the assistance of Dr. Bob Armstrong, Dr. Sam Sheps and Ruth Milner for the many hours they spent answering questions, reviewing drafts and providing helpful comments and suggestions during the development of this thesis. Thanks also go to many people in the Pregnancy Outreach Projects. Mary Heatherington, Arlene White and Melinda Grey provided assistance with for record tracking and data abstracting. Lisa Forster-Coull contributed greatly to my understanding of the Pregnancy Outreach Project and was always available to answer questions. I wish to also thank all the women who participated in the Pregnancy Outreach Projects, for their willingness to participate and their contribution to research. My most heartfelt thanks goes to my husband Tom Ramsay, without whose support, and countless hours of solo child minding, this thesis would not have been completed. Thanks also to my daughters Lara and Margo, who have added fulfillment to my life and years to my thesis.  OVERVIEW The past two decades has been a period of unprecedented growth in the area of prenatal care for women at risk for adverse pregnancy outcomes.  With  increased knowledge of the multiplicity of risk factors for perinatal complications, prenatal care programs for high risk women have evolved from single focus programs to multidimensional care. Currently, comprehensive prenatal care programs that encompass prenatal education, infant care, lifestyle assessment, social support and nutritional supplements are viewed as the most effective means of enhancing perinatal outcomes in high risk populations.  In order to determine the effectiveness of comprehensive prenatal care, many programs undergo formal program evaluations.  On the whole these  evaluations are process oriented and concerned with program implementation and consumer satisfaction. Outcome analysis is usually limited to maternal behavioral changes during pregnancy and infant birth weight and gestational age.  In general, infant birth outcomes are compared with city, province, or  national averages. Although these evaluations provide valuable information on program clients, they do not reveal the degree of impact possible for high risk women. True program impacts can only be demonstrated through the use of comparison groups.  The BC Pregnancy Outreach Projects provide a vehicle for completing this type of evaluation. Three Vancouver Island Pregnancy Outreach Project sites were chosen to participate in a study to document the impact of comprehensive prenatal care programs.  The primary questions of interest in this study were:  i. What were the characteristics of the women who attended a Pregnancy Outreach Project during 1990-1991?  ii. Were any statistically significant differences in maternal outcomes between Pregnancy Outreach Project clients and controls?  iii. Were there any statistically significant differences in infant outcomes between the infants of Pregnancy Outreach Project clients and infants of controls? The secondary questions of interest in this study were:  iv. Were there any differences in maternal and infant outcomes between women who entered the Pregnancy Outreach Projects prior to mid pregnancy and their matched controls?  v. Were any differences in maternal and infant outcomes between women who entered the Pregnancy Outreach Projects after mid pregnancy and their matched controls?  vi. Were there any subgroups of women at risk, for who measurable program effects were shown?  This report, which describes the above study, is organized as follows:  Chapter 1: reviews low birth weight as a problem i n Canada and very briefly describes the impact of low birth weight i n terms of infant mortality and morbidity; economic and social costs; and manifestations of the complication. It describes the shift i n focus from treatment of low birth weight infants to prevention, through comprehensive prenatal care programs that target high risk pregnant women,  Chapter 2: presents a critical review the literature pertaining to comprehensive prenatal programs for socially disadvantaged women, with respect to maternal, infant and longterm outcomes,  Chapter 3:  describes the purpose, goal, objectives and structure of the B C  Pregnancy Outreach Projects. Client specific information is provided about the three Vancouver Island sites participating i n the study,  Chapter 4: describes the rational for and the methods used in the study,  Chapter 5: presents the results of the study,  Chapter 6: discusses the results and limitations of the study and recommends further actions to be considered to clarify the issues raised.  CHAPTER 1 INTRODUCTION  Birth weight is an important indicator of the health status of a country's population.  A n infant's weight at birth is the single strongest predictor of  survival. Thus, low birth weight is widely considered to be the most important risk factor for infant mortality and childhood disability. Worldwide, 17% of all infants are born with low birth weight ( U N I C E F , 1991). In developed countries the rate of low birth weight is considerably lower.  In 1989, 4.7% of single live  infants born in Canada weighed less than 2500 grams at birth (Statistics Canada, 1991).  T w o thirds of infants who die i n the first year of life are low birth weight infants, the majority of these infants die during the neonatal period; low birth weight infants are 40 times more likely to die than infants of normal birth weight.  The risk of early death is nearly 200 times greater for infants weighing  less than 1,500 grams (Levitt et al., 1993). In C a n a d a / l o w birth weight accounts for almost 75% of early neonatal mortality (Levitt et al., 1993) and 70% in the U.S. (Casiro et al., 1993)  Steady reductions in infant mortality over the past 20  years have been accomplished through improved survival of very low birth weight  infants made possible by technological and medical advances  in  neonatal care, based on a better understanding of neonatal physiology and pathology.  In this century, the infant mortality rate decreased from 100/1000 live births at the beginning of the century (Shapiro et al., 1968); to 8/1000 live births i n 1988  for Canada and 9.7/1000 for the U S (Statistics Canada, 1991;  A n n u a l Vital  Statistics Report, 1990). A recent B.C. study that examined trends i n low birth weight mortality since 1952 found birth weight specific declines i n the rate of infant mortality. Since the early 1950's infants weighing 2000-2499 grams have experienced a steady decline i n mortality from 69/1000 single live births to 29/1000 in 1988.  For infants weighing 1000-1499 grams, the decline i n infant  mortality started in the m i d 1960's and decreased from 583/1000 to 149/1000 in 1988. The biggest gains have been made for infants weighing 750-999 grams. In 1952, the rate of infant mortality for this group of infants was 862/1000.  This  rate remained virtually unchanged until the early 1970's. Since then, the infant mortality rate has declined sharply to a rate of 382/1000 in 1988 . 1  The trend towards better survival of infants with birth weights between 1000 and 1999 grams has been reported since the early 1980's. Since this time period significant reductions in mortality have been shown for infants weighing 15012000 grams at birth (Casiro et al, 1993; Kitchen et al, 1992) or having reached 34 weeks of gestation Qijon & Jijon-Letort, 1995). One recent U.S. study found that once infants reached a threshold of 1600 grams, the rates of both morbidity and mortality declined sharply (DePalma et al., 1992).  C o m p a r e d with the decrease i n infant mortality d u r i n g this century, the decrease in the rate of low birth weight has been very modest.  In 1950, the  national rate of low birth weight i n the U S was 7.5% and 7.2% in Canada. For the next thirty years the incidence of low birth weight declined slowly and steadily to a rate of 5.4% i n Canada, and 6.8% i n the U S . However, since 1980,  Personal communication, Sandi Wiggins.  this rate has remained virtually unchanged i n both countries (Public Health Service, 1988; Statistics Canada, 1991).  The economic costs associated with low birth weight care are high. Birth weight has a direct impact on hospital length of stay and therefore health care costs. Quebec hospital data show the average length of stay for infants weighing over 2,500 grams was 4.5 days, compared with 8 days for infants weighing 2,000-2,499 grams, 21 days for infants weighing 1,500-1,999 grams, and 33 days for infants 1,000-1,499 grams at birth (Lepage et al.,1989). One recent Canadian study found the m i n i m u m cost of hospital care for low birth weight infants was $873 per day (Casiro, et al, 1992). Other Canadian studies have estimated the per diem cost of caring for a low birth weight infant ranges from $500-2,500 (Heleva & Heaman, 1989; D'Alton 1988; Creasy, 1988).  In the U S the price of care is even higher,  admission alone to a neonatal intensive care unit is estimated at $9,600 and per diem costs range from $776-1,918 (Kay et al., 1991).  While infants weighing  between 500 and 1499 grams make up about 2% of the neonatal population, they consume over one-third of the total neonatal care budget (Schwartz, 1989).  The longterm social costs of low birth weight are even more profound. L o w birth weight infants have more birth complications and are more likely than other infants to have deficits i n their physical and mental development.  Low  birth weight survivors have an increased incidence of disability from a broad range of conditions, including: congenital anomalies, respiratory illnesses, neurodevelopmental handicaps and complications from neonatal care treatment  (Millar, et al., 1993; A y l w a r d  et al., 1989;  intensive  Dunn,  1981;  Fitzhardinge, 1976; Hack & Fanaroff, 1984 & 1989; Kramer, 1987; Ramey et al., 1978; Shapiro et al., 1980, Teberg et al., 1988).  While technological advances will no doubt continue to improve the survival of low birth weight infants, real improvements i n the health status of all infants w i l l be realized through a reduction i n the rate of low birth weight. Further reductions i n low birth weight appear to be possible as many western nations have a smaller percentage  of low birth weight infants than does  Canada. In Sweden, Finland, Norway and Ireland only 3-4% of infants are born with low birth weight.  The W o r l d Health Organization states that a low birth  weight rate of 3/100 live births is the probable threshold achievable (World Health Organization,1986). In the U S a national objective has been set to reduce the incidence of low birth weight to 5% of all live births and no higher than 9% in any subpopulation (Institute of Medicine, 1985).  In Canada, Ontario has  targeted the reduction the incidence of low birth weight to less than 4% by the year 2000 (Ministry of Health, Government of Ontario, 1989).  Infant birth weight is determined both by intrauterine growth and gestational age at birth.  Major contributors to low birth weight,  therefore,  include  intrauterine growth retardation, preterm birth, and a combination of these two. In Canada and other developed countries, intrauterine growth retardation accounts for about one third of low birth weight infants, and preterm birth about two thirds. Incomplete understanding of the underlying mechanisms of both intrauterine growth retardation and premature labour have been major obstacles in the prevention of low birth weight  In place of adequate specific  information about causes, extensive research has been conducted to determine risk factors associated with low birth weight.  Risk factor studies i n both developed and developing countries demonstrate that causality of low birth weight is multifactorial, and that many risk factors are interrelated (Kramer, 1987; Institute of Medicine, 1985; Silins et al., 1985). W h i l e the  u n d e r l y i n g etiologies of intrauterine growth retardation and  premature  labour are different,  the  risk  factors  associated  with  these  complications overlap.  In developed countries, where preterm birth accounts for the majority of low birth weight infants, medical attention for the prevention of preterm birth has focused on arresting preterm labour via early detection. frequent  cervical examination  and  ultrasonography,  Methods include ambulatory  home  monitoring, patient and provider education, hospitalization, bed rest, fluids, tocolytic drug therapy and cervical cerlage (Creasy, 1988; Holbrook et al., 1987; lams, 1989; Morrison et al., 1987; Papiernik et al., 1985). These approaches are costly, carry inherent risks and have met with variable success. Where success in prevention has been shown, it remains unclear if medical care was the most effective intervention, or if results were achieved from changes i n maternal behavior and increased social support. Although traditional medically oriented prenatal care has drastically reduced the rate of adverse outcomes for low risk women, it has done little to improve infant outcomes for high risk women.  Currently, attention is shifting away from medically oriented prevention programs towards health promotion programs. The traditional medical model is not sufficient on its o w n to prevent low birth weight, because of uncertain and multiple etiologies of intrauterine growth retardation and premature labour (lams, 1989). encompass  The traditional approach to prenatal care does not  biological, behavioral and socioeconomic  factors that  influence  infant birth weight.  New programs that enhance prenatal care through the  provision of health promotion approaches including: prenatal education, environmental support, behavior modification and nutrition supplementation are now viewed as more effective means of reducing the incidence of low birth weight.  REFERENCES  A n n u a l Vital Statistics Report (1990). Hyattsville, Maryland: National Center for Health Statistics. A y l w a r d , G . , Pfeiffer, S., Wright, A . , & Verhulst, S. (1989). Outcome studies of low birth weight infants published in the last decade: A metaanalysis. The Journal of Pediatrics. 115(4). 515-520 Casiro, O . G . , McKenzie, M . E . , McFadyen, L . , Shapiro, C , Seshia, M . M . , MacDonald, N . , Moffatt, M . & Cheang, M.S. (1992). Cost analysis of extremely low birthweight infants: an update. Journal of Paediatrics & C h i l d Health. 28(5). 410. Creasy, R. (1988). Prevention of preterm labour. Presented at the Fourth National Conference on Regionalized Perinatal Care and Prevention of Handicap. Ottawa. D'Alton, M . (1988). Prevention and management of preterm labour in Canada. Presented at the Fourth National Conference on Regionalized Perinatal Care and Prevention on Handicap. Ottawa. DePalma, R.T., Leveno, K.J., Kelly, M . A . , Sherman, M . L . & Carmody, T.J. (1992). Birth weight threshold for postponing preterm birth. American Journal of Obstetrics & Gynecology. 167 (4, part 1), 1145-1149. Dunn, H . (1981). Residual Handicaps in children of low birth weight. Canadian Institute of C h i l d Health. Fitzhardinge, P . M . (1976). Follow-up studies of the low birth weight infant. Clinical Perintology. & 503-516. Hack, M . & Fanaroff, A . (1984). The outcome of growth failure associated with preterm birth. Clinical Obstetrics and Gynecology. 27(3). 647-661. Hack, M . & Fanaroff, A . (1989). Outcomes of extremely-low-birthweight infants between 1982 and 1988. The New England Tournal of Medicine. 321(24). 1642-1647. Heleva, M . & Heaman, M . (1989). Preterm birth prevention program conceptual document. St. Boniface General Hospital. Winnipeg.  Holbrook, R , Falcon, J., Herron, M . , Lirette, R., Laros, J., & Creasy, R. (1987). The evaluation of the weekly cervical examination i n a preterm birth prevention program. American Journal of Perinatology. 4(3), 240-244. lams, J. (1989). Current status of prematurity prevention. T A M A . 262(2), 265266. Institute of Medicine (1985). Preventing low birth weight. Washington, D C : National Academy Press. Jijon, C R . & Jijon-Letort, F.X. (1995). Perinatal predictors of duration and cost of hospitalization for premature infants. Clinical Pediatrics. 34(21. 79-85. Kay,T3.J., Share, D . A . , Jones, K . , Smith, M . , Garcia, D . & Yeo S.A. (1991). Process, Costs, and Outcomes of Community-Based Prenatal Care for Adolescents. Medical Care. 29 (6). 531-542. Kitchen, W . H . , Bowman, E . , Callahan, C , Campbell, N . T . , Carse, E . A . , Charton, M . , Doyle, L . W . , Drew, J., Ford, G.W., Gore, J. (1993). The cost of improving the outcome for infants of birthweight 500-999 g i n Victoria. The Victorian Infant Collaborative Study Group. Tournal of Paediatrics & C h i l d Health. 29(1). 56-62. Kramer, M.S. (1987). Intrauterine growth and gestational duration determinants. Pediatrics, 80, 502-511. Lepage, M . , Levasseur, M . , Colin, C , Beaulac-Baillargeon, L . , & Goulet, L . (1989). Perinatal and infant mortality and morbidity: quality of life for infants and for parents. Quebec: ministere de la Sandte et des Services sociaux. Levitt, C , Watters, N . , Chance, G . , Walker, R. & A v a r d , D . (1993). Low-birthweight symposium: summary of proceedings. Canadian Medical Association Tournal. 148C5L 767-771. McCormick, M . et al. (1985). The contribution of low birth weight to infant mortality and childhood morbidity. The N e w England Journal of Medicine. 31(2}, 82-90. Millar, W.J., Strachan, J. & Wadhera, S. (1993). Trends In L o w Birth Weight. Canadian Social Trends, Statistics Canada, Spring, 26-29. Ministry of Health, Government of Ontario (1989). Mandatory Health Programs and Services Guidelines.  Morrison, J., Martin, J., Martin, R., Gookin, K . , & Wiser, W . (1987). Prevention of preterm birth by ambulatory assessment of uterine activity: A randomized study. American Journal of Obstetrics and Gynecology. 156, 536-543. Papiernik, E . et al. (1985). Prevention of preterm births: A perinatal study i n Haguenau, France. Pediatrics. Z6_(2), 154-158. Public Health Service (1988). Progress toward achieving the 1990 objectives for pregnancy and infant health. Morbidity and Mortality Weekly Report. 37,405-412. Ramey, C.T., Stedman, D.J., Borders-Patterson, A . & Mengel, W . (1978). Predicting school failure from information available at birth. A m e r i c a n Journal of Mental Deficiencies. 82. 525-534. Schwartz, R . M . (1989). What price prematurity? Family Planning Perspectives. 21(4). 170-174. Shapiro, S., Schlesinger, E . & Nesbitt, R. (1968). Infant, perinatal, maternal and childhood mortality i n the United States. Cambridge Mass: Harvard University Press. Shapiro, S. , McCormick, M . C . , Starfield, B . H . , et al. (1980). Relevance of correlates of infant deaths for significant morbidity at one year of age. American Journal of Obstetrics and Gynecology. 126* 363-373. Silins, J., Semenciw, R., Morrison, H . , Lindsay, J., Sherman, G . , M a o , Y . & Wigle, D . (1985). Risk factors for perinatal mortality i n Canada. Canadian Medical Association Journal. 133.1214-1219. Statistics Canada (1991). Trends i n low birth weight i n Canada 1971-1989. Health Reports. 3(4), 311-325. Teberg, A . et al. (1988). Outcomes of S G A infants. Seminars in Perinatology. 12(1). 84-87. U N I C E F , (1991). The State of the World's Children. U . K . Oxford University Press. W o r l d Health Organization (1986). Infant mortality and birth weight. A n n u a l Statistics.  CHAPTER 2 T H E IMPACT OF COMPREHENSIVE P R E N A T A L C A R E PROGRAMS FOR SOCIALLY DISADVANTAGED PREGNANT W O M E N A REVIEW O F T H E LITERATURE  INTRODUCTION  The effectiveness of prenatal care in reducing low birth weight and other adverse pregnancy outcomes has been the subject of many decades of research, with somewhat mixed results.  A number of early studies reported finding no  association between prenatal care and low birth weight, prematurity or neonatal mortality (Drillien, 1957; Shwartz, 1962; Terris & G o l d , 1969).  In  1966,  Abramowicz and Kass presented an overview of the results of studies that had been published over the past decade, and concluded that the published evidence supporting a positive relationship between prenatal care and birthweight was inconclusive. Conversely, studying 130,000 births in upstate N e w York i n 1973, Stickle and M a (1977) found that prenatal care, initiated in the first trimester was associated with better pregnancy outcomes.  Inadequate or absent prenatal care has often been cited as a risk factor for low birth weight.  Using data from the 1988 National Maternal and Infant Health  Survey, Sanderson et al (1991) found the absence of prenatal care to be associated with low birth weight.  After controlling for social and demographic  confounding variables, Greenberg (1983) found that women who failed to seek prenatal care were at increased risk for delivery of low birth weight infants. Several studies have also found that women who delayed initiating prenatal care or who obtained few prenatal care visits were more likely to experience adverse pregnancy outcomes, in particular, low birth weights, preterm birth and  14 neonatal mortality (Sanderson et al 1991; Schlesinger & Kronebusch, 1990; Leveno et al, 1985).  The finding of an association between inadequate prenatal care and low birth weight has been clouded by issues of confounding.  Characteristics of women  who obtained adequate prenatal care, rather than care itself, may have increased their chances of delivering a normal birth weight infant.  H a l l et al (1980)  initiated this debate when they questioned whether the purported benefits of prenatal care were attributable to the care per se, or rather social, demographic and medical determinants of the utilization of prenatal services.  W o m e n who  obtained adequate prenatal care generally were of optimal childbearing age, had high levels of education, were married, and had incomes above the poverty line. Conversely, women who d i d not receive adequate prenatal care may have delivered infants of low birth weight because they were characterized by other risk factors, including childbearing at extremes of the age spectrum, poor education, single marital status, minority status and low income.  Studying  birth certificate data, Taffel (1978 & 1980) found that maternal education and ethnic  group were more important than initiation of prenatal care  determining low  birthweight.  Recent studies have  confirmed that  in the  association between minority race (black) and low birth weight is significantly higher than the association between prenatal care and low birth weight (Hulsey et al., 1991; Collins et al., 1990; Miller & Jekel, 1987).  Several studies have demonstrated that the benefits of prenatal care varied by social status.  Mothers at highest risk for low birth weight, who received  adequate prenatal care, benefited the most from preventive efforts. Gortmaker (1979) found that prenatal care was associated with a significant reduction i n the  rate of low birth weight and infant mortality, even after social and demographic influences were considered.  Greenberg (1983) demonstrated an association  between prenatal care and higher birthweight that varied somewhat persisted after controlling for mother's race and education.  but  Prenatal care had  the greatest observed impact for socially disadvantaged women, because of their high overall risk for delivery of low birth weight infants. H e concluded that the efficacy of prenatal service was modified by social situation. (1984),  after  controlling  for:  prenatal  care,  length  of  Showstack et al gestation  and  sociodemographic factors, found that adequate prenatal care was associated with a significant and substantial increase in birth weight, especially for infants of black women.  Data from the 1988 U.S. National Maternal and Infant Health Survey found a relative decline in the rate of entry into early prenatal care (Sanderson et al., 1991). In fact, since 1980, the percentage of births to women with no prenatal care or care only in the third trimester of pregnancy had increased, especially among black women, whose rate of late entry rose from 8.8% in 1981 to 10.3% in 1985 (Brown, 1988).  Young women, minority women and w o m e n with low  incomes, i n particular, were delaying prenatal care.  Researchers have conducted numerous studies to determine the barriers to prenatal care access. prenatal care.  Socioeconomic factors were related to the utilization of  However, other factors were often as important or more  important than financial barriers. Young et al (1989) i n a study of 201 women who entered prenatal care during their third trimester, found the psychosocial aspects of obtaining care, including denial and concealment of pregnancy, family crisis and lack of child care, often delayed initiation of prenatal care. In a  study of 227 women with no prenatal care, Scupholme et al (1991) found that barriers to prenatal care were system related, patient related and financial. Lack of  information about  service  providers, problems  completing  required  paperwork and the negative attitude of staff and providers were the main barriers to prenatal care services.  Harvey and Faber (1993) in a study of 236  w o m e n who received inadequate or no prenatal care, found 13 financial, personal and organizational barriers to access. In a logistic regression analysis that controlled for social and demographic characteristics, six barriers were significant predictors of inadequate care: poor understanding o f / o r low value given to prenatal care; financial difficulties; difficulty scheduling appointments; excessive physical or psychological stress; lack of information on providers of services; and ambivalence or fear regarding the pregnancy.  Shlessinger &  Kronenbusch (1990) found that barriers affecting prenatal care access were age related.  Finances, difficulty scheduling appointments, transportation barriers  and child care difficulties were barriers for older women.  Ambivalence about  the pregnancy, belief that care was not important and lack of knowledge about prenatal care were barriers for younger women.  The literature w o u l d suggest  that removal of financial barriers and improved access to traditional medical prenatal care alone, has not generally improved the pregnancy outcomes of socially disadvantaged high risk women.  Issues in Studying the Effects of Prenatal Care  The impact of prenatal care on birth weight has received considerable attention, in part because, among the major risk factors, prenatal care is probably the most amenable  to change  through deliberate health programs.  However,  the  effectiveness of prenatal care i n reducing low birth weight and neonatal  mortality, has been complicated by a number of data and methodological limitations.  The major constraint to this line of inquiry has been the inability of researchers to apply experimental protocols, specifically the use of random assignment to prenatal care treatment (and nontreatment), for obvious ethical  reasons.  Researchers have been forced to use nonrandomized studies, relying on differences in comparison groups or before-after studies to examine prenatal care effectiveness.  A second limitation, not independent of the first, is the presence of selection bias i n the study of natural populations of prenatal care users and nonusers. The self-selection  of some w o m e n into prenatal care and others to avoid  prenatal care confounds the causal relationship between care use and outcomes. W o m e n who entered care early in pregnancy and who obtained many prenatal visits may have (or have had) problematic pregnancies or medical risks which required additional care, or they may have been healthy low-risk women who simply want to obtain good prenatal care.  Conversely, women who entered  care late i n pregnancy may have had very healthy previous pregnancies and, consequently, d i d not perceive a need for early care, or they may have been women who d i d not know they were pregnant or d i d not understand the need for prenatal care (two common reasons for insufficient prenatal care). of this lack of homogeneity  w i t h i n categories  Because  of prenatal care use,  the  relationships between traditional measures of prenatal care and birth outcomes remains unclear.  Another methodological concern is the adequate measurement of prenatal care use.  Popular measures of prenatal care, such as trimester of entry and the  number of prenatal visits are more prevalent than those based on the content of prenatal care.  However, these indicators may be misleading.  W o m e n who  deliver prematurely often had fewer prenatal visits than those who deliver at term, even if they followed the recommended visit schedule until delivery. Unless a statistical adjustment was made, early deliveries were almost always associated with fewer prenatal visits.  This confounding of cause and effect has  been most commonly addressed through use of the Kessner Index of Prenatal care.  The index, is an algorithm that includes trimester of pregnancy prenatal  care began, number of prenatal visits and length of gestation, compared to the expected norm for visits (Kessner, 1973).  Another concern revolves around the content of prenatal care.  Although most  researchers try to solve the quantitative issue of prenatal care, the problem of defining the content of care, equally as important, has infrequently been addressed.  Traditionally, the content of prenatal care has been tailored to the  individual w o m a n and her pregnancy, thus each w o m a n may have received somewhat different care.  Some of the variables that influence the receipt of prenatal care also pregnancy outcome.  influence  In order to determine the independent effects of prenatal  care, confounding variables must have been controlled.  Most studies control  for some sociodemographic and obstetrical variables, however, it is impossible to know or include all confounding variables.  Finally, the validity of information recorded in data sources remains a problem. In large studies, the biggest source of data on prenatal care and low birth weight comes from vital statistics reports. birthweight  information  was  Several researchers have found that while  generally  well  recorded,  information  on  gestational age, initiation of prenatal care and number of prenatal visits was somewhat unreliable.  These  and other reasons hamper the production of conclusive  concerning the effectiveness of prenatal care.  evidence  Currently, attention is shifting  away from medically oriented prenatal programs towards health promotion programs.  The traditional medical model has not been sufficient on its o w n to  prevent low birth weight, because of uncertain and multiple etiologies of intrauterine growth retardation and premature labour (lams,  1989).  traditional approach to prenatal care does not encompass the  biological,  behavioral and socioeconomic factors that influence infant birth weight. programs  that enhance prenatal  care  through  the  provision  The  of  New health  promotion approaches including: prenatal education, environmental support, behavior modification and nutrition supplementation are now viewed as more effective means of reducing the incidence of low birth weight.  T H E EMERGENCE OF COMPREHENSIVE  PRENATAL CARE  The  care  traditional  approach  to  prenatal  with  its  focus  on  medical  management doesn't address the psychological and social barriers to prenatal care.  In addition, traditional prenatal care does not focus o n the social and  behavioral  problems  of of high risk w o m e n ,  problems  that impact  on  pregnancy.  Since the m i d 1970's there has been an emphasis on programs that  alter the content of prenatal care by addressing medical, social and behavioral issues during pregnancy. In 1986, the Expert Panel on Prenatal Care, conveyed by  the  Department of H e a l t h and H u m a n  Services  L o w Birth Weight  Prevention W o r k G r o u p , reviewed prenatal care content and recommended changes.  The panel determined that the definition of prenatal care had to  change and proposed a more contemporary definition of care: Prenatal care consists of health p r o m o t i o n , risk assessment, and intervention linked to the risks and conditions uncovered. These activities require the co-operative and coordinated efforts of the woman, her family, her prenatal care providers, and other specialized providers. Prenatal care begins w h e n conception is first considered and continues until labour begins. The objectives of prenatal care for the mother, infant, and family relate to outcomes through the first year following birth (Mortimer et al., 1991, pg. 783).  The Expert Panel on Prenatal Care defined the basic components of prenatal care as: " early and continuing risk assessment, health promotion, and medical and psychosocial interventions and their follow-up.  Prenatal care should add  to the traditional medical concerns a new emphasis  on the  psychosocial  dimensions of that care, maintaining a balance among factors" (Mortimer et al., 1991, pg. 783) . Further, they stated that prenatal care programs should have preset, specific purposes and activities for defined populations and groups. Positive outcomes of prenatal care programs should include the newborn's gestational age and birth weight and the mother's medical condition and health habits.  Comprehensive programs are multidimensional that foeus on pregnancy and labour education, lifestyle behavior modification, nutritional education and psychosocial support, in addition to traditional medical care. Empirical studies have been completed and the literature as a whole suggests that adequate, comprehensive prenatal care contributes to healthy birth outcomes among high risk women.  Research has shown that programs w h i c h provided outreach to  pregnant women or free prenatal care, reduced perinatal morbidity (Moore et al,1986); i m p r o v e d access to prenatal care and birth outcomes (Corman & Grossman, 1985; Norris & Williams 1984; Schlesinger & Kronebusch 1990); and were cost effective (Institute of Medicine, 1985; Moore et al 1986). Although the exact mechanism(s)  through w h i c h comprehensive prenatal care impacted  infant health remains u n k n o w n , studies suggest that i m p r o v e d nutrition (Geronimus, 1986), preterm delivery education (Harris, 1982) and screening of risk factors that arose during pregnancy (Institute of Medicine, 1988) were some of  the mechanisms through w h i c h prenatal care for low income  women  improved infant outcomes.  ANALYSIS O F T H E IMPACT O F COMPREHENSIVE P R E N A T A L C A R E PROGRAMS: M A T E R N A L A N D INFANT OUTCOMES  This literature review examined the quality and results of published primary evaluations  of  comprehensive  prenatal  care  programs  for  socially  disadvantaged women intended to optimize the health of mothers and infants. Comprehensive care refers to preventive services provided between conception and delivery, consisting of pregnancy and labour education, lifestyle behavior modification, nutritional education and psychosocial support, i n addition to traditional medical care.  The programs had preset, specific purposes and  activities for defined populations and groups.  The outcomes of these studies  were evaluated relative to indicators for high risk status, clarity of program description, adequacy of comparison group, inclusion of other risk factors i n design a n d / o r analysis, appropriateness of the analysis and clarity of the results. Outcomes of comprehensive programs included gestational age, birth weight, rates of low birth weight, preterm birth and infant mortality.  If available,  maternal outcomes and longterm impacts were also included.  The studies were published i n English between January 1980 and December 1982. 1980 was chosen as a starting point to have approximately a decades worth of  data reflecting  the  shift  i n focus  to, and increased  comprehensive prenatal care begun in the 1970's. methods  from clinical and social  accessibility  of,  The evaluations adapted  science research to produce  credible  information about the efficacy and effectiveness of interventions and programs. O n l y evaluations of programs that aimed to integrate medical and social services to improve the health outcomes of mothers and infants were included.  There were 33 studies that adequately met the criteria.  The vast majority of  these were case-control studies that assessed program effectiveness.  Nine  randomized trials that assessed program efficacy were included i n this analysis. Five randomized trials assessed the overall efficacy of comprehensive care, while four trials had more specific aims.  The studies were characterized  according to program focus and mode of delivery as follows: multidisciplinary provision - large programs; multidisciplinary provision - small programs; case management and home visitation.  23  Multidisciplinary Provision - Large Programs  Programs targeted at mothers and infants have traditionally occupied a basic position in the evolution of health services in the United States.  The premise  has been that intervention early in the life cycle offered maximum preventive health benefits.  Maternity and Infant Care (MIC) programs and Improved  Pregnancy Outcome (IPO) projects were two major federal  comprehensive  prenatal care initiatives.  In 1963 a renewed national commitment to prental care programs followed amendment of Title V of the Social Security A c t , resulting i n more than 50 maternity and infant care (MIC) projects that extended to all parts of the country (Kessner et al, 1973).  M I C projects were originally intended to reduce the  incidence of mental retardation and other handicapping conditions associated with pregnancy.  In 1967, the explicit goal of reducing infant mortality was  added to the legislation.  The hypothesis of M I C was that accessible and  attractive services w o u l d encourage w o m e n to receive early and regular prenatal care which, in turn, would contribute to fewer prenatal complications, low birth weight infants and fetal and neonatal deaths.  Projects were located i n  low-income areas and provided routine prenatal care and multidisciplinary supportive services.  There have been two published evaluations of M I C  projects since 1980.  In a large case-control study, Sokol et al (1980) evaluated the social, medicalobstetrical and perinatal impact of Cleveland Metropolitan General Hospital's Maternity and Infant Care (MIC) Project during 1976 and 1977._MIC registrants  were women who resided i n specific underprivileged and impoverished target areas.  The comparison group was comprised of women who resided outside  the catchment area of the M I C Project. W o m e n who were referred to Cleveland Metropolitan General Hospital for high risk obstetrical pregnancies, and those who received no antenatal care, were excluded from the comparison group, to decrease allocation bias in the study. A l l study participants received the medical aspects of prenatal care in the same teaching hospital clinic and delivered in the same hospital, thus source and nature of prenatal care were the biggest differences between the groups.  In addition to routine antepartum medical care, each M I C registrant received an organized multidisciplinary assessment, which was followed by counseling for needed paramedical services.  The registrants received health education,  nutritional counseling and parenting guidance.  If indicated, home visits were  made by community health aides and professional staff. The comparison group received traditional prenatal care, but not multidisciplinary assessment or counseling.  Results were available for 3307 M I C registrants, and after exclusions (430), 1679 comparison mothers.  M I C registrants and the comparison group were similar  w i t h regard to sociodemographic socioeconomic status) characteristics.  (race, age,  parity, marital status and  M I C registrants were seen significantly  earlier for prenatal care, on average at 16 weeks gestation compared with 19 weeks.  Over 47 percent (47.6%) of M I C registrants initiated care i n the first  trimester compared with 34.0% of the comparison group.  The number of  prenatal visits made by each w o m a n was not available for study.  MIC  registrants had significantly higher antepartum risk scores for intrauterine  growth retardation. There were no differences between the study groups with regard to risk scores for preterm birth.  The intrapartum courses were similar  for both groups.  Fewer preterm infants (less than 38 weeks) were born to M I C mothers (17.9%) than to comparison group mothers (23.9%). This 25% reduction was clinically and statistically  significant.  A l t h o u g h the authors f o u n d a  statistically  significant difference in the rate of low birth weight between the two groups, 11.7% for M I C infants compared with 14.0% for the comparison group, the 16% reduction was not clinically significant.  The most striking differences between  the two groups occurred in the rate of perinatal mortality. Comparison group infants  experienced  a  2.6  fold  increase  in  total  perinatal  mortality.  Subcomponent evaluation revealed that both antepartum stillbirth (6.3% versus 14.8%) and neonatal death (8.7% versus 24.0%) occurred at higher rates in comparison group infants.  The authors observed:  "The key finding of this study is that with similar social and medical-obstetric risks, patients who received care from the M & C projects experienced a significantly lower perinatal mortality than those who did not. Given that all study patients received the same care during labor and delivery, it is reasonable to infer that the observed differences i n outcome may have been related to differences i n care d u r i n g the antepartum period. The major difference i n antepartum care lies i n the ancillary services of the M & C project. Thus, the authors consider it more likely that the entire M & C ancillary support system...plays an important role i n achieving these results" (Sokol et al., 1980, pgs. 185-186).  While the two groups were balanced with respect to various biosocial factors they were significantly different with respect to antepartum risk scores.  Sixty  two percent (62.7%) of M I C registrants had high risk scores compared with 56.3% of the comparison group.  A n analysis that compared w o m e n with  similar risk scores w o u l d have provided a more refined method of determining maximim program effects.  G i v e n the large sample size and the number of  variables collected, the use of multivariate analysis to isolate the independent effect of the M I C program would have greatly added to the analysis.  In their excellent review, Peoples and Seigel (1983) used more discerning methods of data analysis to study the impact of the North Carolina M I C Project. The birth results of 5822 M I C registrants i n three counties, who gave birth between 1970  and 1977  were compared w i t h  11,447 births from  three  comparison counties with similar socioeconomic status, health resources and perinatal status. The data were analyzed by means of a weighted least squares procedure, after controlling for race; marital status; education; age/parity and reproductive risk. Group membership (MIC or comparison) was included as an independent variable i n the analysis to test the significance of the relationship between group membership and infant outcomes.  M I C participants had significantly higher proportions of risk factors for low birth weight than d i d the comparison group. Unadjusted results showed that M I C services improved use of prenatal care but had no impact on the incidence of low birth weight.  In a refined analysis, adjusting for maternal characteristics,  the authors found differential M I C effects.  There was no evidence of M I C  influence on subpopulations at lowest risk, but improvements in use of care and birth weight were found among women characterized as high risk.  In  addition, the greater the number of risk characteristics, the greater the impact MIC.  Black adolescents had a significantly lower rate of low birth weight  (12.6%) than blacks i n the comparison group (14.4%). quantitative  sufficiency  of  prenatal  care  in  Improvements i n the  conjunction  with  MIC  comprehensive services appeared to contribute to a reduction i n the incidence of low birth weight for high risk subpopulations.  The authors suggested two  alternative explanations for their findings: (1) that high risk clients were more responsive to M I C services; or (2) that M I C services were provided differentially to high and low risk clients.  A l t h o u g h the authors found a statistically  significant difference i n the rate of low birth weight between the two groups of black adolescents, the 12.5% reduction was not clinically significant.  In 1976, the federal Improved Pregnancy Outcome (IPO) project was initiated by the Bureau of Community Health Services to foster the development of statebased systems of maternity care.(Committee on Perinatal Health, 1976).  The  ultimate goal of the IPO project was to improve maternal and pregnancy outcome i n states that had contributed heavily to the incidence of infant mortality.  In contrast to M I C , states were encouraged to develop their o w n  methods for using IPO funds.  By 1980, 34 states had received IPO funding.  North Carolina and Florida completed evaluations of their IPO projects.  In N o r t h Carolina, IPO funds were used to develop a comprehensive care program in two counties with inadequate maternity services.  Certified nurse  midwives provided prenatal, intrapartum and postpartum care to low income and adolescent mothers.  The local health departments worked i n conjunction  with midwives to provide nutritional counseling, social services, and health education.  Interdisciplinary teams planned, coordinated, and monitored  patient care. Peoples et al (1984) evaluated IPO effects for the period July 1979 to August 1981, by comparing the pregnancy outcomes of (1) all black women i n the two counties served by the IPO program (N=1254) with those of all black w o m e n i n two neighboring counties  (N=1063) with similar  socioeconomic  composition; (2) all black women IPO registrants (N=648) with all black women in the comparison counties; (3) all black adolescents (age 10-19) IPO registrants (N=297) with all black adolescents (N=318) in the comparison counties.  Adequacy of prenatal care and low birth weight information were obtained from vital statistics.  After controlling for group membership, maternal age,  education level, parity, marital status and reproductive risk using a weighted least squares procedure, the investigators  reported significantly more IPO  registrants (41.2%) than controls (30.3%) received adequate prenatal care. The magnitude of difference was greatest among adolescents as 37.8% of IPO adolescents (10-19 years) obtained adequate prenatal care compared with 18.3% of control adolescents.  Despite apparent improvements i n the adequacy of  prenatal care, there was no significant difference in the incidence of low birth weight between IPO groups and comparison groups. The authors suggest three possible reasons why the IPO program d i d not meet its basic goal of decreasing low birth weight.: (1) the program d i d not include specific protocols for managing high risk women (such as preterm labour education or smoking interventions); (2) the intensity of care was inadequate to the degree of risk; or (3) the comparison group women were at less risk and this was not completely controlled in the analysis.  In Florida, IPO funds were initially used to establish a system of prenatal care delivery for low income w o m e n i n five rural counties services. women  w i t h no prenatal  In 1982, the program was expanded statewide to cover all low income i n Florida.  IPO services focussed  retardation and preterm birth.  o n both intrauterine  growth  Services included case management by nurse  practitioners with ancillary service provision by multidisciplinary teams.  The  components  of care i n c l u d e d preterm labour screening,  pregnancy  and  parenting education, home visits, postpartum and well baby care, and family services.  Participants were also referred to the Special Supplemental Food  Program For Women, Infants and Children (WIC)2 services.  Clarke et al (1992) evaluated the statewide IPO program for the period 1985 to 1988, by compared rates of low birth weight and neonatal mortality-^ for program participants and a synthetically matched comparison group.  The  comparison group consisted of nearly poor women, matched to IPO participants on race, maternal care, education, marital status and number of prenatal visits.  For the study years, IPO participants i n both racial groups (white and black) experienced lower rates of low birth weight than their matched comparison group. The difference in the rate of low birth weight ranged from 7.5%-7.7% for whites to 7.6%-11.7% for blacks.  None of the differences were statistically or  clinically significant.  IPO participants i n both racial groups experienced lower rates of neonatal mortality than their matched comparison group. For black IPO participants the rate of neonatal mortality declined dramatically over the study years decreasing from 9.1 to 6.8/ 1000 deaths.  In contrast, the neonatal mortality rates for the  comparison group during the study period increased overall, from 9.1 to 9.8 deaths. The 30% difference in mortality rates was both statistically and clinically significant.  A l t h o u g h the  relative  rates  for  neonatal  mortality  were  WIC program provides food supplements and nutrition education to low income families with children, with particular focus on pregnant women. 3 Although Florida's IPO program also focussed on preterm birth prevention, this outcome was not analyzed due to poor documentation of gestational age infaormation and lack of provider compliance with preterm birth prevention protocols. Personal communication, Dr. L. Clarke. 2  considerably lower for whites than for blacks, the trends i n neonatal mortality over time were similar for both racial groups.  In 1985, the rate of neonatal  mortality among white IPO participants (7.06) was 12% less than the white comparison group rate (8.05).  By the end of the study period, the rate of  neonatal mortality was 27% lower for white IPO participants (4.66 compared with 6.42). Again, this difference was both clinically and statistically significant.  The nonsignificant difference i n the rate of low birth weight may have partially been attributed to by study design. The authors appeared to have overmatched the comparison groups, by retrospectively matching o n five variables.  By  overmatching they may have created two groups that were too similar, therefore diminishing the observed impact of the IPO program.  The results  may also have been confounded by differences in socioeconomic status between the two groups. Given the fact that IPO participants were medicaid eligible, and that medicaid income cutoff was well below the poverty line, IPO participants could have been significantly economically disadvantaged over their nearly poor comparison group.  In response to limited federal funding and increasing rates of poor infant outcomes, many states initiated their own programs to improve maternal care coordination and thereby infant outcomes.  California, N o r t h Carolina and  Kentucky have evaluated state sponsored initiatives.  In the late 1970's, an increase in physicians refusing to accept M e d i - C a l patients, coupled with increases i n the number of M e d i - C a l eligible and other pregnant women reporting difficulties i n obtaining prenatal care, led to the development  of O B Access. The Obstetrical Access Pilot Project (OB Access) was piloted from 1979 until 1982 i n 13 counties i n California.  The project's goals were (1) to provide better access to comprehensive obstetrical services for M e d i - C a l eligible mothers i n areas with inadequate obstetrical care resulting from the lack of a resident obstetrician or from the decision of providers not to participate i n Medic-Cal; and (2) to reduce the incidences of low birth weight, perinatal morbidity and perinatal and infant mortality. O B Access offered a specified maternity benefit package that included psychosocial and nutritional assessment and counselling, perinatal and parenting education and counselling, prenatal vitamins, laboratory tests and well baby exams,  in  addition to routine prenatal, intrapartum, postpartum care.  T w o evaluations of the O B Access project were undertaken.  Korenbrot_(1984)  evaluated the impact of O B Access for the 5244 participants who delivered between 1979 and 1982.  A comparable group of M e d i - C a l reimbursed clients  who gave birth in 1978, was matched on race/ethnicity, age, parity, multiple gestation, infant sex and county of residence.  Information was retrieved from  O B Access records for project participants and from Vital Statistics record for the comparison group. The comparison group was chosen from women who gave birth in 1978, as this was the only year for which M e d i - C a l indicators were available. There was only a slight reduction in the low birth rate for California from 1978 until 1982.  N o information is provided regarding the similarity of the two study groups after  the  matching process.  A d e q u a c y of prenatal care and maternal  complications were not included in the analysis.  The overall rate of low birth  weight was  4.7% for O B Access participants and 7.1% for the  comparison group. significant. infants  matched  This 33% reduction was both clinically and statistically  The rate of very low birth weight was reduced 61% for O B Access  (0.5%) c o m p a r e d w i t h  (1.3%) for matched  comparison  infants.  A l t h o u g h the 61% reduction in the rate of very low birth weight w o u l d be considered clinically significant, the rates were very low i n both groups of infants.  Lennie et al (1987) reevaluated the O B Access Project results for 2825 women who participated in the full package of O B Access services in order to determine program effectiveness.  Full-care O B Access was defined by a m i n i m u m of 8  prenatal exams including a comprehensive initial exam, a psychosocial health education and nutritional needs assessment, at least one birth education class and prenatal vitamin supplements.  Sociodemographic information was not  provided for Full-care or remaining O B Access clients.  This information would  have been useful in determining if there were any differences in the two groups of O B Access mothers.  The rate of low birth weight was 3.1% for infants of Full-care O B Access mothers compared with 7.7% for infants of matched M e d i - C a l cases. This 60% reduction i n the rate of low birth weight was clinically and significant.  statistically  The rate of very low birth weights was more profound with only  one infant (less than 0.1%) of Full-care O B Access mothers experiencing birth weight less than 1500 grams, compared with 1.3% of matched comparison group infants.  Although the O B Access studies made an effort to reduce self-selection bias by obtaining the comparison group from a time period prior to the introduction of project, they were prone to several methodological errors. Overmatching was a problem as the comparison group was matched to O B Access participants on six variables, instead of the recommended four variables.  Overmatching w o u l d  tend to decrease the differences between the groups and diminish the observed impact of the intervention.  In addition, given that data were obtained from  two data sources, systematic differences i n recording could have occurred, introducing another source of bias into the results.  The authors d i d not  document the maternal characteristics of the study groups after matching, therefore it remains unclear if there were significant differences between the two groups of study participants.  Thus, it was feasible that the results could  have occurred due to confounding.  In 1984 the Guilford County Health Department in N o r t h Carolina piloted a comprehensive case managed prenatal care program for indigent women who were not receiving adequate prenatal services. receive  M e d i c a i d because their incomes  This group of women d i d not  were above the M e d i c a i d limit.  Buescher et al (1987) evaluated the effects of pilot program inl984 by comparing the live birth outcomes for 396 health department participants to 362 Medicaid eligible women who received traditional prenatal care from obstetricians.  The  Guilford  County  health  department  provided  a  comprehensive,  coordinated system of prenatal care. Each woman was evaluated on entry and an individual prenatal care plan developed. screen  and educate  women  Special provisions were made to  about preventing  preterm labour.  Health  education, counseling and other health department services were also provided  for the women.  W o m e n at nutritional risk were referred to the W I C program,  administered through the health department.  Both study groups received  intrapartum and post partum care in the same hospital.  Data was obtained from Vital Statistics and W I C records. A significantly greater proportion of the  control group were single  and black.  Fewer  department participants received adequate prenatal care than controls.  health Further  analysis revealed that health department participants started prenatal care later but obtained more visits than controls.  Health department participants had a  significantly (clinically and statistically) significantly lower rate of low birth weight than d i d controls, at 8.3% compared with 19.3%.  The low birth rate  among health department participants was only slightly higher than for the remaining Guilford county population at 7.1%.  The program's success i n  preventing low birth weight was almost entirely through reducing preterm births rather than reductions  i n intrauterine growth retardation.  The  proportion of births that were term and low birth weight was similar for both groups. W o m e n who didn't obtain an adequate quantity of prenatal care made the most significant gains in decreased low birth weight rates.  The authors appropriately used logistic regression analysis to determine the impact of comprehensive  prenatal care (health department) on low birth  weight, after controlling for the effects of race, marital status, W I C , quantity of prenatal care and other measured risk factors. Controls were on average twice a likely (RR=2.1) to have a low birth weight infant.  Other characteristics with a  strong, independent effect on the probability of a low birth weight infant were: multiple birth, inadequate prenatal care, W I C non-participation, single marital status, adolescence and previous poor obstetrical history.  The authors note that the respective low birth weight results could have been influenced  by selection  bias.  Medicaid  controls may have been  more  socioeconomically disadvantaged than health department participants as the income level for Medicaid i n North Carolina was extremely low.  There were  also substantial differences between the two study groups with regard race and marital status that were not controlled i n the analysis.  These  differences  decrease the credibility of the findings.  The authors cleverly chose to address these concerns by examining a subset of study  participants  backgrounds.  that  had  similar  sociodemographic  and  economic  They examined the birth results of 138 health department  participants w h o were M e d i c a i d eligible and compared them w i t h other M e d i c a i d women.  The proportion of the health department subgroup that  experienced low birth weight was significantly and clinically lower at 9.4% for health department participants compared with 19.4% for the comparison group. The authors concluded that a coordinated comprehensive approach to prenatal care was essential for women in poverty, and that the ancillary services of the health department program appear to be most beneficial among those women who start prenatal care late or have an insufficient number of visits.  In October 1987, i n response to concerns over N o r t h Carolina's high infant mortality r a t e / a n d the preliminary results of the Guilford C o u n t y pilot, the state Medicaid program (Baby Love Program) was expanded statewide to offer maternity care coordination services for women at 100% of Federal poverty level.  In order to ensure that a comparison group of nonrecipients could be  obtained, the Baby Love Program was evaluated before the statewide system was fully developed.  Buescher et al (1991) examined all the Medicaid live births i n 1988 and 1989 to evaluate the effect of the program.  For the study years, live birth results were  available for 15,526 Medicaid women who received coordinated care and 34,463 women who received traditional care.  The authors compared birth outcomes  for the two study groups overall, and by length of participation i n coordinated care.  W o m e n receiving coordinated care were slightly more likely to be black, single, less than 18 years and poorly educated.  Due to large numbers of participants,  these differences were statistically significant, though not clinically so. risk factors i n c l u d i n g poor obstetrical history, inadequate  Other  prenatal care,  maternal complications of pregnancy and percentage of smokers were similar between the groups.  The authors found significant differences i n all measured outcomes favoring women who received coordinated care. The rate of low birth weight rate was 8.7% among infants of coordinated care w o m e n compared with 10.5% for control infants; the rate of very low birth weight was 1.2% compared with 2.0 %; and the infant mortality rate was 9.9% compared with 12.2%.  None of these  differences appear to have been clinically significant.  In order to determine the direct impact of coordinated care, the authors once again utilized logistic regression analysis.  They found that w o m e n  who  received standard care were 1.2 times as likely to have a low birth weight infant  than w o m e n w h o received coordinated care.  G i v e n this relative risk, the  program can only be said to have had a modest impact on low birth weight. Although the effect on low birth weight was modest, the authors confirmed a positive financial value of coordinated prenatal care.  For each $1.00 spent on  coordinating services, Medicaid saved $2.02 in medical cost for infant up to 60 days of age.  Kentucky implemented a similar program to N o r t h Carolina's Baby Love Program for Medicaid eligible women in 1984.  U s i n g a similar study design,  Buescher and W a r d (1992) examined all Medicaid single live births in 1985 and 1986.  Birth results were available for 4,978 women who received coordinated  care through health departments and 18,083 women who received traditional care.  Results of the study were stratified by race.  Black controls had slightly more  sociodemographic risk factors for low birth weight than coordinated care recipients.  There were no significant differences in maternal characteristics  between the Caucasian study participants. Regardless of race, significantly more coordinated care recipients obtained W I C benefits and received prenatal care.  adequate  The rate of low birth weight and very low birth weight was  significantly lower for coordinated care recipients of both races.  Caucasian  coordinated care infants had rates of low birth weight and very low birth weight of 6.4% and 0.58% compared with 8.2% and 1.3% for traditional care infants. The rates were higher for black infants at 8.3% and 0.98 % for coordinated care infants compared with 11.7% and 1.87% for traditional care infants. outcomes for black infants appear to have been clinically significant.  Only the  Results of the logistic regression analysis, controlling for selected risk factors, showed the relative risk of having a low birth weight infant for women who received traditional care was 1.26 for Caucasians and 1.37 for blacks. The relative risks again indicate differential program impacts with blacks obtaining the most benefit from coordinated care.  Although six of the eight studies found statistically significant differences i n the rates of low birth weight favoring comprehensive prenatal care participants, only three of these differences w o u l d be considered clinically significant.  All  five studies that examined very low birth weight found differences that were statistically significant.  These results were questionable in three studies as the  rates of very low birth weight were similar to population norms.  It was  impossible to determine if comprehensive care programs exerted their greatest impact on preventing preterm delivery or intrauterine growth retardation, as many studies examined only differences in low birth weight.  The effect of  comprehensive care on infant mortality appears more conclusive, as all three studies found statistically significant results.  Comprehensive prenatal programs appeared to have differential effects based on severity of risk. W h e n results were stratified by race, black comprehensive care participants obtained clinically and statistically favorable low birth weight rates compared to non participants. blacks  confirmed  a moderately  Results of logistic regression analysis for elevated  relative  risk  for n o n p r o g r a m  participants. The effects of comprehensive care for white participants appear to have been modest at best.  In summary, although the comprehensive care studies were sufficiently large to have adequate power, there were a number of factors which may have affected how the above findings were interpreted. studies were susceptible  to selection bias, and this w o u l d be a  alternative explanation for the findings. the design and analysis.  W i t h one exception, all the major feasible  The use of vital statistics data limits  M a n y confounding variables, especially behavioral  variables, were not collected in vital statistics data. Therefore, they could not be controlled i n the analysis.  U s i n g vital statistics information also limited  program measures mainly to birth outcomes, therefore the extent to w h i c h comprehensive measured.  programs effects were  seen i n other areas could not be  Both studies that utilized matching were prone to overmatching,  potentially decreasing the differences between the groups, therefore negating the observed impact of comprehensive programs.  Multidisciplinary Provision - Small Programs  The examination of smaller studies can provide more detailed analysis of program impacts for both mothers and infants at birth and beyond. Six studies c o m p a r i n g the  maternal and infant outcomes  of w o m e n  multidisciplinary comprehensive care have been reported.  who  received  T w o studies had  very small sample sizes, less than 60 per group, and their results are not reported.  Results of the remaining four are detailed here.  O n e program  enrolled low income w o m e n within a health maintenance organization, the other programs were adolescent focussed.  Ershoff et al (1983) reported the results of maternal behavioral changes and infant outcomes for low income w o m e n i n and out of a health education  program i n a health maintenance organization.  A l l w o m e n presenting for  prenatal care between December 1980 and M a r c h 1981 at Hawthorne Health Center who were English literate and less than 24 weeks gestation, received enhanced prenatal care that focussed on nutritional assessment, counselling and  smoking cessation.  C o m p a r i s o n group w o m e n , subject to the same  inclusion criteria, were chosen from two groups; (1) a r a n d o m sample of women who received prenatal service at Hawthorne Health Center i n the four months preceding program initiation; or (2) a random sample of women who delivered in other facilities affiliated with the health maintenance organization (HMO).  The comparison group received standard H M O prenatal care that  included medical care and optional health education programs.  Study results were presented for a subpopulation of women who were smoking on initiation of prenatal care; 57 program participants and 72 controls.  The  program participants were at greater sociodemographic risk, as a greater proportion were black or hispanic, less educated and poor. Significantly more program participants decreased cigarette consumption, ceased smoking and attained adequate weight gain than d i d controls. Adequacy of prenatal care was not reported.  There was  a large birth weight difference  w i t h program  participant infants weighing on average 170 grams more than control infants. Program participant infants obtained a 28% reduction i n low birth rate, 7.0% compared w i t h 9.7% for control infants.  A l t h o u g h this result was  not  statistically significant, likely due to small sample size, the result was clinically significant.  W h e n low birth weight was analyzed according to intrauterine  growth retardation and preterm birth, most of the difference was due to reductions in the rate of preterm birth as 1.7% of program infants were born preterm versus 6.9% of control infants. The 75% reduction was clinically  significant.  T h e p r o g r a m results  were  somewhat  unexpected  intervention d i d not include preterm labour prevention or education.  as  the  There is  general agreement i n the literature, that enhanced weight gain and smoking reduction influence have the largest impact on intrauterine growth retardation not preterm birth.  This evaluation was particularly useful as it controlled for smoking, the confounding variable with the biggest impact on infant birth weight. as the  analysis  was  not stratified  according to differences  However,  i n maternal  demographic variables, other variables may have confounded the results.  In  addition, both groups of study participants had access to educational and nutritional counseling, therefore both groups may have received enhanced care.  Lack of adequate power due to small sample size may have led to the  statistically nonsignificant results.  Smith et al (1978) conducted a case-control study to evaluate the impact of an ongoing comprehensive, multidisciplinary, psycho-social educational program on the medical outcome of pregnant indigent adolescents.  One hundred and  twenty six (126) program adolescents who delivered between 1970 and 1974 were matched to 126 controls based on age, parity, race and month of delivery.  Coincident  with  routine  obstetrical  examinations,  p r o g r a m adolescents  attended weekly classes on nutrition, contraception, child development, labour preparation,  and  psychosocial  multidisciplinary team.  aspects  of  pregnancy,  presented  by  a  Adolescents who attended at least two sessions were  included in the analysis. Controls received standard prenatal care.  Although program adolescents obtained statistically significantly more prenatal care visits (6.3) than controls (5.1), this difference was not clinically significant. Program  adolescents had significantly  fewer  prenatal and intrapartum  complications, and were more likely to return for postpartum exams than were controls.  Infants of program mothers were significantly heavier than control  infants, on average 164 grams. L o w birth weight and preterm birth rates were not  provided.  Differential p r o g r a m effects  were also f o u n d , p r o g r a m  adolescents less than 16 years had infants of significantly higher birth weights and better apgar scores than young adolescent controls. The number of young adolescents (less than 16 years) was not reported.  Elster et al (1987) compared maternal, infant and parenting results for 125 adolescents  who  received  care  from a comprehensive  pregnancy  and  parenthood program between January 1983 and July 1984, with 135 adolescents who received care from community health providers.  The Teen Mother and C h i l d Program ( T M C P ) provided routine medical care, pregnancy education, psychosocial, nutritional and financial counseling, infant health education and parenting skills.  T M C P participants and controls were  eligible for the study if they were less than 19 years of age, English speaking, and free  from major chronic diseases.  Both groups  were also eligible  for  supplemental food coupons through the local W I C program.  Prenatal, labour and delivery, and infant data were obtained from hospital record abstraction.  Subjects were interviewed at home d u r i n g scheduled  intervals for 26 months post delivery to ascertain child health and parenting data.  43  The  two groups were similar with regard to most demographic, behavioral,  medical variables and psychosocial adjustment scores. differed  from  controls by coming from families  T M C P adolescents  w i t h somewhat  higher  socioeconomic status, and they were more likely to currently be attending school, have completed school or be employed at the time of conception. Although significantly less T M C P adolescents began prenatal care in the first trimester, 33% compared to 46% of controls, a significantly greater proportion of T M C P ^adolescents obtained the expected number of prenatal visits, 87% versus 70%. The rates of maternal complications were similar between the two groups with one exception, T M C P adolescents had a significantly increased rate of treated pregnancy induced hypertension. This may however, been due to closer monitoring of T M C P adolescents.  There were no statistically significant differences between the two groups with regard to infant outcomes.  However, 7% of T M C P infants were born preterm  compared to 10% of controls.  This 30% reduction was clinically significant.  Using log-linear analysis to account for differences i n socioeconomic status and vocational educational status, the data was reevaluated.  A l l maternal outcomes  excluding the rate of pregnancy induced hypertension remained significant.  The authors suggest that although the comprehensive program d i d not impact birth outcomes, it had a significant effect on events that occurred during the first two postpartum years. A t 12 and 26 months postpartum, T P C P adolescents scored significantly better on composite  measures  encompassing  medical,  psychosocial and parenting events than d i d the control group, even after accounting for possible confounding factors.  44  This study was both well designed and analyzed.  The authors chose a novel  way of attempting to reduce selection bias by examining the psychological (and motivational) status of study participants. A thorough examination of maternal and infant outcomes were conducted, and differences between the study groups were controlled i n the analysis.  Perhaps most importantly, the  documented the longterm impacts of the program.  Lack of  authors  statistically  significant infant results may have been due to lack of power versus lack of program effects.  Kay  et al (1991) conducted an effectiveness evaluation for adolescents who  attended the Young Adults' Health Center ( Y A H C ) between January 1981-June 1988.  One hundred and eighty (180) Y A H C clients were matched to 180 control  adolescents based on age and year of delivery.  Both study groups obtained at  least three prenatal care visits and delivered at the same hospital.  With a  sample size of 180 per group the authors concluded that study had 90% power to detect a 125 gram difference in birth weight.  The study groups were similar with regard to medical and sociodemographic backgrounds.  Y A H C clients obtained significantly more prenatal care visits  (12.8) compared with their matched controls (9.8).  Significantly more Y A H C  clients reduced or stopped smoking during their pregnancy (27.6%) than d i d controls (9.5%). There was no difference i n the rate of maternal complications between the two groups.  Analysis of variance and covariance was used to compare the pregnancy outcomes of the two study groups. Maternal age, race, insurance coverage, and  smoking status were adjusted for each dependant variable.  After adjustment,  there were no statistically significant differences in birth weight, gestational age, neonatal intensive care admission, infant morbidity, A p g a r scores or rates of low birth weight and preterm birth between the study groups.  However, the  rate of low birth weight was 15.5% for Y A H C clients compared to 10.8% for controls.  This 30% difference was clinically significant.  Postpartum, program  adolescents were more likely to use contraceptives and less likely to become pregnant again after delivery.  This case control study also employed a good design and analysis. Although the authors stated the sample size had adequate power to detect a 125 gram difference in birth weight, it is doubtful that it had adequate power to detect small differences i n rates of low birth weight.  A g a i n , nonsignificant results  may have been due to power issues rather than lack of program effects.  In summary, all four small comprehensive care studies were consistent i n finding that women who participated in comprehensive care obtained both clinically and statistically significantly more prenatal care.  They were able to  show that participants i n comprehensive programs smoked less and gained more weight than d i d controls. outcomes were mixed.  The impact of these changes on infant  T w o studies found significant birth weight differences  favoring comprehensive program participants, especially young adolescents, but none found the rate of low birth weight to be significantly reduced for program infants.  However, it was doubtful that any of the studies had  sufficient sample sizes, thus power to detect differences i n adverse pregnancy outcomes including low birth weight, preterm birth and intrauterine growth retardation.  Perhaps more importantly, both studies examining maternal  behaviors postpartum found program effects for medical, psychosocial and parenting outcomes.  Case Management  In programs that employed case management, one caregiver, generally a NurseMidwife was responsible for planning and reviewing care plans after every client appointment.  Case managers  either w o r k e d i n conjunction w i t h  multidisciplinary teams or referred clients to ancillary service providers when need arose.  T w o randomized trials evaluated the effects of case managed care  for low income women.  Seven case-control studies evaluated the effects of case  managed care for adolescents.  The results of four case-control studies are  detailed here as the remainder had small sample sizes, less than 60 per group.  In a multicenter trial Heins et al (1990) randomized 1458 women at high risk for low birth weight or preterm birth between July 1983 and October 1987. W o m e n were eligible for randomization if they had a score of 10 points or more on their first visit using the risk screening tool developed by Papiernik-Berkhauer (1980) and modified by Creasy et al (1980). In addition, women with a previous history of low birth weight were eligible.  Women were excluded from the trial if they  had a history of medical or pregnancy complications or multiple pregnancy. The authors determined that the trial had a 90% chance of detecting a statistically significant decrease in the rate of low birth weight from 13% to 8%.  W o m e n randomized into the intervention group received case-managed care from  nurse-midwives.  Controls received  pregnancies from obstetricians.  standard  care  for high-risk  W o m e n in the intervention group were seen  every 1-2 weeks throughout their pregnancy and were assessed with regard to lifestyle modification, nutritional attainment, social support, activity level and if required they were taught preterm labour signs and symptoms, uterine palpation and activity restrictions.  Care plan modification and referral as  required occurred at each meeting.  W o m e n in the control group received  standard prenatal care from obstetricians, with less emphasis on personal life style issues and individualized social support. They were seen less frequently for prenatal visits and received cervical examinations only if symptoms of preterm  labour appeared.  Both groups  of w o m e n  h a d access to  the  WIC.program, nutritionists, and public health nurses.  Comparability  was  achieved  between  the  two  groups  in  terms  of  sociodemographic variables risk scores, smoking and clinic site. There was no statistical or clinical difference in smoking cessation rates between intervention and control groups. There was no statistical difference i n initiation of prenatal care. Very few study participants began care in the first trimester, almost half of each group initiated care after 20 weeks gestation.  Information on the number  of prenatal visits obtained was not provided.  Infants of intervention mothers had slightly lower rates of low birth weight, very low birth weight, preterm birth and very preterm birth.  N o n e of the  results were statistically or clinically significant. The rates of adverse pregnancy outcomes remained extremely high i n both groups.  Subgroup analyses were  performed to assess whether the overall non-significant results were obscuring subgroup differences. with high risk scores.  Program effects were shown for infants of black women The rate of very low birth weight was 2.6% for black  intervention infants compared with 6.7% for black controls.  48  In spite of the excellent design format, the trial fell short in several areas. study groups tended  to initiate prenatal care late i n their  Both  pregnancies.  Therefore, the lack of program results may have been confounded by this factor. Lack of program results may also have occurred because the two study groups may have received very similar care. Both groups of women had access to W I C , nutritional counseling and public health nursing services.  A l t h o u g h the  intervention group was seen at regular intervals, there was no way to ascertain that frequent contact d i d not occur for controls. been  Thus, the results could have  contaminated by increased intervention among the  Limited infant outcome information was another issue.  control group.  By failing to provide  information on intrauterine growth retardation, it was impossible to determine what proportion of low birth weight was attributed to this outcome.  Given that  smoking cessation and enhanced nutrition generally effect infant weight gain during each week of gestation, this outcome measure w o u l d have assisted in determining true program impacts.  McLaughlin  et  al  (1992) r a n d o m i z e d 428  low  income  women  into  a  comprehensive prenatal care program. W o m e n qualified for the program by being less than 28 weeks gestation and having a high risk score for child maltreatment.  The program was designed to test the effects of comprehensive  care on birth weight, child development and child maltreatment.  The  comprehensive  care  multidisciplinary team  group  received  of nurse-midwives,  prenatal care  provided by  social workers,  paraprofessional home visitors, and a psychologist.  a  nutritionists,  The team focussed  on  psychosocial support for mothers, education about self-care and promotion of  healthy behaviors d u r i n g pregnancy.  Comprehensive care mothers were  offered individual meetings with the psychologist until they reached 28 weeks gestation, then they attended prenatal support groups led by the psychologist. The comparison group received standard prenatal care by obstetrical residents. Both study groups delivered in the same hospital.  The randomization process succeeded i n producing similar sociodemographic groups.  Birth results were available for 308 women.  care mothers weighed  o n average  Infants of comprehensive  84 grams heavier than infants i n the  comparison group. This difference was not clinically or statistically significant. W h e n stratified by parity, significant program effects on birth weight were seen for  p r i m i p a r o u s women.  Intervention  infants  of p r i m i p a r o u s  mothers  averaged 144 grams heavier than infants of primiparous controls, a modest difference.  C o m p r e h e n s i v e prenatal care d i d not  result  in  significant  differences in the rate of low birth weight for intervention infants as a whole, or for primiparas and multiparas separately.  Multiple regression analysis using treatment group, race and age groupings showed no effect of intervention for the overall sample.  A similar analysis of  variance was performed based on parity. For primiparous mothers, standard care and maternal age were strong predictors of low birth weight.  There were  no predictor variables for multiparous mothers.  Once again, although the analytical design was excellent, the analysis scanty.  was  The authors failed to include maternal outcomes that w o u l d have  enhanced the results.  A s the cutoff for entry into the trial was late (up to 28  weeks), and no information was provided on initiation of care,  nonsignificant  results could have been due to late entry effects rather than lack of program effects.  N o information was p r o v i d e d o n maternal behavioral change, or  prenatal care as a proxy for this.  Therefore it is feasible to question if the  intervention d i d not succeed because women d i d not attend. By only providing infant outcomes, the analysis was severely limited.  L a G u a r d i a et al (1989) examined the impact of intensive social  services,  behavior modification and pregnancy education p r o v i d e d i n a  sheltered  environment, on the incidence of low birth weight among indigent urban adolescents between 1984 and 1986.  One hundred and twelve (112) adolescents  who lived in, and received care from a maternity shelter were compared with 113 adolescents who lived at home.  Both groups of adolescents received  medical care services from the same provider.  Inclusion criteria other than  maternal age less than 19 years at conception, were not stated.  The two study groups had dissimilar backgrounds. Controls were slightly older and more likely to smoke. adolescents  were  of  A significantly greater proportion of intervention  minority  status,  p r i m i p a r o u s , single  and  poor.  Intervention adolescents registered on average a month later for prenatal care. There  were  no  differences  i n weight  gain, prenatal  intrapartum courses between the two groups.  complications  or  Information on adequacy of  prenatal care or number of prenatal visits was not provided.  Intervention  infants were only slightly heavier than controls, the average difference being 40 grams. There were no group differences in the rates of low birth weight infants. However, intervention infants were significantly less likely to be preterm and low birth weight (2.6%) than were controls (9.7%). This difference was clinically significant.The validity of the findings has been confounded b y failure to  account for differences in maternal risk factors, especially smoking, in either the design or analysis.  Using a case-control format, H a r d y et al (1987) assessed the impact of the Johns Hopkins Adolescent Pregnancy Program (JHAPP) over a six year period from 1876 to 1981.  Adolescents i n the J H A P P participated i n a case managed,  comprehensive  program of care.  Controls were chosen from  adolescent  w o m e n who participated i n other Hopkins programs and received standard prenatal care. The program was limited to adolescents less than 18 years of age at the time of conception.  The J H A P P consisted of a defined program of medical care, prenatal and infant care education, behavioral lifestyle assessment, psychosocial support services, and community referral. A multidisciplinary team approach coordinated by an individual case management system was employed by the program. Weekly team meetings and chart review after each visit enabled the J H A P P to monitor appropriateness of care and plan for future needs.  Due to difficulties in ensuring an appropriate control group for the first years of the study, the results were divided into two 3 year blocks from 1976-1978 and 1979-1981. In the first block, results were provided for 930 J H A P P participants and 2028 controls. Program participants were predominantly black, single and on social assistance. Controls differed somewhat by being older, less likely to be black and twice as likely to be multiparas.  Results showed  experienced significantly less maternal complications.  that J H A P P  However, the rates of  low birth weight, preterm birth and perinatal death were similar for both groups. Stratified analysis based on maternal age showed a program effect for  adolescents less than 15 years at delivery.  Program adolescents experienced  significantly less maternal complications, low birth weight infants and preterm births.  For the block 1979-1981 controls were matched to p r o g r a m participants according to: age at delivery; educational attainment; obstetrical history; prepregnancy weight and length of gestation at first prenatal visit.  Maternal  and infant outcomes for the entire group, stratified by age and length of time i n the program were reported for 744 program participants and 744 controls. The matching strategy produced groups that had very similar sociodemographic backgrounds. The proportion of smokers was similar i n both groups. Program participants had significantly better pregnancies, they gained o n average five more pounds than controls; attended on average 0.5 more prenatal visits, and experienced a 30% reduction i n anemia and a 40% reduction in preeclampsia. W i t h the exception of number of prenatal visits, these differences were also clinically significant.  J H A P P infants were only slightly larger than control infants, o n average 45 grams heavier. They had significantly better rates of low birth weight and very low birth weight than controls at 9.9% versus 16.4% and 1.9% versus 3.9% respectively.  The rate of preterm birth was lower among program participants,  but was not statistically or clinically significant. based o n maternal age were reported.  Differential program effects  Both nulliparous and m u c i p a r o u s  adolescents less than 16 years of age had, on average, the largest infants and the lowest frequency of low birth weight.  Differential program effects based on  length of contact with the program were also reported.  Adolescents  who  entered the program during their first trimester gained on average six (6)  pounds more prior than controls.  However, this additional weight gain was  not transferred to their infants who averaged only 64 grams heavier.  This was  an excellent case-control  study  that  appropriately controlled  differences i n study groups through the design, and conducted a thorough analysis of available data.  By matching on five maternal variables, the study  groups were perhaps overmatched, diminishing the observed impact of the intervention.  Piechnik and Corbett (1985) analyzed infant results for the Adolescent Obstetric Clinic ( A O C ) over a five year period 1974-1978, using a case-control analysis. Adolescents who participated in the A O C were cared for by a multidisciplinary team featuring nurse-midwife managed care. Adolescents less than 18 years of age without serious medical or obstetrical complications were eligible for the program.  Adolescents, subject to the same inclusion criteria, who received  standard prenatal care.comprised the control group.  Comprehensive care  included prenatal screening and management, patient education, psychosocial evaluation  and  counseling,  nutritional  assessment  and  counselling,  intrapartum care and postpartum followup.  Results were reported on 738 A O C participants and 2018 controls.  Program  participants were predominantly black, single, and on social assistance. controls were older, Caucasian and married.  More  A g e and race were controlled in  the analysis. The authors state that marital status was not controlled because it was not identified as a factor affecting pregnancy outcome among adolescents. There was both a statistically and clinically significant difference in the overall rate of low birth weight between infants of A O C participants (9.2%) and infants  of controls (12.7%). When stratified by age this difference remained significant. W h e n stratified by race, the low birth weight remained significant for blacks of all ages.  The greatest program impact was found in young black adolescents.  For young black A O C participants (<15 years) the rate of low birth weight was 10.9% compared with 22.3% for control infants.  For the older age group (15-17  years) the rate of low birth weight was 10.2% and 13.6% respectively.  The measured level of significance varied throughout the analyses from 0.1 to 0.15.  When this author reanalyzed the results using the more common level of  significance of 0.05, program impacts were only significant for blacks less than 15 years of age.  Once again, the impact of case-managed comprehensive prenatal care on infant outcomes was somewhat  mixed.  suggested that comprehensive  Results from the two randomized trials  care d i d not have a significant impact o n  improved infant outcomes overall. However, these findings should be viewed with caution, i n one trial the study groups received similar interventions, i n the other trial, serious omissions clouded the results.  Results of the two large  case-control studies found significant differences in the rates of low birth weight favoring comprehensive care infants. risk found significant differences  A l l studies that examined subgroups at i n infant outcomes favoring w o m e n  at  greatest risk. Generally, significant program effects were documented for young adolescents (less than 16 years), black women and primiparous women.  Only  one of the randomized trials and two of the case-control studies had adequate power to detect differences in rates of low birth weight.  Again, the case-control  studies were subject to selection bias and in some studies lack of control over confounding variables  55  H o m e Visitation  The results of 7 randomized trials and 1 retrospective case-control study of prenatal interventions, which included home visits supplemental to prenatal care, were reviewed. social  support  Four randomized trials focussed primarily on enhancing  through  home  visitation,  while  other  components  of  comprehensive care were not actively reinforced^ . In three randomized trials^ and the case-control study^ , home visitors actively intervened and reinforced all aspects of comprehensive prenatal care.  The South Manchester Family Worker Project (SMFWP) r a n d o m i z e d 1227 women at risk for low birth weight between 1982 and 1985 (Spencer et al, 1989). The authors stated that "The project aimed to provide additional social support to pregnant women at above average risk of giving birth to a low birth weight infant.  It was intended that this support w o u l d reduce the level of stress,  thereby improving the well-being of the pregnant w o m e n and ultimately the health of their babies" ( Spencer, et al., 1989, pg. 281).  W o m e n presenting for  routine prenatal care who were less than 20 weeks gestation and identified at risk for having a low birth weight  infant, via a sociodemographic and  obstetrical screening tool?, were included in the trial.  The authors estimated  that the study had a 76% chance of detecting a difference i n birth weight of 77 grams.  4  5  6  7  Spenser et al, Oakley et al, Bryce et al, Dawson et al. Olds et al, Graham et al and Villar et al. Polland et al. Screening tool documented on page 283.  The family workers, were lay women with no formal qualifications i n health care or social services.  They visited each client once or twice a week and  assisted with a variety of tasks including child care, shopping, promotion of health/social services and assisted with housing and state benefit obtainments. O f the women randomized into the intervention group, only 41.4% utilized the services of a family worker. The analysis was based on an intention to treat principle, therefore all women randomized into the intervention group were included in the study results.  The two study groups were almost identical in terms of infant outcomes.  There  was very small difference i n birth weight between the groups, intervention infants were on average 35 grams lighter than controls. The differences in rates of low birth weight, small for gestational age and preterm birth were neither clinically not statistically different.  W h e n results were reanalyzed using only  women who accepted a family worker as the intervention group the results remained nonsignificant.  A subgroup analysis of primigravid adolescents  showed differences in the rate of low birth weight and preterm birth, favoring p r o g r a m infants, but the  differences  were not statistically  or clinically  significant. The size of this subgroup was not large enough to constitute a powerful test of these outcomes.  Study design likely contributed to nonsignificant results. the  project was  to  enhance  social  support,  there  Although the aim of was  no  measure  of  psychological support i n the entry criteria. Given the fact that only 41% of the intervention group received the intervention, it is likely that the study d i d not adequately target those most likely to benefit.  The study may also have been  hampered by a very limited interpretation of social support. The study design  appears to have interpreted assistance with household maintenance as some form of psychological support.  The rationale for this has not been identified.  A s home visitors were actively discouraged from reinforcing any aspect of prenatal care, it is doubtful that this intervention could be called social support.  Oakley et al (1990) randomized 509 women with a history of low birth weight infants to receive either, a social support intervention from midwives addition to standard antenatal care or standard antenatal care.  in  Social support  intervention consisted of a "minimum package" of three home visits plus two telephone contacts.  Midwives provided a listening service for the women to  discuss any topic of concern to them, gave practical information and advice when asked, and carried out referrals to other health professionals and agencies.  Virtually all the intervention mothers obtained at least one home visit, 92% of them received at least three visits. prenatal  care  visits.  Control  Both groups obtained a similar number of mothers  experienced  significantly  hospitalizations  d u r i n g pregnancy than intervention mothers.  received  interventions  more  d u r i n g labour and delivery.  more  They also Postnatally,  intervention mothers reported less anxiety regarding parenting and infant health.  There was virtually no clinical or statistical differences between the two study groups i n any measure of infant outcome.  Intervention infants o n average  weighed 38 grams more than control infants. The rates of low birth weight and preterm birth were extremely high in both groups. infants required additional postnatal care.  Slightly fewer intervention  This trial appears to have been prone to design problems that may have limited the impact of the intervention.  Firstly, the level of psychological support was  not addressed in the entry criteria. Therefore the psychological risk status and the need for social support remains unknown. Secondly, there does not appear to be very much difference between the two groups with regard to prenatal care. The only difference being social support. Even though 90% of the intervention group received 3 or more home visits, they w o u l d have been seen on average every two months.  It is doubtful that this level of servicing w o u l d significantly  affect social support, maternal health or infant outcomes.  Thirdly, the study  was constrained by a very narrow definition of social support. Midwives acted as confidants but were quite restricted in their ability to reinforce prenatal care. Finally, it was doubtful that the trial had adequate power to detect differences i n the rate of preterm birth or low birth weight.  Bryce et al (1991) randomized 1970 women between October 1984 and December 1987.  W o m e n were eligible for the program if they had previously experienced  one or more poor obstetrical outcomes**.  W o m e n were excluded if they were  non-English speaking; previously had been enrolled i n the trial; were more than 25 weeks gestation or were carrying a dead fetus. The program group was offered additional social support by a trained midwife and routine prenatal care, the control group was offered routine prenatal care. The authors conclude that the trial had a 60% chance of detecting a true reduction in preterm birth by 25%.  H o m e visits were scheduled every 4-6 weeks by the midwife, who acted as confidant and listener.  Antenatal care, advice and information were provided  Poor obstetrical outcome included: previous LBW infant or PD; one or more perinatal deaths; three or more first trimester miscarriages; one or more second trimester miscarriages; or antepartum haemorrhage in a previous pregnancy. 8  to program participants only on request. Ninety percent (90%) of the women in the program group accepted the intervention, and 80% obtained at least one home visit.  Program infants experienced a nonsignificant overall reduction in  preterm birth, 12.8% compared with 14.9%.  H o w e v e r , significantly more  extremely preterm infants (20-27 weeks gestation) were born i n the program group (2.8% ) compared with the control group (1.4%).  G i v e n the small  number of infant born extremely preterm, this result could have occurred due to chance. groups.  The rate of low birth weight was virtually the same between the  Subgroup analysis showed a program effect only for women i n the  highest socioeconomic class.  N o program effect was apparent for women with  limited existing social support.  Post hoc stratification indicated a positive  program effect for women with a history of previous preterm singleton births.  This trial was also prone to design errors.  Like the previous two trials,  psychosocial risk status was not an entry criterion for an intervention that focussed on providing psychosocial support.  Therefore the intervention could  have targeted the wrong group. Once again, there d i d not appear to be tangible differences between the prenatal care received by the two groups, as midwives were restricted from reinforcing prenatal care.  It remains unclear what  proportion of the intervention group received more than one home visits, therefore the level/intensity of servicing could not be assessed.  Although the  program appeared to be more effective for w o m e n of high  socioeconomic  status, it is not known if other variables not controlled i n the analysis could hgave been responsible for this.  In a small trial, Dawson et al (1991) randomized 170 low income between July 1977 and March 1978.  women  W o m e n were eligible for the trial if they  were expecting their first or second child, 20-26 weeks pregnant, at least 16 years of age, and could speak English. risk.  W o m e n were not selected for psychosocial  W o m e n were stratified by race and parity and randomized according to  race-parity subgroups.  Women  in  the  intervention  group  were  offered  the  services  paraprofessional home visitor in addition to routine prenatal care. visitors  p r o v i d e d emotional  support,  transport,  household  of  a  The home assistance,  emotional guidance and responded to questions regarding pregnancy, nutrition and health behaviors. home visits.  Health and health services were a minor component of  Control mothers received prenatal care that including social and  nutritional services and occasional home visits by public health nurses.  Thus,  the difference between.the two study groups was the mainly the provision of paraprofessional home visitors.  Ninety-two percent (92%) of intervention mothers accepted the intervention, and 90% of the intervention group had at least two months of home visits prenatally.  There was no difference between the intervention and control  groups i n obstetrical or intrapartum complications.  Infant data were analyzed  with Kruskal-Wallis and Mann-Whitney tests because they were not normally distributed.  There was no statistical difference between the two groups of  infants with regard to gestational age, birth weight, or preterm birth.  However,  4% of intervention infants experienced preterm birth compared with 12.5% of controls.  This 3 fold difference was clinically significant.  infants with low birth weights was not reported.  The proportion of  The non significant results may alternatively be explained by issues of study design and power. Again, psychosocial risk was not a criterion for entry in the study, increasing the speculation about the appropriateness of the intervention. Although 90% of the intervention group received two months of home visits, it is not clear how often these visits occurred. The intensity of servicing, as a result of study design, may not have been sufficient to effect maternal health and infant outcomes.  A l t h o u g h the study results indicated a nonsignificant  preterm birth rate, 4% of intervention infants and 12.5% of control infants experienced preterm birth.  This nonsignificant result may have been due to  inadequate power rather than lack of program effect.  Olds et al (1988) randomized 400 women into a comprehensive program of prenatal and postnatal nurse home visits between A p r i l 1978 and September 1980.  Primiparous women less than 25 weeks gestation were eligible for the  program if they were 18 years old or younger, single, or of low socioeconomic status.  The program aimed at improving the outcomes of pregnancy, early  childrearing, and life-course development.  Intervention w o m e n  received  home visits every 1-2 weeks during pregnancy and care focused on pregnancy and  parent education, behavior modification, enhancement  support  systems  and linkage  with  community  services.  of All  informal women  randomized into the intervention group participated in the program.  The  control group received traditional medical prenatal care and well child visits.  Participation in the program was significantly associated with a wide range of positive maternal effects.  By the end of the pregnancy intervention women  were significantly different from control women: they h a d utilized more community services; attended more prenatal care classes; made greater lifestyle  modifications; increased their reliance on social supports; and experienced less kidney infections  However, these improvements were not transferred to their infants.  There  were no clinical or statistical differences i n birth weight or rates of low birth weight or preterm birth between intervention and controls overall.  However,  subgroup analyses showed positive program effects for young adolescents (14-16 years) and smokers.  In contrast to the control group, infants born to young  adolescents were significantly heavier, weighing on average 395 grams more. There  were no  low  birth weight or preterm birth infants  among  the  intervention young adolescents^, compared with 11.8% of the infants of young control adolescents. percentage  Program smokers experienced  a significantly  of preterm birth at 2.1% compared w i t h 9.8%  for  lower  controls.  Postpartum followup showed program participants were less anxious about parenting, used less restrictive and punishing behaviors and provided more educative and stimulating toys for their infants. Infants of program participants had significantly better mental development at 12 and 24 months postpartum. This excellent trial was able to clearly show both short term and longterm outcomes for mothers and infants.  The L a t i n A m e r i c a n N e t w o r k for Perinatal and Reproductive Research conducted a randomized trial of home visitations at four centers i n Argentina, Brazil, C u b a and Mexico between January 1989 and M a r c h 1991 (Villar et al, 1992).  W o m e n were i n c l u d e d i n the  trial  if they  h a d one  or  more  Had the authors chosen to evoke the 'Rule of Three's" by substituting 3 (the upper 95th confidence limit for a null value) for the null value, 10.7% of the intervention infants would have been born preterm and low birth weights. Therefore, there would have been no clinical or statistical difference between the intervention and control infants for these outcome measures. 9  sociodemograhic or physical risk f a c t o r s ^ for delivering a low birth weight infant. had  W o m e n who met the criteria, were less than 23 weeks gestation and  no  history  of  major  mental  illness,  cervical  celclage  or  Rh  isoimmunization were recruited.  A total of 2235 women participated i n the study.  W o m e n i n the intervention  group received four to six home visits, about one a month, from a social worker or nurse.  In addition to the provision of direct emotional support, the home  visitor was actively involved i n education and counselling regarding medical recommendations, nutrition and health behaviors.  The control group received  standard prenatal care.  Eighty three percent (83%) of the women assigned to the intervention group received at least four home visits, and 90% were visited at least once.  Study  results were recorded according to intention to treat principle, therefore results were  reported  on  all w o m e n  randomized  to  the  intervention  group.  Intervention w o m e n had significantly greater prenatal care knowledge and reduced their smoking consumption more than d i d controls. Both groups experienced a similar number of prenatal care visits.  N o clinical or statistical  differences were observed between the groups overall i n rates of maternal and neonatal morbidity, low birth weight, preterm birth and intrauterine growth retardation, length of stay and postnatal hospitalization.  Stratified analysis  based on demographic risk factors, psychological distress and study site were conducted.  Program effects were seen for women with pshychological risks.  Infants of women with high base line levels of psychological distress and low ™ Risk factors included: previous LBW infant or PD; previous fetal or infant death; age <19 years; body weight <51kg; low family income; less than 3 years of school; cigarette or alcohol consumption; or single parenthood.  levels of social support had consistently  lower rates of low birth  weight,  preterm birth and intrauterine growth retardation, though not significantly so. The remaining stratified analyses revealed no trends and were not clinically or statistically significant.  Several limitations on the methods Villar et al., may have precluded a fair test of the hypothesis pregnancies.  that social support was not beneficial d u r i n g high risk  First, the selection criteria may not have been adequate  to  determine a high risk group, as the rate of low birth weight i n controls (9.4%) was similar to the unselected population rate of 10.6% for Mexico city (Barros, 1992).  Second, it w o u l d appear that the providers of social support had no  previous close relationship with the recipients.  T o be effective, social support  may have to be provided by persons with w h o m the recipients have had time to develop a relationship. Third, the providers saw the women roughly once a month during the intervention, which may have been insufficient to provide meaningful social support.  In a small trial, Graham et al (1992) randomized 154 high risk, low income women between March 1987 and September 1989.  W o m e n who were between  17 and 28 weeks gestation, who had a low family function score, and had experienced at least one stressful life event prior to registration were eligible for the intervention.  Other risk factors included smoking, low maternal weight-  height ratio, maternal age over 27 or previous preterm birth.  The goal of the  program was to strengthen intrafamilial and interpersonal support systems of the women. of interest.  Prenatal care utilization and low birth weight were the outcomes  Both the experimental and the control group received routine prenatal care. The experimental group also received home intervention from trained peers, that provided psychosocial support, nutritional education, prenatal care and childbirth education, lifestyle health risk education and community service referral.  Sixty three percent (63%) of the experimental group obtained four  home visits, 74% obtained at least one home visit.  Rates of prenatal care utilization indicated that women i n the experimental group, regardless of the number of home visits, had significantly more prenatal visits than those in the control group.  W o m e n who obtained at least four  home visits had a lower rate of low birth weight than those who obtained only one visit. However, the difference was not clinically or statistically significant.  Study design may have contributed to the nonsignificant infant  results.  A l t h o u g h w o m e n could enter the trial between 17 and 28 weeks gestation, information on when they entered was not provided. With such a late cutoff, sufficient time may not have been available for social support to affect maternal behavior.  In addition, as the number of home visits were limited to 4, the  program may have lacked intensity to effect change.  One of the weaknesses of  the program was the limited involvement of family support persons, therefore the home visitor could do very little to influence familial support.  Although  the authors contend they had adequate power to detect a 15% absolute reduction i n low birth weight, their estimates may have been overly liberal, and the sample size too small.  Polland et al (1992) reported the results of The Maternal C h i l d Health Advocate ( M C H A ) case-control study to assess the effectiveness of paraprofessional  advocate services on participation in prenatal care and infant birthweight.  One  hundred and eleven (111) low income black w o m e n who received three or more advocate contacts and delivered singleton births were matched to 111 w o m e n from the same prenatal clinic who received traditional prenatal care. Advocates were peers of similar educational background and ethnicity who counselled  and assisted pregnant w o m e n w i t h health and social  service  referrals, housing, shopping, transportation and other basic necessities.  Matching criteria included trimester of prenatal care initiation; parity and race. Results of the match indicate that the comparison group was slightly more advantaged sociodemographically than were program participants.  Results of  the analysis indicated program effects for both prenatal care participation and birth weight.  Program participants obtained more prenatal visits than d i d their  comparison group with an average of 8.0 visits versus 6.5 visits. was both clinically and statistically significant.  The difference  Although the authors found a  statistically significant difference in birth weight favoring program infants, the 148 grams average difference was modest.  The rates of low birth weight and  preterm birth were not included. The effects of the intensity of advocate contact on amount of prenatal care and birthweight were examined by stepwise linear multiple regression within the program group.  Intensity of advocate contact  was the only predictor variable significantly associated with prenatal care participation. N o predictor variable contributed significantly to birthweight.  Self selection  may bias limits the findings, as the study d i d not use a  randomized design.  A l t h o u g h program participants were more likely to be  socially disadvantaged, by virtue of participating i n an intervention program, they may have been motivationally advantaged.  This confounder  among  many others may have complicated the results.  Power calculations were not  included in the results, but with the small sample size, the power was likely not sufficient to rule out chance as an alternate explanation of the results.  The results of comprehensive programs that included home visits, once again found that women who participated in comprehensive prenatal care improved their health habits by reducing their smoking consumption and increasing their weight gain compared to nonparticipants.  In general, significantly more  women i n comprehensive care obtained the expected number of prenatal visits for length of gestation when compared to nonparticipants.  In general, the  improvement i n maternal health habits, as measured by increased prenatal visits, d i d not however lead to overall improvements i n infant outcomes. few  studies found that comprehensive  A  care participants had significantly  heavier infants than hon participants, however, the difference from a clinical perspective was modest.  A l l seven randomized trials of home visitation failed  to demonstrate significant effects on low birth weight, preterm birth or other infant outcomes.  O f the five randomized trials with sufficient power, three  detected program effects for high risk subgroups; primigravid adolescents, y o u n g adolescents, smokers and w o m e n with high baseline  psychological  distress or low levels of social support.  In summary, the lack of significant differences  in overall rates of adverse  outcomes for comprehensive care participants who received home visitation interventions could possibly be explained by design issues. In general the aim of home visitation programs was to provide additional social, educational, and home support to women and their families.  The hypothesis being that women  who feel supported would engage in healthy behaviors, and this would lead to  healthy infant outcomes.  W i t h the exception of one or two trials, the study  designs were seriously flawed.  They failed to utilize measures of psychological  risk i n their entry criteria, leading to speculation about the appropriateness of the target group. Lack of acceptance of the intervention, infrequent contact and lack of involvement of family supports characterized all but two trials.  Four  trials defined enhanced care so narrowly that there was virtually no difference between the intervention and control group.  The one trial that provided an  excellent design d i d find subgroup program effects.  CONCLUSION  W h e n compared to standard prenatal care, comprehensive care for socially disadvantaged women has been shown to be effective in assisting women to make lifestyle behavioral changes.  There was some evidence to show that  comprehensive care participants were more successful in reducing or quitting smoking, improving their nutritional status and i n gaining adequate  weight.  Comprehensive care also facilitated participants ability to make appropriate use of community resources, especially WIC.  W h e n initiation of prenatal care, number of prenatal visits a n d / o r adequacy of prenatal care^l were analyzed as a proxy for maternal behavior, mixed results were found in the literature. In general, both comprehensive care participants and nonparticipants initiated prenatal care during their second trimester of pregnancy.  A s many measures of prenatal care adequacy depend on first  trimester initiation of prenatal care, very few comprehensive care participants or nonparticipants obtained adequate care. The effectiveness of comprehensive  1 1  As measured most commonly by Kessner's Prenatal Care Index.  prenatal programs tended to be shown in differences i n the number of prenatal visits between participants and nonparticipants.  Published studies tended to  find that w o m e n in comprehensive care obtained significantly more prenatal visits, even when they came later in their pregnancies.  Conclusions were harder to draw with respect to the effects of comprehensive care o n birth outcomes.  Randomized trials that examined program efficacy  found no overall program effects for low birth weight or preterm delivery. The reason for the nonsignificant infant results i n many of the trials could quite possibly be accounted for by poor study designs.  O n l y two trials utilized  appropriate design, however, they tended to have small sample sizes. In spite of the overall nonsignificant results, many trials found favorable program effects for subgroups of w o m e n outcomes.  at increased risk of adverse  pregnancy  Primiparous women, young adolescents (less than 19 years), and  blacks were the most commonly reported groups for w h o m program effects were seen.  In case-control studies with adequate power, the majority found program effects for low  birth weight,  randomized  trials,  comprehensive care. targeting  preterm birth and neonatal  similar  subgroups  were  mortality.  found  to  As  benefit  from  There were several studies that evaluated programs  adolescents and found p r o g r a m participants had infants  significantly  with  higher birth weights and lower rates of low birth  with  weight,  particularly young adolescents. This was supported by several smaller studies not detailed i n this r e v i e w ^ .  Case-controlled studies, essentially  provide  information on a highly selected group of participants and, therefore, provide  1 2  Daniels & M a n n i n g , 1983; Felice et al, 1981; Smoke & Grace, 1988  weaker  evidence  than  r a n d o m i z e d trials, a n d  their  results  were  not  generalizable.  In both randomized trials and case-control studies the literature found that longer participation i n prenatal care was an important factor i n achieving positive maternal and infant outcomes.  W o m e n who participated longer  gained more weight, obtained more prenatal care, and had infants that were heavier and at decreased risk for low birth weight.  Older adolescents and  Caucasian primips were the subgroups most likely to obtain longer participation i n comprehensive care.  Comprehensive care also appeared to have an impact after the program was complete.  Participation in comprehensive care was associated with longterm  maternal and infant changes.  Improved postpartum psychosocial adjustment,  better parenting skills, lengthened pregnancy intervals were found i n women who participated in comprehensive care. Improved cognitive ability was found for comprehensive care infants and children up two years of age.  Several methodological problems with both randomized trials and case-control studies were found in this literature review.  The majority of the randomized  trials failed to include in their selection criteria information that was crucial to determining the appropriate target group for their intervention.  Lack of  participation i n the intervention, u n k n o w n or infrequent contact w i t h the intervention and very restrictive interventions were methodological problems common to the majority of the trials.  Selection bias is by far the biggest methodological problem i n case-control studies.  The self-selection of some women into comprehensive care and others  to avoid it confounded the causal relationship between comprehensive care and outcomes.  Attempts to diminish the bias of self selection were rarely  reported in the literature.  T w o studies addressed this issue, one by choosing  controls from a time period prior to initiation of the intervention, the other, more creatively, by attempting to obtain controls that were motivationally similar.  Lack of control for confounding variables, (especially behavioral  variables) characterized every large study that utilized vital statistics data. The quality and type of the data available in secondary data bases, especially vital statistics data bases, severely restricted these analyses.  M a n y studies that  utilized vital statistics data were unable to utilize gestational age data due to missing and outlying data.  This was reflected i n lack of intrauterine growth  retardation information for infants.  Several methodological problems were common to both randomized trials and case-control studies.  Socially disadvantaged women, for w h o m most of these  interventions were targeting, tended to enter care during their second or even third trimester.  By virtue of entering late, they were unable to obtain full  benefits of comprehensive care.  Nonsignificant differences may well have  occurred due to late entry effects and not lack of program effects.  Another methodological problem commonly encountered was lack of adequate power to determine true program effects. Thireteen studies simply had sample sizes too small to detect differences in rates of adverse infant outcomes.  Equally  as common were studies that calculated the power of their study based on very  modest differences in birth weight or, conversely extreme differences in adverse outcomes.  The final methodological concern was the adequate measurement of prenatal care.  Popular measures of prenatal care, such as trimester of entry and the  number of prenatal visits, were limited by the fact that women at either end of these continuums d i d not compose homogeneous groups.  W o m e n who enter  care early in pregnancy and who obtain many prenatal visits may have (or have had) problematic pregnancies or medical risks which required additional care, or they may have been healthy w o m e n who simply wanted to obtain good prenatal care.  Conversely, women who entered care late in pregnancy may  have h a d very healthy previous pregnancies and, consequently, d i d not perceive a need for early care, or they may have been w o m e n who d i d not know they were pregnant or d i d not understand the need for prenatal care. Because of this lack of homogeneity within categories of prenatal care use, the relationships between traditional measures of prenatal care and birth outcomes remain unclear.  REFERENCES  Abramowicz, M . & Kass, E . H . (1966). Pathogenesis and prognosis of prematurity. N e w England Tournal of Medicine. 275. 878-1059. Barros, F . C , Huttly, S.R., Victoria, C . G . Kirkwood, B.R. & Vaughan, J.P. (1992). Comparison of the causes and consequences of prematurity and intrauterine growth retardation: a longitudinal study i n southern Brazil. Pediatrics. 90. 238-244. Brown, S.S. (ed.) (1988). Prenatal care: Reaching mothers, reaching infants. Washington, D C : National Academy Press. Bryce, R.L., Stanley, F.J. & Garner, J.B. (1991). Randomized controlled trial of antenatal support to prevent preterm birth. British Journal of Obstetrics and Gynaecology. 98,1001-1008. Buescher, P . A . , Smith, C , Holliday, J.L. & Levine, R . H . (1987). Source of prenatal care and infant birthweight: The case of a North Carolina county. American Journal of Obstetrics & Gynecology. 156(1), 363-264. Buescher, P . A . , Roth, M.S., Williams, D . & Golforth, C M . ( 1991). A n Evaluation of the Impact of Maternity Care Coordination o n Medicaid Birth Outcomes i n North Carolina. American Journal of Public Health. gl/12), 1625-1629. Buescher, P . A . & W a r d , N.I. (1992). A Comparison of L o w Birth Weight A m o n g Medicaid Patients of Public Health Departments and Other Providers of Prenatal Care i n North Carolina and Kentuky. Public Health Reports. 107(1). 54-59. Clarke, L . L . , Millar, M . K . , Vogel, W.B., Davis, K . E . & Mahan, C S . (1993). The Effectiveness of Florida's :"Improved Pregnancy Outcome" Program. Tournal of Health Care for the Poor and Underserved. 4(2). 117-132. Collins, J.W. & David, R.J. (1990). The differential Effect of Traditional Risk Factors o n Infant Birthweight among Blacks and Whites i n Chicago. American Tournal of Public Health. 80 (6). 679-681. Committee on Perinatal Health (1976). Toward improving the outcome of pregnancy. White Plains, N Y : The National Foundation-March of Dimes.  74 Corman, H . & Grossman, M . (1985). Determinants of neonatal mortality rates in the U.S. Tournal of Health Economies. 4. 213-236. Daniels, M . B . & Manning D . (1983). A Clinic for Pregnant Teens. A m e r i c a n Journal of Nursing. January, 68-71. Dawson, P., V a n Doorninick, W.J. & Robinson, J.J. (1989). Effects of HomeBased, Informal Social Support o n C h i l d Health. Developmental and Behavioral Pediatrics. 10(2), 63-67. Drillien, C M . (1957). The social and economic factors affecting the incidence of premature birth. Part 1. Premature births without complications of pregnancy. Journal of Obstetrics & Gynecology of the British Empire. 64. 161-165. Elster, A . B . , Lamb, M . E . , Tavare, J. & Ralston, C . W . ( 1987). The Medical and Psychosocial Impact of Comprehensive Care on Adolescent Pregnancy and Parenthood. T A M A . 258.1187-1192. Ershoff, D . H . , Aaronson, N . K . , Danaher, B.G. & Wasserman, F . W . (1983). Behavioral, health and cost outcomes of a H M O - b a s e d prenatal health education program. Public Health Reports. 98(6). 536-547. Felice, M . E . , Granados, J.L., Ances, I.G., Hebel, R , Roeder, L . M . & Heald, F.P. (1981). The Young Pregnant Teenager: Impact of Comprehensive Prenatal Care. Journal of Adolescent Health Care. 1(3). 193-197. Geronimus, A . T . (1986). The effects of race, residence, and prenatal care on the relationship of maternal age to neonatal mortality. A m e r i c a n Journal of Public Health. 69(7). 653-657. Gortmaker, S. (1979). The effects of prenatal care upon the health of the newborn. American Journal of Public Health 69(7). 653-657. Graham, A . V . , Frank, S.H., Zyzanski, S.J., Kitson, G C . & Reeb, K . G . (1992). A Clinical Trial to Reduce the Rate of L o w Birth Weight i n an Inner-city Black Population. Family Medicine. 24(6). 439-446. Greenberg, R.S. (1983). The impact of prenatal care i n different social groups. Obstetrics & Gynecology. 145(7). 797-801. Hall, M . H . , Chng, P.K. & MacGillivray, I. (1980). Is routine antenatal care worth while? Lancet. 2. 78-80.  75 Hardy, J.B., King, T . M . & Repke, J.T. (1987). The Johns Hopkins Adolescent Pregnancy Program: A n Evaluation. Obstetrics and Gynecology. 3(1). 300305. Harris, J.E. (1982). Prenatal medical care and infant mortality. In: Fuchs V , ed. Economic aspects of health. Chicago: University of Chicago Press. Harvey, S.M. & Faber, K.S. (1993). Obstacles to Prenatal Care Following Implementation of a Community-Based Program to Reduce Financial Barriers. Family Planning Perspectives. 25(1). 32-36. Heins, H . C , Nance, N . W . , McCarthy, B.J. & Efird, C M . (1990). A randomized trial of nurse-midwifery prenatal care to reduce low birth weight. Obstetrics & Gynecology. 75.341-345. Hulsey, T . C , Levkoff, A . H . , Alexander, G.R. & Tompkins, M . (1991). Differences in Black and White Birth Weights: The role of Maternal Demographic Factors and Medical Complications of Pregnancy. Southern Medical Tournal. 84 (4). 443-446. Institute of Medicine (1985). Preventing L o w Birth Weight. Washington, D C : National Academy Press. Institute of Medicine (1988). Prenatal Care: Reaching Mothers, reaching Infants. Washington, D C : National Academy Press. Kay, B.J., Share, D . A . , Jones, K . , Smith, M , Garcia, D . & Yeo, S.A. (1991). Process, Costs, and Outcomes of Community-based Prenatal Care for Adolescents. Medical Care. 29(6). 531-541. Kessner, D . M . , Singer, J., Kalk, C E . & Schlesinger, E.R. (1973). Infant Death: an analysis by maternal risk and health care. Contrasts i n Health Status. V o l 1. Washington, D C : Institute of Medicine. National Academy of Sciences. Korenbrot, C , Showstack, J., Loomis, A . & Brindis, C . (1989).. Birth weight outcomes i n a teenage pregnancy case management project. Journal of Adolescent Health Care. 10.(2), 97-104. Korenbrot, C . C (1984). Risk Reduction i n Pregnancies of Low-Income Women: Comprehensive Prenatal Care Through the O B Access Project. Mobias. 4(3). 34-43. LaGuardia, K . D . , Druzin, M . L . & Eades, C . (1989). Maternity shelter care for adolescents: Its effect pn incidence of low birth weight. American Tournal of Obstetrics & Gynecology. 161(2). 303-306.  76  Lennie, J.A., K l u n , J.R. & Hausner, T. (1987). Low-Birth-Weight rate reduced by the obstetrical access project. Health Care Financing Review. 8(3). 83-86. Leveno, K J . Cunningham, F . G . , Roark, M . L . et al. (1985). Prenatal care and the low birth weight infant. Obstetrics & Gynecology. 66. 599-605. McLaughlin, F.J., Altmeier, W . A . , Christensen, M.J., Sherrod, K.B., Deitrich, M.S. & Stern, D.T. (1992). Randomized Trial of Comprehensive Prenatal Care for Low-Income Women: Effect on Infant Birth Weight. Pediatrics, mXh 128-132. Miller, H . C . & Jekel, J.F. (1987). The effects of Race on the Incidence of L o w Birth Weight: Persistence of Effect After Controlling for Socioeconomic, Educational, Marital and Risk Status. The Yale Journal of Biology and Medicine. 60. 221-232. Moore, T. R., Origer, W . & Key, T . C . (1986). The perinatal and economic impact of prenatal care i n a low-socioeconomic population. American Journal of Obstetrics & Gynecology. 154(1), 29-33. Mortimer, G . , Merkatz, I.R. & H i l l , J.G. (1991). Caring for our future: A report by the Expert Panel on the content of prenatal care. Obstetrics & Gynecology. 77(5). 782-787. Norris, F.D.& Williams, R.I. (1984). Perinatal outcomes among Medicaid recipients i n California. American Journal of Public Health. 74(10). 11121117. Oakley, A . , Rajan, L . & Grant, A . (1990). Social support and pregnancy outcome. British Tournal of Obstetrics and Gynaecology, 97,155-162. Olds, D . M . (1990). The Prenatal/Early Infancy Project: A Strategy for Responding to the Needs of Fligh-Risk Mothers and Their Children. Early Childhood Programs, (pg 59-87). Haworth Press. Olds, D . L . , Henderson, C.R., Tatelbaum, R. & Chamberlin, R. (1986). Improving the delivery of prenatal care and outcomes of pregnancy: A randomized trial of home visitation. Pediatrics. 77.16-28. Peoples, M . D . , & Siegal, E . , (1983). Measuring the Impact of Programs for Mothers and Infants o n Prenatal Care and L o w Birth Weight: The Value of Refined Analyses. Medical Care, 21(6), 586-605.  Peoples, M . D . , Grimson, R.C., & Daughty, G . L . (1984). Evaluation of the effects of the North Carolina Improved Pregnancy Outcome Project: Implications for State-Level Decision-Making. American Tournal of Public Health. 74(6). 549-554. Piechnik, S.L. & Corbett, M . A . (1985). Reducing L o w Birth Weight A m o n g Socioeconomically High-Risk Adolescent Pregnancies: Successful Intervention with Certified Nurse-Midwife- Managed Care and a Multidisciplinary Team. Journal of Nurse-Midwifery. 30(2). 88-98. Polland, M . L . , Giblin, P.T., Waller, J.B. & Hankin, J. (1992). Effects of a H o m e Visiting Program on Prenatal Care and Birthweight: Case Comparison Study. Tournal of Community Health. 17(4). 221-229. Sanderson, M . , Placek, P.J. & Keppel, K . G . (1991). Results of the 1988 National Maternal and Infant Health Survey. Birth. 18 (1). 26-32. Schlesinger, M . & Kronebusch, K. (1990). The failure of prenatal care policy for the poor. Health Affairs. 9(4). 91-111. Scupholme, A . , Robertson, E . G . & Kamons, A.S. (1991). Barriers to prenatal care in a multiethnic urban sample. Tournal of Nurse-Midwifery. 36(2). 111116. Showstack, J.A., Budetti, P.P. & Minkler, D . (1984). Factors Associated with Birthweight: A n Exploration of the Roles of Prenatal care and Length of Gestation. American Tournal of Public Health. 74(9). 1003-1008. Shwartz, S. (1962). Prenatal care, prematurity and neonatal mortality. American Tournal of Obstetrics & Gynecology. 83(5). 591-598. Smith, P.B., Wait, R.B., Mumford, D . M . , Nenney, S.W. & Hollins, B.T. (1978). The Medical Impact of an Antepartum Program For Pregnant Adolescents: A Statistical Analysis. American Journal of Public Health, 68(2). 169-172. Smoke, J & Grace, M . C (1988). Effectiveness of Prenatal Care and Education for Pregnant Adolescents: Nurse-Midwifery Interventions and Team Approach. Journal of Nurse-Midwifery. 33(4). 178-184. Sokol, R.J., Woolf, R.B., Rosen, M . G . , & Weingarden, K . (1980). Risk, antepartum care, and outcome: impact of a maternity and infant care project. Obstetrics & Gynecology, 56,150-156. Stickle, G . & M a , P. (1977). Some social and medical correlates of pregnancy outcome. American Journal of Obstetrics & Gynecology, 127,162-166.  Spencer, B., Thomas, H . & Morris, J. (1989). A randomized controlled trial of the provision of a social support service during pregnancy: The South Manchester Family Worker Project. British Tournal of Obstetrics and Gynaecology. g £ , 281-288. Taffel, S. (1978). Prenatal Care United States, 1969-1975. Vital and Health Statistics, Series 21, No. 33, D H E W Pub. N o . P H S 78-1911. Hyattsville, M D . National Center for Health Statistics. Taffel, S. (1980). Factors associated with low birthweight United States, 1976. Vital and Health Statistics, Series 21, N o . 37, D H E W Pub. N o . P H S 801915. Hyattsville, M D . National Center for Health Statistics. Terris, M . , & G o l d , E . M . (1969). A n epidemiologic study of prematurity. American Tournal of Obstetrics & Gynecology. 103(3). 371-379. Villar, J., Farnot, U . , Barros, F., Victora, C , Langer, A . , & Belizan, J . M . (1992). A Randomized Trial of Psychosocial Support During High-Risk Pregnancies. The N e w England Journal of Medicine. 327(18). 1266-1271. Young, C . L . , M c M a h o n , J., Bowman, V . M . & Thompson, D . (1989). Adolescent third trimester enrollment in prenatal care. Journal of Adolescent Health Care. 10. 393-397.  79 CHAPTER 3 P R E G N A N C Y O U T R E A C H PROJECTS  Background  In 1988, the British Columbia Ministry of Health piloted Pregnancy Outreach Projects (POP) i n eight B . C . c o m m u n i t i e s ^ .  These prenatal programs were  developed to address adverse pregnancy outcomes, especially low birth weight among high risk populations, and were initiated i n response to concerns from both the public health and medical communities.  While health and social  service authorities recognized that low birth weight was  associated  demographic, medical, obstetrical, behavioral and environmental traditional prenatal care only addressed medical and obstetrical (Institute of Medicine, 1985).  with  factors, concerns  In addition, most women who participated i n  community based prenatal services were not at risk for having low birth weight infants.  H i g h risk women frequently delayed access to traditional prenatal care  and d i d not access community based prenatal services.  W i t h these factors in  m i n d , the P O P was developed to enhance the prenatal care of high risk pregnant women through community directed services.  Pregnancy Outreach Projects  are c o m m u n i t y based  programs aimed  at  identifying high risk women in the community, engaging them in prenatal care services, and supporting them i n making behavioral changes to reduce their risk of having low birth weigh infants or other adverse pregnancy outcomes (Pregnancy Outreach Projects: Project Handbook,1993).  The focus of the  Pregnancy Outreach Programs were piloted i n Cranbrook, Terrace, Prince George, W i l l i a m s Lake, Surrey, Port A l b e r n i , Duncan a n d Nanaimo. Currently there are 22 communities participating i n the P O P s . 1 3  programs is on behavioral modification.  Particular attention is p a i d to  nutrition, emotional support, smoking, alcohol and drug use.  Funding for P O P is provided by contract to community agencies which have well established links to high risk women. typically not health care agencies.  These sponsoring agencies are  Support and direction is provided to the  agencies t h r o u g h mandatory advisory professionals and community leaders.  committees  composed  of  health  Local health units maintain close ties  through representation on advisory committees.  The programs are staffed by nurses, nutritionists, outreach workers and volunteers.  A health professional coordinates the program, while the outreach  workers are the primary service providers. services to both clients and staff.  Volunteers provide support  Peer support through the use of outreach  workers is an integral component of the programs.  Objectives  The goal of the P O P is to promote positive health practices that contribute to the health of infants and mothers. interventions.  This goal is achieved through lifestyle oriented  Individual health goals are developed and clients are assisted in  achieving these by modifying their behavior.  Six broad objectives are used to  measure program success, these are presented in Table 3.1.  Table 3.1 Pregnancy Outreach Projects Objectives  Objective 1  To increase food intake to meet the m i n i m u m recommended servings i n the "B.C. G u i d e for Pregnancy".  Objective 2  To decrease the number of cigarettes smoked per day i n those pregnant women w h o smoke.  Objective 3  To decrease the use of alcohol w i t h a v i e w to abstinence.  Objective 4  To reduce d r u g use to only those approved b y a physician.  Objective 5  To ensure that there is at least one consistent source of emotional support for the client To encourage breast feeding so that 70% of clients are breast feeding o n discharge from hospital.  Objective 6  Service M o d e l  The  P O P utilize  a multidimensional  service  model  to  achieve  program  objectives. This model is shown i n Figure 3.1.  Figure 3.1 Pregnancy Outreach Projects Service M o d e l  Outreach Assessment Referral to P O P  • Risk Identification •T-ACE  Care Plan  •Group Sessions •Food Supplements • Individual Counselling • Referral  Reassessment •Case Conferencing  * Acknowledgements and outcomes of referral are provided to referral source.  Referral from POP  82  Pregnancy Outreach Project service begins with referral; most  commonly  through self referral, health unit, community agency or p h y s i c i a n ^ .  During  the first few client contacts, the program staff determine specific risks through the use of the Individual Prenatal Risk Identification Tool and T - A C E (alcohol screening) questionnaire.  In order to be eligible for the program a client must  be less than 28 weeks gestationl^  a  n  d have at least one major risk or three  minor risk factors.  Client screening and assessment for program eligibility are conducted by all program staff.  In consultation with the client, the program staff develop a  client specific care plan.  One staff member, usually the outreach worker, is  chosen as the key worker for each client; however the team approach to service remains strong.  Program Components  There are four essential components to program service.  These components  were selected based on supporting evidence i n the literature and experience with high risk counselling programs. The four components that comprise the program: group sessions, food supplements, individual counselling and health care referrals are detailed in Table 3.2.  Referrals to the program come from many sources including: family physicians, public health professionals, social services, alcohol and d r u g programs, community agencies, prenatal instuctors, native friendship centers, a n d self via: friends; newspaper articles; pamphlets a n d bulletins. 15 Pregnancy Outreach Programs are limited to w o m e n less than 28 weeks gestation as lifestyle interventions are unlikely to alter the course of pregnancy thereafter. 1 4  Table 3.2 Program Components  Group Sessions  D r o p - i n sessions are required to be h e l d at least once every two weeks. Bi-weekly o r w e e k l y sessions are the standard. Each session a client selected topic o n pregnancy o r infant care is p r e s e n t e d . G r o u p d i s c u s s i o n f o l l o w s the presentation. L o w cost nutritional snacks a n d recipes are p r o v i d e d at each d r o p i n session. T h e sessions p r o v i d e clients w i t h the o p p o r t u n i t y to gain k n o w l e d g e a n d take control of their lives.  Food Supplements  F o o d supplements are p r o v i d e d to clients based o n financial need a n d n u t r i t i o n a l assessment. M i l k , juice, eggs a n d cheese are the most c o m m o n l y offered supplements. F o o d vouchers m a y serve as a "hook" for clients w h o might not otherwise become i n v o l v e d w i t h the program.  Individual Counselling  N u r s i n g , nutrition and lay counsellors are available for each client. Clients are encouraged to see a counsellor as often as needed. Based o n experience of h i g h - r i s k c o u n s e l l i n g programs, a m i n i m u m of 5 counselling sessions are used to define adequate program participation. Sessions take place at the program site, client's home or other site.  Referral  Referrals are made to both c o m m u n i t y a n d g o v e r n m e n t agencies d u r i n g pregnancy a n d after birth, d e p e n d i n g o n i n d i v i d u a l client needs. Y o u n g single clients are c o m m o n l y referred to "Nobody's Perfect" parenting programs, and the H e a l t h U n i t s contact each client f o l l o w i n g d e l i v e r y , for post-natal assessment and followup.  At  three points d u r i n g the program the client's progress is monitored: at  program intake, two months after intake, and the last visit prior to due datel6. D u r i n g these assessments the client and counsellor discuss and evaluate the client's goal achievements and alter the care plan as required.  Post delivery,  clients often visit and attend drop i n sessions, at w h i c h time infant outcomes The last prenatal visit generally occurs between 36 and 38 weeks of gestation.  are obtained. personnel.  In some sites, infant outcomes are also obtained from health unit Client focussed  quantitative evaluations are conducted yearly.  These formative evaluations document changes i n client behavior related to the six program objectives. c o m p i l e d into  Information from all program sites is  annual evaluation  reports  (Pregnancy  then  Outreach Projects:  Quantitative Evaluation Report, 1990 & 1991). Gestational age and birth weight for P O P client infants are presented and compared to provincial statistics.  Assessment Tools  Pregnancy Outreach Projects utilize two tools to determine eligibility for the programs and client specific risks.  These are the Individual Prenatal Risk  Identification Tool and the T - A C E questionnaire.  Individual  Prenatal Risk Identification Tool  In 1988 the Burnaby Health Department undertook the task of developing the Individual Prenatal Risk Identification Tool (IPRIT).  The goal to produce an  assessment tool that would assist community health service providers in early identification and care of high risk pregnant women.  A t completion of the  process, the B.C. Ministry of Health piloted the tool i n the newly formed P O P . Since A p r i l 1989, the IPRIT has been utilized by all programs.  The IPRIT is both a screening and assessment tool (Appendix A ) .  It is a  multidimensional tool that assesses physical, socio-economic, substance abuse and emotional risk factors.  The tool provides simple decision rules for  inclusion of a risk factor i n the pregnant woman's profile, and further quantifies the risk factors as being either major or minor in nature.  The IPRIT not only determines if the client qualifies for intervention by the program, but also provides the direction for subsequent counselling.  Once a  client is enrolled in the program, lifestyle interventions are determined by the assessed risk factor profile.  The  reliability, validity, sensitivity, specificity and predictive values of the  IPRIT have not been determined.  Therefore, it is difficult to know  how  accurate this tool is in selecting only those women who are at increased risk of adverse pregnancy outcomes.  T-ACE Questionnaire  The  T-ACE  questionnaire is an alcohol screening questionnaire that  was  developed in 1987 by Dr. Robert Sokol, Chairman, Department of ObstetricsGynecology at Wayne State University School of Medicine (Sokol, et al., 1989). The purpose of the instrument is to develop a brief questionnaire that provides physicians with a simple, quick tool to assist them in identifying risk drinkers in their practice. The T - A C E questionnaire is a screening tool composed of four simple questions  (Appendix B).  One of the questions  addresses  alcohol  tolerance, while the other three focus on drinking behavior and perceptions.  The T - A C E questionnaire is administered to all women who admitted to ever having had alcohol. The questionnaire determines a woman's risk score based on preconception drinking practices. For each of the questions i n the T - A C E , a  score is assigned.  The T - A C E score has a range of 0-5.  The first question is  assigned a m a x i m u m score of 2 and the remaining three questions have maximum scores of 1.  A t the end of the questionnaire the client's score is  totaled and a score 2 or more is considered to be indicative of risk drinking. A n y woman with a score of 2 or more is eligible for the program. For a T - A C E score of 2 or more the sensitivity^ of the tool is 69%, positive predictive v a l u e l  9  specificity 18  89% and  23% (Sokol, et al., 1989).  Vancouver Island Pregnancy Outreach Projects  There are three programs Alberni and Nanaimo.  o n Vancouver Island, located i n D u n c a n , Port  These were among the eight pilot sites for the P O P and  all commenced operation during the fall of 1988.  In Duncan, the P O P serves  Cowichan Valley, Kuper Island, Chemanius, Shawnigan Lake and Cobble H i l l . It is sponsored by the Cowichan Valley Native Friendship Centre.  The Port  A l b e r n i P O P serves the town of Port A l b e r n i and surrounding areas, it is sponsored by the A l b e r n i Health Outreach for Parents and Infants.  The  Nanaimo P O P serves Nanaimo, Nanoose and the reserves in the area.  It is  sponsored by the Tillicum Haus Friendship Centre.  U Sensitivity refers to the ability of a test to correctly identify those w h o have the disease or health problem i n question. W h e n sensitivity is high, the number of false negatives (those w h o have the disease but have a negative test result) is l o w . Specificity refers to the ability of a test to correctly identify those w h o d o not have the disease or health problem i n question. W h e n specificity is high, the number of false positives (those w h o do not have the disease but have a positive test result) is l o w . 19 Positive predictive value refers to the probability that the i n d i v i d u a l has the disease or health problem given a positive test result. 1 8  Although the P O P do not select or target any high risk populations, aboriginal and adolescent w o m e n are frequently referred to the program..  A l l three  Vancouver Island P O P are affiliated with aboriginal communities and service a large number of aboriginal women.  M a n y social service and community  agencies also refer adolescent women for P O P services.  A l l three sites provides serves for a similar high risk pregnant population. Previous  formative  evaluations  (Pregnancy  Outreach  Projects,  Reports,1991) have demonstrated that P O P clients are significantly  Site  different  from the general population of women giving birth i n Central Vancouver Island.  P O P clients are predominantly young, single, poorly educated and on  social assistance.  About half of the clients have at least one child, and more  than half are aboriginal. The most common major risk factors are: inadequate nutrition, smoking and inadequate pre-pregnancy weight.  The most common  minor risk factors are: financial problems, inadequate housing, low self-esteem, unstable relationship, family history of abuse or neglect, limited learning ability and isolation.  88  REFERENCES  Pregnancy Outreach Projects: Program Handbook (1993). Victoria, B.C: Community & Family Health, Nutrition Branch, B . C . Ministry of Health. Pregnancy Outreach Projects: Quantitative Evaluation Report (1990). Victoria, B.C: Research & Evaluation B r a n c h , , B.C. Ministry of Health. Pregnancy Outreach Projects: Quantitative Evaluation Report (1991). Victoria, B.C: Research & Evaluation B r a n c h , , B . C . Ministry of Health. Pregnancy Outreach Project: Duncan Site Report (1991). Victoria, B.C: Research & Evaluation Branch, , B.C. Ministry of Health. Pregnancy Outreach Project: Nanaimo Site Report (1991). Victoria, B.C: Research & Evaluation B r a n c h , , B.C. Ministry of Health. Pregnancy Outreach Project: Port Alberni Site Report (1991). Victoria, B.C: Research & Evaluation B r a n c h , , B.C. Ministry of Health. Institute of Medicine (1985). Preventing L o w Birth Weight. National Academy Press.  Washington, D C :  Sokol, R.J., Martier, S. & Ager, J. (1989). The T - A C E questions: Practical prenatal detection of risk-drinking. American Journal of Obstetrics and Gynecology. 160(4). 863-870.  89  CHAPTER 4 RATIONALE A N D METHODS  RATIONALE  The  Pregnancy Outreach Projects have undergone three overall provincial  evaluations since their inception. areas:  implementation,  These evaluations were concerned with four  effectiveness,  acceptance  and satisfaction.  The  quantitative reports focused on implementation issues as well as impact of intervention  o n the  clients  (Pregnancy  Evaluation Report, 1990 & 1991). acceptance  and  satisfaction  Outreach Projects:  Quantitative  The qualitative reports focussed o n client  (Pregnancy  O u t r e a c h Projects:  Qualitative  Evaluation Report, 1990 & 1991). The P O P evaluations to date have been very favorable.  Each year the programs were enrolling more high risk pregnant  women, many who had never participated in any prenatal care.  Clients were  entering the programs earlier i n their gestation, and were increasingly being integrated into existing prenatal and postnatal community services.  The clients  were consistently modifying their behavior i n relation to the six objectives, and satisfaction with the programs among clients was great.  C o m m u n i t y and  sponsoring agency support for the programs has remained very high.  A l t h o u g h both quantitative and qualitative evaluations had shown that the programs were very sucessful  i n meeting their stated objectives, program  managers were interested i n measuring the impact of maternal behavior change on maternal and infant outcomes.  In addition, interest was expressed in  determining the subgroups that received the most benefit from the programs. Although these questions were beyond the mandate of the programs, they were  of interest to managers and staff.  In order to document these results, a case-  control study, comparing P O P clients to a high risk comparison group was required.  Questions  The primary questions of interest i n this study were:  1. What were the characteristics of the women who attended a P O P during 1990-1991? 2. Were there any statistically significant differences in maternal outcomes between P O P clients and controls? 3. Were there any statistically significant differences i n infant outcomes between infants of P O P clients and infants of controls?  The secondary questions of interest in this study were:  4. Were there any differences in maternal and infant outcomes between women who entered the P O P prior to m i d pregnancy and their matched controls? 5. Were there any differences in maternal and infant outcomes between women who enter the P O P after m i d pregnancy and their matched controls? 6. Were there any subgroups of women at risk, for who measurable program effects were shown?  METHODS Study Design  The questions were addressed by means of a quasi-experimental matched casecontrol analysis, This design was chosen in order to produce two study groups that were similar with respect to certain sociodemographic variables that were associated  w i t h adverse  pregnancy outcomes.  Through  matching, the  confounding effects of the sociodemographic variables was controlled, allowing the observed impact of the programs to be determined.  Maternal and infant  outcomes for P O P clients and controls were then compared.  Sample Size and Power  A t the design stage, it was anticipated that approximately 125 P O P clients would be included in the case group. A s newborn outcomes were of primary interest in this study, a dichotomous definition of low birth weight was used to determine the power of this study. With a sample size of 125 women per group (1:1 match), this study w o u l d have 56% power (two-tailed test) to detect a difference in the proportion of low birth weight from 15% (high risk rate) to 5% (population rate for B.C.).  In order to achieve 90% power this study w o u l d  require a sample size of 226 women per group (Cohen,1988).  G i v e n the fixed number of cases, the only means available to increase the power of the study was to increase the number of controls. By matching each case to three controls, the sample size of 125 cases and 425 controls provided an  85% chance of detecting the specified difference in the rate of low birth weight (Schlesselman, 1982). Data from this study were analyzed using a bi-directional (two-tailed) test of significance rather than a uni-directional test for two reasons.  First, it was  unclear if matching P O P clients to controls on sociodemographic  variables  p r o v i d e d a comparison group that was sufficiently similar to P O P clients. Although the use of sociodemograhic characteristics to obtain a control group is common among case-control studies, it has not been determined if this was the best method of obtaining a comparable comparison group.  Secondly,  it  appeared likely that while the programs could positively impact P O P infants, the extremely high risk status of clients could negate program impacts. risk status and p r o g r a m influence  Thus,  could exert their effects i n opposite  directions, hense no program effects could possibly be seen. G i v e n that the risk status of P O P clients may be higher that controls, and that P O P impacts on infant birth weight and gestational  age may have been diminished by the  extremely high risk status of P O P clients, bi-directional tests were chosen for the analyses.  Data Sources  Maternal, birth, and birth outcome data were obtained from the University of British  C o l u m b i a ( U B C ) Perinatal  retrospectively charts.  Study.  The  U B C Perinatal  Study  abstracted data from maternal and infant hospital medical  Hospital chart abstraction included, but was not limited to; prenatal  record, physician orders, physician record of progress (intrapartum), nursing admission record, nursing progress notes, social work progress notes, laboratory  and radiology findings, infant birth record, infant progress notes, and discharge information.  Study Subjects  There are two groups of subjects i n this study, P O P clients (cases) and the comparison group (controls).  Pregnancy Outreach Project Clients  A l l w o m e n who participated i n a Vancouver Island P O P and gave birth between July 23, 1990 and July 21, 1991 constituted the client group for this study.  P O P clients who obtained at least one counseling visit and had known  birth outcomes were included. In addition to the stated criteria, infants must have been singleton births of known birth weight, gestational age and gender.  Comparison  group  A high risk comparison group was required for this study. drawn from the the U B C Perinatal Study.  This group was  The U B C Perinatal Study was a  population based study that ocurred between July 23, 1990 and July 21, 1991 within the Central and North Vancouver Island health regions.  The objective  of the U B C Perinatal Study was to evaluate the introduction of a program for identifing pregnant women at risk for increased alcohol consumption, and to determine their incidence at risk drinking and the association w i t h infant outcomes (Armstrong, et al.,1994).  Detailed pregnancy, delivery and newborn  information were collected o n 3659 women who gave birth during this one year  period.  A l l women who resided in the Central Vancouver Island region, the  same region as P O P clients, formed the pool from which controls for this study were drawn (n=2345).  The same selection criteria were applied i n order to  determine eligibility into the control pool.  Identification of P O P Clients  A l l P O P clients i n this study gave birth during the time period of the U B C Perinatal Study, and were captured among that study's results. Although it was known that P O P clients were part of the larger U B C Perinatal Study, there were no easy means of identifying who the P O P clients were.  The U B C Perinatal  Study d i d not collect information on additional sources of prenatal care and no unique identifier was common to both P O P clients and U B C Perinatal Study participants.  Therefore, a probalistic linkage method was needed to identify  P O P clients within the U B C Perinatal Study.  Computerized record linkage was used to identify P O P clients within the U B C Perinatal Study data base.  Five maternal and infant variables were utilized:  maternal birth date; infant birth date; gestational age; infant weight; and infant sex.  T o ensure a high probability of success, the information from both the P O P  data base and the U B C Perinatal Study data base had to agree on all five variables.  Maternal birth date, infant birth date and infant sex were absolute  measurements. measurements.  Infant  b i r t h weight  and  gestational  age  were  softer  Birth weight within 250 grams and gestational age within two  weeks were deemed to be acceptable limits for the data link. In situations were there was more than one possible match, birthing hospital was used to narrow the possible link.  95  Independent Variable  The independent variable i n this study was participation in a P O P at any of the three Vancouver Island sites: Duncan, Nanaimo or Port Alberni during the time period of July 23,1990 to July 21,1991.  Matching Variables  Maternal age, maternal race, parity and family income were the matching variables used in this study. order  to  create  sociodemographic  a  The four demographic variables were chosen i n  comparison characteristics  group of  that  would  P O P clients.  closely Many  mirror studies  the have  investigated the association of these sociodemographic variables w i t h birth outcomes.  A l t h o u g h cigarette consumption is also strongly associated with  adverse birth outcomes, this variable was not utilized in this study.  Smoking  information was poorly documented, and it was impossible to ascertain for individual clients, if the documented cigarette consumption was the rate prepregnancy, mid-pregnancy or at delivery. Table 4.1 shows the categorization of the sociodemographic matching variables  Table 4.1 Maternal Matching Variables  Age Race Parity Family Income  <19 years Aboriginal zero low  20 - 34 years Non-Aboriginal one to four adequate  > 35 years > five  U s i n g a simultaneous distribution matching strategy, each P O P client was matched to three controls based on the four sociodemographic variables.  The  result of this simultaneous matching procedure was the creation of the two study groups: P O P clients and comparison group.  Age at Delivery: In this study clients were matched i n three age categories: 14-19 years, 20-34 years, and 35 or more years. Within the youngest age group (13-19 years) the values were dichotomized into 14-16 years and 17-19 years.  Many  studies supported the hypothesis that young maternal age itself was a risk factor for both preterm birth and low birth weight infants (Blondel et al., 1987; Kramer, 1987; McCormick et al., 1984; Institute of Medicine, 1985) Furthermore, adolescents with low gynecological age (conception within 2 years of menarche) were at greater risk than older adolescents for preterm birth and low birth weight (Kitzes, 1986; Scholl et al., 1989). Advanced maternal age (over 34 years) has also been shown to adversely affect infant outcome, increasing mortality, low birth weight and preterm birth (Cnattingius et al., 1992; Friede et al., 1988).  Maternal birth date was obtained from either the prenatal record or the hospital admission record. Maternal age at delivery was calculated by subtracting infant date of birth from maternal date of birth.  Race: In this study, clients  were matched according to aboriginal status.  Aboriginal status was important for reasons related to birth weight, gestation and birth defects.  Aboriginal status is associated with heavier birth weights at  each week of gestation and with a higher incidence of preterm birth (Buck et al., 1992; Kierans et al., 1993; Thomson, 1990).  Also, the incidences of alcohol  related birth defects and infant mortality are higher among aboriginal infants in  97  British C o l u m b i a  (Bray et al., 1989; Fetal A l c o h o l Syndrome i n British  Columbia, 1983; Kierans et al, 1993; Robinson et al., 1987).  Aboriginal status was determined through either documentation on the study participants prenatal record, or through personal health number (PHN).  A  P H N ending with an R2 was used to identify people.of aboriginal descent. This includes status, non-status and metis people, both on and off reserves (British Columbia Ministry of Health, 1988).  Parity:  M a n y studies have shown that parity was risk factor for adverse  pregnancy income independent  of maternal age.  Infants  b o r n to both  primiparous and grand multiparous (>5) women were at higher risk for low birth weight (Kramer, 1987; Preventing L o w Birthweight, 1985).  Grand  multiparous w o m e n also have a higher risk for perinatal deaths due  to  placental complications (Brunner et al., 1992).  Parity was determined through either documentation on the prenatal record or hospital admission record, if no prenatal record was available.  Family Income: Socioeconomic status is commonly measured by variables such as occupation, education, income, or a composite generated from the weighted sum of a number of variables.  In this study, family income was chosen to  represent socioeconomic status. The choice of family income was supported by studies that suggest that income, rather than education or occupation correlates best  with  the  socioeconomic  differences  in  adverse  pregnancy  outcomes.(Binsacca et al., 1987; Egbuonu et al., 1982; Starfield et al., 1982; Stein et al., 1987; Wigle et a l , 1980).  98  Measures of family income are related to the definition of a family.  According  to the 1986 Census Dictionary (Statistics Canada, 1987), a household refers to a person or group of persons who occupy a private dwelling and do not have a usual place or residence elsewhere in Canada.  A census family includes a  husband or wife with children who have never married, or a lone parent with one or more children who have never married, living in the same dwelling. Groups of two or more persons who live i n the same dwelling and are related to each other by blood, marriage, common-law relationships, or adoption are referred to as an economic family.  The choice of either census family or  economic family would have been appropriate given that both are more likely than household to be based on the notion of economic dependency.  Economic  family was the definition selected because it was utilized in previous Canadian studies (Thomson, 1990; Wilkins et al., 1989; Wilkins et al., 1991).  Size of economic family unit was calculated by adding the number of adults i n the family to the mother's parity and current pregnancy, to determine the total number of individuals within the family unit. Family income was determined by first coding maternal and paternal occupations, and then determining income by occupation for 1985, the latest year available (Standard Occupational Classification,1981; Population and D w e l l i n g Characteristics: E m p l o y m e n t Income by Occupation, 1989).  For families where one or both adult members  were unemployed, social assistance rates were utilized.  The family's socioeconomic status was deemed adequate if family income was above the Statistics Canada low income cut-off, and inadequate if it was below the Statistics Canada low income cut-off (Income Distributions by Size i n  Canada: L o w Income Cut-offs, 1991). The Statistics Canada income calculations were based o n individual family size and population size of 30,000 - 99,999 residents.  The latter is intended to adjust for differences in the cost of living  that are a consequence of the size of the city.  A n attempt was made to match each control to P O P client on all four variables. W h e n this was not possible, controls were matched o n a m i n i m u m of three variables. Maternal age at delivery was the best matched variable and with each subset (1 case: 3 controls) matched according to age categories. was inconsistently recorded i n the U B C Perinatal Study.  Maternal race  For 1022 (45.8%)  w o m e n i n the U B C Perinatal Study, information o n maternal race was not documented. Information from B.C. Division of Vital Statistics^ i n addition to U B C Perinatal Study home interviews 21 indicated that women of u n k n o w n racial origin were likely to be non-aboriginal.  Data Analysis  A l t h o u g h the study design employed matching, the necessity of maintaining matching i n the analysis is a statistical grey area.  Several authors strongly  support maintaining matching i n the analysis, because if matching is dropped the results may be biased towards the n u l l hypothesis Schlesselman, 1982).  (Feinstein, 1985;  In order to maximize the strength of the analysis, given  the small sample size, matching was maintained in the analysis.  Jennifer Gait, V i t a l Statistics, B . C . M i n i s t r y of Health, personal communication. U B C Perinatal Study home interviews assessed 82.1% of w o m e n w i t h u n k n o w n racial origin to be caucaian. 2 0  2 1  The analyses were performed for the total study sample, then two subanalyses were performed. The first subanalysis determined if there were any differences in maternal and infant outcomes based o n intensity of service, the second subanalysis determined if there were any subgroups for who measurable program effects were shown.  A l t h o u g h previous studies have shown that the impact of intensity of service on infant outcomes was mixed, intensity of program service was measured in this study.  Intensity of service is commonly measured either by number of  program contacts or length of time in the program.  Although the P O P have  defined intensity of service by number of visits, this measure was not used in the study for two reasons. First, the P O P have determined that five counselling visits constitutes "program success", this figure is based o n information from the Montreal Diet D i s p e n s a r y ^ , whose program is quite different from the P O p 2 3 therefore not necessarily applicable.  Secondly, the three P O P sites  differed i n their definition of what constituted a counseling visit.  For this study intensity of program service was measured by length of contact with the program. completed  W o m e n who entered a P O P prior to midpregnancy (20  weeks) formed one  subgroup, and those w h o  entered  past  midpregnancy (21-28 weeks)24 formed the other subgroup. This analysis was done because it p r o v i d e d a means for separately evaluating the impact of differences in program contact o n maternal behavior modification and infant outcomes.  Several evaluations of comprehensive prenatal care programs have  22 Lisa Forster-Coull, N u t r i t i o n Branch, B . C . M i n i s t r y of Health, personal communication. 23 Sheila Dubois, acting Executive Director,Montreal Diet Dispensory, personal communication. 24 W o m e n over 28 weeks of gestation are not eligible for the projects.  shown that w o m e n w h o enter prior to midpregnancy h a d better infant outcomes (Leveno, et al. 1985;. Scholl, et al. 1987; Alexander, et al. 1987).  Maternal Outcomes of Interest  Initiation of prenatal care, number of prenatal visits, adequacy of prenatal care and maternal morbidity were the maternal outcomes of interest i n this study. For the continuous variables, initiation of prenatal care and number of prenatal visits, values were pooled across each control subset.  Then matched 1:3 T-tests  were performed between the case (POP client) and pooled control values within each subset to determine if there were any differences between the two study groups (Miettinen, 1969).  The Cochran-Mantel-Haenszel ( C M H ) procedure for obtaining a point estimate of the odds ratio and a C M H chi-square test of significance were utilized to analyze the dichotomous variables.  W i t h this procedure, outcome measures  for the controls within each subset were also pooled.  Each case and its  corresponding set of matched controls were regarded as a separate subset within a 2 x 2 table. C M H odds ratios were used to measure differences i n adequacy of prenatal care and maternal morbidity. The C M H chi-square test and confidence intervals were then calculated to determine whether the odds ratios were significant.  Initiation of Prenatal Care:  The time interval between last menstural period  and initial contact w i t h physician i n weeks, initiation of prenatal care.  was used to determine the  Number of Prenatal Visits: Total number of prenatal visits was obtained from the prenatal record. The hospital portion of the prenatal record may not show the total number of prenatal visits, as these records were often forwarded to the admitting hospital at 37 completed weeks of gestation. Therefore the actual number of prenatal visits may have been higher than recorded.  Adequacy of Prenatal Care:  The time interval between the expectant mother's  initial contact with her physician and her infant's birth was used to determine adequacy of prenatal care. A dichotomous variable for prenatal care was created using Peoples et al, (1984) adaptation of Kessner's Adequacy of Care Index Levels (Appendix C ) .  The Kesnner Index is an algorithm that  includes  trimester of pregnancy prenatal care began, number of prenatal visits and length of gestation, compared to the expected n o r m for visits.  U s i n g the  adapted method, prenatal care was determined to be either adequate or inadequate.  Maternal Morbidity:  Maternal morbidity refers to diseases or conditions that  developed during pregnancy, as well as obstetrical complications resulting from labour and delivery.  Variables utilized to determine maternal morbidity were  obtained from both the prenatal record and the physician's record of progress (intrapartum). placenta  Maternal morbidity included any one of the following: anemia,  previa, placenta  abruptio, p o l y h y d r a m n i o s ,  oligohydramnios,  pregnancy induced hypertension, gestational diabetes, preeclampsia, toxemia, threatened premature labour and post partum hemorrhage.  103  Infant Outcomes of Interest  M e a n birth weigh, gestational age, head circumference and length as well as rates of preterm birth, low birth weight, small for gestational age, large for gestational age, infant morbidity and congenital anomalies were the infant outcomes of interest i n this study.  Once again, the continuous variables were  analyzed using matched 1:3 T-tests.  The Cochran-Mantel-Haenszel ( C M H )  procedure was again utilized to analyze dichotomous variables.  O d d s ratios  were used to measure differences in the rates of low birth weight, preterm birth, small for gestational age, large for gestational age, congenital anomalies, as well as perinatal morbidity for study groups. The C M H chi-square test and confidence intervals were then calculated to determine whether the odds ratios were significant.  Gestational Age:  Gestational age was abstracted from infant birth record  according to: last menstrual period date (date of delivery minus patient reported last menstrual period).or ultrasound date (date of delivery minus date of ultrasound plus number of weeks gestation at ultrasound). Priority was placed on determining gestational age by last menstrual period date.  Birth Weight:  Birth weight was abstracted from two different sources: infant  birth record or physicians record of progress (intrapartum report). Priority was placed on obtaining birth weight from the infant birth record.  Head Circumference: record.  H e a d circumference was abstracted from infant birth  104  Length:  Infant length was abstracted from infant birth record.  Preterm Delivery:  Infants born prior to 37 completed weeks (less than 259 days)  of gestation were determined to be preterm (British C o l u m b i a Ministry of Health, 1990).  Low Birth  Weight:  Infants with birth weights less than 2500 grams were  categorized as low birth weight (British Columbia Ministry of Health, 1990).  Small for  Gestational  Age:  Small  for  gestational  age  (also k n o w n  as  intrauterine growth retardation) was determined using Canadian comparisons for singleton births (Arbuckle et al.,1989 &1993).  Infants with birth weights  below the 10th percentile for gestational age were considered to be small for gestational age.  Large for Gestational Age:  Large for gestational age was determined using  Canadian comparisons for singleton births (Arbuckle et al., 1989 &1993). Infants with birth weights above the 90th percentile for gestational age were considered to be large for gestational age.  Perinatal Conditions:  Perinatal conditions refered to diseases or conditions that  developed during pregnancy or were preexisting and aggravated by labour and delivery. trauma,  They included one or more of the following: birth infections, birth seizures,  respiratory distress,  hemorrhage, narcotic  abstinence  syndrome, fetal alcohol syndrome, and neonatal intensive care admission.  Information was obtained from either infant birth record or infant progress notes.  Congenital  Anomalies:  Congenital anomalies  describe  any important  structural defects (both internal and external) present i n an infant at birth that were not caused by birth injury. Congenital anomalies were determined using International Classification of Diseases  codes 740-759 (ICD 9 C M , 1989).  Information was obtained from either infant birth record or infant progress notes.  Data for this study were analyzed using the statistical software package, SPSS-PC version 4.0.  106  REFERENCES  Alexander, G.R. & Comely, D . A . (1987). Prenatal Care Utilization: Its Measurement and Relationship to Pregnancy Outcome. A m e r i c a n Tournal of Preventive Medicine. 3(5). 243-253. Arbuckle, T.E. & Sherman, G.J. (1989). A n analysis of birth weight by gestational age i n Canada. Canadian Medical Association Journal. 140. 157-165. Arbuckle, T . E . , Wilkins, R. & Sherman, G.J. (1993). Birth Weight Percentiles by Gestational Age i n Canada. Obstetrics and Gynecology. 81(1). 39-48. Armstrong, R , Loock, C . & Robinson, G . (1994). University of British Columbia Perinatal Study. (Unpublished data). Binsacca, D.B., Ellis, J., Martin, D . G . & Petitti, D.B. (1987). Factors Associated with L o w Birthweight i n an Inner-City Population: The role of Financial Problems. American Journal of Public Health. 77(4), 505-506. Blondel, B., Kaminski, M . Saurel-Cubizolles, M.J. & Breart, G . (1987). Pregnancy Outcome and Social Conditions of Women under 20: Evolution i n France from 1972 to 1981. International Journal of Epidemiology. 1£, 425430. British Columbia Ministry of Health (1988). Medical Services Plan References (Section 5-13). Victoria, B. C : Queen's Printers. British Columbia Ministry of Health (1990). Vital Statistics A n n u a l Report. Victoria, B. C : Queen's Printers. Bray, D . L . & Anderson, P.D. (1989). Appraisal of the Epidemiology of the Epidemic of Fetal Alcohol Syndrome A m o n g Canadian Native Peoples. Canadian Tournal of Public Health. 8J2, 42-45. Brunner, J., Melander, E . , Krook-Brandt, M . & Thomassen, P . A . (1992). Grand multiparity as an obstetric risk factor; a prospective case-control study. European Tournal of Obstetrics & Gynecology and Reproductive Biology. 47,201-205. Buck, G . M . , Mahoney, M . C . , Michalek, A . M . , Powell, E.J. & Shelton, J.A. (1992). Comparison of Native American births i n upstate N e w York with other race births, 1980-86. Public Health Reports. 107C5L 569-575.  Cnattingius, S., Forman, M.R., Berendes, H . W . & Isotalo, L . (1992). Delayed childbearing and risk of adverse pregnancy outcome. J A M A . 268(7). 886890. Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. Hillsdale, N e w Jersey: Lawrence Erlbaum Associates. Egbuonu, L.,& Starfield, B. (1982). C h i l d health and social status. Pediatrics. 69. 550-557. Fetal Alcohol Syndrome i n British Columbia (1983). The Health Surveillance Registry of British Columbia, (unpublished data). Feinstein, A . R . (1985). Clinical Epidemiology: The Architecture of Clinical Research. Toronto, Ontario: W . B . Saunders Company. Friede, A . , Baldwin, W . , Rhodes, P . H . , Buehler, J.W. & Strauss, L . T . (1988). Older maternal age and infant mortality i n the United States. Journal of Obstetrics and Gynecology. 72(2). 152-157. I C D 9 C M (1989). A n n Arbor, Michigan: Commission o n Professional and Hospital Activities. Income Distributions by Size in Canada: L o w Income Cut-offs (1991). Ottawa, Ontario: Statistics Canada (Catalogue N o . 13-207). Kierans, W.J., Collinson, M . A . , Foster, L . T . & U h , S-H. (1993). Charting Birth Outcome i n British Columbia: Determinants of Optimal Health and Ultimate Risk. Victoria, B.C.: Division of Vital Statistics, B.C. Ministry of Health. Kitzes, J. (1986). Having a baby....Under 16. Emergency Medicine. 18(3). 28-44. Kramer, M.S. (1987). Intrauterine growth and gestational duration determinants. Pediatrics. 80. 502-511. Leveno, K.J., Cunningham, F . G . , Roark, M . L . , Nelson, S.D. & Williams, M . L . (1985). Prenatal Care and the L o w Birth Weight Infant. Obstetrics & Gynecology. 66(5). 599-605. McCormick, M . C.,Shapiro, S. & Starfield, B. (1984). H i g h risk young mothers: Infant mortality and morbidity i n four areas i n the United States. American Tournal of Public Health. 74(1), 8-23. Miettinen, O.S. (1969). Individual matching with multiple controls i n the case of continuous responses. Biometrics. 25. 339-355.  108  Miettinen, O.S. (1969). Individual matching with multiple controls i n the case of all or nothing responses. Biometrics. 25. 321-338. Peoples, M . D . , Grimson, R.C. & Daughtry, G . L . (1984). Evaluation of the Effects of the North Carolina Improved Pregnancy Outcome Project: Implications for State-Level Decision-Making. A J P H . 74(6). 549-554. Population and Dwelling Characteristics: Employment Income by Occupation (1989). Ottawa, Ontario: Statistics Canada. Pregnancy Outreach Projects: Quantitative Evaluation Report (1990). Victoria, B.C: Research & Evaluation B r a n c h , , B . C . Ministry of Health. Pregnancy Outreach Projects: Quantitative Evaluation Report (1991). Victoria, B.C: Research & Evaluation B r a n c h , , B.C. Ministry of Health. Pregnancy Outreach Projects: Qualitative Evaluation Report (1990). Victoria, B.C: Research & Evaluation B r a n c h , , B.C. Ministry of Health. Pregnancy Outreach Projects: Qualitative Evaluation Report (1991). Victoria, B.C: Research & Evaluation Branch, , B.C. Ministry of Health. Institute of Medicine (1985). Preventing L o w Birthweight. National Academy Press.  Washington, D C :  Robinson, G . C . , Conry, J.L. & Conry, R.F. (1987). Profile and Prevalence of Fetal Alcohol Syndrome i n an Isolated Community i n British Columbia. Canadian Medical Association Journal. 137. 203-207. Schlesselman, J.J. (1982). Case-Control Studies: Design. Conduct Analysis. N e w York: Oxford University Press. Scholl, T . O . , Hediger, M . L . , Salmon, R.W., Belsky, D . H . & Ances, L G . (1989). Association between low gynecological age and preterm birth. Paediatric and Perinatal Epidemiology. 3_, 357-366. Scholl, T . O . , Miller, L . K . , Salmon, R.W., Cofsky, M . C . & Shearer, F . (1987). Prenatal Care Adequacy and the Outcome of Adolescent Pregnancy: Effect on Weight Gain, Preterm Delivery, and Birth Weight. Obstetrics & Gynecology. 69(3). 312-316. S P S S / P C + Statistics 4.0 (1990). Chicago, Illinois: SPSS Inc. S P S S / P C + 4.0 Base Manual (1990). Chicago, Illinois: SPSS Inc.  109 Standard Occupational Classification 1980 (February, 1981). Ottawa, Ontario: Statistics Canada, Minister of Supply and Services Canada. Starfield, B . H . , Shapiro, S., McCormick, M . C . , & Bross, D. (1982). Mortality and morbidity i n infants with intrauterine growth retardation. Journal of Pediatrics. 101.978-983. Statistics Canada. Census dictionary (1987). Catalogue no. 99-101E. Ottawa, Ontario: Minister of Supply and Services Canada. Stein, A . , Campbell, E.A-, Day, A . , McPherson, K . & Cooper, P.J. (1987). Social adversity, low birth weight, and preterm delivery. British Medical Tournal. 295. 291-293. Thomson, M . (1990). Heavy birthweight i n Native Indians of British Columbia. Canadian Tournal of Public Health. 81,443-446. Thomson, M . (1990). Association between mortality and poverty. Tournal. 32(8), 337-338.  B . C . Medical  Wigle, D . & M a o , Y. (1980). Mortality by income level i n urban Canada. Ottawa, Ontario: Minister of National Health and Welfare. Wilkins, R., Adams, O . & Brancker, A . (1989). Changes i n mortality by income in urban Canada from 1971 to 1986. Health Reports. 1(2), 137-174. Wilkins, R , Sherman, G.J. & Best, P. (1991). Birth outcomes and infant mortality by income i n urban Canada, 1986. Health Reports. 3(1), 7-31.  110 CHAPTER 5 RESULTS  POP D A T A BASE D E V E L O P M E N T Pregnancy Outreach Projects Intake  There were 212 women with due dates between July 1, 1990 and July 31, 1991 referred to the three Vancouver Island P O P . A s shown i n Table 5.1, source of referral was k n o w n for 147 (69.3%) of these women.  O f the k n o w n referrals,  almost half (44.9%) of the women referred themselves to a POP, having become aware of the program from friends, family, other clients, community groups, or through various media sources.  M a n y of the remaining referrals (51%) arose  from contact with community based health professionals and agencies. few referrals came from physicians.  Table 5.1 Sources of Referral  Source of Referral Self Health Unit Other H e a l t h Professionals Community Agencies Physician A l c o h o l and D r u g Programs Social Services a n d H o u s i n g Self v i a P h y s i c i a n  N=147  %  66 26 22 17 8 4  44.9 17.7 15.1 11.6 5.4 2.7 2.0 0.7  3 1  Very  Client Retention  Figure 5.1 provides an overview of program retention. referred to the program, 16 were not assessed.  O f the 212 w o m e n  The m a i n reasons for non  assessment were inability to contact and disinterest i n the program. O f the 196 women who were assessed, 54 d i d not met the eligibility criteria.  The most  common reason for exclusion from the program was gestation greater than 28 weeks.  One hundred and thirty six of a possible 142 eligible clients were enrolled in the program, for an overall participation rate of 95.8%.  A review of the available  data for the 6 nonparticipants suggests their risk profiles for adverse pregnancy outcomes d i d not differ in any systematic manner from clients who participated in the program.  O f the 136 clients that started the program, 91 (66.9%) stayed i n the program until delivery.  Nineteen clients m o v e d from Vancouver Island and their  results were lost to this study.  Twenty one clients were lost to followup from  the P O P , however, infant information was available for these women and they were included as study participants. Therefore, a total of 112 P O P women were eligible for this study.  Figure 5.1 Client Retention  Referred to POP N=212  Couldn't Contact N=9  Refused to Be Screened N=4  Not Eligible (>28 weeks) N=l  Moved From Area N=2  High Risk N=186  Low RisV N=10  Not Eligible For Intervention (>28 weeks At Registration) N=44  Abortion N=2  Assessments Completed N=196  Eligible For Intervention N=142  Not ^ Interested N=6  Began Program N=136  From Area N=19  Until Delivery N=91  Followup N=21  N=3  113  Lost to Followup  Table 5.2 provides a comparison of 91 clients w h o stayed i n a P O P until delivery, and the 21 who were lost to followup. were very similar.  Both groups of program clients  Clients who completed the program entered half a week  ahead of those that dropped and on average, had a greater number of major risks and fewer minor risks at program entry. Demographically, clients who did not complete the program were slightly younger, had fewer children and more likely to be in a relationship.  They were also less likely to have completed high  school or be employed, however they were less likely to rely on social assistance as their source of income.  They were equally as likely to be aboriginal as  nonaboriginal.  Table 5.2 P O P Clients-Completed and Lost to Followup  Characteristics of Pregnancy Outreach Project Clients Completed Program (N=91) N % Average Major Risk M i n o r Risks Age Parity Gestation at Intake Single Aboriginal N o t C o m p l e t e d H i g h School Employed Receiving Income Assistance  0.98 4.5 22.8 1.4 17.4 49 63 81 10 92  43.8 56.3 72.3 8.9 82.1  Lost to Followup (N=21) N %  0.87 4.7 21.2 1.1 17.9 9 12 19 0 18  37.5 50.0 79.1 0 75  114  Study Eligible P O P Clients  A s shown in Figure 5.2, data were available for a total of 112 P O P clients who delivered between July 23, 1990 and July 21, 1991. subsequently excluded from the analysis. delivered twins.  Three clients (2.7%) were  T w o women were excluded as they  Information on 16 infants required manual review as they  were either missing gender, birth weight or gestational age information or the recorded values were outside of acceptable parameters. For 1 client, infant birth information was not obtained and the client was subsequently excluded from the analysis. The remaining 15 infants were successfully assigned.  A total of 109 P O P clients were eligible for this study.  Computerized record  linkage was then used to identify P O P clients within the U B C Perinatal Study data base. Eighty six (78.9%) P O P clients were initially identified using the five maternal and infant linking variables. Manual review of P O P charts was then conducted for the remaining 23 (21.1%) clients.  Twenty P O P clients were  successfully identified following the manual review.  The loss of only three  P O P clients (2.8% ) using the data link, compared favorably with previous investigations in w h i c h a similar approach was used (Peoples et al., 1983; Beuscher et al, 1991).  Incomplete linking of the P O P clients has led to the  possibility that some P O P clients may have been counted amongthe comparison group.  T o the extent that P O P d i d improve birth outcomes, this incomplete  linking w o u l d lead to an underestimate of the true differences between the two groups. study.  A total of 106 (97.2%) P O P clients comprised the case group for this  115  Figure 5.2 Study Eligible P O P Clients  POP Population N=112  Missing Birth Information N=l  Participants N=109  Computer Link N=86  M a n u a l Review of Records N=23  Records not Linked N=3  P O P Study Participants N=106  Power Calculation  A t the design stage it was predicted that 125 P O P clients w o u l d be available for this study.  G i v e n 1:3 matching adequate power w o u l d have been achieved to  detect a decrease in the proportion of low birth weight from 15% among to 5% among P O P clients.  W i t h only 106 P O P clients available for this study, the  power to detect this change in low birth weight has decreased to 70%.  Development of Comparison Group  Each of the 106 P O P clients was matched to 3 controls based on the demographic matching variables.  Eighty nine (83.9%) P O P clients were matched to three  controls on all four variables.  For the remaining 17 P O P clients (16.1%)  incomplete matches occurred.  The relatively small number of aboriginal  women^S in the U B C Perinatal Study necessitated incomplete aboriginal race matching. Eleven aboriginal P O P clients were matched with 1 control who was nonaboriginal and 6 were matched with two controls who were nonaboriginal. Twelve P O P clients with between 1 and 4 children were matched with one control who was primiparous. Thirteen low income P O P clients were matched with one control who had adequate income.  Missing Data  Generally  information o n  maternal matching variables  was  less  well  documented than were maternal and infant outcome variables. Information on maternal age and parity were available for each study participant. For 97 (22.9 2 5  A b o r i g i n a l w o m e n accounted for 8.1% of the U B C Perinatal Study population.  %) participants, information of maternal race was not documented o n the prenatal record.  Based o n information from Vital S t a t i s t i c s ^ w o m e n who  lacked race information were assumed to be nonaboriginal. Sixty seven (15.8%) study participants were either missing or had incomplete information on occupation.  Prenatal records were missing for 24 (5.7%) study participants. Hospital chart documentation indicated that these w o m e n d i d not visit physicians during their pregnancy. incomplete.  For an additional 4 (1%) participants, prenatal records were  For 38 (8.9%) participants, information on last menstrual period  was either missing (n=5) or grossly out of range (n=33).  For these women,  gestational age was calculated by using ultrasound date.  For three (0.7%) infants gestational age, birth weight or gender were not documented, these three variables were required for calculating both large and small for gestational age. For ninety six (22.6%) infants, head circumference was not documented, and 124 (29.2%) were missing information on length.  Seven  (1.7%) infants had no documentation regarding congenital anomalies and perinatal conditions.  ANALYSIS O F M A T E R N A L A N D INFANT OUTCOMES  Maternal Characteristics  Descriptive data on maternal characteristics of the study groups is presented in Table 5.3. The mean age at delivery was 22.8 years for P O P clients and 23.2 years for controls. 2 6  Both study groups were proportionately represented among the  Jennifer Gait, V i t a l Statistics, B . C . M i n i s t r y of Health, personal communication.  three age categories, with 33% of the study sample aged 13-19 years, 63.2% aged 20-34 years and 3.8% were 35 years or older.  Compared to the 1990 provincial  data (British Columbia Ministry of Health, 1990), the study groups had a much greater proportion of adolescents (5.7% provincially) and a smaller proportion of those aged 35 or older (10.7% provincially).  The large proportion of  aboriginal women comprising the adolescent age group may explain some of the elevated proportion of births i n this age group.  F r o m a provincial  standpoint, among the aboriginal population 20% of births occurred to women aged 19 years or less (Tuk, 1995).  There were significant differences i n the racial composition of the two study groups, with regard to both the numbers of aboriginal women and women of unknown race.  Whereas 56.6% of P O P clients were aboriginal, they accounted  for only 43.1% of the controls.  This occurred because although a large  proportion of the U B C Perinatal Study aboriginal w o m e n were selected as c o n t r o l s ^ , the overall total proportion of aboriginals i n the U B C Perinatal Study was small.  The proportion of aboriginal women in the study greatly  exceeded provincial numbers.  According to recent studies (Kierans et al., 1993;  T h o m s o n , 1990; T u k , 1995)  aboriginal w o m e n of childbearing age i n the  province accounted for between 2.5% -2.75% of the birthing population.  There were also significant differences between the two study groups with regard to marital status, 51.9% of the P O P clients were single compared with 59.1% of the controls.  Provincially 24.3% of all live births occurred to  unmarried women (British Columbia Ministry of Health, 1990).  137 or a possible 181 (75.7%) of the documented aboriginal population were selected as controls using the matching criteria. 2 7  The vast majority of study participants had low incomes.  A slightly greater  proportion of P O P clients (74.5%) had low family incomes than d i d controls (70.8%).  The proportion for which income information was u n k n o w n was  similar for both groups.  P O P clients and controls had the same average parity at 1.40 births. Primiparas comprised 34.9% of the P O P clients and 39.3% of the controls.  The proportion  of grand multiparous women (5 or more children) was similar between the two groups at 6.6% for P O P clients and 5.3% for controls.  Compared to provincial  data, the proportion of primiparous women was lower (43.6% provincially), while the proportion of grand multiparas were similar (6.5% provincially).  Twenty nine percent of the multiparous P O P clients had experienced  adverse  obstetrical outcomes, mainly preterm births, in previous pregnancies compared with 21.6% of multiparous controls.  The proportion of underweight  and  overweight women was small i n both groups with slightly more controls at pregravid weight extremes. Similar proportions of P O P clients and controls had existing medical conditions, 15.1% and 18.2% respectively.  Both groups were similar with regard to modifiable behaviors although the proportion of smokers was slightly higher among the controls (48.1%) than the P O P clients (44.2%). P O P clients were at greater risk for alcohol and illicit drug use, though not significantly so.  Table 5.3 Study Clients-Maternal Profile  POP Clients (N=106) N %  Comparison Group (N=318) N %  Maternal Age Average * < 19 years 20-34 years > 35 years  22.8 ± 5.6 35 33.0 67 63.2 4 3.8  23.2 ± 5.8 105 33.0 63.2 201 12 3.8  Race Aboriginal Caucasian Unknown  60 29 16  56.6 27.4 15.0  137 95 82  43.lt 29.9 25.8t  Marital Status Single Married  55 51  51.9 48.1  188 130  59.lt 40.9+  Family Income Low income Adequate Unknown  79 21 6  74.5 19.8 5.7  225 73 20  70.8 23.0 6.3  Parity Average* Primipara Multipara Parity > 5  37 62 7  34.9 58.5 6.6  125 176 17  39.3 55.3 5.3  Previous Poor OB Outcome  18  29.0  38  21.6  Physical Underweight(<50kg) Overweight (>80kg) Existing Medical Condition  2 1 16  1.9 .09 15.1  10 4 58  3.1 1.3 18.2  Behavioral Smoker T-ACE > 2 Drug  49 18 5  44.2 17.0 5.0  153 47 9  48.1 14.8 2.8  1.4 ± 1.7  A  * Mean ± Standard Deviation Denominator is Multiparous Women + Statistically Significant (p < 0.05) A  1.4 + 1.6  Primary Analyses  Maternal  Outcomes  Maternal outcomes for the overall analysis are shown in Tables 5.4 and 5.5.  It  can be seen that P O P clients experienced slightly better results i n all measured maternal outcomes when compared with their matched controls.  P O P clients  initiated prenatal care on average 1.2 weeks ahead of controls. The matched T test value of 1.36 (df=96, C.L.=-2.9; 0.54) showed that there was no statistically significant  association between P O P participation and early initiation of  prenatal care.  It is also highly unlikely that a 1.2 week mean difference i n  initiation of prenatal care was clinically significant.  P O P clients obtained on average 7.9 prenatal visits, compared with 7.2 visits for controls. This difference of 0.5 prenatal visits produced a significant T-value of 2.05 (df=96, C.L.=0.21; 1.3) indicating as association between P O P participation and an increased number of prenatal visits obtained.  O n l y a small proportion of both P O P clients and controls obtained adequate prenatal care, 28.3% and 21.1% respectively.  The difference in the proportion  who received adequate care resulted in an odds ratio of 1.11 (df=96, C.L.=0.66; 1.9).  The proportion of women who experienced morbidity was nearly identical in the two study groups with 20.7% of P O P clients and 19.5% of controls having experienced one or more pregnancy/delivery complications.  For three of the four measures of maternal outcomes, initiation of prenatal care, adequacy of prenatal care and maternal morbidity, the overall analysis showed no program effects for P O P clients.  However, P O P clients achieved  slightly more prenatal visits than their matched controls, indicating modest program effects for this maternal outcome.  Table 5.4 Maternal Outcomes-Initiation and Visits  POP Clients (N=106)  Controls (N=318)  Mean Difference  T-Test & 95% C.L.00  Initiation of Prenatal Care (weeks)*  14.4 ± 6.8  15.6 ± 4.9  -1.2  1.36 (-2.9; 0.54)  Number of Prenatal Visits*  7.9 ± 2.8  7.2 ± 1.8  0.64  2.05t (0.21; 1.3)  * Mean ± Standard Deviation oo Confidence Limits t Statistically Significant.(P < 0.05)  Table 5.5 Maternal Outcomes-Adequacy and Morbidity  POP Clients (n=106)  Controls (n=318)  Odds Ratio & 95% C.L.oo  Adequacy of Prenatal Care § (%)  30 (28.3)  67 (21.1)  1.11 (0.66; 1.9)  Maternal Morbidity i(%)  22 (20.7)  62 (19.5)  0.81 (0.48; 1.4)  oo Confidence Limits § As measured by modified Kessner Index (Appendix C) j Maternal Morbidity included any one of the following: anemia, placenta previa, placenta abruptio, polyhydramnios, oligohydramnios, pregnancy induced hypertension, gestational diabetes, preeclampsia, toxemia, threatened premature labour and post par turn hemorrhage.  Infant  Outcomes  Results of infant outcomes are presented i n Tables 5.6 and 5.7.  It can be seen  (Table 5.6) that there were virtually no differences i n measures of growth between the two groups of infants.  P O P infants had slightly shorter gestations  than control infants, with a mean difference of 0.46 weeks. The matched T-test value of 1.67 (df=105, C.L.=-1.0; 0.08) showed lack of a statistically significant association between P O P participation and gestational age. P O P infants were on average 11 grams lighter that control infants. head circumferences centimeters).  (0.33  centimeters)  and slightly  shorter length  (0.55  None of these differences were statistically significant.  Analysis of infant growth and development significant statistical differences controls.  They had slightly smaller mean  indicated that there were no  between infants b o r n to P O P clients and  A closer look at the data showed that there was m u c h greater  variability among all measures of growth for infants of P O P clients than for control infants, likely due to smaller numbers of P O P clients.  125 Table 5.6 Infant Growth  POP Clients (N=106)  Controls (N=318)  Mean Difference  T-Test & 95% C.L.oo  Gestational Age (weeks)*  38.6 ±2.5  39.0 ± 1.4  -0.46  1.67 (-1.0; 0.08)  Birth Weight (grams)*  3 3 9 ±693  3,361± 372  -11.5  0.16 (-155.9; 132.8)  Head Circumference (centimeters)*  34.4 ± 2.0  34.7 ±1.5  -0.33  1.21 (-0.87; 0.22)  Length (centimeters)*  50.7 ± 3.4  51.3 ± 3.0  -0.55  0.97 (-1.7; 0.60)  * Mean ± Standard Deviation "° Confidence Limits  Table 5.7 describes and evaluates the impact of P O P on infant morbidity. Infant morbidity was defined in terms of preterm birth, low birth weight, small for gestational age, large for gestational age, perinatal complications and congenital anomalies.  In contrast to measures of infant growth there were  differences  between the two groups. With the exception of small for gestational age, P O P infants experienced more morbidity than control infants.  The rate of preterm birth was 15.1 per 100 births for P O P infants compared with 8.8 for control infants.  The odds ratio of 2.39 (C.L.=1.2; 4.7) indicated a  statistically significant association between a higher rate of preterm birth and P O P participation.  P O P infants also experienced a higher rate of low birth  weight at 11.4 per 100 births compared with 6.6 for control infants.  The  difference resulted in an odds ratio of 1.48 (C.I.=0.74; 3.0) and was not statistically significant.  126  Results of the small for gestational age measure are the reverse of those shown previously, and indicated favorable program effects for P O P infants. The rate of small for gestational age was 8.5 per 100 births for P O P infants and 12.3 for control infants.  The difference produced an odds ratio of 0.53 (C.L.=0.26; 0.99),  indicating a statistically significant association between P O P participation and lower rate of small for gestational age. The rate of large for gestational age was 17.0 per 100 births for P O P infants compared with 10.4 for control infants.  This  difference resulted an odds ratio of 1.24 (C.I.=0.71; 2.2).  The number of infants experiencing one or more perinatal conditions relatively low for both groups of infants.  was  The rate was 10.5 per 100 births for  P O P infants compared with 6.5 for control infants (O.R.=1.41; C.L.=0.69; 2.9). Considerably more P O P infants were born with one or more anomalies than were control infants. respectively.  congenital  The rates were 14.2 per 100 births and 4.7  The resulting odds ratio of 3.31 (C.L.=1.6; 6.9) was statistically  significant.  Statistical results of all measures of infant morbidity for the overall sample were mixed.  P O P infants had significantly higher rates of preterm birth and  congenital anomalies. gestational age.  However, they had a significantly better rate of small for  There were no statistically significant differences for rates of  low birth weight and large for gestational age between the two comparison groups.  Table 5.7 Infant Morbidity  POP Clients (N=106)  Controls (N=318)  Odds Ratio & 95% C L . oo  Preterm Birth °°(%)  15.1  8.8  2.39t (1.2; 4.7)  Low Birth Weight /(%)  11.4  6.6  1.48 (0.74; 3.0)  Small for Gestational Age  8.5  12.3  0.53t (0.26; 0.99)  Large for Gestational Age  17.0  10.4  1.24 (0.71; 2.2)  Perinatal Conditions * (%)  10.5  6.5  1.41 (0.69; 2.9)  Congenital Anomalies ^(%)  14.2  4.7  3.31t (1.6; 6.9)  °o Confidence Limits °° Less than 37 completed weeks of gestation. / Less than 2500 grams. d Birth weight less than 10th percentile for gestational age. i Birth weight greater than 90th percentile for gestational age. * One or more of the following conditions:birth infections, birth trauma, seizures, respiratory distress, hemorrhage, narcotic abstinence syndrome, fetal alcohol syndrome, and neonatal intensive care admission. tf One or more condition under ICD-9 codes 740-759. t Statistically Significant.(p < 0.05)  Subanalysis-Program Entry  To determine if there were any differential program effects based on length of contact with the program (a measure of intensity of service) a subanalysis was performed.  W o m e n who entered a P O P prior to 20 completed weeks of  gestation (early entry) formed one group, and those who entered between 21 and 28 weeks gestation (late entry) formed the other group.  Multivariate analysis comparing early and late entry clients could theoretically have been utilized i n this subanalysis.  However, given the small subgroup  sizes (60 early entry and 46 late entry clients) and the dramatic differences in risk profiles (Table 5.8), utilization of this measure w o u l d not have yielded accurate results.  For this reason, a comparison of each subgroup of P O P clients with  their matched controls was performed.  By comparing the P O P subgroups to  their matched controls, sociodemographic similarities were maintained i n the analysis, allowing contact with the P O P to be the main difference between the comparison groups.  Descriptive information on P O P clients based u p o n entry to the programs is presented in Table 5.8.  Although both early and late entry clients were similar  w i t h regard to mean risk scores u p o n entry to the programs, they were significantly  different i n every other respect.  W o m e n w h o entered  the  programs later were at greater sociodemographic risk for having a low birth weight infant.  They were more likely to be adolescent, unmarried,.  and experiencing their first child.  Caucasian,  Table 5.8 P O P Clients-Early and Late Entry  Early Entry N=60 Maternal Age * < 20 Years(%) Single(%) Aboriginal(%) Parity * Primiparity(%) Major Risk * Minor Risk *  Late Entryl N=46  6  21.2 ± 5.9t 43.5t 58.7t 43.5+ 1.0 ± 1.2+ 44.5+ 1.0 ± 1 . 2 4.48 ± 2.4  24.0 ± 5 . 3 25.0 45.0 55.0 1.8 ± 1.9 28.3 1.0 ±0.98 4.5 ± 1.7  Early Entry was defined as program entry prior to 20 completed weeks of gestation. 1 Late Entry was defined as program entry between 21 and 28 weeks of gestation. * Mean ± Standard Deviation + Statistically Significant (p < 0.05) e  Early Entry Maternal Outcomes  Maternal  outcomes for w o m e n who entered the programs prior to m i d  pregnancy (early entry) and their matched controls are shown in Tables 5.9 and 5.10.  For every measure of maternal outcome, early entry P O P clients obtained  better results than their controls. O n average, early entry P O P clients initiated prenatal care during their first trimester of pregnancy while their controls initiated care during their second trimester.  Early entry P O P clients initiated  prenatal care on average 3.5 weeks ahead of their controls (C.L.=-5.7; -1.1). This difference indicated a significant association between P O P participation and early initiation of prenatal care. Early entry P O P clients obtained on average 1.2 more prenatal visits than their controls (C.L.=0.3; 1.9). was statistically significant.  Again, this difference  Once again, a far greater proportion of early entry P O P clients (36.7%) obtained adequate  prenatal care28, compared to their controls (22.5%). This 14.2%  difference resulted in a statistically significant odds ratio of 2.47 (C.L.=1.2; 5.0).  The proportion of early entry P O P clients who experienced one or more pregnancy/delivery complications was slightly lower than for their controls. Approximately sixteen percent (16.6%) of P O P clients and 19.0% of controls experienced morbidity.  For three of the four measures of maternal outcomes, initiation of prenatal care, number of prenatal visits and adequacy of prenatal care, early entry P O P clients obtained statistically better results that their controls indicating positive program effects.  For the remaining measure, maternal morbidity, there was no  significant difference between the two groups.  Based o n a n adaptation of the Kessner Index of Prenatal care. This index is an algorithm that includes trimester of pregnancy prenatal care began, number of prenatal visits a n d length of gestation, compared to the expected n o r m for visits (Appendix E). 2 8  Table 5.9 Early E n t r y Maternal Outcomes-Initiation and Visits e  Initiation of Prenatal Care (weeks)* Number of Prenatal Visits*  P O P Clients (N=60)  Controls (N=180)  Mean Difference  95% Confidence Limits  12.6 ± 5.9  16.1 ± 4.8  -3.6  ("5.7; "1.1)+  8.5 ± 2.6  7.3 ± 1.7  1.1  (0.3; 1.9)+  Early Entry was defined as program entry prior to 20 completed weeks of gestation. * M e a n ± Standard Deviation + Statistically Significant (p < 0.05)  e  Table 5.10 Early E n t r y Maternal Outcomes-Adequacy and Morbidity e  P O P Clients (N=60)  Controls (N=180)  Odds Ratio  95% Confidence Limits  Adequacy of Prenatal Care § (%)  22 (36.7%)  40 (22.2%)  2.47+  (1.2; 5.0)  Maternal  10 (16.6%)  34 (19.0%)  0.92  (0.61; 1.4)  M o r b i d i t y (%) !  Early Entry was defined as program entry prior to 20 completed weeks of gestation. § A s measured by modified Kessner Index (Appendix C) j Maternal M o r b i d i t y included any one of the following: anemia, placenta previa, placenta abruptio, polyhydramnios, oligohydramnios, pregnancy induced hypertension, gestational diabetes, preeclampsia, toxemia, threatened premature labour and post partum hemorrhage. + Statistically Significant (p < 0.05) e  Early Entry Infant Outcomes  Results of the early entry infant subanalysis are presented i n Tables 5.11 and 5.12.  A l t h o u g h early entry P O P clients had significantly better maternal  outcomes than their controls, the improvements were not transferred to P O P infants.  Table 5.11 shows there were virtually no differences in any measure of  infant growth between early entry P O P infants and their control infants.  Early  entry P O P infants had slightly shorter gestations than control infants, with a mean difference of 0.3 weeks.  Although early entry P O P infants were slightly  younger, they were slightly heavier. Early entry P O P infants were on average 11 grams heavier that their control infants.  Both groups of infants were identical  in head circumference (34.5 cm) and length (51.0 cm).  The subanalysis for measures of infant growth showed no significant statistical differences between early entry P O P infants control infants. Again, P O P infants experienced m u c h greater variability among all measures control infants.  of growth than  133 Table 5.11 Early E n t r y Analysis-Infant Growth e  P O P Clients (N=60)§  Controls (N=180)§  Mean Difference  95% Confidence Limits  Gestational A g e (weeks)*  38.8 ± 2.2  39.1 ± 1.3  -0.3  (-0.97; 0.38)  Birth Weight (grams)*  3,408 ± 6 3 2  3,396 ± 3 5 2  11.4  (-183.4; 206.2)  H e a d Circumference (centimeters)*  34.5 ± 1.8  34.5 ± 1.2  -0.21  (-0.68; 0.68)  Length (centimeters)*  51.0 ± 3 . 7  51.0 ± 3.3  -0.35  (-2.1; 1.4)  Early Entry was defined as program entry prior to 20 completed weeks of gestation. * M e a n + Standard Deviation § For gestational age a n d birth weight only e  Similar to the overall infant analysis, although the two groups of infants looked similar with regard to growth measurements, they were very different with regard to measures of morbidity (Table 5.12).  Once again, with the  exception of small for gestational age, early entry P O P infants experienced more morbidity than their control infants.  The differences, when compared with the  overall analysis were not as pronounced.  The rate of preterm birth was 10.0 per 100 births for early entry P O P infants compared with 8.3 for their control infants. 3.4) was not statistically significant.  The odds ratio of 1.23 (C.L.=0.45;  Early entry P O P infants also had a higher  rate of low birth weight at 8.3 per 100 births compared with 5.6 for their control infants. The corresponding odds ratio of 1.50 (C.L.=0.53; 4.3),was nonsignificant.  A g a i n , the results of small for gestational  age were the reverse of those  previous, and showed early entry P O P infants had a lower rate of small for gestational age than d i d their control infants.  The rate of small for gestational  age was 6.7 per 100 births for P O P infants and 10.6 for control infants (O.R.=0.63, C.L.=0.22; 1.8).  The rate of large for gestational age was 16.6 per 100 births for  early entry P O P infants compared with 11.7 for their control infants (O.R.=1.33, C.L.=0.79; 2.1).  The number of infants experiencing one or more perinatal conditions again relatively low for both groups of infants.  was  The rate was 8.3 per 100 births  for early entry P O P infants compared with 6.2 for their control infants. Considerably more early entry P O P infants were b o r n with one or more congenital anomalies than were their control infants.  The rates were 13.3 per  100 births and 4.0 respectively. The resulting odds ratio of 3.8 (C.L.=1.4; 10.5) was statistically significant.  The morbidity results for early entry P O P infants were mixed. Early entry P O P infants had considerably higher rates of large for gestational age and congenital anomalies than their control infants.  Early entry P O P infants, however, had a  considerably lower rate of small for gestational  age.  For the remaining  measures of morbidity, the two groups of infants were not very different.  Table 5.12 Early E n t r y Analysis-Infant Morbidity e  P O P Clients (N=60)  Controls (N=180)  Odds Ratio  95% Confidence Limits  Preterm Birth °°(%)  10.0  8.3  1.23  (0.45; 3.4)  L o w Birth W e i g h t /(%)  8.3  5.6  1.50  (0.53; 4.3)  Small for Gestational A g e ^(%)  6.7  10.6  0.63  (0.22; 1.8)  Large for Gestational A g e  16.6  11.7  1.33  (0.79; 2.1)  Perinatal C o n d i t i o n s * (%)  8.3  6.2  1.40  (0.47; 4.2)  Congenital A n o m a l i e s tf(%)  13.3  4.0  3.8+  (1.4; 10.5)  e  Early Entry was defined as program entry prior to 20 completed weeks of gestation.  °° Less than 37 completed weeks of gestation. / Less than 2500 grams. ^ Birth weight less than 10th percentile for gestational age. i Birth weight greater than 90th percentile for gestational age. * O n e or more of the following conditions:birth infections, birth trauma, seizures, respiratory distress, hemorrhage, narcotic abstinence syndrome, fetal alcohol syndrome, and neonatal intensive care admission. tf One or more condition under ICD-9 codes 740-759. + Statistically Significant.(p < 0.05)  136  Late Entry Maternal Outcomes  Maternal outcomes for women who entered the programs after m i d pregnancy (late entry) and their matched controls are shown in Tables 5.13 and 5.14. For three of the four measures of maternal outcome, late entry P O P clients were slightly disadvantaged when compared to their controls.  Both late entry P O P clients and their controls tended to initiate prenatal care during their second trimester. Late entry P O P clients initiated prenatal care on average 2.2 weeks (C.L.=-0.59; 4.9) behind their controls. Although P O P clients initiated care later, they obtained on average the same number of prenatal visits (7.1) as their controls.  O n l y 17.4% of late entry P O P clients received adequate prenatal care compared to 21.8% of their controls.  The proportion of late entry P O P clients w h o  experienced one or more pregnancy/delivery complications was slightly higher than for their controls. Approximately twenty six percent (26.1%) of P O P clients and 22.6% of controls experienced morbidity. None of these differences were statistically significant.  137  Table 5.13 Late Entryl Maternal Outcomes-Initiation and Visits  Initiation of Prenatal Care (weeks)* Number of Prenatal Visits*  P O P Clients (N=46)  Controls (N=138)  Mean Difference  95% Confidence Limits  17.1 ± 7.1  14.9 ± 5.0  2.2  (-0.59; 4.9)  7.1 ± 2.9  7.1 ± 1.9  -0.004  (-0.08; 0.79)  Late Entry was defined as program entry between 21 and 28 weeks of gestation. * M e a n ± Standard Deviation 1  Table 5.14 Late Entryl Maternal Outcomes-Adequacy and Morbidity  P O P Clients (n=46)  Controls (n=138)  Odds Ratio  95% Confidence Limits  Adequacy of Prenatal Care § (%)  8 (17.4%)  30 (21.8%)  0.78  (0.54; 1.1)  Maternal  12 (26.1%)  31 (22.6%)  1.27  (0.58; 2.8)  M o r b i d i t y i(%)  Late Entry was defined as program entry between 21 and 28 weeks of gestation. § A s measured b y modified Kessner Index (Appendix C ) i Maternal M o r b i d i t y i n c l u d e d any one of the following: anemia, placenta previa, placenta abruptio, polyhydramnios, oligohydramnios, pregnancy induced hypertension, gestational diabetes, preeclampsia, toxemia, threatened premature labour and post p a r t u m hemorrhage. 1  Late Entry Infant Outcomes  Results of the late entry subanalysis are presented in Tables 5.15 and 5.16.  Table  5.15 showed both groups of infants were very similar with regard to measures of growth.  Late entry P O P infants had slightly shorter gestations than their  control infants, with a mean difference of 0.6 weeks.  In addition to being  slightly younger, late entry infants were on average 42 grams lighter (C.L.=273.6; 190.7).  Late entry P O P infants h a d slightly  smaller mean head  circumferences (0.63 centimeters) and were slightly shorter (0.75  centimeters)  than their control infants.  There were no statistically significant differences i n any measure of infant growth between late entry P O P infants and their control infants.  Once again,  late entry P O P infants experienced greater variability in their outcomes.  139  Table 5.15 Late Entryl Analysis-Infant Growth  P O P Clients (N=46)§  Controls (N=138)§  Mean Difference  95% Confidence Limits  Gestational A g e (weeks)*  38.3 ± 2.8  39.0 ± 1.5  -0.67  (-1.6; 0.26)  Birth Weight (grams)*  3,273 ± 7 6 5  3,315 ± 3 9 5  -41.5  (-273.6; 190.7)  H e a d Circumference (centimeters)*  34.3 ± 2.3  34.9 ± 1 . 7  -0.63  (-1.5; 0.24)  Length (centimeters)*  50.5 ± 3.1  51.3 ± 2.8  -0.75  (-2.3; 0.80)  Late Entry was defined as program entry between 21 and 28 completed weeks of gestation. * M e a n ± Standard Deviation § For gestational age a n d birth weight only 1  Similar to both previous infant analyses, although the two groups of infants look similar with regard to growth measurements, they were very different w i t h regard to measures of morbidity (Table 5.16).  Once again, with the  exception of small for gestational age, late entry P O P infants experienced more morbidity than their control infants.  The differences, when compared with the  overall analysis were more extreme.  The rate of preterm birth was 21.7 per 100 births for late entry P O P infants compared with 9.4 for their control infants. The difference (O.R.=2.7, C.L.=0.98; 6.6) was borderline significant.  Late entry P O P infants born also had a higher  rate of low birth weight at 15.2 per 100 births compared with 8.0 for their control infants. The difference resulted in an odds ratio of 2.11 (C.L.=0.76; 5.8).  140  A g a i n , the results of small for gestational  age were the reverse of those  previous, and showed late entry P O P infants had a lower rate of small for gestational age than their control infants.  The rate of small for gestational age  was 10.9 per 100 births for late entry P O P infants and 14.5 for control infants (O.R.=0.72, C.L.=0.26; 2.0). The rate of large for gestational age was 17.4 per 100 births for late entry P O P infants compared with 8.7 for their control infants (O.R.=2.88, C.L.=0.87; 7.2).  The number of infants experiencing one or more perinatal conditions was 13.0 per 100 births for late entry P O P infants compared with 6.7 for control infants. This difference resulted in an corresponding odds ratio of 2.13 (C.L.=0.75; 5.0). A g a i n , considerably more late entry P O P infants were born with one or more congenital anomalies than their control infants. births and 5.9 respectively.  The rates were 15.2 per 100  The resulting difference was borderline significant  with an odds ratio of 2.86 (C.L.=0.99; 8.1).  Statistical results of infant morbidity showed that late entry P O P infants were not statistically different from their control infants with regard to measures of infant morbidity.  However, for two measures, preterm birth and congenital  anomalies the differences bordered on significant.  141 Table 5.16 Late Entryl Analysis-Infant Morbidity  P O P Clients Controls (N=46) (N=138)  Odds Ratios  95% Confidence Limits  Preterm B i r t h °°(%)  21.7  9.4  2.7  (0.98; 6.6)  L o w Birth W e i g h t /(%)  15.2  8.0  2.11  (0.76; 5.8)  Small for Gestational A g e ^(%)  10.9  14.5  0.72  (0.26; 2.0)  Large for Gestational A g e  17.4  8.7  2.5  (0.87; 7.2)  Perinatal C o n d i t i o n s *(%)  13.0  6.7  2.13  (0.75; 5.0)  Congenital A n o m a l i e s ^(%)  15.2  5.9  2.86  (0.99; 8.1)  1  Late Entry was defined as program entry between 21 and 28 completed weeks of gestation.  °° Less than 37 completed weeks of gestation. / Less than 2500 grams. d Birth weight less than 10th percentile for gestational age. i B i r t h w e i g h t greater than 90th percentile for gestational age. * O n e or more of the following conditions:birth infections, birth trauma, seizures, respiratory distress, hemorrhage, narcotic abstinence syndrome, fetal alcohol syndrome, a n d neonatal intensive care admission. ^ One or more condition under ICD-9 codes 740-759.  Subanalysis-Groups  at Risk  To determine if there were any differential program effects based on risk status, a second subanalysis was performed.  This subanalysis examined program  impacts for women with risk characteristics for low birth weight: adolescence; primiparous  women; single women and smokers.  A s it was  questionable  whether aboriginal race, in and of itself, is a risk factor for preterm birth and large for gestational age, women were stratified by race. Select infant outcomes were measured, these were: low birth weight, preterm birth and small  for  gestational age.  A s women were not matched o n two of these risk factors,  smoking and single marital status, the matching technique was eliminated for this subanalysis. The results of the subgroup analysis is presented in Tables 5.17 and 5.18.  The results of the subanalysis for aboriginal women are presented in Table 5.17. Across all risk categories infants of aboriginal P O P clients had lower birth weights than infants of aboriginal controls. For both P O P clients and controls, infants born to adolescents  had the heaviest average birth weights.  The  difference in birth weight ranged from 122 grams for smokers to 187 grams for primiparous women.  For three  of the  four risk  groups  (adolescents,  primiparous, single) i n the subanalysis, infants of aboriginal P O P clients experienced higher rates of adverse outcomes than d i d infants of aboriginal controls.  More aboriginal adolescent P O P clients (15.0%) had infants born with low birth weight than d i d aboriginal adolescent controls (2.4%). This difference produced an odds ratio of 7.41 and statistically significant confidence limits of 1.1 and 38.1. Just over twenty one percent (21.4%) of aboriginal primiparous P O P infants were small for gestational age compared with 4.4% of aboriginal primiparous control infants. This difference was statistically significant ( O.R.=6.14; C.L.=1.2; 32.6).  The results are somewhat mixed, but generally reversed for aboriginal smokers. A l t h o u g h confidence limits were not statistically significant, program effects were seen for aboriginal P O P smokers.  There was a 2.1 fold difference i n the  143 rate of small for gestational age for infants born to aboriginal P O P smokers (7.1%) and aboriginal smoking controls (15.4%).  Table 5.17 Subanalysis-Aboriginals  R i s k Characteristic  N  Birth Wt.  LBW/  (grams)  (%>  PB°° (%.)  SGA (%)  9  Adolescent P O P Clients Controls  20 42  3348 3495  15.0+ 2.4  20.0 4.8  10.0 2.4  Primiparous P O P Clients Controls  14 45  3191 3380  14.3 4.4  21.4+ 4.4  14.3 6.7  Single P O P Clients Controls  29 61  3307 3434  6.9 3.3  10.3 4.9  6.9 3.3  Smoker P O P Clients Controls  14 39  3234 3356  7.1 10.3  14.3 10.3  7.1 15.4  °° Preterm Birth, less than 37 completed weeks of gestation / L o w Birth Weight, less than 2500 grams ^ Small for Gestational Age,birth weight less than 10th percentile for gestational age + Statistically Significant (p< 0.05)  The results o f the subanalysis for Caucasian women a r e presented in Table 5.18. In  all risk  categories infants of  Caucasian  POP  clients h a d  higher  mean  birth  weights than infants of controls. Once again, for both P O P clients and controls, infants b o r n t o adolescents  h a d t h e heaviest average birth weights.  The  difference i n birth weight ranged from 58 grams for primiparous women to 181 grams  for  smokers. In contrast  to  aboriginal women, across  all  four  risk g r o u p s ,  infants of Caucasian P O P clients experienced lower rates o f adverse outcomes than infants  of  Caucasian c o n t r o l s .  144  N o n e of the i n f a n t o u t c o m e m e a s u r e s h a d statistically s i g n i f i c a n t confidence limits.  H o w e v e r , there were t w o i n f a n t outcomes  b e t w e e n the c o m p a r i s o n g r o u p s w e r e l a r g e .  for w h i c h  differences  There was a 3.07 fold d i f f e r e n c e i n  t h e p r o p o r t i o n of s m a l l f o r g e s t a t i o n a l age infants b o r n to C a u c a s i a n a d o l e s c e n t s .  Just o v e r s i x p e r c e n t (6.7%) of C a u c a s i a n a d o l e s c e n t P O P i n f a n t s w e r e s m a l l f o r gestational  a g e c o m p a r e d w i t h 20.6% of C a u c a s i a n a d o l e s c e n t  control infants.  There w a s a 3.22 fold d i f f e r e n c e i n the p r o p o r t i o n of p r e t e r m births to s i n g l e women.  Four p e r c e n t of C a u c a s i a n s i n g l e P O P c l i e n t s g a v e b i r t h p r e m a t u r e l y  c o m p a r e d w i t h 12.9 % of C a u c a s i a n s i n g l e c o n t r o l s .  Table 5.18 Subanalysis-Caucasians  R i s k Characteristic  N  Birth W t .  LBW/  PB°°  SGA  (grams)  (%)  (%)  (%)  Adolescent P O P Clients Controls  15 68  3438 3292  6.7 10.3  6.7 8.8  6.7 20.6  Primiparous P O P Clients Controls  21 87  3320 3262  13.6 11.5  9.1 11.5  13.6 19.5  Single P O P Clients Controls  25 93  3373 3245  8.0 10.8  4.0 12.9  16.0 24.7  Smoker P O P Clients Controls  29 88  3347 3166  6.9 11.4  6.9 12.5  17.3 23.8  °° Preterm Birth, less than 37 completed weeks of gestation / L o w Birth Weight, less than 2500 grams ^ Small for Gestational A g e , birth weight less than 10th percentile for gestational age  3  SUMMARY  The stated aim of the Pregnancy Outreach Projects is to identify women at risk in the community, engage them in prenatal care services and support them in making behavioral changes to reduce their risk of having low birth weight infants or other adverse pregnancy outcomes ( Pregnancy Outreach Projects: Project H a n d b o o k , 1993).  The focus of care is on behavior modification.  Appropriate nutritional intake, enhanced emotional support, smoking, alcohol and drug reduction are the primary behaviors of interest i n the programs.  W i t h the focus of care on behavior modification, it was expected that maternal program effects could be realized i n three areas, attainment of prenatal care services,  monitoring of prenatal condition and attainment  weight gain.  of appropriate  Given the nature of long term participation, it was also expected  that women who entered the programs early would have more opportunity to modify their behavior and therefore obtain better outcomes.  Infant program effects were more problematic to determine for two reasons; the late cutoff point for entry into the programs (28 weeks) and the time lag between knowledge attainment, and attitudinal and behavioral change.  A s the  programs focused on maternal behaviors k n o w n to effect primarily infant weight gain, infant program effects could be expected i n gestational  age  appropriate weight gain, particularly for infants of women who entered the programs early (prior to 20 weeks gestation).  G i v e n the present program  structure, maternal behavioral change may have occurred too late in pregnancy for women who entered the programs late (after 20 weeks gestation), to transfer  146  benefits of the intervention to their infants.  Therefore it was unclear if infants  of women who entered late w o u l d have significantly different outcomes than controls.  Results of the analysis for the two study groups overall showed that P O P clients, on average, initiated prenatal care slightly earlier and obtained significantly more prenatal visits than their matched controls. A slightly greater proportion of P O P clients received adequate prenatal care and a slightly smaller proportion experienced morbidity. (Due to limitations of the data base, maternal weight gain could not be measured.)  It was expected that the programs might not be  able to influence initiation and adequacy of prenatal care given that P O P clients entered the programs at any time during their first 28 weeks of pregnancy. The programs may have influenced the number of prenatal visits attained, though the difference between the two study groups was modest.  In measures of growth (gestational age, birth weight, head circumference and length) the two groups of infants were similar. infant morbidity the results were mixed.  However, for measures of  A greater proportion of P O P infants  experienced preterm births, low birth weight, large for gestational age, perinatal • conditions and congenital anomalies.  The differences were significant for rates  of preterm birth and congenital anomalies.  However, P O P infants had a  significantly lower rate of small for gestational  age.  G i v e n the  strong  nutritional component of the programs, and emphasis on smoking cessation, it is likely that program effects were shown in the differential rate of small for gestational age between the two groups of infants.  In the early entry subanalysis, results indicated that women who entered the programs early may have had more time to modify their behavior, as their maternal outcomes were better than those of their controls. W i t h the exception of maternal morbidity, where their was no difference, early entry P O P clients had statistically better maternal outcomes than their controls.  For P O P clients who entered after m i d pregnancy, the programs likely d i d not have the opportunity to influence maternal behavior change. slightly worse maternal outcomes than their controls.  P O P clients had  By virtue of coming late,  the programs could not influence initiation or adequacy of prenatal care. The programs may have been able to influence the number of prenatal visits. Although P O P clients initiated care on average 2 weeks later than their controls, they obtained the same average number of visits.  None of the differences for  the late entry subanalysis were statistically significant.  The infant subanalysis based on entry into the programs followed similar patterns as the overall analysis.  In the early entry analysis, although the two  groups of infants looked similar with regard to measures of infant growth, they were very different with regard to measures of morbidity. Once again, with the exception of small for gestational age, early entry P O P infants experienced more morbidity than their control infants.  Once again it was possible that program  effects were being shown for P O P infants, who obtained a lower rate of small for gestational age than control infants. similar for late entry P O P subanalysis.  The pattern of infant results was again P O P infants again, had a lower rate of  small for gestational age, but higher rates of other adverse outcomes, compared to their control infants.  For two morbidity outcomes,  preterm birth and  congenital anomalies, POP infants had outcomes that were borderline significant.  It was interesting to note that although POP clients who entered early looked very different than those who entered late (Table 5.9), the analyses based on time of entry showed little difference from the overall analysis. Early versus late entry into prenatal care had little effect on actual outcomes, the effect of time of entry was instead shown in augmented or diminished differences in outcome measures between the two subgroups of POP clients and controls.  Results of the subanalysis based on subgroups at risk show that both aboriginal and Caucasian adolescent women (both POP clients and controls) had infants with the heaviest average birth weights.  Although some researchers have  found that socioeconomically advantaged adolescents have heavy infants, poor socioeconomic circumstances in addition to young maternal age has not been associated with heavy infants.  Measures of infant morbidity were very different for aboriginals and Caucasians. With one exception, aboriginal POP clients had infants that were at greater risk of low birth weight, preterm birth and small for gestational age than were infants of aboriginal controls.  Program effects may have been seen for  aboriginal smokers whose infants had lower rates of low birth weight and small for gestational age. Caucasian POP clients experienced the opposite results. Their infants were at decreased risk of adverse outcomes, across all risk categories, than were Caucasian control infants.  Although none of the  differences were statistically significant, large differences favoring POP clients  149 were seen in the rate of small for gestational age for infants of adolescents, and the rate of preterm birth for single women.  150  REFERENCES  British Columbia Ministry of Health (1990). Vital Statistics A n n u a l Report. Victoria, B. C : Queen's Printers. Buescher, P.A., Roth, M.S., Williams, D . & Golforth, C M . ( 1991). A n Evaluation of the Impact of Maternity Care Coordination o n Medicaid Birth Outcomes in North Carolina. A T P H . 81Q2), 1625-1629. Kierans, W.J., Collinson, M . A . , Foster, L . T . & U h , S-H. (1993). Charting Birth Outcome i n British Columbia: Determinants of Optimal Health and Ultimate Risk. Victoria, B.C.: Division of Vital Statistics, B . C . Ministry of Health. Peoples, M . D . , Grimson, R . C . & Daughtry, G . L . (1984). Evaluation of the Effects of the North Carolina Improved Pregnancy Outcome Project: Implications for State-Level Decision Making. A T P H . 74 (6), 549-554. Pregnancy Outreach Projects: Program Handbook (1993). Victoria, B.C: Community & Family Health, Nutrition Branch, B . C . Ministry of Health. Thomson, M . (1990). Heavy birthweight i n Native Indians of British Columbia. Canadian Tournal of Public Health. 81,443-446. Tuk, T. (1995). Health Indicators for Status Indians i n B.C. 1987-1992. (Unpublished data).  151 CHAPTER 6  DISCUSSION O F R E S U L T S A N D I M P L I C A T I O N O F S T U D Y  DISCUSSION The association between socially disadvantaged pregnant women and adverse infant outcomes, has been consistently documented in studies over the past five decades.  In an attempt to prevent low birth weight and other adverse  outcomes, comprehensive prenatal care programs that identify and provide specialized services to socially disadvantaged women have evolved since the 1970's.  C o m p r e h e n s i v e prenatal care programs that  provide  lifestyle,  nutritional and psychosocial interventions have been developed to improve maternal health and therefore, infant outcomes.  A l t h o u g h several studies  have described programs that attempt to reduce low birth weight a n d / o r preterm birth, the results of the available studies have been inconsistent.  The  findings of the present study will be compared to previous reported studies to examine the impact of P O P on maternal and infant outcomes.  Maternal Outcomes  The ability of comprehensive prenatal care programs to improve maternal compliance with prenatal visits, has been well documented i n the literature. This study found an association between participation i n comprehensive care (POP) and improvement i n measures of prenatal care outcomes relative to the matched comparison group. Although P O P clients initiated care earlier and a greater proportion obtained adequate care than controls, neither result is statistically or clinically significant.  N o clear trends regarding the impact of  comprehensive prenatal care on these measures were shown i n the literature. This could be expected, as comprehensive programs have little control over when w o m e n enter prenatal care (other than setting a m a x i m u m length of gestation limit), therefore, their ability to influence initiation and adequacy of prenatal care is severely limited.  O n the other hand, comprehensive prenatal care programs have consistently been shown to increase the number of prenatal care visits participants obtain. This study found a statistically significant difference i n the number of prenatal visits obtained by P O P clients.  Although the size of the difference (0.5 visits)  was the same as that found by Hardy et al. (1987), it was markedly smaller than the 3 visit difference found by Elster et al. (1987), and somewhat smaller than the 1.5 visit difference found by Polland et al (1992) and the 1.1 visit difference found by Smith et al (1978). Opinions of prenatal care experts cast doubt on the clinical significance of this 0.5 visit difference^.  Examining the number of prenatal care visits, in and of itself, as a measure of maternal health is somewhat risky.  W o m e n who obtain several visits may  have more pregnancy problems that require closer monitoring or may simply be responding to recommended number of visits.  Conversely, w o m e n who  obtain few visits may have uncomplicated pregnancies and not require close monitoring or simply not know the need to obtain early and regular care. However, even with these divergent views, investigators still consider the number of prenatal visits to be an important indicator of prenatal care management.  Drs. M . C o x & B. R i d y a r d , Obstetrician & Gynecologist, personal communication.  Very  few  studies  examined  pregnancy  and/or  labour  and  delivery  complications for measures of maternal morbidity, and for those that d i d , no clear trend emerged. comprehensive composite  care participants suffered  measures)  comparison groups. participants  Both Olds et al (1986) and H a r d y et al (1987) found that less maternal morbidity (using  and had better managed pregnancies  than d i d their  Smith et al. (1978) found that comprehensive  experienced  m o r e morbidity.  A l t h o u g h a finding of more  morbidity seems counterintuitive, this could occur if intervention were more closely monitored than controls.  care  women  The rate of maternal morbidity  was very similar between P O P clients and controls in this study. Like measures of prenatal care, this measure can be somewhat misleading as increased rates of reported maternal morbidity may result from either better monitoring or increased risk, the reverse is also true.  Infant Outcomes  W i t h the exception of diminished infant mortality, no clear trends emerged from the literature with regard to the impact of prenatal care on infant outcomes.  comprehensive  None of the randomized trials reviewed  found intervention impacts, while approximately half of the studies found  have  case-control  comprehensive care participation improved infant outcomes.  Some studies found improvements in the rate of low birth weight but not i n preterm birth, while others found the reverse.  This study found that the impact of comprehensive prenatal care o n infant outcomes was mixed. The true impact of the programs may have been shown in the statistically and clinically lower rate of small for gestational age infants  between P O P clients and control.  G i v e n the emphasis placed o n improving  nutritional status and decreasing risky lifestyle behaviors, (especially smoking) in the program, this result is likely.  Unfortunately, comparisons of this  outcome measure with the literature are hard to make. none  of the reported studies  W i t h one exception,  documented rates of intrauterine growth  retardation or term low birth weight.  C o m p a r i s o n s between  small for  gestational age and low birth weight are not appropriate because both term and preterm births are included.  Although a significantly (both statistically and clinically) greater proportion of P O P infants were born preterm and with congenital anomalies (correlated outcomes), these outcomes are likely not a reflection of a lack of program effects.  There are two plausible explanations for these results: First, because  preterm labour education and management were not included in the program, program participation w o u l d not be expected to influence the rate of preterm birth.  Second, a greater proportion of P O P multips (17%) had experienced at  least one previous preterm infant compared with 11.9% of control multips. M a n y of these P O P clients had experienced three or more preterm births. Therefore, the programs d i d not address measures to combat preterm birth and P O P clients as a whole, were at greater risk for this outcomes.  P O P infants also experienced higher rates of low birth weight, large for gestational age and perinatal conditions.  Although, none of the differences  were statistically significant, the 1.6 to 1.7 fold differences in these rates of adverse outcomes makes them all clinically significant.  155  Impact of Program Entry on Maternal and Infant Outcomes  In the literature, two measures were utilized to determine if there  were  differences i n maternal and infant outcomes as a result of intensity of service; (1) the number of comprehensive visits or, (2) the length of gestation covered by  comprehensive  care.  D u e to  lack  of  consistency  in  measuring  comprehensive care visits for P O P clients, the latter measure was chosen i n an attempt to measure differential impacts of the program. A s a measure of length of gestation covered by comprehensive care, the utilization of number of months of prenatal coverage is controversial, particularly w h e n evaluating infant outcomes.  W o m e n who deliver preterm infants bias the results to  indicate that less gestational coverage leads to poorer infant outcomes.  The  present study avoided that methodological pitfall b y measuring gestational age at enrollment categorized as before or after midgestation.  Similar to the trend found i n the literature, this study found that women who entered care early obtained significantly (both statistically and clinically) better prenatal care coverage and experienced less morbidity than their controls. A s both groups of women had very similar sociodemographic characteristics, the reasons for this difference was not clearly apparent. Some sort of motivational bias  that favors P O P clients  to enter care early a n d c o m p l y w i t h the  recommended number of visits is a plausible explanation.  Alternately, some  unmatched variable, more common among P O P clients than controls, could have influenced early program participation.  The impact of the programs for women who entered prenatal care late is harder to measure.  Looking for program effects by measuring initiation of prenatal  care and adequacy of care w o u l d be misleading, as the programs have no influence over these outcomes for late entering clients.  O n l y by examining the  number of prenatal visits obtained in conjunction with initiation of prenatal care can the influence of the program be shown.  It is doubtful that the  programs had any impact on maternal outcomes for late entering clients as they obtained the same number of visits as their controls and started care only two weeks later.  In contrast to trends i n the literature, the present study found that infants of women who initiated prenatal care early d i d not have significantly lower rates of adverse outcomes compared to their controls. morbidity outcome outcomes.  measures,  In fact, with for most of the  P O P infants h a d higher rates of  adverse  For one measure, congenital anomalies, d i d the 3 fold difference  reach both statistical and clinical significance.  Reversing the trend was the outcome measure small for gestational  age.  Although the difference was not statistically significant, the 1.6 fold difference is clinically significant.  Published reports found greater intensity of service to be  associated with better infant outcomes.  Both Olds et al (1986) and H a r d y et al  (1987) found that women who registered early for prenatal care had infants that were significantly heavier than those of the comparison group. Although, the birth weight results in this study follow the same general pattern, the impact of P O P participation on birth weight differences was minimal.  The reasons for  lack of statistically significant program effects for women with greater prenatal care contact is unknown.  Perhaps variables not controlled i n the  design  confounded the results.  Alternatively, the study might not have had enough  power to detect true differences.  Similar to the trend i n the literature, infants of P O P clients who initiated comprehensive care late were not statistically different from infants of their controls. From a clinical perspective, P O P infants had significantly higher rates of most adverse outcomes, again with the exception of small for gestational age. However, the differences between the two groups in this subanalysis may have nothing to do with program impact and everything to do with late entry effects. By virtue of entering the programs between 21 and 28 weeks of gestation, clients may not have had adequate time to transfer benefits of behavior modification onto their infants. to  rule  out  Once again, the sample size of this subanalysis was to small  chance  as  an  alternative  explanation  for  the  statistically  nonsignificant results.  Impact of Program on Subgroups at Risk  The present study found differential program effects on infant outcomes when stratified by maternal race and risk characteristics.  A l t h o u g h i n general,  aboriginal P O P clients had higher rates of adverse infant outcomes than aboriginal controls, clinically significant program effects were clearly shown in the rate of small for gestational age for infants of P O P smokers.  While many  studies found infants of minority women who received comprehensive care, mainly blacks, to have better infant outcomes, no studies examined program impacts for women of aboriginal race. Therefore comparisons to the literature are impossible.  Caucasian P O P clients regardless of risk category had m u c h lower rates of adverse infant outcomes  than Caucasian controls.  Clinically  significant  program effects were shown for infants of both adolescents and single women. This finding is in general agreement with trends i n the literature that found infants of adolescents, particularly y o u n g adolescents obtained significant benefits from comprehensive care.  In this study, infants of P O P adolescents  were on average 148 grams heavier than control infants and experienced a 3.07 fold decrease in their rate of small for gestational age.  Olds et al (1986) found  that young adolescents had infants that were on average 395 grams heavier than infants of the comparison group.  Infants of program adolescents also had  significantly lower rates of preterm birth and low birth weight.  Peoples et al  (1983) found that program adolescents had significantly fewer infants born with low birth weight.  H a r d y et al. (1987) found that adolescents had the heaviest  infants and the lowest rate of small for gestational age infants.  A s the size of  the adolescent population in this study was very small, there is no way of knowing if the results were due purely to chance, however, they are consistent with the findings of other studies.  One interesting finding was that regardless of race or intervention group, infants of adolescents had the highest mean birth weight.  The trend i n the  literature has been that infants of adolescents have h a d lower mean birth weight than non-adolescent mothers. While some studies have found age to be an independent risk factor, others have found increased socioeconomic risk status to be an independent risk factor.  Given that all the adolescents in this  study were socioeconomically disadvantaged, the finding of heavier birth weights is contrary to current trends.  159 Other Issues  Two other issues deserve comment.  First is the finding that, i n every outcome  measure, the standard deviations for P O P clients showed more variability than those of controls. There are two possible explanations for this difference.  First,  this difference i n variability may be due to differences i n the degree of risk between the two groups.  P O P clients were at risk for adverse pregnancy  outcome based o n a broad range of social, demographic, economic, medical or obstetrical  risks.  Risk  sociodemograhic variables.  characteristics  for  controls  were  limited  to  A s controls were more homogeneous than P O P  clients, they may have been at less risk.  Second, the variability i n standard  deviations for outcomes measures may simply have reflected the small size of the P O P client group compared to the comparison group.  The  second issue relates to the use of the screening/assessment  tools.  The  ability of the T - A C E and IPRIT screening tools to accurately determine a high risk population are questionable.  The IPRIT tool has not been tested or  evaluated against any measure, therefore its predictive values are unknown. Although the T - A C E tool is used by some physicians in clinical practice, it has mediocre sensitivity and a poor predictive value. appropriately select a "high risk" group is questionable.  Therefore it's ability to  160  LIMITATIONS  Despite careful planning, this study was not immune to problems that limit the validity, reliability and generalizability of the statistical anlayses.  The most  critical threats are discussed below.  The most serious limitation i n this study is selection bias.  Referrals to the  programs were almost equally split between self referral and community health agency referral.  It is not known if health attitude and behavior or motivational  levels differed between those that self selected and those that were agency referrals.  It may be that individuals who self selected had higher levels of  motivation than other potential clients w h i c h c o u l d have impacted the likelihood of entry and retention in the program.  Individuals could have been  referred to P O P either because they also were highly motivated, or, conversely, had lifestyles and circumstances that put them at risk for adverse pregnancy outcomes.  Thus as a group the P O P clients may have had differences i n  susceptibility.  It is also likely that the P O P clients as a group could have differed from controls. It may be that individuals at extreme risk for adverse pregnancy outcomes were referred to the programs, either through self or agency referral while those with modest risks were not.  A s there were no means available to  evaluate  differences in health beliefs and behaviors between the two study groups, it is possible that selection bias could have resulted in differences i n susceptibility between these groups. Selection bias would have acted to diminish the impact of the programs. The impact of selection bias could have been reduced if it had  been possible to select controls from a time period prior to the start of the programs.  The power of this study to detect a 10% absolute difference in the rate of low birth weight (from 15% to 5%) was lower than predicted. In the design stage it was estimated that 125 P O P clients w o u l d be available for this study.  Matched  1:3 the study w o u l d have had an 85% chance of detecting a true difference. W i t h the P O P group limited to 106 clients, the power decreased to 70%. Although not optimal, this level of power was comparable (if not better) than many of the studies reviewed in the literature.  This study was also subject to misclassification bias.  Computerized record  linkage was used to identify P O P clients within the U B C Perinatal Study data base.  F o l l o w i n g this and a manual review, three P O P clients were  identified.  not  It is possible that the missing women could have been selected as  part of the control group. The impact of misclassifying these three P O P clients as controls w o u l d be small, as the missing women accounted for only 0.7% of the study sample.  It is not k n o w n if the matching strategy employed i n this study adequately selected a control group of comparable risk to P O P clients.  Stratification and  matching are the design techniques commonly used to counter selection bias and ensure that desired similarities occur between the compared groups. In this study the matching technique was employed and three controls were matched to each P O P client based o n four demographic risk factors.  Although, a  demographic matching process has historically been used i n public health  epidemiological studies, demographic variables may not be the best predictors of risk i n this group.  The study w o u l d have been enhanced if controls could have been matched to each P O P client, based on the client's particular risks.  Matching on client  specific risks would have ensured that each subset had similar risks. Social and physical variables i n c l u d i n g reproductive risk, maternal height, pregravid weight, weight gain and smoking status have higher attributable risks for adverse pregnancy outcomes than the variables used.  However, information  on the former variables was either not collected or poorly recorded. A s a result the statistical process of demographic matching may have omitted the crucial variables needed to predict or identify program impacts.  A l t h o u g h several experts have stated that in a retrospective matched casecontrol analysis, study groups should be matched on a m a x i m u m of four variables, there remains the possibility that the study groups were overmatched (Feinstein,  1985;  Schlesselman,1982;  Miettinen,  1970).  The  result  of  overmatching is the creation of two groups that are too similar, diminishing the ability of the study to isolate the effects of the programs and thereby reducing validity and statistical efficacy of results.  A s mentioned previously, the necessity of maintaining matching i n the data analysis is controversial.  Although matched analyses are intuitively harder to  understand, a matched analysis was completed because it was more likely to produce a statistically significant value than an unmatched analysis.  G i v e n the  small sample size, this appeared to be the most appropriate action. In addition,  matching controlled for many of the variables that w o u l d likely confound the results.  A l t h o u g h control for certain known risk factors was conducted through the matching process, there remained several uncontrolled variables that were possible confounders.  Preexisting health status was a potential confounder as  information pertaining to participants physical and psychological health was not available.  A greater proportion of multiparous P O P clients than controls  experienced a previous poor pregnancy outcome, this is also a potentially serious confounder.  M a n y other uncontrolled variables including health care  practices, reproductive risk, environmental influences, nutritional status could have affected birth outcomes.  lifestyle  habits and  Therefore, the study  results are confounded to the extent by w h i c h differences i n uncontrolled variables between the two study groups existed.  Prenatal care was another confounder in this study.  There was no way to  ascertain if controls received any additional prenatal care.  Although the P O P  was the only comprehensive intervention i n the area, controls could have received interventions from other social service agencies.  The possibility of  this occurring is highly unlikely, never the less, any additional prenatal care received by controls would diminish the differences between the study groups and could minimize the impact of the programs.  P O P content and servicing also pose limitations. For the purposes of this study the three Vancouver Island programs are treated as being homogeneous i n their service provision. This is not likely to be true, as there may be systematic differences in program delivery between the sites. However, it is impossible to  identify program differences that may have influenced maternal and infant outcomes.  This study was  unable to quantify or qualify the amount and type of  intervention p r o v i d e d to i n d i v i d u a l clients i n the P O P .  Therefore, it is  impossible to ascertain the most effective aspects of the intervention efforts, nor which clients were the most or least responsive to intervention efforts.  Finally, as only the Vancouver Island P O P have been evaluated in this study, and  the target  groups and services differ among all the programs, the  conclusions reached are limited to the Vancouver Island P O P and may not be applicable to all programs.  The use of secondary data from the U B C Perinatal Study limited both the study design, scope of research questions and conclusions reached.  Despite these  limitations the study d i d provide a confirmation of the relationship between P O P and maternal and infant outcomes.  Given these relationships there are a  number of policy and research recommendations that can be made.  IMPLICATIONS Pregnancy Outreach Project Implications  In this study P O P participation was found to be associated overall with slightly improved maternal outcomes.  However, neither the role of P O P participation  per se, not the impact of any comprehensive care program i n improving infant outcomes has been clearly delineated.  G i v e n this, there are some areas i n  which changes in program content would assist in the clarification of program impacts.  A careful examination of the screening tools should be undertaken. A n effort should be made to determine the sensitivity and specificity of the IPRIT if it is to remain as a screening tool. In addition, an examination of the content of the IPRIT should be undertaken as there are some variables for w h i c h the association with adverse pregnancy outcome has not been proven, or has i n fact been disproved. A l t h o u g h the author of the T - A C E questionnaire found it to rate highly against the gold standard, the Michigan Alcohol Scrrening Test^O this view is biased.  G i v e n the dismal predictive positive value (23%) and the  mediocre sensitivity (69%), consideration should be given to it's replacement.  A s program effects were seen for subgroups of w o m e n with increased risk, aboriginal  smokers, adolescent and single  Caucasian  w o m e n i n particular,  consideration should be given to targeting these subpopulations.  By targeting  specific subpopulations, referral agencies could more easily identify potential high risk w o m e n and refer them to the programs earlier. target  subpopulations  Clearly defining  may increase the likelihood that P O P services  reaching more women in need.  are  The subpopulations traditionally targeted for  comprehensive care programs are adolescents,  minority w o m e n and those  living i n poverty. A t this time it is unclear if the subpopulations most likely to benefit from P O P services are specifically targeted.  Selzer, M . L . (1971). The M i c h i g a n A l c o h o l Screening Test: the quest for a n e w diagnostic instrument. A m e r i c a n H o u r n a l of Psychiatry, 127,1653-1658. 3 0  In conjunction with targeting specific subpopulations, every effort should be made to encourage early entry into the programs. A s behavior modification is a large component of the programs, and behavior change is a slow process, early entry maximizes program benefits.  There were, for example,  significant  differences in maternal outcomes and the rate of small for gestational age for w o m e n w h o entered the programs early, compared to their controls. significant differences were found for women who entered late.  No  The time lag  from program entry to delivery likely was too short for w o m e n who entered late (and their infants) to gain the benefit of behavior changes.  Therefore it  w o u l d be prudent to encourage early entry in order to maximize maternal behavior change and improve infant outcomes.  A broader range of endpoints should be considered when determining program objectives.  A s the P O P include interventions that go beyond the pregnancy  period, these should be reflected i n the objectives.  Program objectives should  be expanded to include improvements i n obstetrical and perinatal outcomes, improvements in appropriate parental behaviors, and improvements in social and emotional adjustment to the parental role.  It may also be important to consider expanding the scope of interventions to include the first postnatal year. M a n y studies ( Elster et al., 1987; Heins et al., 1990; Spencer et al., 1989; Peoples et al., 1984; K a y et al., 1991) have found that comprehensive prenatal services d i d not impact pregnancy outcomes compared with traditional prenatal care.  when  In contrast, comprehensive care's  major impact was on events that occurred during the first two postnatal years. Parental psychosocial health and parenting behaviors were found to be the major gains of comprehensive programs. This underscores the need to include  both pregnancy and parenthood intervention services i n programs.  comprehensive  Expanding the scope of P O P services could be achieved either  through additional postnatal programming or affiliation with existing infant programs.  Research Implications  Although this study d i d find an overall improvement i n one infant outcome, and p r o g r a m impacts  for w o m e n  based  o n initiation of care and risk  categorization, it is important to clarify that this study was not exhaustive. Therefore, a number of areas in which further research could be explored have been presented.  A s one of the major limitations of this study was the small sample size, an analysis covering a longer period of time with a larger sample size might produce more conclusive findings.  U s i n g a similar design but increasing the  number of P O P clients w o u l d provide a large enough case pool to ensure adequate power.  This could be accomplished through pooling the birth results  of all the programs over a 12 -18 month period.  The availability of a  comparable control group remains an issue. Controls would need to be selected based on predetermined medical, behavioral or psychosocial risks.  Although  increasing the sample size would increase power, the tradeoff may be i n limited outcomes measures.  Outcomes of interest would be limited to those available  on hospital or vital statistics records.  G i v e n the difficulty i n determining a similar comparison group, a withinmother (before-after)  study design could be utilized to determine program  impacts.  Infants of women who participated i n a P O P during their second (or  subsequent) pregnancy would be compared to the first (or previous) infant. By comparing sibling pairs, infants born to the same mothers, each w o m a n becomes her o w n control.  Thus selection bias and other confounders are  eliminated, allowing true program impacts to be shown.  It may also be prudent, given limited financial resources, to determine if there are specific risk factors within the P O P population that contribute greatly to adverse pregnancy outcomes.  A multivariate analysis of women with adverse  pregnancy outcomes could be conducted to determine which risk factors in this population contributed to adverse outcomes.  This w o u l d again require pooling  all P O P sites, or collecting several years of data from a few sites. If certain risk factors could be identified, it may be possible to target potential clients most likely to benefit from comprehensive prenatal services.  There are also research issues related to the longterm impacts of the programs. A n analysis that looks beyond the scope of traditional maternal and infant outcomes w o u l d assist i n determining other program impacts.  A n evaluation  that focused on family and infant outcomes during the first year of life w o u l d determine if there are any longterm impacts of the programs. Family outcomes could  include  social  support  d u r i n g pregnancy,  social  and  emotional  adjustment to parenting, parent-infant interaction, and understanding of infant growth and development.  Infant outcomes could include social adjustment  and attatchment, growth and development measures, emergency room visits, hospitalizations and immunizations.  169  REFERENCES  Elster, A . B . , Lamb, M . E . , Tavare, J. & Ralston, C . W . (1987). The Medical and Psychosocial Impact of Comprehensive Care on Adolescent Pregnancy and Parenthood. T A M A . 258 (9), 1187-1192. Feinstein, A . R . (1985). Clinical Epidemiology: The Architecture of Clinical Research. Toronto, Ontario: W . B . Saunders Company. Graham, A . V . , Frank, S.H., Zyzanski, S.J., Kitson, G . C . & Reeb, K . G . (1992). A Clinical Trial to Reduce the Rate of L o w Birth Weight i n an Inner-city Black Population. Family Medicine. 24(6). 439-446. Hardy, J.B., King, T . M . & Repke, J.T. (1987). The Johns Hopkins Adolescent Pregnancy Program: A n Evaluation. Obstetrics and Gynecology. 3(1). 300305. Heins, H . C , Nance, N . W . , McCarthy, B.J., Efird, C M . (1990). A randomized trial of nurse-midwifery prenatal care to reduce low birth weight. Obstetrics & Gynecology. 75.341-345. Kay, B.J., Share, D . A . , Jones, K . , Smith, M . , Garcia, D . & Yeo, S. (1991). Process, Costs, and Outcomes of Community-Based Prenatal Care for Adolescents. Medical Care. 29 (6), 531-541. Miettinen, O.S. (1969). Individual matching with multiple controls i n the case of all-or-none responses. Biometrics. 25. 321-338. Olds, D . L . , Henderson, C R , Tatelbaum, R. & Chaimberlin, R. (1986). Improving the Delivery of Prenatal Care and Outcomes of Pregnancy: A Randomized Trial of Nurse Home Visits. Pediatrics. 77(1). 16-28. Peoples, M . D . , & Siegal, E . , (1983). Measuring the Impact of Programs for Mothers and Infants o n Prenatal Care and L o w Birth Weight: The Value of Refined Analyses. Medical Care. 21(6). 586-605. Schlesselman, J.J. (1982). Case-Control Studies: Design. Conduct Analysis. N e w York: Oxford University Press. Spencer, B., Thomas, H . , & Morris, J. (1989). A randomized controlled trial of the provision of a social support service during pregnancy: The South Manchester family worker project. British Journal of Obstetrics & Gynecology. 96.281-288.  170  APPENDIX A  Ill Province of British Columbia  Ministry of Health and Ministry Responsible for Seniors  Individual Prenatal Risk Identification  Location:  Date:  Client ID:  SEE GUIDE FOR DEFINITIONS AND EXPLANATION. Code Description  Yes Explanation  Physical Factors  PF1 PF2 PF3 PF4 PF5 PF6 PF7 PF8 PF9 PF10 PF11 PF12  • • •  Previous pregnancy loss Illness/condition with impact on pregnancy Pre-pregnancy weight - body mass index (BMI) Rate of weight gain Inadequate nutrition Previous child with anomaly Previous child requiring neonatal intensive care Multiple pregnancy Birth interval Grand multipara - 5 or more pregnancies Established genetic risk Age 17 and younger/ 36 and older  • • •  • • • •  • •  Substance Abuse/Misuse  SA1 SA2 SA3 SA4  Cigarette smoking Alcohol use Inappropriate use of over the counter and prescription drugs Other drug use  • • •  •  Psychosocial & Economic Factors  PE1 PE2 PE3 PE4 PE5 PE6 PE7 PE8 PE9 PE10 PE11 PE12 PE13  Single parenthood Delayed access to prenatal care Refusal of/resistance to appropriate services Isolation - ethnic, language and social Limited learning ability/illiterate Marital problems/unstable relationship/family violence Mental health problems Low self-esteem Inability to cope/anxiety regarding pregnancy and baby Unrealistic expectations Unwanted pregnancies/denial of pregnancy Financial problems Inadequate housing  tLTH XXXX93/03  • • • • • • • • • • • • •  WHITE COPY - HEADQUARTERS  YELLOW COPY - COORDINATOR  A GUIDE FOR THE USE OF INDIVIDUAL PRENATAL RISK IDENTIFICATION T O O L  PURPOSE The purpose of this form is to provide a tool which will identify some of the major factors that can influence the outcome of the pregnancy and at a quick glance provide the risk factors specific to the individual client. Program staff can use it as a checklist when determining the care plan for the client. It is intended to complement the prenatal assessment of the physician by highlighting lifestyle factors in particular. The guide is not meant to be an all inclusive source of information ofrisksin families and pregnancies. It compiles in a single document basic information to assist professionals in the early identification of risks with the ultimate goal of reducing perinatal morbidity and mortality. Personal experience, knowledge and intuition oh the part of the professionals are as important, if not more, than whatever guide or form is used. The guide should be used with the knowledge and understanding of risks, situations and their effect on health to arrive at a decision for appropriate intervention. The comprehensive multidisciplinary approach to care should be a sound principle to adopt. It will ensure that all points of intervention are covered and appropriate preventive measures are taken through community outreach and other family health programs of the health agency. The lists of risk factors noted on the forms are not meant to be all inclusive. They are intended to cover the most frequent problems producing risk.  DEFINITIONS In general, the risk factors that will increase the chances of morbidity and mortality are of a physical, nutritional, mental/emotional, socio-economic or occupational nature. For the purpose of this guide, the following definitions have been adopted: risk:  an increased probability of adverse outcomes  high risk groups:  groups with increased probability of adverse outcomes  high risk families:  families whose circumstances indicate high risk factors which may interfere with optimum family life and functioning  high risk pregnancy:  a pregnancy in which the mother and/or the fetus has an increased probability of maternal and fetal morbidity or mortality prenatally and intranatally  high risk infant:  newborn or infant with familial maternal and perinatal factors that may lead to an increased probability of morbidity and subsequent disabilities  The risks are provided as a check list for coordinators to ensure they are discovering the risks that may -be encountered with the perinatal client. A brief description of each risk is provided to help understand the risk factors.  PHYSICAL FACTORS YES NO Previous pregnancy loss Illness/condition with impact on pregnancy Pre-pregnancy weight - body mass PF3 index (BMI) Rate of weight gain PF4 Inadequate nutrition PF5 Previous child with anomaly PF6 Previous child requiring PF7 neonatal intensive care Multiple pregnancy PF8 Birth interval PF9 PF10 Grand multipara - 5 or more pregnancies PF11 Established genetic risk PF12 Age 17 and younger/ 36 and older at time of delivery  PF1 PF2  PF1:  • • • • • • • • • • • •  • • • • • • • • • • • •  Previous pregnancy loss  Previous pregnancy loss - abortion (both spontaneous and elective), stillbirth, neonatal and infant death (up to 365 days old), such as SIDS are significant factors. Depending on the cause of such loss the same conditions may be either present or occur again for another reproductive loss. The level of risk depends on the causative factor.  A Guide for the Use of Individual Prenatal Risk Identification Tool  PF2:  Illness/condition with impact on pregnancy  The following conditions would lead to unfavourable outcome of pregnancy if close medical surveillance is not provided: poorly controlled diabetes or hypertension, chronic renal failure, congenital or rheumatic heart disease, and very rapid weight gain. Other conditions may have an impact on pregnancy if not controlled by routine medical care, eg. mild hypertension, gestational diabetes, and urinary tract infections. Many conditions may lead to premature labour, congenital anomalies, intrauterine growth retardation, and other associated morbidities. These include infections (rubella, STD, toxoplasmosis, genital herpes), abnormal presentation, surgical procedure during pregnancy, uterine and associated malformations, toxemia, anemia, bleeding, diabetes, hypertension, obesity, renal disease, isoimmunization, etc. The risk and its effects are related to the severity of the condition. Other conditions such as blindness, deafness and physical handicaps can affect the mother in pregnancy. The level of risk will depend on the individual's abilities, compensating mechanisms, and support structure.  PF3:  Pre-pregnancy weight  Body Mass Index (BMI) = wt (kg) ht (m ) 2  2  The underweight woman has a BMI under 19.8. A BMI of over 29 indicates obesity. A woman's nutritional status prior to and during pregnancy are important factors that influence the health of the fetus and the baby. The mother's pre-pregnancy weight and weight gain during pregnancy are two factors which affect the-infant's birth weight and thus the infant's health. "No widely accepted standards of weight for height exist for adolescents. Except for very young girls or those who conceive within 2 years of menarche, adult BMI recommendations may be used provisionally to classify girls as underweight, moderate weight, overweight and obese." Nutrition During Pregnancy. National Academy of Sciences. 1990. Note:  1 pound 1 inch 1 foot  = 0.45 kilograms s= 2.54 centimetres = .3048 meters  Page 2  PF4:  Rate of weight gain  Inadequate weight gain: 2nd and 3rd trimester ! • if weight gain less than 1 kg/month for women beginning pregnancy with an acceptable BMI (BMI = 19.8 - 26) • if weight gain is less than 0.5kg/month for obese women (BMI > 29) Rapid weight gain: 2nd and 3rd trimester • if weight gain is greater than 3 kg/month Measurement should be carefully evaluated to avoid measurement or recording errors, or differences due to " clothing, boots, shoes, etc. Inappropriate rate of weight gain may lead to low birthweight infants and related problems. Underweight women (BMI < 19.8) are certainly at risk if their weight gain is less than 1 kg/month and overweight women (BMI > 26-29) if their weight gain is less than 0.5 kg/month. The literature does not identify specific guidelines for these populations. Rapid weight gain may indicate fluid retention, multiple gestations, or excessive food intake. For the underweight woman (BMI < 19.8) with a weight gain > 3 kg/month, clinical judgement is required to determine whether this represents a health risk or is a result of 'catch-up' weight gain.  PFSt Inadequate Nutrition Consistently less than the minimum recommended servings in 1 or more food groups, as outlined in the "B.C. Food Guide for Pregnancy": less than 8 servings of Grain Products less than 6 servings of Vegetables and Fruit less than 3 servings of Milk Products less than 2 servings of Meat and Alternatives  The Baby's Best Chance: Parents' Handbook of Pregnancy and Baby Care provides essential information with regards to nutrition requirements for the pregnant woman. The "B.C. Food Guide for Pregnancy" outlines the appropriate numbers of food group servings for adequate calories and nutrients. A deficiency can represent a serious risk to the development of the fetus and to the mother's health. The assessment of the four food groups should be based on the client's reporting of her typical dairy intake. It is recommended that the consulting nutritionist be involved in the nutrition screening aspect of the initial interview.  PF6:  Previous child with anomaly or disorder  This includes conditions with impact on development of the child; eg. chronic heart disease, neural tube defects (i.e. spina bifida), cleft palate, fetal alcohol syndrome, fetal alcohol effects; and conditions which are more readily corrected or have only minor functional impairment, eg.  A Guide for the Use of Individual Prenatal Risk Identification Tool  Page 3  ventral-septal defects with spontaneous closure, minor orthopaedic abnormalities, uncomplicated pyloric stenosis, etc. Cerebral palsy, mental retardation, congenital anomalies ... if the same perinatal conditions still exist, they may lead to the same risk in the present pregnancy. Established genetic risk • either from previous pregnancies or from a familial history i.e., muscular dystrophy, cystic fibrosis, etc. is significant.  PF7:  Previous high risk infant  High risk infants that were premature (<37 weeks), postmature (>42 weeks), or had a low birthweight (<2500 grams).  SA1: Smoking Cigarette smoking has been shown to decrease infant birth weight in direct proportion to the amount smoked. Cigarette smoking increases the risks of perinatal morbidity and mortality. The growth-retarding effect of cigarette smoking and higher incidence of spontaneous abortions, stillbirths or placental complications among women who smoke during pregnancy may be due to several factors including direct toxicity of carbon monoxide, nicotine and/or other constituents of tobacco, reducing blood flow to the uterus affecting transfer of nutrients to the fetus, or suboptimal maternal food intake. Passive smoking (secondhand smoke) may also be a cause of concern during pregnancy due to the oxygen depleting effect of carbon monoxide.  SA2: PF8:  Multiple pregnancy  Prenatal mortality resulting from twin births is as high as 14%, the greatest mortality resulting from premature birth. Special emphasis should be placed on nutritional counselling for multiple pregnancy.  PF9:  Birth interval  Although the optimum birth interval has not been defined, the incidence of fetal growth retardation and prematurity is consistently high when the birth interval is less than two years. Spacing allows time for the mother's body to recover and to be in optimal health before becoming pregnant again.  Use of the T - A C E questions is recommended to determine the risk of alcohol misuse.  PF10: Grand multipara Parity alone or combined with maternal age is significant. Higher risk of morbidity occurs at the first pregnancy and at the fifth pregnancy or more.  Note:  PF11: Age 17 and underlage 36 and over at time of delivery Pregnant women 17 years of age and younger risk low birth weight infants. Pregnant women 36 years of age and over risk infants with chromosomal abnormalities.  SUBSTANCE ABUSE/MISUSE YES SA1 SA2 SA3 SA4  Cigarette smoking Alcohol use Inappropriate use of over the counter and prescription drugs Other drug use  Alcohol use  There is no known safe level of alcohol consumption for pregnant women. It is not possible at this time to say what is the minimum level of alcohol consumption that may endanger the fetus. Heavier alcohol misuse (such as maternal dependency) may lead to the fetal alcohol syndrome: low birth weight, failure to thrive, mental handicap, facial congenital anomalies, developmental delays, hyperactivity, etc. Alcohol (2 or more drinks per day or binge drinking) and other drug use (including tobacco and cocaine), may independently increase the risk of spontaneous abortion and low birth weight infants. When combined, fetal risk is greatly increased.  • • • •  NO  • • • •  1 Drink  Binge  SA3:  = 12 oz beer = 5 oz wine = 1 mixed drink (1.5 oz. or 'hard' liquor) = consuming 5 or more alcoholic drinks on any one occasion  Inappropriate use of over the counter and treatm drugs  Drugs may affect the intake, absorption, metabolism and/or utilization of nutrients in the body, thereby influencing maternal nutrition status. The effect that a drug has oh the fetus depends on many factors including the type of drug, the amount taken by the mother, the stage of pregnancy at which it is taken, and the frequency and duration of its use. Some drugs are known to have or strongly suspected of having any teratogenic effect in humans. Women should discuss with their family physician before taking any medications. Determine the pregnant woman's use of any drugs, including the use of herbs.  SA4:  Other drug use (including cocaine, opiates, solvents, Statistically, pregnancy complications occur more frequently in unmarried than in married women. The increased and poly-drug use)  Any needle drug use, any use of cocaine or crack, poly drug use, daily use of other drugs, for example tylenol #3 (codeine), hash, marijuana is to be considered a significant risk to the infant.  PSYCHOSOCIAL & ECONOMIC FACTORS Social Environment: The effects of maternal social environment on the outcome of pregnancy are recognized to be both multiple and profound. 'Social environment* is described as the summation of numerous factors, including the family's standards of health and hygiene, housing and financial status, emotional and social support and so on. The effects may be direct or indirect and may be difficult to separate within the context of socio-economic status. It is the inter-relationship of these factors, rather than any single factor, that works to affect the outcome of the pregnancy. , ' ;  PE1 PE2 PE3  PE12 PE13  Single parenthood Delayed access to prenatal care Refusal of/resistance to appropriate services Isolation - ethnic, language and social Limited learning ability/illiterate Marital problems/unstable relationship/family violence Mental health problems Low self-esteem Inability to cope/anxiety regarding pregnancy and baby Unrealistic expectations Unwanted pregnancies/denial of pregnancy Financial problems Inadequate housing  PE1:  Single parenthood  PE4 PES PE6 PE7 PE8PE9 PE10 PE11  YES NO  • • • • • • • • LI- • LT • • • • • • • • • • • • • • •  The frequency of cases of low birth weight infants and the perinatal mortality rates of infants born to unmarried mothers is higher than those of children of married women. Marital status alone is not necessarily an indicator of potential risk for mother and fetus so much as it is an indicator of an unwanted/unplanned pregnancy. These pregnant women, especially if unwed or teenagers, tend to neglect antenatal care and leave advice unheeded.  amount of risk can be associated with multiple social problems. Single parenthood still has an influence, but to a decreased amount, if financial and emotional support is present. PE2:  Delayed access to prenatal care  Early access to medical care and return follow-up visits are essential forriskidentification and monitoring. Some of the factors to consider are no medical care by 20 weeks, frequent missed appointments, no follow-up on medical advice and no attendance at prenatal classes in a primipara. PE3:  Refusal of/resistance to appropriate services  Refusal of or resistance to appropriate services, such as Ministry of Social Services, poses obvious threats to the client's receiving appropriate medical care and support for the mother and the fetus. This refusal or resistance can be due to a lack of trust on the part of the pregnant woman due to past experiences within her family or community or previous requests for help may have been unmet in the past. PE4: Isolation—Ethnic, language, social, and/or geographical  Ethnic or language isolation can tend to deprive mothers of available information and resources. This can apply to immigrant and refugee status women as well as Aboriginal women. Social isolation i.e., lack of supports, possibly new to area, can create a void in resources, either classes or physicians, which can put a mother at risk of not being assessed early and receiving adequate care and attention. Social isolation in itself is a stress and must be dealt with in conjunction with the stress of pregnancy. Geographic isolation can be an issue in remote areas as well as for mothers with limited transportation options and the location of facilities and programs. PES:  Limited learning ability /illiterate  Limited learning ability/Illiteracy especially if associated with other risks is significant. Problems can range from severe communicative disability to a limited ability to understand. These people may not have access to information nor an understanding of the importance of education regarding pregnancy, childbirth and child care. PE6:  Marital problems/unstable relationship(family violence  Marital problems/unstable relationship: Marital discord, lack of partner support, lack of extended family support may lead to a higher incidence of reproductive loss, low birth weight (preterm, small for dates) nutritional problems, absence of maternal child bonding, neglect and abuse resulting in developmental delays and other associated morbidities.  A Guide for the Use of Individual Prenatal Risk Identification Tool  Page 5  Family violence/abuse: Determine if the woman is currently in an abusive relationship, if there are affects on the emotional or physical health of the woman, or if there is a possibility of repetition during pregnancy or shortly thereafter.  difficulty accepting pregnancy and developing a relationship with the growing fetus may present with extreme anxiety about the condition of the baby and will be hypervigilant in looking for signs that 'something is wrong' with the pregnancy.  Evidence of neglect - history of abuse/neglect, for example lack of positive parenting in the past, history of negative foster home placements.  PE10: Unrealistic expectations Unrealistic expectations of roles of mother and or father, baby and significant others can lead to frustration, stress, neglect and abuse. Another psychosocial maladaptation of pregnancy is failure to make adequate, concrete plans for postnatal care of the baby. The absence of family members or friends to assist in the care of the baby or, at the other extreme, passivity and over reliance on family members are signs of difficulty in adapting to pregnancy, as is unrealistic planning or inadequate preparation for managing the baby at home. -  A family history of abuse/neglect (emotional or physical) tends to repeat itself from generation to generation and where there is abuse present in the home, the new baby is in high risk of being abused and neglected. PE7: Mental health problems Mental health problems, current and previous occurrence(s), may shed light on one's family background, coping mechanisms, self-esteem and reactions to stress or crisis. As the pregnant woman strives to develop a degree of comfort with the many changes in social context and psychologic equilibrium, there often occurs a surfacing of old conflicts that were never adequately resolved in earlier developmental periods. For example, pregnant clients may experience conflicts of autonomy with their mothers, renewed rivalry with siblings, or active uncertainty about sexuality and disturbing fantasies about past relationships, each of which had been adequately dealt with prior to pregnancy but which now result in troubling family interactions or marital discord. Manifest problems in adjustment prior to pregnancy, such as marital discord, economic difficulties, poor self-concept, and neuroticism may be exacerbated by pregnancy. Anxiety allowed to go unallayed may lead to maladaptive mother-child interaction. PE8: Low self-esteem Low self-esteem can manifest itself in a pregnant woman having no confidence in herself, her body, her decisionmaking choices. Exhibition of depression, lack of self-worth or motivation, and uncaring of self and other people. She may even choose to be in an abusive relationship or refuse to avail herself of advice and information.  PEU: Unwanted pregnancy /denial of pregnancy Pregnant women who have an unwanted pregnancy or unplanned pregnancy and/or who deny the pregnancy, can tend to neglect antenatal care and leave advice unheeded. The stresses in these women are very high. PE12: Financial problems Unemployment, very low income, and/or receiving social assistance may lead to a higher incidence of reproductive loss, low birth weight, nutritional problems, neglect and abuse resulting in developmental delays and other associated morbidities. PE13: Inadequate housing While this can be a difficult risk to assess, some of the features to be considered may be: lack of facilities (bathroom, cooking, bedroom, etc.), space/overcrowding, hazardous living conditions, pest infestation, etc. For 'street people', this is a significant risk, as well as for others with an unstable functional household unit - where there is significant moving of the family and/or many people coming and going out of the house. This can be a high stress factor for the pregnant woman and her family.  PE9: Inability to cope/anxiety regarding pregnancy and baby Coping potential is the ability of the individual and family to adapt to stress. When individuals experience stress, they may use a variety of methods to cope. With an intense perception of threat, defense mechanisms such as denial, projection, rationalization, displacement and intellectualization may occur. The prolonged denial of the high-risk status of the pregnancy may result in failure to comply with therapeutic regimes. Anxiety regarding the pregnancy and baby may manifest itself in many expressed irrational fears and distortions. Women who are having  Acknowledgements to: Ottawa Health Department and Ontario Ministry of Health FORM 5070 REV88 OCTOBER NA-RISK2.CHP  . 177  APPENDIX B  T-ACE Measurement T - A C E is a measurement tool of four questions that are significant identifiers of risk drinking (i.e., alcohol intake sufficient to potentially dame the embryo/fetus). For the Pregnancy Outreach Program the T - A C E is completed at intake. The T - A C E score has a range of 0-5. The value of each answer to the four questions is totalled to determine the final T - A C E score.  1.  How many drinks does it take to make you feel high? 0 2  2.  Have people annoyed you by criticizing your drinking? 0 1  3.  no yes  ^  nnoyance  Have you felt you ought to cut down on your drinking? 0 1  4.  olerance  less than or equal to 2 drinks more than 2 drinks  no yes  Have you ever had a drink first thing in the morning to steady your nerves or to get rid of a hangover? 0 1  Source:  0 ut Down  {] ye Opener  no yes  Sokol, Robert J., "Finding the Risk Drinker in Your Clinical Practice" in  Alcohol and Child/Family Health: Proceedings of a Conference with Particular Reference to the Prevention of Alcohol-Related Birth Defects, edited by Robinson, G . and Armstrong R., Vancouver, B . C . , December 1988. Note:  For the purposes of the Pregnancy Outreach Program Evaluation - a client is at risk for alcohol use if she has a positive T - A C E (a score of 2 or greater).  Rev93/06/24  179  APPENDIX C  Criteria for Adequacy of Care Index Levels* , Adequacy ol Care  Tnmesier o( First Prenatal Visit  Adequate  First (1-3 Months)  Less-tnan-Adequate  All Other Combinations  Gestation (Weeks) 13 or less 14-17 18-21 22-25 26-29 30-31 32-33 34-35 36 or more  Number ol Prenatal Visits 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or  more or not stated more more more more more more more more  * Source: Peoples et al, (1984). Evaluation of the Effects of the N o r t h Carolina Improved Pregnancy Outcome Project: Implications for State-Level DecisionMaking. A J P H , 74(6), 549-554.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0098977/manifest

Comment

Related Items