UBC Faculty Research and Publications

Design and implementation of a dental caries prevention trial in remote Canadian Aboriginal communities Harrison, Rosamund; Veronneau, Jacques; Leroux, Brian May 13, 2010

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

Item Metadata


52383-13063_2009_Article_465.pdf [ 524.18kB ]
JSON: 52383-1.0221603.json
JSON-LD: 52383-1.0221603-ld.json
RDF/XML (Pretty): 52383-1.0221603-rdf.xml
RDF/JSON: 52383-1.0221603-rdf.json
Turtle: 52383-1.0221603-turtle.txt
N-Triples: 52383-1.0221603-rdf-ntriples.txt
Original Record: 52383-1.0221603-source.json
Full Text

Full Text

TRIALSHarrison et al. Trials 2010, 11:54http://www.trialsjournal.com/content/11/1/54Open AccessS T U D Y  P R O T O C O LStudy protocolDesign and implementation of a dental caries prevention trial in remote Canadian Aboriginal communitiesRosamund Harrison*1, Jacques Veronneau2 and Brian Leroux3AbstractBackground: The goal of this cluster randomized trial is to test the effectiveness of a counseling approach, Motivational Interviewing, to control dental caries in young Aboriginal children. Motivational Interviewing, a client-centred, directive counseling style, has not yet been evaluated as an approach for promotion of behaviour change in indigenous communities in remote settings.Methods/design: Aboriginal women were hired from the 9 communities to recruit expectant and new mothers to the trial, administer questionnaires and deliver the counseling to mothers in the test communities. The goal is for mothers to receive the intervention during pregnancy and at their child's immunization visits. Data on children's dental health status and family dental health practices will be collected when children are 30-months of age.The communities were randomly allocated to test or control group by a random "draw" over community radio. Samplesize and power were determined based on an anticipated 20% reduction in caries prevalence. Randomization checkswere conducted between groups.Discussion: In the 5 test and 4 control communities, 272 of the original target sample size of 309 mothers have been recruited over a two-and-a-half year period. A power calculation using the actual attained sample size showed power to be 79% to detect a treatment effect. If an attrition fraction of 4% per year is maintained, power will remain at 80%. Power will still be > 90% to detect a 25% reduction in caries prevalence. The distribution of most baseline variables was similar for the two randomized groups of mothers. However, despite the random assignment of communities to treatment conditions, group differences exist for stage of pregnancy and prior tooth extractions in the family. Because of the group imbalances on certain variables, control of baseline variables will be done in the analyses of treatment effects.This paper explains the challenges of conducting randomized trials in remote settings, the importance of thoroughcommunity collaboration, and also illustrates the likelihood that some baseline variables that may be clinicallyimportant will be unevenly split in group-randomized trials when the number of groups is small.Trial registration: This trial is registered as ISRCTN41467632.BackgroundThe poor dental health of Aboriginal children in Canadais a major public health issue. Early childhood caries orECC is the term used for dental caries in the young child.This condition can affect children before their first birth-day [1]. ECC can be a painful condition, influencing achild's ability to eat properly, sleep through the night [2],grow and develop normally [3] and thus achieve fullpotential. Furthermore, caries in the primary (baby) teethhas a significant and positive association with caries andmalalignment of the permanent teeth [4]. In some Cana-dian Aboriginal communities, the prevalence of ECCexceeds 90% [5]. Furthermore, the financial burden oftreating ECC is enormous, and even more so for young* Correspondence: rosha@interchange.ubc.ca1© 2010 Harrison et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative CommonsAttribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction inany medium, provided the original work is properly cited.Aboriginal children from remote communities who often Division of Pediatric Dentistry, Faculty of Dentistry, University of British Columbia, 2199 Wesbrook Mall, Vancouver, V6T 1Z3, CanadaFull list of author information is available at the end of the articleHarrison et al. Trials 2010, 11:54http://www.trialsjournal.com/content/11/1/54Page 2 of 9must travel vast distances for comprehensive treatment[6].Dental caries (tooth decay) occurs when "cavity-caus-ing" bacteria, foods usable to the bacteria and susceptibleteeth are in contact with each other long enough to allowbacterial by-products to demineralize the enamel of theteeth [4]. A universally effective, caries-prevention pro-gram with predictable long-term results for youngAboriginal children has yet to be found. Because ECC is adisease that is multi-factorial in origin, any preventiveprogram must include a variety of strategies [7]. Themicroorganisms implicated in the initiation of dental car-ies are transmitted from mother to child [8-10]. The inoc-ulation of bacteria received by the child appears to berelated to maternal oral health status, diet and oralhygiene practices [10]. Once the baby's primary teethbegin to erupt, brushing the teeth regularly with tooth-paste containing fluoride is essential. Another means ofintroducing fluoride is regular application of fluoride var-nish[11]. Given that poor dietary behaviours also contrib-ute to ECC, dietary modification is an importantcomponent of an infant oral health promotion strategy. Asleep-time bottle, constant daytime sipping from a bottleor sippy cup containing anything other than water andfrequent snacking are practices linked to the develop-ment of extensive caries [12].Most caries-prevention strategies require that a parentchange an existing behaviour, for example bottle-feeding,or adopt a new behaviour, for example regular tooth-brushing. However, providing knowledge alone to par-ents rarely leads to long-term changes in preventivebehaviours [13]. Conversely, behavioural techniques thatstructure change and assist parents in the process ofchange may lead to long-term change in behaviours.The purpose of this report is to provide the scientificrationale and methodological approach that wereemployed in a randomized controlled trial, involvingAboriginal (Cree) mothers and infants in eastern JamesBay in Quebec, Canada. The goal of the trial was to testthe effectiveness of an intensive one-on-one preventivecounseling intervention, Motivational Interviewing (MI).Issues related to design of the trial, training and recruit-ment of staff, articulation with the existing organizationalstructure of health care delivery and implementation ofthe trial in these remote communities will be discussed.The aim of this report is to enlighten other communitiesand public health researchers who may be planning simi-lar interventions.MethodsStudy-designby the Cree Board of Health and Social Services of JamesBay (CBHSSJB).The trial was designed as a single-blind study with clus-ter randomization by community and two treatmentgroups. Randomization of individual mothers withineach community was inappropriate for this study becauseof the close-knit nature of the communities and the riskof contamination. Therefore, participants were allocatedto treatment conditions by community using cluster ran-domization. A cluster randomization design reduced therisk of cross-contamination but did not eliminate it com-pletely. A characteristic of this project that reduced con-tamination was the fact that the nine "clusters"(communities) were distinct and well separated by con-siderable distance in many cases. Furthermore, eachcommunity had its own health clinic. In addition, fewmothers of young children traveled much between com-munities because of the prohibitive cost and the difficultyof traveling with young children.Study-research questionsPrimary questionIs there any difference in the dental health status of youngCree children whose mothers have participated in a cli-ent-centred, one-on-one, preventive counseling interven-tion, Motivational Interviewing (MI), compared withchildren whose mothers received oral health informationin the form of an educational pamphlet?This question will be answered by testing the hypothe-sis that the prevalence of caries in 30 month old childrenwill be lower in the experimental communities than inthe control communities.Secondary questionsAre Cree mother's knowledge and beliefs about childdental health issues, their dental health practices andchild feeding and comforting practices altered by partici-pation in an intervention based on principles of Motiva-tional Interviewing (MI)?Additional questionsWill children whose mothers participate in the MI inter-ventions have fewer negative health outcomes related topoor dental health, for example pain or problems eating,than control children? Does mothers' participation inthese interventions decrease the probability of their chil-dren requiring dental treatment under general anesthesiaor with sedation?These questions will be answered by testing the nullhypothesis that there will be no difference between thetwo groups.Study populationThe research was approved by the Behavioural EthicsReview Board of the University of British Columbia andThe Cree, an Aboriginal nation of North America, are thelargest First Nations group in Canada with over 200,000members. The Quebec Cree nation is called EeyouHarrison et al. Trials 2010, 11:54http://www.trialsjournal.com/content/11/1/54Page 3 of 9Istchee - Cree for Land of the People, and lies to the eastand southeast of James Bay. The Quebec Cree numbersome 14,000 people and live in nine distinct settlements.In 1985, a survey of children in the eastern James Baycommunity of Chisasibi revealed that the mean numberof decayed, extracted and filled primary tooth surfaces orthe "defs" for 4-6 year old girls was 19.3 and for boys was24.2 [14]. A more recent community-wide survey of 1079Cree children between the ages of 12 months and 12years [15] found that 30.4% of 12-24 month olds had den-tal caries. By comparison, only 4% of similarly aged non-Aboriginal children in Quebec had caries [15]. Theamount of dental disease in these young Cree children isdisturbing particularly given the fact that each EeyouIstchee community has a dental clinic. It is unlikely thatthere will ever be enough dental clinics and dental practi-tioners in Eeyou Istchee to manage the amount of disease.Furthermore, Eeyou Istchee community water supplieshave no added fluoride. Clearly, existing health servicesneed to be supplemented with a population-basedapproach to promote improved child oral health.Inclusion/exclusion criteriaOver a period of about 2 1/2 years, all women in the 9Cree communities who recently had given birth or werebetween about the 12th and 34th week of their pregnancywere asked to participate in the study. Any woman know-ing of an impending, permanent move out of her commu-nity was excluded. Even children born with a medicalconcern or a congenital anomaly, for example cleft palatewere included, if that was the family's wish. Informationabout the health of the child will be gathered at the out-comes assessment and will be managed in the statisticalanalysis.Study intervention: Motivational Interviewing (MI)Patient-centred, personalized approaches that avoiddirect persuasion have been shown to produce goodresults in promotion of behaviour change [16,17] Theintervention in this trial followed the principles of Moti-vational Interviewing or MI [18], a client-centred butdirective counseling style. With this approach, the moti-vation for change comes from the client, but the coun-selor helps create, by questioning and reflection, theexpectation of change. Feedback and advice are offeredwithin the context of acknowledgement of the client'sright to choose. Many possible paths to a solution areprovided. Client and counselor agree upon a menu ofeffective behaviours. This strategy fits well with the phi-losophy of the Cree who are more comfortable if some-one suggests ways to think of taking a different approacheffective, women should be given the opportunity "toreflect on their roles as women and mothers, caregiversand providers" [19].MI has been successful in the management of addictivebehaviours such as smoking and non-addictive behav-iours associated with conditions like diabetes [16,17].Furthermore, brief MI interventions have produced goodresults [20]. MI has been previously successfully appliedin a trial to reduce ECC in 6-18 month old IndoCanadianchildren in western Canada [21]. At the end of the 2-yeartrial period, the MI children had a 46% lower rate of toothsurfaces affected by caries than did control children.In this project called "I wish my child would have beau-tiful teeth" or, in the Cree language, Kimaa MiywaapitetNitawaashiim experimental group mothers had an MIsession during pregnancy and, ideally, participated in sev-eral more MI sessions until their child was two years ofage. Mothers received appropriate resources at each MIvisit to enable them to implement selected behaviours(infant toothbrushes, toothpaste, sippy cups.) Fluoridevarnish was offered after the age of one-year.The control group mothers received a culturally-appro-priate educational pamphlet describing healthy dentalcare practices for young children. Pamphlets were mailedto mothers when their child was 6 months of age andagain at 18 months of age. The pamphlet titled "ProtectBaby Teeth: Circle of Smiles" had been previously pro-duced in 2000 by the Nursing Caries Committee of the St.Theresa Point First Nation of Manitoba, Canada and isavailable from them on request [22]. Fluoride varnish wasavailable to control children at local dental clinics.RandomizationThe advantages and disadvantages of testing the inter-vention as a randomized controlled trial were discussedand debated at length during a 2-year community consul-tation process. Those who participated in the discussionsduring the consultation's phase were reassured that therandomization process would be open, impartial andunbiased. Therefore, the randomization was done pub-licly during a daytime broadcast over community radio.The territory's community radio station with its extensivelistening audience was an accepted and trusted way ofconveying local news and information in Eeyou Istchee.A community radio broadcast of the randomization pro-cess was well-suited to the remoteness of Eeyou Istcheeand the distances between communities.There were two "rounds" of a constrained randomiza-tion process: one round for the 2 larger communities anda second round for the 7 smaller communities. Two bas-kets had been prepared for the "on-air" randomization:rather than tells them directly how to act. It also fits wellwith a recommendation from a gestational diabetes pro-gram in Eeyou Istchee that, for an intervention to beone basket for the large communities and another basketfor the smaller communities. Each basket containedenvelopes marked "test" or control"; the larger commu-Harrison et al. Trials 2010, 11:54http://www.trialsjournal.com/content/11/1/54Page 4 of 9nity's basket contained 2 envelopes (1 test; 1 control) andthe smaller community's basket contained 7 envelopes (4test; 3 control). Communities were randomized in eachround by alphabetically ordering the communities'names. For example, for each round the first name on thealphabetical list of communities was announced followedby the drawing of an envelope from the basket; the nextname was announced followed by another draw until allenvelopes were allocated. The draw was done "live" onafternoon radio by a radio station employee who was notassociated with the research. Of the 9 communities, 5were allocated to the test condition and 4 to the controlcondition. The decision was made to allocate one moretest than control community to allow a more robustexploration of intervention effects e.g. analyses accordingto number of MI sessions attended. Communities werenot aware of their allocation until their name was drawn.No concerns about the randomization process have beenexpressed by any community.Study personnelCree dental assistants who worked part-time in commu-nity dental clinics were ideal personnel to work on thisproject. The study protocol suggested that some of thesedental assistants be recruited and employed in a newcapacity as Dental Health Representatives, or DHRs, towork on the project in test communities and also torecruit mothers from neighboring control communities.Training and calibration in recruitment procedures andin the MI technique were to be provided to the DHRs inworkshops facilitated by an expert in MI. DHRs wouldlearn how to frame questions, help structure change,apply fluoride varnish and demonstrate toothbrushing.Instruction and practice in administering the surveyinstruments would also take place.InstrumentsQuestionnaires on oral health practicesWomen enrolled in the study completed instruments thatincluded items on demographics, their personal oralhygiene practices and dental knowledge. An instrumentcalled the Readiness Assessment of Future Parents con-cerning Infant Dental Decay or RAFPIDD was also com-pleted. This 47-item instrument was a modification of avalidated instrument developed by Weinstein and Riedyto assess a mother's stage of change with regard to herchild's dental health [23]. The instrument was enhancedwith items specific to Cree mothers and pre-tested forinternal consistency with pregnant women in EeyouIstchee.Child dental health and dental health behaviours at end-from outside of Eeyou Istchee who have no other associa-tion to the trial will do the assessments. The clinicaldetection of caries will involve visual/tactile examinationswith explorers to remove plaque, front surface mirrors,cotton rolls and a dental head light. Criteria for cariesdetection will be those described by Pitts [24]. Ten per-cent of the children will be re-examined on a randombasis for assessment of reliability.Information about dental health knowledge, home-carebehaviours and caries-related health impacts will be col-lected by survey instruments administered at the assess-ment visit. The survey questions have been validated in aprevious survey of 301 children undertaken in commu-nity health clinics in Quebec [15].Economic evaluationThe objectives of the economic evaluation will be to esti-mate the net incremental cost of the intervention strategyand to estimate an incremental cost-effectiveness ratio.Estimation of the net incremental cost of the programwill involve a comparison of costs of dental resources uti-lized by each arm of the study. The cost analysis willinvolve a study of the time and materials required to carryout each of the intervention visits. Salary costs for theDHRs will be assessed based on their compensation rateand disposable materials will be valued at acquisitioncosts. Value of the mother's time to take her child to theclinic for well-baby care and for MI will be determined byage and gender-matched wages and subjected to sensitiv-ity analyses. If the intervention is effective, children in theintervention arm should have lower rates of use of dental-related health services during the period of the trial.Information concerning types, amount and cost of carereceived as well as the time lost by caregivers in pursuingthis care will be collected. The net cost of the program,then, will include the intervention costs minus any sav-ings in dental services. The incremental cost-effective-ness ratio at follow-up will be based on these netincremental costs and the difference in the primary out-come measure, number of decayed, extracted or filledprimary tooth surfaces.Sample sizeThe target sample size of 309 mothers was determinedbased on the need to have high power to detect a 20%reduction in caries prevalence. This magnitude of reduc-tion was similar to that observed in a community-basedoral health promotion program implemented in NativeAmerican villages that demonstrated a 25% decrease in"baby bottle tooth decay" in participating communitiesafter 3 years [25].The power was calculated using the method of adjust-ment for intra-class correlation by the variance inflationpointClinical data will be collected by a dental examination ofeach child at 30 ± 3 months of age. Calibrated examinersfactor [26] applied to the standard formula for calculatingpower for comparison of proportions from independentsamples. The intra-class correlation coefficient was esti-Harrison et al. Trials 2010, 11:54http://www.trialsjournal.com/content/11/1/54Page 5 of 9mated to be 0.0090 by applying the analysis of variancemethod to the preliminary data on caries prevalence byvillage, which produced a variance inflation factor of 1.35.The control-group caries prevalence was estimated to be0.86 using a weighted average of these individual preva-lences weighted by sample size per village. Based on theabove projections, a total sample size of 265 mother andchild pairs would yield power of 82% to detect a 20%reduction in caries from 0.86 to 0.69.A reported infant mortality rate of 15 deaths per 1000births was considered [27]. Overall loss to follow-up,including mortality, was estimated to be 5% per year forthe 3 years that mother and baby were in the study. Thus,the sample size was determined to be 309 mother andchild pairs. Based on a 75% participation rate, 412 birthswere required to achieve a sample size of 309. Otherresearchers working in Eeyou Istchee with new mothershave reported an 80% participation rate which was simi-lar to our expectations. The anticipated number of chil-dren enrolled in each community was estimated using thebirth data from 2001 and a 75% participation rate in eachvillage, for a total sample size of 258 accrued per year.Thus, it was estimated that 14 months would be requiredto enroll 309 mothers.Statistical analysisPrimary statistical analysisThe primary statistical analysis will be a comparison ofcaries prevalence in intervention and control groups,using a permutation test[28] with test statistic equal tothe difference between caries prevalences in the twogroups. A significance level will be determined using theexact permutation distribution of the test statistic, whichwill be computed by enumerating all possible randomassignments of villages to intervention or control condi-tions according to the randomization scheme. The use ofthe permutation test accounts for intra-class correlationbetween outcomes on children in the same village andalso addresses concerns over the small-sample perfor-mance of statistical methods such as Generalized Esti-mating Equations [29], which rely on asymptotic theory.An exact confidence interval for the treatment effect (dif-ference in caries prevalences) will be computed using theusual procedure for inverting the permutation test. Peter-son et al provide an illustration of the application of thisprocedure to a group-randomized trial [30]. Additionalanalyses will be conducted using the same procedures forboys and girls separately.The main analysis will be done at the end of the studywhen data collection at 30+/-3 months is complete. Nointerim analyses will be performed because it is consid-analysis of the baseline data was done to help the investi-gators develop an understanding of the characteristics ofthe sample.Subgroup analysesWithin the experimental communities, outcomes will becompared looking at factors such as total number of MIsessions attended, number of fluoride varnish applica-tions, individual DHR and individual community. Withinthe entire study sample outcomes will be compared look-ing at factors such as number of fluoride varnish applica-tions and reported frequency of tooth brushing.Progress of recruitment and powerTwo hundred and seventy-two mothers (131 test; 141control) have been recruited over 2 1/2 years. Because ofthe cluster-randomized design of the trial, the power ofthe trial can be determined not only by the total samplesize but also by the number of communities and by hav-ing sufficient numbers of subjects recruited in each of thecommunities. Whereas the original projected sample sizeof 309 women would have yielded 82% power to detect a20% reduction in caries prevalence from 0.86 to 0.69,power calculations using the actual sample sizes showthat the power will be 79% to detect this treatment effect.These calculations assume a loss to follow-up of 5% peryear. If we are able to maintain an attrition fraction of 4%per year, the power will be 80%. Power will be very high (>90%) to detect a 25% reduction in caries prevalence.Baseline characteristicsFor the entire group of women recruited, the mean (SD)age was 25.6 (6.1) years; stage of pregnancy was 20.2(10.7) weeks. The majority or 65.1% of the women hadother children and 42.0% of these mothers had a childwho had previously had a tooth extracted. The women'smean (SD) score out of 5 for the knowledge questions was3.0 (1.3). As far as oral health behaviours, 92.3% of moth-ers brushed their teeth with toothpaste at least once a dayand 72.7% of mothers had been to a dentist within the lasttwo years.Treatment group comparisons at baselineRandomization checks were conducted on possibledemographic and behavioral differences between groups(Table 1). The distributions of most variables was verysimilar for the two randomized groups of mothers (spe-cifically, for age, knowledge score, other children, tooth-brushing, and seeing the dentist in the prior two years).However, despite the random assignment of communitiesto treatment conditions, group differences existed forstage of pregnancy and prior tooth extractions for otherered extremely unlikely for there to be evidence of a treat-ment effect prior to collection of the final outcome dataon the entire sample. However, a preliminary frequencychildren in the family. Mothers in the test or MI commu-nities had later stages of pregnancy than mothers in con-trol communities (22.4 vs. 17.7 weeks), and mothers withHarrison et al. Trials 2010, 11:54 Page 6 of 9http://www.trialsjournal.com/content/11/1/54other children in the MI communities were less likely tohave had another child with a tooth extraction (34 vs.49%). Testing of differences between randomized groupson baseline variables is not particularly informative,because any differences found must by definition be typeI errors. However, such tests were performed here forillustrative purposes. Note that the difference in preg-nancy stage did reach statistical significance (Z-statisticlarger than 2 corresponds to a p-value < 0.05), whereasthe difference in tooth extraction prevalence did not.These results illustrate that a difference that could bedeemed clinically important may result even within therange of normal sampling variability in the cluster-ran-domized setting. In contrast, such a difference might beless likely to occur in an individual-randomized designwith the same number of participants. Statistical signifi-cance aside, what matters is that the group differences(particularly for tooth extraction) may be large enough tobe associated with clinically meaningful differences inhealth outcomes at follow-up in the enrolled children.Therefore, secondary analyses of outcomes will be donewith regression adjustment to control differences in stageof pregnancy and presence of other children with toothextractions as covariates. An unadjusted analysis will beperformed as the primary analysis; the results of theadjusted and unadjusted analyses will be compared todetermine the impact of the baseline group differences onthe results.DiscussionThis project with Aboriginal mothers aimed to evaluatethe effectiveness of a specific behavioural intervention,the trial in remote communities in northern Canadaadded to the project's challenges.Community consultationThe process of community consultation was one of sev-eral complex issues that arose during the planning of thetrial. Prior to undertaking the trial, extensive discussionsthroughout the community were undertaken. While dis-turbances in the balance between bacteria, substrate andhost are the local factors associated with ECC, healthdeterminant like economics, social norms and housingconditions also have a substantial impact on the develop-ment of the disease [31]. Therefore, any intervention hadto be planned with early and continuous community col-laboration and consultation and needed to be sensitive toexisting beliefs and traditions about parenting practices,child comforting and infant health. Oral health interven-tion strategies must be sensitive to the intended targetpopulation because of the important role that cultureplays in shaping health-related attitudes and behavioursin Aboriginal communities [32]. The project investiga-tors, one of whom was the public health dentist for theterritory, undertook two years of consultation with com-munity leaders, health care workers and families fromEeyou Istchee to discuss appropriate strategies prior tobeginning the trial. In addition, the Research Committeeof the CBHSSJB collaborated on this project from its con-ception.When the project began, relatively few community-based initiatives to improve the oral health of young chil-dren had been undertaken [25,33] and scant few had beenconducted as randomized controlled trials [34,35]. Theconcept, advantages and disadvantages of a randomizedTable 1: Randomization effectiveness: baseline characteristics of study sample mothers by treatment group.Variable N Test N Control Z1Age (years)2 130 25.5 (6.4), 15 - 44 137 25.6 (5.8), 15 - 39 -0.11Stage of pregnancy (weeks)2 102 22.4 (8.7), 6 - 38 90 17.7 (12.1), 0 - 36 2.25Knowledge score2 127 3.1 (1.2), 0 - 5 140 2.9 (1.4), 0 - 5 0.57Has other children2 129 83 (64.3%) 140 92 (65.7%) 0.32Other child had prior tooth extraction3,482 28 (34.1%) 92 45 (48.9%) 1.09Brush with fluoride toothpaste3 130 120 (92.3%) 141 130 (92.2%) 0.04Saw dentist < 2 years ago3 130 93 (71.5%) 141 104 (73.8%) -0.421Z = observed Z-statistic for comparison of MI and Control groups using Generalized Estimating Equations to account for clustering of subjects within communities2Results shown are mean (SD), minimum - maximum3Results shown are number positive for the variable (% positive).4 Only responses of mothers with other children are includedMotivational Interviewing, to decrease the prevalence ofearly childhood caries in their children. The location of controlled trial were discussed at length to ensure thatHarrison et al. Trials 2010, 11:54http://www.trialsjournal.com/content/11/1/54Page 7 of 9communities understood that, during the trial, "test"communities were to have a more intensive program than"control" communities. The consensus from the consulta-tion was that, since no infant oral health promotion pro-gram was currently in place, it was acceptable toparticipate in a trial to definitively determine the mosteffective approach.Existing organizational structureAnother matter that arose during the planning of the trialwas how the project was to articulate with the structureand organization of health care services within EeyouIstchee. Surveillance, promotion, prevention, protection,regulation, research and training relating to the health ofthe Cree population in Eeyou Istchee are managed by thePublic Health Department of the CBHSSJB. The Depart-ment's long-term goal related to this specific project wasthat, if successful, the MI approach would become a com-ponent of the Department's health promotion activities.To that end, their inclination was for the day-to-day workof the project to be undertaken by existing HealthDepartment staff rather than new staff, the so-called"Dental Health Representatives" (DHRs), who were to bespecific to the research project. After considerable delib-eration, it was decided that existing department staff,women currently employed as Community Health Repre-sentatives (CHRs) would do the recruiting of mothersand be the project interveners. The Project Manager whowas hired was a dental hygienist from the Cree nationwho was born, raised and had worked for several years inEeyou Istchee.Training of project staff and the MI interventionOne of the first steps in launching the project was staffengagement ("buy-in") and training. Because of the deci-sion to engage existing staff (CHRs) as recruiters andinterveners in the project, job-postings and hiring did notslow down the start of the project. However, 3 of the 9communities did not have a CHR on staff at the project'sstart; therefore, local women (as per the original pro-posal) were hired to begin recruitment. From the begin-ning of the project and throughout, the Project Managervisited the communities and met with the project's staffto problem-solve challenges and review procedures. Shealso maintained regular telephone contact with the staffin each community who were working on the project.As detailed in the proposal for the trial, a two-day train-ing workshop for CHRs in the 5 intervention or "test"communities was held within the first year of the project.The project's MI consultant provided a template for theMI script and menu that would be used in the counselingCree. Following the workshop, the menus were finalized,printed on flipcharts and sent with explanatory notes toeach CHR in each of the test communities. The MI con-sultant followed up some months later with an "MI-coaching conference call" to problem-solve MI with theCHRs. A second follow-up workshop was held the nextyear. Following this workshop, the PM visited each of thecommunities individually to problem-solve recruitingand MI challenges.The MI scripts were based on the work of Wein-stein[36] and on scripts developed for a previous trial[21,37,38] One script was created for pregnant and newmothers (Additional file 1) and another slightly modifiedscript was developed for mothers after their infant's firsttooth had erupted until their child was about 2-years ofage. Mothers of newborn, pre-dentate infants wererecruited to the trial in addition to pregnant women; thiswas a variation from the original proposal but was doneto ensure that every newborn infant and mother had theopportunity to participate. An important aspect ofimproving infant oral health is enhancing the oral healthof the mother to prevent transmission of cavity-causingbacteria from mother to infant. With this in mind, preg-nant and new mothers were given "privilege cards" thatallowed them expedited dental services at their commu-nity's dental clinic.For a variety of reasons, CHRs in many of the commu-nities were challenged by the process of subject recruit-ment and delivering the intervention. The negative side ofengaging these existing personnel in a research projectwas their sense of being overworked and not sufficientlyrewarded for this addition to their daily workload. Thisproject was yet another burden for the already busy andin-demand CHRs. Furthermore, the CHRs may havefound that expectant women and new mothers were toobusy to spend an additional amount of time completingproject documents or were simply not interested in par-ticipating in a research project. Even though the projectinvestigators had tried to be diligent in following recom-mended principles of involving Aboriginal communitiesin research, recruitment of participants was a struggle.Thus, in 2 of the 5 test communities and in 2 of the 4 con-trol communities, recruitment and, for the test commu-nities, delivery of the intervention was eventuallycompleted by the Project Manager. One of the project-specific DHRs achieved success in one test-communityand eventually took over the work of a retiring CHR inanother test-community. A CHR, aided by the ProjectManager, completed and continued the project's work inthe 5th test community. Local women who were not CHRscompleted recruitment in the remaining 2 control-com-sessions with the mothers. By engaging the CHRs in dis-cussion and critique, both the script and the menus wereextensively modified to suit the language and style of themunities. Recruitment was closed after 2 1/2 years whenthe number of mothers recruited, though short of theHarrison et al. Trials 2010, 11:54http://www.trialsjournal.com/content/11/1/54Page 8 of 9original goal, was sufficient to maintain the power of thetrial.Administrators may request that in order for researchtrials to be implemented within their organization, exist-ing personnel should work on the project. However, theseindividuals may simply not have the time, the interest orthe desire to participate, despite the provision of exten-sive in-service training and ongoing support. Therefore,it can be concluded that in these instances, the trial'shuman and financial resources need to be efficiently andquickly redistributed to ensure success of the project.Despite the initial and ongoing challenges involved withimplementing this clinical trial, the results will enhanceour understanding of the details of implementation ofrandomized trials in remote communities. In addition,the role of a client-centred, directive counseling style likeMotivational Interviewing in enhancing positive oralhealth behaviours in Aboriginal families will be betterunderstood.Additional materialCompeting interestsThe authors declare that they have no competing interests.Authors' contributionsRH and JV obtained funding for the study. All authors contributed to thedesign of the study. RH and JV developed the instruments, adapted an MIcounseling protocol to the study and supervised the collection of baselinedata. JV oversaw all phases of the community consultation. BL performed thesample size calculations and the analysis of the baseline data. RH wrote themanuscript. All authors have read and approved the final manuscriptAcknowledgementsThe authors thank Ms. Juliana Snowboy-Matoush, the Project Manager, for her skillful and dedicated project management of this trial. The project progressed because of the work of the dental health representatives and the community health representatives, and the involvement of the participating families from Eeyou Istchee. The support of Ms. Jill Torrie and the Research Committee of the CBHSSJB is gratefully acknowledged. The involvement of Dr. Phil Weinstein, consultant to the Motivational Interviewing intervention, is acknowledged.This work is funded by a Canadian Institute of Health Research Randomized Controlled Trial Grant #FRN 67817 awarded to Drs. Harrison and VeronneauAuthor Details1Division of Pediatric Dentistry, Faculty of Dentistry, University of British Columbia, 2199 Wesbrook Mall, Vancouver, V6T 1Z3, Canada, 2Public Health Department, Cree Board of Health and Social Service of James Bay, 200 Sam Awashish Street, Mistissini, Quebec, G0W 1C0, Canada and 3Department of Biostatistics, Department of Dental Public Health Sciences, Box 359460, University of Washington, Seattle WA 98195 USAReferences1. Ripa L: Nursing caries: a comprehensive review.  Pediatr Dent 1988, 3. Acs G, Lodolini G, Kaminsky S, Cisneros GJ: Effect of nursing caries on body weight in a pediatric population.  Pediatr Dent 1992, 14:302-305.4. Li Y, Wang W: Predicting caries in permanent teeth from caries in primary teeth: An eight-year cohort study.  J Dent Res 2002, 81(8):561-566.5. Schroth RJ, Harrison RL, Lawrence HP, Peressini S: Oral health and the aboriginal child: a forum for community members, researchers and policy-makers.  J Can Dent Assoc 2008, 74(5):429-432.6. Milnes AR, Rubin CW, Karpa M, Tate R: A retrospective analysis of the costs associated with the treatment of nursing caries in a remote Canadian aboriginal preschool population.  Community Dent Oral Epidemiol 1993, 21(5):253-260.7. Tinanoff N, Douglass JM: Clinical decision making for caries management in children.  Pediatr Dent 2002, 24(5):386-392.8. Berkowitz RJ: Streptococcus mutans - establishment and transmission in infants.  ASDC J Dent Child 1976, 43(3):192-195.9. Li Y, Caufield PW: The fidelity of initial acquisition of mutans streptococci by infants from their mothers.  J Dent Res 1995, 74(2):681-685.10. Wan AKL, Seow WK, Purdie DM, Bird PS, Walsh LJ, Tudehope DI: Oral colonization of streptococcus mutans in six-month-old predentate infants.  J Dent Res 2001, 80(12):2060-2065.11. Tinanoff N, Kanellis MJ, Vargas CM: Current understanding of the epidemiology mechanisms, and prevention of dental caries in preschool children.  Pediatr Dent 2002, 24(6):543-551.12. Shiboski CH, Gansky SA, Ramos-Gomez F, Ngo L, Isman R, Pollick HF: The association of early childhood caries and race/ethnicity among California preschool children.  Public Health Dent 2003, 63(1):38-46.13. Benitez C, O'Sullivan D, Tinanoff N: Effect of a preventive approach for the treatment of nursing bottle caries.  ASDC J Dent Child 1994, 61(1):46-49.14. Shaw L, Clark DC, Edger NP: The oral health status of Cree children living in Chisasibi, Quebec.  J Can Dent Assoc 1987, 53(3):201-205.15. Véronneau J: Cree Board of Health and Social Service of James Bay.  Chisasibi, QC: Report on Cree dental health in Eeyou Istchee (Northern Quebec); 2002:81. 16. Burke BL, Arkowitz H, Menchola M: The efficacy of motivational interviewing: a meta-analysis of controlled clinical trials.  J Consult Clin Psychol 2003, 71(5):843-861.17. Rubak S, Sandbaek A, Lauritzen T, Christensen B: Motivational interviewing: a systematic review and meta-analysis.  Br J Gen Pract 2005, 55(513):305-312.18. Miller WR: Motivation for treatment: a review with special emphasis on alcoholism.  Psych Bull 1985, 98:84-107.19. Gray-Donald K, Robinson E, Collier A, Kinga D, Renaud L, Rodrigues S: Intervening to reduce weight gain in pregnancy and gestational diabetes mellitus in Cree communities: an evaluation.  CMAJ 2000, 163:1247-1251.20. Berg-Smith SM, Stevens VJ, Brown KM, Van Horn L, Gernhofer N, Peters E, Greenberg R, Snetselaar L, Ahrens L, Smith K: A brief motivational intervention to improve dietary adherence in adolescents. The Dietary Intervention Study in Children (DISC) Research Group.  Health Educ Res 1999, 14(3):399-410.21. Harrison RL, Benton T, Everson-Stewart S, Weinstein P: Effect of motivational interviewing on rates of early childhood caries: a randomized trial.  Pediatr Dent 2007, 29(1):16-22.22. St Theresa Point First Nation Nursing Caries Committee: Protect Baby Teeth: Circle of Smiles.  Winnipeg MB: Vision CyberMedia Group Inc; 2000. 23. Weinstein P, Riedy CA: The reliability and validity of the RAPIDD scale: readiness assessment of parents concerning infant dental decay.  ASDC J Dent Child 2001, 68(2):129-135. 14224. Pitts NB, Evans DJ, Nugent ZJ: The dental caries experience of 5-year-old children in Great Britain. Surveys coordinated by the British Association for the Study of Community Dentistry in 1999/2000.  Community Dent Health 2001, 18(1):49-55.25. Bruerd B, Jones C: Preventing baby bottle tooth decay: eight-year results.  Public Health Rep 1996, 111(1):63-65.Additional file 1 Motivational Interviewing Script.Received: 31 December 2009 Accepted: 13 May 2010 Published: 13 May 2010This article is available from: http://www.trialsjournal.com/content/11/1/54© 2010 Harrison et a ; licensee BioMed Central Ltd. is an Open Access articl  distributed under he 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.rials 2010, 11:5410:268-282.2. Low W, Tan S, Schwartz S: The effect of severe caries on the quality of life in young children.  Pediatr Dent 1999, 21:325-326.26. Murray DM: Design and Analysis of Group-Randomized Trials.  New York: Oxford University Press; 1998. Harrison et al. Trials 2010, 11:54http://www.trialsjournal.com/content/11/1/54Page 9 of 927. Torrie J: How healthy are the Eeyouch in 2002? An update.  Cree Board of Health and Social Services of James Bay 2003.28. Gail MH, Mark SD, Carroll RJ, Green SB, Pee D: On design considerations and randomization-based inference for community intervention trials.  Stat Med 1996, 15(11):1069-1092.29. Liang K-Y, Zeger S: Longitudinal data analysis using generalized linear models.  Biometrika 1986, 73:13-22.30. Peterson AV, Kealey KA, Mann SL, Marek PM, Sarason IG: Hutchinson Smoking Prevention Project: long-term randomized trial in school-based tobacco use prevention--results on smoking.  Natl Cancer Inst 2000, 92(24):1979-1991.31. Reisine S, Douglass JM: Psychosocial and behavioral issues in early childhood caries.  Community Dent Oral Epidemiol 1998, 26(1 Suppl):32-44.32. LeMaster PL, Connell CM: Health education interventions among Native Americans: a review and analysis.  Health Educ Q 1994, 21(4):521-538.33. Harrison RL, Wong T: An oral health promotion program for an urban minority population of preschool children.  Community Dent Oral Epidemiol 2003, 31(5):392-399.34. Kowash MB, Pinfield A, Smith J, Curzon ME: Effectiveness on oral health of a long-term health education programme for mothers with young children.  Br Dent J 2000, 188(4):201-205.35. Sgan-Cohen HD, Mansbach IK, Haver D, Gofin R: Community-oriented oral health promotion for infants in Jerusalem: evaluation of a program trial.  Public Health Dent 2001, 61(2):107-113.36. Weinstein P: Motivate your dental patients: a workbook. The public health/pediatric edition.  In. Seattle WA: University of Washington; 2002:75. 37. Weinstein P, Harrison RL, Benton T: Motivating parents to prevent caries in their young children: one-year findings.  J Am Dent Assoc 2004, 135(6):731-738.38. Weinstein P, Harrison R, Benton T: Motivating mothers to prevent caries: confirming the beneficial effect of counseling.  J Am Dent Assoc 2006, 137(6):789-793.doi: 10.1186/1745-6215-11-54Cite this article as: Harrison et al., Design and implementation of a dental caries prevention trial in remote Canadian Aboriginal communities Trials 2010, 11:54


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