UBC Faculty Research and Publications

Psychological health of family caregivers of children admitted at birth to a NICU and healthy children:… Klassen, Anne F; Lee, Shoo K; Raina, Parminder; Lisonkova, Sarka Dec 14, 2004

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

Item Metadata


52383-12887_2004_Article_51.pdf [ 541.68kB ]
JSON: 52383-1.0221626.json
JSON-LD: 52383-1.0221626-ld.json
RDF/XML (Pretty): 52383-1.0221626-rdf.xml
RDF/JSON: 52383-1.0221626-rdf.json
Turtle: 52383-1.0221626-turtle.txt
N-Triples: 52383-1.0221626-rdf-ntriples.txt
Original Record: 52383-1.0221626-source.json
Full Text

Full Text

ralssBioMed CentBMC PediatricsOpen AcceResearch articlePsychological health of family caregivers of children admitted at birth to a NICU and healthy children: a population-based cross-sectional surveyAnne F Klassen*†1, Shoo K Lee†2, Parminder Raina†3 and Sarka Lisonkova†2Address: 1Centre for Community Child Health Research, Department of Pediatrics, University of British Columbia, Vancouver, BC, Canada, 2Centre for Healthcare Innovation and Improvement, Dept of Pediatrics, University of British Columbia, Vancouver, BC, Canada and 3Evidence-Based Practice Centre, Dept of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, ON, CanadaEmail: Anne F Klassen* - afk@interchange.ubc.ca; Shoo K Lee - shool@interchange.ubc.ca; Parminder Raina - praina@mcmaster.ca; Sarka Lisonkova - slisonkova@cw.bc.ca* Corresponding author    †Equal contributorsAbstractBackground: There is little information in the research literature on how parents of children whospend time in a neonatal intensive care unit (NICU) adapt psychologically to the demands ofcaregiving beyond the initial hospitalization period. Our aim was to compare parents of NICUchildren with parents of healthy full-term children, looking specifically at the relationship betweenparental psychosocial health and child characteristics, as well as the relationship between importantpredictor variables and psychosocial health.Methods: A cross-sectional survey was sent to parents as their child turned 3 1/2 years of age.The setting was the province of British Columbia, Canada. The sample included all babies admittedto tertiary level neonatal intensive care units (NICU) at birth over a 16-month period, and aconsecutive sample of healthy babies. The main outcome was the SF-36 mental componentsummary (MCS) score. Predictor variables included caregiver gender; caregiver age; marital status;parental education; annual household income; child health status; child behavior; birth-related riskfactors; caregiver strain; and family function.Results: Psychosocial health of NICU parents did not differ from parents of healthy children. Childhealth status and behavior for NICU and healthy children were strongly related to MCS score inbivariate analysis. In the pooled multivariate model, parental age, low family function, high caregiverstrain, and child's internalizing and externalizing behavioral symptoms were independentlyassociated with lower psychosocial health. In addition, female gender was associated with lowerpsychosocial health in the NICU group, whereas lower education and child's problem with qualityof life indicated lower psychosocial health in the healthy baby group.Conclusions: Overall, parental gender, family functioning and caregiver strain played influentialroles in parental psychosocial health.Published: 14 December 2004BMC Pediatrics 2004, 4:24 doi:10.1186/1471-2431-4-24Received: 29 June 2004Accepted: 14 December 2004This article is available from: http://www.biomedcentral.com/1471-2431/4/24© 2004 Klassen et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Page 1 of 11(page number not for citation purposes)BMC Pediatrics 2004, 4:24 http://www.biomedcentral.com/1471-2431/4/24BackgroundNeonatal intensive care is associated with a range of long-term health problems such as cerebral palsy, mental retar-dation, deafness, blindness and milder but more commonproblems such as learning disabilities and behavioralproblems [1-13]. Although these problems create chal-lenges for the parent responsible for the day-to-day provi-sion of care to their child at home, the impact ofcaregiving on the health of parents of children dischargedfrom neonatal intensive care units (NICUs) remains anunder-explored research topic. There is a literature thatfocuses on the early hospitalization period. These studiesshow that mothers of preterm infants experience moresevere levels of psychological distress in the neonatalperiod than do mothers of healthy full-term infants [14-17]. In the few studies that compare the impact of caregiv-ing on parents of children discharged from NICUs withparents of healthy full-term children, the addition of apreterm infant into the family has been shown to havenegative repercussions for the family in some studies [18-21], but not in others [22-24]. In one of the few NICUstudies where parental mental health was the primary out-come measure, mothers of high and low-risk very lowbirth weight infants were compared with parents ofhealthy full-term infants [17]. The authors report thatearly differences between the groups at one month andtwo years were no longer apparent by the age of three,although parenting stress remained high throughout.In the present study, we sent a questionnaire booklet tomothers of all children admitted to a level III neonatalintensive care unit in the province of British Columbia(Canada) over a 16 month period to collect data on arange of factors in order to examine both neonatal andcaregiver outcomes. Our study differs from other NICUfollow-up studies in that it is population-based, focuseson preschool aged children and examines the full spec-trum of NICU graduates. The aims of this paper are two-fold: (1) to compare psychosocial health of parents ofNICU children with parents of healthy full-term children,looking specifically at the relationship between parentalpsychosocial health and child characteristics (i.e., healthstatus, behavior problems, and birth-related risk factors);and (2) to identify predictors of parental psychosocialhealth (i.e., socioeconomic and demographic variables,child characteristics, caregiver strain, and familyfunction).MethodsSampleEthical approval was gained from the University of BritishColumbia and participating hospitals. Our NICU sampleincluded 2221 surviving babies admitted for more thanJune 1997). These 3 hospitals (Children's and Women'sHealth Centre of BC, Royal Columbian Hospital, VictoriaGeneral Hospital) provided 100% of the tertiary careNICU beds in the province. The birth mothers' name andcontact details were obtained from the health recordsdepartment at two hospitals and manually extracted fromledgers of the third hospital. Our list of babies wasmatched with provincial mortality records to exclude anybabies that had died after discharge from the NICU andthereby prevent questionnaires being sent to bereavedparents.A comparison group of 718 healthy singleton full-termbabies was recruited from the two hospitals with a hospi-tal-based primary care unit (i.e., Children's and Women'sHealth Centre of BC and the Royal Columbian Hospital).This sample included all babies delivered over an 11month period (March 1996 and January 1997) by pri-mary care physicians at these two clinics. Babies with asibling in the NICU sample and babies subsequentlyadmitted to an NICU for more than 24 hours wereexcluded. Contact details for the mother were obtainedfrom the health records department at one hospital, anddirectly from the primary care unit at the other.We excluded from the sample 150 babies (123 NICU; 27healthy children) who did not meet our inclusion criteriafor the following reasons: parent did not speak English (n= 95); baby died (n = 34); mother died (n = 6); and notapplicable (n = 1). In addition, we excluded cases wherethe questionnaire was completed on the wrong child (n =7) and where a comparison baby was subsequently admit-ted to a NICU (n = 7). The overall response rate (afterexclusions), was 55% (54.3% NICU, 56.9% healthy babygroup). The response rate for located families (82.8% ofthe sample was located) was 67.4% (n = 1140) for theNICU group, and 66.4% (n = 393) for the comparisongroup. Five NICU respondents returned a signed consentform without a completed questionnaire and weredropped from the analysis. Seventy-five percent of parentsprovided permission for data linkage between the ques-tionnaire data and CNN database. The NICU sampleincluded 181 children that were part of a multiple birthgroup: 171 twins; and 10 triplets. Table 1 contains samplecharacteristics. Most questionnaires (98%) were com-pleted by a biological parent, most often the mother(96%). The NICU sample was composed of 1.8% fewerbiological parents; 2.6% more male respondents, and11.9% more families who earned less than $50,000 peryear.MaterialsOur main measure of outcome was the SF-36 mental com-Page 2 of 11(page number not for citation purposes)24 hours to one of three level III NICUs in British Colum-bia (BC), Canada over a 16 month period (March 1996 toponent summary (MCS) score [25,26]. The SF-36 is a wellvalidated generic measure of adult physical andBMC Pediatrics 2004, 4:24 http://www.biomedcentral.com/1471-2431/4/24psychosocial health related quality of life (HRQL), whichis composed of 36 items that measure 8 health domains.The MCS is computed from the following four domains:mental health (5 items); vitality (4 items); social function-ing (2 items); and role limitations due to emotional prob-lems (3 items). It has a mean of 50 and standard deviationof 10 and represents the mean and standard deviation ofthe general population (USA).Child health status was measured using the Health StatusClassification Preschool Version (HSCS-PS) [27]. Thismeasure asks about twelve health status (HS) problemsthat we have grouped into the following 4 categories: neu-rosensory (i.e., seeing and hearing); motor development(i.e., getting around, using hands and fingers, taking careof self); learning (i.e., speaking, learning/rememberingand thinking/solving problems); and quality of life (i.e.,pain/discomfort, feelings, behavior and general health).Each attribute has 3 to 5 levels of severity ranging fromnormal function to severe functional limitations. For eachcategory of health problems, we recoded the data into theChild behavior was measured with the Child BehaviorChecklist 1.5–5 (CBCL/1.5–5) [28]. This questionnairemeasures internalizing, externalizing and total problems,and scales can be scored categorically to indicate normal,borderline or clinical range scores.Data for birth-related risk data were obtained from theCanadian Neonatal Network Study [29] for the NICUchildren whose parents provided written consent for datalinkage. The following variables were examined: birth-weight; gestational age; small for gestational age; multiplebirth, apgar score less than 7 at 5 minutes; congenitalanomalies; the presence of a major morbidity (i.e., a com-posite score for the presence of at least one of the follow-ing: chronic lung disease (at 36 weeks); severeintraventricular hemorrhage (≥ grade 3); nosocomialinfection; necrotizing enterocolitis; retinopathy of prema-turity (≥stage 3)); and neonatal illness severity score [30].Caregiver strain was measured using the Parental Impact-Time (PTT) scale from the Infant Toddler Quality of LifeTable 1: Characteristics of study sampleGroup; no. (%) of subjectsNICUN = 1135ComparisonN = 393Biological parent1 1091 (97.7) 389 (99.5)Female1 1070 (95.4) 383 (98.0)Married/common-law 962 (85.9) 344 (87.8)Age of parent, years19–29 195 (17.8) 61 (15.7)30–39 704 (64.1) 265 (68.3)≥ 40 199 (18.1) 62 (16.0)Education levelUniversity 373 (33.4) 146 (37.4)Trade/technical school or community college 494 (44.3) 176 (45.1)High school graduation 185 (16.6) 50 (12.8)No high school diploma 64 (5.7) 18 (4.6)Household income, $2 511 (48.1) 136 (3)<30,000 247 (23.3) 58 (15.5)30 – 49,999 264 (24.9) 78 (20.8)50 – 79,999 333 (31.4) 145 (38.7)80 > 218 (20.5) 94 (25.1)Male children in the sample 633 (55.8) 198 (50.6)Age of child, years3 years 784 (69.3) 253 (64.4)4 years 328 (29.0) 134 (34.1)5 years 19 (1.7) 6 (1.5)1p < .05 (chi-square, Fischer's exact test); 2p = .0018 (chi-square)Page 3 of 11(page number not for citation purposes)following: no problem; a mild problem; or a moderate orsevere problem.Questionnaire [31]. This 7-item scale asks parents to indi-cate limitations in the amount of time in the past 4 weeksBMC Pediatrics 2004, 4:24 http://www.biomedcentral.com/1471-2431/4/24they had for their own personal needs due to problemswith their child's health (e.g., physical, emotional, cogni-tive, behavior, temperament). Scores on these scales canrange from 0 to 100, with lower scores indicating greatercaregiver strain.Family function was measured using the Family Assess-ment Device (FAD) [32]. Scores for this 12-item question-naire can range from 0 to 36, with higher scores indicativeof greater family dysfunction.ProcedureA questionnaire booklet, which included the question-naires described above, was sent to the address of the birthmother as her child turned 3 1/2 years of age. A consentletter was included to obtain permission to link the ques-tionnaire data with hospital birth records. The primarycaregiver in our study was defined as the person who, tothat point in the child's life, had spent the most amountof time with the child. This could include the mother orfather or another parent (e.g., grandparent, foster parent,guardian). We asked the primary caregiver (referred to inthis paper as parent) to complete the questionnaire book-let and consent form. Non-respondents were sent areminder letter, additional copies of the questionnairebooklet and a phone call as necessary. If the telephonenumber was not in service or reassigned, or a question-naire booklet was returned to us from the post office, weimplemented a comprehensive search strategy thatinvolved searching the Internet and contacting the moth-ers' primary care physician.Data analysisTo address the first objective, we compared the psychoso-cial summary score for the SF-36 questionnaire for parentsof NICU children and parents of healthy children usingstudent's T-test. T-test, ANOVA and the equivalent non-parametric tests, and Spearman correlation were used toexplore relationships between MCS score and variouschild characteristics, including health status, behavior andbirth-related risk factors. For health status and childbehavior, we computed an effect size (mean differencedivided by standard deviation of the group with no prob-lems (health status) or with scores in the normal range(behavior)), to look at the magnitude of the difference inMCS score between subgroups for the NICU and healthybaby samples, and used the Cohen's guidelines for inter-pretation (0.2 is small, 0.5 is medium, 0.8 is large) [33].To address the second objective, multiple regression anal-ysis was used to examine the independent effects of, andproportion of variance in MCS scores explained by ourpredictor variables. For the analysis we examined a pooledexamine the contribution of each predictor variable forthe two samples. Variables with significant (p < .05) orborderline p-values in bivariate analysis were included inthe model. Certain birth-related risk factors (i.e., birth-weight, congenital anomalies, illness severity score, andgestational age) were entered into the model on the basisof clinical rather than statistical importance, however, noeffects were found. Potential predictor variables includethe following: caregiver's gender; caregiver's age (continu-ous); marital status (married or common-law versusother), caregiver's education (less than high school gradu-ation vs. other); annual household income (< or >$30,000); child health status (i.e., neurosensory; motordevelopment; learning; and quality of life problems);child behavior; caregiver strain (continuous); and familyfunction (continuous). For child health status and behav-ior variables, no problem (health status) and scores in thenormal range (behavior) were the reference categories,with mild and moderate/severe (health status) or border-line and clinical range scores (behavior) entered sepa-rately, or combined and entered as dichotomousvariables. We computed effect sizes to interpret the signif-icance of beta coefficients.ResultsPsychosocial health comparing NICU and healthy childrenThe unadjusted mean MCS score for parents of NICU chil-dren did not differ from parents of healthy children (48.2versus 48.8; p = .305). We also compared MCS scores afteradjusting for the three sample characteristics that differedbetween the two groups (i.e., proportion of biological par-ents; gender of subject; and those with lower householdincome), and no differences were found in the outcomevariable.Psychosocial health by child health status problemOn the HSCS-PS, 55.2% of healthy children had no healthproblems in any area, compared with 39.8% of NICUchildren (p < .001 on Chi-square). Table 2 shows the jointdistribution of health status problems across the four cat-egories for the NICU and healthy sample. These resultsshow that the NICU sample had a higher proportion ofchildren with more health status problems, as well as ahigher proportion with moderate/severe versus mildproblems.For parents of NICU children, for all 4 health status cate-gories, parental MCS scores decreased as severity of thechild health problem increased (see Table 3). Effect sizescomparing parents of children with no health status prob-lems with parents of children with a moderate or severehealth status problem were all moderate to large indicat-ing important differences in parental mental healthPage 4 of 11(page number not for citation purposes)model and a model where we stratified by group member-ship (i.e., NICU vs. healthy baby sample) to separatelyaccording to Cohen's benchmarks. The results for parentsBMC Pediatrics 2004, 4:24 http://www.biomedcentral.com/1471-2431/4/24of healthy children show similar trends, with mainlymoderate to large effect sizes.Psychosocial health by child behavior problemChild behavior was strongly related to parental psychoso-cial health in both groups of parents (see Table 4). Parentswhose child scored in the clinical range for internalizingand externalizing symptoms and the total problem scoreon the CBCL/1.5–5 had the lowest mean (i.e., poorest)MCS scores. The differences between this group and thegroup with children scoring in the normal range resultedParental psychosocial health by birth-related risk factorsWithin the NICU sample, MCS score did not vary by anybirth-related risk factor (i.e., gestational age; small for ges-tational age; apgar score; multiple birth; the presence of amajor morbidity; and neonatal illness severity score),with the exception of the presence of a congenital anom-aly. For this variable, MCS scores were significantly lowerin parents of children with versus without a congenitalanomaly (mean difference = -3.8; p = .017; effect size = -.37). Children with a congenital anomaly (n = 87) hadproportionally more mild and moderate/severe healthTable 2: Distribution of children with health status problems across the 4 health status categories for NICU and healthy childrenHSCS problems by domains NICU (N = 1104) Comparison (N = 386)no problem N 438 215% 39.67 55.71 mild problem N 309 111% 27.99 28.762+ mild problems N 183 37% 16.58 9.591 moderate/severe problem only N 40 7% 3.62 1.811 moderate/severe problem + any mild N 69 15% 6.25 3.892–3 moderate/severe problems N 60 0% 5.43 04 moderate/severe problems in all domains N 5 1% 0.45 0.26p < 0.0001, chi-squareTable 3: Parental mental health summary score, 95% confidence intervals, number of subjects, p-value and effect size for child health status categorySample Type of HS problem None Mild Moderate or Severe p-value* Effect sizeNICU children Neurosensory 48.4 (47.7, 49.0) n = 975 46.3 (41.7, 51.0) n = 33 41.9 (36.0, 47.8) n = 17 .023.040.63Motor development 49.1 (48.5, 49.8) n = 789 45.8 (44.0, 47.6) n = 174 42.0 (38.8, 45.2) n = 63 <.001<.001.74Learning/remembering 49.3 (48.6, 50.0) n = 623 47.5 (46.3, 48.8) n = 298 43.4 (40.9, 45.9) n = 109 <.001<.001.63Quality of life 49.8 (49.0, 50.5) n = 659 46.3 (45.1, 47.4) n = 313 39.4 (35.7, 43.2) n = 59 <.001<.0011.11Healthy childrenNeurosensory 48.9 (47.9, 49.8) n = 361 57.0 n = 1 31.1 n = 1 .120.1541.89Motor development 49.2 (48.2, 50.3) n = 333 45.6 (42.6, 48.7) n = 31 37.9 (22.6, 53.2) n = 3 .018.0021.18Learning/remembering 49.6 (48.5, 50.7) n = 266 46.9 (44.7, 49.1) n = 89 45.8 (39.1, 52.5) n = 14 .037.025.41Quality of life 50.1 (49.0, 51.2) n = 271 45.6 (43.5, 47.8) n = 86 36.2 (23.8, 48.6) n = 8 <.001<.0011.58* first based on Anova, second based on Kruskal-Wallis non-parametric test (in italics)Page 5 of 11(page number not for citation purposes)in large effect sizes, indicative of clinically important dif-ferences in parental psychosocial health.status problems in all 4 categories (see Table 5).BMC Pediatrics 2004, 4:24 http://www.biomedcentral.com/1471-2431/4/24Correlates of psychosocial health in generalIn general, variables significantly associated with the MSCscore in bivariate analysis were as follows: any health sta-tus problems (mean difference = -3.8; p < .001);neurosensory problems (mean difference = -3.7; p =0.04); motor development problems (mean difference = -4.4; p < .001); learning/remembering problems (meandifference = -2.9; p < .001); poorer quality of life (meandifference = -4.8; p < .001); more internalizing behavioursymptoms (mean difference = -8.3; p < .001); moreexternalizing behavior symptoms (mean difference = -9.9;p < .001); household income below $30,000 per year(mean difference = -2.6; p < .001); female gender (meandifference = -2.6; p < .001);not living as common-law ormarried (mean difference = -3; p = .03); more caregiverstrain (r = .41; p < .001); and lower family function (r = -.44; p < .001). Borderline significance was also found forless than high school education (mean difference = -2; p =.08).We examined a pooled model (both groups together) fora direct comparison of the NICU and healthy groups afteradjustment for other variables. Due to the low number ofPredictors significantly associated with the outcome werethe following: parental age (Beta = 0.15; p = 0.001); inter-nalizing behavior (Beta = -2.06; p = 0.017); externalizingbehavior (Beta = -3.24; p = 0.004); parental strain (Beta =0.15; p < 0.001); and family function (Beta = -0.53; p <0.001). The pooled model also showed an interactioneffect between NICU admission and education (less thanhigh school) (Beta-education = -5.94 with p = 0.009; Beta-interaction = 7.28 with p = 0.005)(see Table 6.) For theNICU group, education did not show any effect in termsof difference in outcome, but for the healthy group, lowereducation was associated with a significantly lower meanMCS score. More specifically, for respondents with lessthan high school education, the healthy group reportedlower MCS scores than did the NICU group. The resultswere not affected by exclusion of multiple births and casesof congenital anomalies from the analysis.Although other interaction terms with NICU status didnot add any more significant results in the pooled model(non-significant partial F-test), we examined separatemodels for the NICU and the healthy baby group to fur-ther explore the association between gender and MCSTable 4: Mean score, p-value and effect size for SF-36 psychosocial summary score comparing CBCL/1.5–5 normal with borderline and clinical groupsCBCL scale Normal Borderline Clinical p-value Effect sizeNICU sample Internalizing 49.5 (48.9,50.2) n = 841 42.5 (39.6,45.5) n = 67 40.5 (37.6,43.4) n = 78 <.001 .95Externalizing 48.9 (48.3,49.6) n = 925 41.6 (38.0,45.2) n = 45 35.0 (29.6,40.2) n = 32 <.001 1.43Total 49.3 (48.6,49.9) n = 831 43.3 (39.7,47.0) n = 34 35.3 (31.0, 39.5) n = 39 <.001 1.45Healthy children Internalizing 49.6 (48.6,50.6) n = 324 40.7 (36.8,44.6) n = 20 40.1 (34.0,46.2) n = 16 <.001 1.03Externalizing 49.3 (48.3,50.3) n = 342 43.4 (37.0,49.8) n = 14 34.1 (22.5,45.8) n = 8 <.001 1.67Total 49.4 (48.4,50.4) n = 330 41.5 (34.0,49.0) n = 12 36 (27.0, 44.0) n = 11 <.001 1.46p-value based on Anova, (non-parametric tests: all p-values < .001)Table 5: Number (%) of NICU children with and without a congenital anomaly to report a problem for each health status category and p-value for Chi-square test of significanceType of HS problem Congenital anomaly None Mild Moderate or Severe p-valueNeurosensory No 715 (95.6) 25 (3.3) 8 (1.1) <.001Yes 70 (83.3) 8 (9.5) 6 (7.1)Motor development No 584 (78.3) 126 (16.9) 36 (4.8) <.001Yes 48 (57.1) 17 (20.2) 19 (22.6)Learning/remembering No 457 (60.9) 222 (29.6) 71 (9.5) <.001Yes 36 (42.9) 30 (35.7) 18 (21.4)Quality of life No 481 (64.2) 233 (31.1) 35 (4.7) <.001Yes 38 (44.2) 34 (39.5) 14 (16.3)Page 6 of 11(page number not for citation purposes)male respondents in the healthy group, we restricted thepooled multivariable analysis to only female respondents.score, and to evaluate the potential influence of congeni-tal anomalies in NICU group.BMC Pediatrics 2004, 4:24 http://www.biomedcentral.com/1471-2431/4/24Correlates of psychosocial health for NICU sampleVariables that were significantly associated with lowerMCS scores at the bivariate level include the following:female caregivers (mean difference = -3.2; p = .037);household income below $30,000 per year (mean differ-ence = -3.3 and p < .001); not living as common-law ormarried (mean difference = -5.1; p < .001); neurosensoryproblems (mean difference = -6.44; p = .011); motordevelopment problems (mean difference = -7.1; p < .001);learning/remembering problems (mean difference = -5.9;p < .001); poorer quality of life (mean difference = -10.4;p < .001); more internalizing behaviour symptoms (meandifference = -9.03; p < .001); more externalizing behaviorsymptoms (mean difference = -13.9; p < .001); the pres-.017); more caregiver strain (r = .411; p < .001); and lowerfamily function (r = -.441; p < .001).Predictors that were significant in the final regressionmodel appear in Table 7. Female gender was an independ-ent risk factor for lower MCS score: females scored onaverage 5.3 points (CI interval 2.5 to 8.0) lower, whichrepresents a moderate effect size of 0.51 (when overallNICU parents group standard deviation (SD) 10.4 forMCS was used as the denominator). Scoring outside thenormal range for internalizing and externalizing childbehavior symptoms independently contributed to lowerMCS scores (-1.9 and -2.8, both with wide confidenceintervals), with the change representing small effect sizesTable 6: Beta coefficients, 95% confidence intervals, standardized beta coefficients and p-values for predictor variables in the multiple regression models for pooled modelPooled modelVariable Beta CI-low CI-high St. beta p-valueIntercept 34.21 30.26 38.17 <.0001NICU -0.38 -1.32 0.55 -0.02 0.500Parental age 0.15 0.08 0.23 0.08 0.001Education -5.94 -9.67 -2.21 -0.13 0.009Internalizing behavior -2.06 -3.47 -0.65 -0.07 0.017Externalizing behavior -3.24 -5.08 -1.40 -0.08 0.004Caregiver strain 0.15 0.13 0.18 0.26 <.0001Family function -0.53 -0.60 -0.46 -0.32 <.0001NICU-education interaction7.28 3.01 11.56 0.14 0.005Table 7: Beta coefficients, 95% confidence intervals, standardized beta coefficients and p-values for predictor variables in the multiple regression models for both samplesNICU sample Healthy baby sampleVariable Beta CI-low CI-high St. beta p-value Beta CI-low CI-high St. beta p-valueIntercept 44.9 40.3 49.5 <.001 35.2 26.3 44.1 <.001Parental age 0.3 0.1 0.5 0.1 .005Female gender -5.3 -2.6 -8.0 -0.11 <.001Education -5.00 -0.8 -9.1 -0.1 .019Internalizing behavior-1.9 -0.1 -3.8 -0.06 .043 -4.0 -0.8 -7.2 -0.1 .014Externalizing behavior-2.8 -0.4 -5.3 -0.07 .025Caregiver strain0.2 0.1 0.2 0.26 <.001 0.1 0.04 0.2 0.2 .003family function -0.5 -0.4 -0.6 -0.32 <.001 -0.6 -0.4 -0.8 -0.4 <.001Quality of life -6.9 -0.4 -13.4 -0.1 .039Page 7 of 11(page number not for citation purposes)ence of a congenital anomaly (mean difference = -3.8; p = of 0.18 and 0.27. More caregiver strain (i.e., lower PTT)was related with poorer MCS scores. A one point changeBMC Pediatrics 2004, 4:24 http://www.biomedcentral.com/1471-2431/4/24in PTT corresponded to a 0.15 (CI: 0.11–0.19) change inMCS score. In NICU parents, the mean PTT was 86.9 andSD was 18.5. Therefore, 2 SD on the PTT would represent5.5 points on the MCS, or an effect size of 0.53. The meanscore for family function (FAD) was 8.1 and the SD was6.4. A one point change in FAD corresponded to a 0.5 (CI:0.62; 0.42) change in MCS. Therefore a 2 SD increase infamily function score (i.e., poorer family functioning)would result in a 6.4 decrease (worsening) in MCS, repre-senting a moderate effect size of .62. Overall, the adjustedR2 was .2884 (F = 73.96; df = 5; p = < .0001), with 5 outof 15 predictors included in the full model.Correlates of psychosocial health for healthy baby sampleVariables that were significantly associated with poorerSF-36 MCS scores at the bivariate level include the follow-ing: younger parental age (r = .19; p < .001); householdincome below $30,000 per year (mean difference = -4.6; p= .005); less than high school education (mean difference= -6.22; p = 0.065); not living as common-law or married(mean difference = -6.1; p = .005); motor developmentproblems (mean difference = -11.3; p = .043); learning/remembering problems (mean difference for any prob-lems versus none = -2.68, p = 0.021); poorer quality of life(mean difference = -13.9; p < .032); more internalizingbehavior symptoms (mean difference = -9.5; p < .001);more externalizing behavior symptoms (mean difference= -15.2; p < .018); more caregiver strain (r = .385; p <.001); and lower family function (r = -.438; p < .001).Predictors that were significant in the final regressionmodel appear in Table 7. The model for parents of healthychildren did not include female gender (because of lownumbers) and externalizing behavior symptoms, andincluded several variables not predictive in the NICUmodel (i.e., parental age; education; quality of life). Bothmodels included internalizing child behaviors, caregiverstrain and family function.In the healthy baby sample, younger parental age wasrelated to poorer MCS score, with a one year change in ageresulting in a 0.26 (CI: 0.08; 0.45) change in MCS. A tenyear difference in age would correspond to a 2.6 differencein MCS, which would represent a small effect size of 0.27(when the overall healthy baby parent group SD for MCS(9.6) was used as a denominator). Education was alsoassociated with MCS. Compared with high school gradu-ates, the MCS score for parents with less than a highschool education were on average 5.0 lower (CI: 0.84;9.1), which represents a moderate effect size of 0.52,although the effect could range from minimal to large dueto lower precision of the beta estimate. Child internaliz-ing symptoms, family function and caregiver strain werelow precision in beta estimates, the effects ranged fromminimal to large. Lower parent-reported child quality oflife was also associated with a lower parental MCS. Par-ents who reported a problem with their child's quality oflife had MCS scores that were 6.9 (CI: 0.37; 13.4) lowerthan parents who reported at least one quality of life prob-lem compared with those who reported at least one prob-lem. Again, due to the small numbers, the effect couldrange from minimal to large. In the final regressionmodel, the adjusted R2 was .3046 (F = 25.97; df = 6; p <.0001), with 6 out of 16 predictors included in the fullmodel.DiscussionThere is little information in the research literature onhow parents of NICU children adapt psychologically tothe demands of caregiving beyond the initial hospitaliza-tion period. We compared the psychosocial health ofparents of NICU children with parents of a group ofhealthy full-term children using the SF-36, a populargeneric measure of psychosocial HRQL. Although chil-dren admitted to a NICU at birth are at increased risk of avariety of long-term health problems, we did not find anydifference in parental psychosocial health when the twogroups were compared. This finding is in agreement withone of the few studies that measured mental health in par-ents of NICU children at preschool age. Singer et al. [17]reported that after the neonatal period, the mental healthof mothers of low-risk infants did not differ from mothersof term infants, and by 3 years, they had lower levels ofdistress, which they suggest may be due to maternal reliefafter an initial period of fear and anxiety. Mothers of high-risk infants, in contrast, had more symptoms of distress at2 years, more negative family impact at 2 and 3 years andmore parental strains and illness stressors at 3 years. Butby 3 years, their reported psychological distress did notdiffer from that of term mothers. The authors suggest thatby 2 years, infant developmental scores are predictive oflater outcomes, and many mothers of high-risk infantsmust relinquish their hopes for their children to "catchup" to healthy born children and that some psychologicaladaptation has taken place despite parental acknowledg-ment of greater family and parenting stressors. With ourcross-sectional design, we are not able to confirm thetrend noted by Singer, but given the lack of relationshipbetween most birth-related risk factors and parental men-tal health, it is possible that the parents of high- and low-risk infants in our sample have adjusted over time.Current health status, in bivariate analysis, was stronglyrelated with parental psychosocial health. In both groupsof parents, those whose child had a neurosensory, motordevelopment, learning/remembering or quality of lifePage 8 of 11(page number not for citation purposes)associated with parental MCS in a similar way as for NICUparents. However, due to lower numbers and resultingproblem had poorer psychosocial health than those withchildren with no problems in these areas. Child behaviorBMC Pediatrics 2004, 4:24 http://www.biomedcentral.com/1471-2431/4/24was also strongly related to parental psychosocial health.More specifically, parents of children who scored in theborderline or clinical range for internalizing, externalizingand total behavior problems on the CBCL/1.5–5 reportedpoorer psychosocial health than parents of children whoscored in the normal range. These findings were consist-ent across both samples of parents. The only birth-relatedrisk factor associated with parental psychosocial healthwas the presence of a congenital anomaly. Here the effectsize was small, but points to the possibility that a congen-ital anomaly may affect parents mental health adversely.Researchers have reached a consensus that a minimallyimportant difference in HRQL is close to one half of astandard deviation [34]. The differences that we found forhealth status and behavior were substantially larger andtherefore represent clinically important differences inparental psychosocial health. However, not all of thesevariables showed a significant effect in the multivariateanalysis, and it is possible that these variables influenceother, more proximal, variables that showed strongereffects on parental psychosocial health.The factors associated with poorer psychosocial health inthe multivariate models provide important informationabout correlates of adjustment for NICU and healthy babyfamilies. In a more general pooled model, parental age,higher caregiver strain, lower family function, and child'sinternalizing and externalizing behavior were independ-ently associated with poorer caregiver's mental healthscore. The effect of lower parental education was modifiedby NICU status of the child. In the healthy baby group,less than high school education indicated lower MCSscore. Child externalizing behavior symptoms and femalegender (parental) were associated with lower MCS scoresin the NICU group, whereas lower parental age, less edu-cation and poorer child quality of life were associatedwith lower MCS in the healthy baby group. For both sam-ples, as it is also seen in a pooled model results, low fam-ily function, high caregiver strain, and child'sinternalizing behavioral symptoms were independentlyassociated with lower parental psychosocial health. Forfamily function and caregiver strain, only a substantialdeparture from mean values (at least 2 SDs) would resultin a clinically important moderate effect size for the NICUgroup. Our interpretation for the healthy baby sample ishampered by wide confidence intervals around the betaestimates, resulting in effect sizes that ranged from mini-mal to large. Internalizing behavior symptoms were asso-ciated with only a small effect on caregiver's MCS score,again with wide confidence intervals around the beta coef-ficients for both samples.A recent publication outlines the integration of a numberThis model includes the following constructs: backgroundand context; child characteristics; caregiver strain; intra-psychic factors; coping/supportive factors; and health out-comes. Fitting our findings within this framework, wefound that poorer psychosocial health in parents wasassociated with background/context variables (i.e., femalegender, younger age, less education); child characteristics(i.e., poorer quality of life, more child behavior prob-lems); caregiver strain; and coping/supportive factors (i.e.,family function). We suggest that future research withNICU parents be conceptually based and measure con-structs found in other research to be important to car-egiver health.Our study has several limitations. Because it is not possi-ble to verify cause-effect using a cross-sectional design, wewere only able to estimate the direct effect of a limitednumber of predictor variables on parental psychosocialhealth. While our study has helped to identify some pos-sibly important caregiving variables, there are other varia-bles important to caregiver health that we did notmeasure. For example, while it is possible that some par-ents of children with severe health problems may havereceived specialized or targeted services (health and/orsocial services) to help them cope with their child's healthproblems, we did not include measures to determine this.Another limitation concerns our response rate. Althoughit is within the range often obtained in a postal survey[36], non-response can introduce bias. Some non-respondents indicated (verbally or in writing) they were"too busy" to participate. It is also likely that some ques-tionnaires returned to us blank were from non-Englishspeakers. Where we had data and were able to exploreresponse bias (NICU sample only), only a few differencesin birth-related sample characteristics and outcome werefound that suggests respondents had sicker babies [37].However, our study findings about health outcomes ofNICU graduates are in agreement with the larger NICU lit-erature, so it is unlikely that the differences we found areentirely due to response bias.ConclusionOur findings would suggest that overall, parental gender,family functioning and caregiver strain played influentialroles in parental psychosocial health. For child character-istics, current behavior was more influential than initialbirth-related risk factors.List of abbreviationsMCS – Mental Component ScoreNICU – Neonatal intensive care unitPage 9 of 11(page number not for citation purposes)of theoretical models into one multidimensional modelthat can be used to describe the caregiving process [35].HS – Health statusBMC Pediatrics 2004, 4:24 http://www.biomedcentral.com/1471-2431/4/24FAD – Family Assessment DevicePTT – Parental Impact TimeCompeting interestsThe author(s) declare that they have no competinginterests.Contributions of each authorAnne Klassen contributed substantially to the study's con-ception and design, acquisition of data, analysis and inter-pretation of data; and she drafted and revised and gavefinal approval of the version to be published.Shoo Lee contributed substantially to the study's concep-tion and design, acquisition of data, analysis and interpre-tation of data; and revised the article critically forimportant intellectual content and gave final approval ofthe version to be published.Parminder Raina contributed substantially to the analysisand interpretation of data; and revised the article criticallyfor important intellectual content and gave final approvalof the version to be published.Sarka Lisonkova contributed substantially to the analysisand interpretation of data; and revised the article criticallyfor important intellectual content and gave final approvalof the version to be published.AcknowledgementsThe Hospital for Sick Children Foundation (Toronto) provided an operat-ing grant for this study. Anne Klassen was recipient of a Killam Postdoctoral Fellowship. From Canadian Institutes of Health Research, Anne Klassen holds a Senior Research Fellowship, and Parminder Raina holds an Investi-gator Award. We wish to thank Jeanne Landgraf, Saroj Saigal, Drs Mike Carkner and Michael Klein, and the Canadian Neonatal Network.References1. Bhutta AT, Cleves MA, Casey PH, Cradock MM, Anand KJ: Cogni-tive and behavioral outcomes of school-aged children whowere born preterm: a meta-analysis. JAMA 2002, 288:728-737.2. Lorenz JM, Wooliever DE, Jetton JR, Paneth N: A quantitativereview of mortality and developmental disability inextremely premature newborns. Arch Pediatr Adolesc Med 1998,152:425-435.3. Escobar GJ, Littenberg B, Petitti DB: Outcome among survivingvery low birthweight infants: a meta-analysis. Archives of Dis-ease in Childhood 1991, 66:204-211.4. Aylward GP, Pfeiffer SI, Wright A, Verhulst SJ: Outcome studies oflow birth weight infants published in the last decade: Ametaanalysis. J Pediatr 1989, 115:515-520.5. Saigal S: Perception of health status and quality of life ofextremely low-birth weight survivors. The consumer, theprovider, and the child. Clin Perinatol 2000, 27:403-419.6. Hack M, Flannery DJ, Schluchter M, Cartar L, Borawski E, Klein N:Outcomes in young adulthood for very-low-birth-weightinfants. NEJM 2002, 346:149-157.7. Victorian Infant Collaborative Study Group: Improved outcomeinto the 1990s for infants weighing 500–999 g at birth. The8. Wolke D, Meyer R: Cognitive status, language attainment andprereading skills of 6-year-old very preterm children andtheir peers: the Bavarian Longitudinal Study. Dev Med ChildNeurol 1999, 41:94-109.9. Vohr BR, Wright LL, Dusick AM, Mele L, Verter J, Steichen JJ, SimonNP, Wilson DC, Broyles S, Bauer CR, Delaney-Black V, Yolton KA,Fleisher BE, Papile LA, Kaplan MD: Neurodevelopmental andfunctional outcomes of extremely low birth weight infants inthe National Institute of Child Health and Human Develop-ment Neonatal Research Network, 1993–4. Pediatrics 2000,105:1216-26.10. Walther FJ, den Ouden AL, Verloove-Vanhorick SP: Looking backin time: outcome of a national cohort of very preterm infantsborn in The Netherlands in 1983. Early Hum Dev 2000,59:175-91.11. Grunau RE, Whitfield MF, Davis C: Pattern of learning disabilitiesin children with extremely low birth weight and broadlyaverage intelligence. Arch Pediatr Adolesc Med 2002, 156:615-20.12. Siagal S, Stoskopf BL, Streiner DL, Burrow E: Physical growth andcurrent health status of infants who were of extremely lowbirth weight and controls as adolescence. Pediatrics 2001,108:407-15.13. Stjernqvist K, Svenningsen NW: Ten-year follow-up of childrenborn before 29 gestational weeks: health, cognitive develop-ment, behaviour and school achievement. Acta Paediatr 1999,88:557-62.14. Brooten D, Gennaro S, Brown LP, Butts P, Gibbons AL, Bakewell-Sachs S, Kumar SP: Anxiety, depression, and hostility in moth-ers of preterm infants. Nurs Res 1988, 37:213-216.15. Doering LV, Moser DK, Dracup K: Correlates of anxiety, hostil-ity, depression and psychosocial adjustment in parents ofNICU infants. Neonatal Network 2000, 19:15-23.16. Gennaro S, Brooten D, Roncoli M, Kumar SP: Stress and healthoutcomes among mothers of low-birth-weight infants. West-ern Journal of Nursing Research 1993, 15:97-113.17. Singer LT, Salvator A, Guo S, Collin M, Lilien L, Baley J: Maternalpsychological distress and parenting stress after the birth ofa very low-birth-weight infant. JAMA 1999, 281:799-805.18. Saigal S, Burrows E, Stoskopf BL, Rosenbaum PL, Streiner D: Impactof extreme prematurity on families of adolescent children. JPediatrics 2000, 137:701-6.19. Taylor HG, Klein N, Minich NM, Hack M: Long-term family out-comes for children with very low birth weights. Arch PediatrAdolesc Med 2001, 155:155-61.20. Ong LC, Chandran V, Boo NY: Comparison of parenting stressbetween Malaysian mothers of four-year-old very low birth-weight and normal birthweight children. Acta Paediatr 2001,90:1464-9.21. Cronin CM, Shapiro CR, Casiro OG, Cheang MS: The impact ofvery low birth-weight infants on the family is long lasting. ArchPediatr Adolesc Med 1995, 149:151-8.22. Lee SK, Penner PL, Cox M: Impact of very low birth weightinfants on the family and its relationship to parentalattitudes. Pediatrics 1991, 88:105-109.23. McCormick MC, Stemmler MM, Bernbaum JC, Farran AC: The verylow birth weight transport goes home: impact on the family.J Dev Behav Pediatr 1986, 7:217-223.24. Tommiska V, Ostberg M, Fellman V: Parental stress in families of2 year old extremely low birthweight infants. Arch Dis ChildFetal Neonatal Ed 2002, 86:F161-F164.25. Ware JE, Snow KK, Kosinski M: SF-36 Health Survey: Manual and Inter-pretation Guide Lincoln, RI: QualityMetric Incorporated; 2000. 26. Ware JE, Kosinski M: SF-36 Physical & Mental Health Summary Scales: AManual for Users of Version 1 2nd edition. Lincoln, RI: QualityMetric;2001. 27. Saigal S, Stoskopf BL, Rosenbaum PL, et al.: Development of amultiattribute preschool health status classification system[abstract]. Pediatric Res 1998, 43:228A.28. Achenbach TM, Rescorla LA: Manual for the ASEBA Preschool Forms andProfiles Burlington, VT: University of Vermont Department ofPsychiatry; 2000. 29. Lee SK, McMillan DD, Ohlsson A, Pendray M, Synnes A, Whyte R,Chien LY, Sale J: Variations in practice and outcomes in theCanadian NICU Network: 1996–1997. Pediatrics 2000,Page 10 of 11(page number not for citation purposes)Victorian Infant Collaborative Study Group. Arch Dis Child FetalNeonatal Ed 1997, 77:F91-94.106:1070-79.Publish with BioMed Central   and  every scientist can read your work free of charge"BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime."Sir Paul Nurse, Cancer Research UKYour research papers will be:available free of charge to the entire biomedical communitypeer reviewed and published immediately upon acceptancecited in PubMed and archived on PubMed Central BMC Pediatrics 2004, 4:24 http://www.biomedcentral.com/1471-2431/4/2430. Richardson DK, Corcoran JD, Escobar GJ, Lee SK: SNAP-II andSNAPPE-II: Simplified newborn illness severity and mortal-ity risk scores. J Pediatr 2001, 138:92-100.31. Klassen AF, Landgraf JM, Lee SK, Barer M, Raina P, Chan HWP, Mat-thew D, Brabyn D: Health related quality of life in 3 and 4 yearold children and their parents: preliminary findings about anew questionnaire. Health and Quality of Life Outcomes 2003, 1:81.32. Cadman D, Rosenbaum P, Boyle M, Offord DR: Children withchronic illness: family and parent demographic characteris-tics and psychosocial adjustment. Pediatrics 1991, 87:884-889.33. Cohen J: Statistical Power for the Behavioural Sciences New York: Aca-demic Press; 1977. 34. Norman GR, Sloan JA, Wyrwich KW: Interpretation of changesin health-related quality of life: the remarkable universalityof half a standard deviation. Medical Care 2003, 41:582-92.35. Raina P, O'Donnell M, Schwellnus H, Rosenbaum P, King G, BrehautJ, Russell D, Swinton M, King S, Wong M, Walter SD, Wood E: Car-egiving process and caregiver burden: conceptual models toguide research and practice. BMC Pediatrics 2004, 4:1.36. Asch DA, Jedrziewski MK, Christakis NA: Response rates to mailsurveys published in medical journals. J Clin Epidemiol 1997,50:1129-36.37. Klassen A, Lee SK, Raina P, Chan HWP, Matthew D, Brabyn D:Health status and health-related quality of life in a popula-tion-based sample of neonatal intensive care unit graduates.Pediatric 2004, 113:594-600.Pre-publication historyThe pre-publication history for this paper can be accessedhere:http://www.biomedcentral.com/1471-2431/4/24/prepubyours — you keep the copyrightSubmit your manuscript here:http://www.biomedcentral.com/info/publishing_adv.aspBioMedcentralPage 11 of 11(page number not for citation purposes)


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items