Barriers and facilitators to implementing practices for prevention of childhood obesity in primary care: A mixed methods systematic review

Summary Primary care providers (PCPs) have an important role in prevention of excess weight gain in pre‐school children. Guidelines exist to support PCPs' practices. This systematic review of PCPs' practice behaviors and their perceptions of barriers to and facilitators of implementation of guidelines was the first step toward the development of an intervention aimed at supporting PCPs. Five databases were searched to identify qualitative, quantitative, and mixed methods studies which examined PCPs' practice patterns and factors influencing implementation of recommended practices. The convergent integrated approach of the Joanna Briggs Institute (JBI) methodology for mixed methods reviews was used for data synthesis. Following analyses, the resultant factors were mapped onto the Capability, Opportunity, and Motivation model of Behaviour (COM‐B). Fifty studies met the eligibility criteria. PCPs inconsistently implement recommended practices. Barriers and facilitators were identified at the provider (e.g., lack of knowledge), parent (e.g., lack motivation), and organization level (e.g., inadequate training). Factors were mapped to all three components of the COM‐B model: psychological capability (e.g., lack of skills), reflective motivation (e.g., beliefs about guidelines), automatic motivation (e.g., discomfort), physical opportunity (e.g., time constraints), and social opportunity (e.g., stigma). These findings reflect the complexity of implementation of childhood obesity prevention practices.

organizations have published guidelines for prevention of childhood obesity in primary care. In England, the National Institute for Health and Care Excellence (NICE) [5][6][7] and Public Health England (PHE) 8 have developed guidelines for PCPs who have a role in prevention of childhood obesity. However, it is widely acknowledged that practitioners do not routinely implement guideline recommended practices. 9 Implementation of guidelines is influenced by a range of factors which may be related to the guideline, the healthcare setting, and the social, cultural, economic, and political context in which PCPs work. These factors are collectively referred to as barriers to and facilitators of implementation, or more broadly as "determinants of clinical behaviours." Identification, appraisal, and synthesis of the existing evidence regarding PCPs' current practices and their perceptions of factors that influence their practice behaviors can inform the development of strategies and interventions to support PCPs' role, service development, and future research into obesity prevention.
A PCP's behavior can be explained and predicted using the same processes and behavioral models that can be applied to human behavior in general. The Capability, Opportunity, and Motivation model of Behaviour (COM-B) model of behavior 10 proposes that interactions between capability, opportunity, and motivation result in the performance of the behavior that in turn influences those three components. Capability is defined as the individual's psychological (e.g., knowledge and communication skills) and physical capability (e.g., physical skills) to engage in the specified behavior. Opportunity refers to factors in the external environment that prompt or enable the performance of the behavior and includes both physical (e.g., resources) and social (e.g., social norms) opportunity. Motivation refers to the brain processes that facilitate the behavior (as a priority over other competing behaviors); they can be reflective (e.g., analytical decision making) or automatic (e.g., habits, and emotional responses, cued by environment). The COM-B model was used in this review to develop a theoretical understanding of the factors that influence practitioners' practice behaviors. This SR aimed to synthesize the evidence on (1) PCPs' current practices to prevent development of obesity in 0-5 year old children; (2) barriers to, and facilitators of guideline recommended practices as perceived by PCPs; and (3) to map these onto the COM-B model. This research was published as an abstract in a special supplement of Obesity Reviews in 2020. 11 2 | METHODS Qualitative and quantitative evidence was included in the review to account for the inherent complexity of implementing clinical practices in primary care. The convergent integrated approach, according to the JBI methodology for mixed-methods reviews, was used for evidence synthesis. 12 This involved simultaneously integrating and synthesizing quantitative and qualitative data through data transformation. The protocol of this review was registered with the International Prospective Register of Systematic Reviews (CRD42017084067). The review is reported in accordance with the updated Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) 2020 reporting guidelines 13 (presented in Table S1).

| Eligibility criteria
For this review, the concept of primary care was based on the World Health Organization (WHO)'s definition of "integrated" primary healthcare: a comprehensive health system which integrates key public health functions (health promotion and preventive care) into existing primary care services. 14 PCPs were defined as practitioners who work in primary care and provide services including health promotion, disease prevention, patient education and counseling. They included doctors (e.g., general practitioners and general pediatricians) and nurses (e.g., practice nurses, health visitors, pediatric nurse practitioners, maternal and child health nurses, and breastfeeding specialist nurses), community midwives, and community dieticians. A barrier was defined as a factor that hindered implementation of guidelines; a facilitator was defined as a factor that promoted implementation. Eligible studies were primary research studies reporting on (i) implementation/non-implementation by PCPs of practices recommended for prevention of excess weight in children aged 0-5 year; (ii) behavioral determinants (e.g., PCPs' knowledge, attitudes, and beliefs); and (iii) barriers and facilitators of implementation of practices as perceived by PCPs. Pre-2002 studies were excluded as UK guidelines for prevention of childhood obesity in primary care were first introduced around this time. 15 Only published peer-reviewed papers in English were included. Table 1 summarizes the inclusion and exclusion criteria.

| Search strategy
A three-step strategy, as recommended by the JBI, was used to identify eligible papers. 16 Figure S1.

| Study screening and selection process
Eligibility screening of titles and abstracts was undertaken by DR. An overview of the study screening and selection process of the original search and the updated search are presented using a tailored PRISMA 2020 flow diagram 17 (see Figure 1). The reasons for exclusion of full text papers were documented by DR and independently verified by a researcher experienced in conducting SRs. The most common reason for excluding full text papers was that the study focused exclusively on treatment of children having obesity.

| Quality appraisal
DR assessed the quality of all the papers using JBI critical appraisal checklists 18,19 that are specific to the research methodology. Coauthors acted as the second reviewers. DR's quality assessment work on 50% of the qualitative and mixed methods papers was checked by FS for accuracy. EM verified DR's quality assessment of 25% of the quantitative papers. Any inaccuracies/discrepancies were resolved through discussion. The assessment process was not used to exclude papers but as a broad guide to provide a context for interpreting the findings.

| Data extraction
Data on aims, study design, participants' characteristics, data collection methods, theoretical framework used (if any), and main findings (i.e., survey results, themes identified by study authors, and participant quotations) were extracted from each paper, using tools available from JBI. 20,21 Data extraction of all the papers was undertaken by DR, and 20% were checked for accuracy by a second reviewer (LE). Any inaccuracies/discrepancies were resolved through discussion.

| Data synthesis
As specified in JBI's convergent integrated approach for mixed methods review, 16 the quantitative data extracted from survey studies and quantitative component of mixed methods studies were "qualitised" through narrative interpretation of the findings into textual descriptions. Subsequently, the "qualitised data" was assembled with the qualitative data extracted from qualitative studies and the qualitative component of mixed methods studies. To guide the synthesis of the evidence related to PCPs' practice behaviors, three "behaviour areas" were identified based on the NICE 5-7 and PHE 8 guidelines for prevention of excess weight development in 0-5 year olds. The behavior areas were developed by grouping the guideline recommendations into themes and identifying the clinical behaviors that are expected to be carried out by PCPs within each area during their interactions with children and parents ( Figure 2). Thematic synthesis of the assembled data was carried out by DR using an iteratively developed coding frame. The themes and categories were refined through discussion with review team members (FS, LE, and EM) at multiple meetings to ensure that they accurately reflected the data. A narrative account of the synthesis was prepared, and quotations were taken from the studies to illustrate the findings. Subsequently, the findings were mapped onto the COM-B model by DR following expert guidance 22 and in consultation with the review team.   of these, 25 studies also reported on PCPs' practices. Data from the studies were tabulated to allow comparison of the country of origin, key objectives, participant characteristics, service user group, study design, and the primary care setting. This information is presented in Table S2.
Only eight studies used a psychological theory or model to guide the research. The majority of the qualitative studies reported using purposive sampling, independent coding by multiple researchers and using consensus meetings to resolve discrepancies. However, only six studies reported on the influence of the researcher(s) on the research.
The majority of survey studies used restricted sampling frames (very few studies used national databases) and convenience (not random) sampling to recruit participants, thus limiting the potential for generalizability of the findings. Most studies provided information on response rates, used appropriate analytic methods, and acknowledged the potential for self-selection bias and self-reporting bias as methodological limitations. A cross-study summary of the quality appraisal is presented in Table 2 (qualitative studies) and Table 3 (survey studies).
The full quality appraisal of the individual studies is available upon request.
Three broad organizing themes emerged from the synthesis:

| Weight and growth assessment
Twenty-one studies reported on weight assessment practices. PCPs generally relied on height and weight growth charts, [25][26][27][28][29][30][31][32][33][34] or simple visual inspection 25,26,32,33,35,36 to assess a child's weight status and monitor weight gain over time. There were geographic differences in the use of reference charts used by PCPs. The reference most commonly cited in the USA-based studies were the CDC (Centre for Disease Control) growth chart 37 (for children over age 2) and the WHO standards 38 (for children under age 2) whereas studies from other countries reported the use of national standards based on WHO standards. Although the BMI chart was regarded as a facilitator of conversations about weight, 25,26,[39][40][41] the routine use of BMI for 2-5 yearolds (and weight-for-length charts for children under 2) was low, with roughly a third of PCPs never using BMI (and weight-for-length for <2), 25,29,32,33,36,[42][43][44][45][46][47][48] or using it selectively, for example, only if PCPs were concerned. 26,36,47 One USA-based study found that the routine use of BMI by pediatricians at well-child visits has increased over the past decade. 49 However, many PCPs who reported they regularly measured BMI were not aware of the guidelines for classifications applied for overweight/obesity. 28,31,33,42,50 Low use of BMI was associated with PCPs' lack of familiarity with BMI, 25

| Breastfeeding support
The data on PCPs' breastfeeding support practices was limited.
Although most PCPs believed that supporting breastfeeding was an important part of their role, 25,55-58 many PCPs did not routinely discuss and provide breastfeeding advice during antenatal and postnatal visits, or assist mothers with specific breastfeeding problems. 57,58 Only a minority reported having observed a new mother breastfeeding (a guideline recommendation) and many had never counseled mothers about infant feeding methods, assisted mothers with breastfeeding techniques, or managed lactation problems. 56,58 All PCP groups reported that they felt unprepared to support the needs of breastfeeding mothers. 56,[58][59][60][61] PCPs attributed their lack of knowledge and skills for managing breastfeeding problems to lack of education and training on breastfeeding management. Importantly, many GPs and pediatricians admitted they lacked competence in key topics (e.g., prescribing to breastfeeding mothers; inadequate weight gain in breastfed infants) where other practitioners (e.g., nurses and midwives) may regard them as an expert for specialist referral. 30,61 Many PCPs acknowledged that they relied on information they had gained anecdotally from colleagues or from their personal or their spouses' breastfeeding experiences. 56,58,61,62 PCPs expressed concern that this could lead to some PCPs offering advice that ran counter to recommendations and result in mothers receiving conflicting and incorrect messages. 55,[59][60][61][62] PCPs stressed the importance of supporting women with their "choice" and not being perceived by mothers and their own peers as being coercive. 61,62 Some PCPs considered breastfeeding as difficult and "exhausting" and believed that bottle feeding was perceived as an easier option by some mothers. 34 PCPs less frequently discussed healthy eating and physical activity with parents of infants (0-2 year olds) and pre-school children (2)(3)(4)(5) year olds) as compared to school age children. 25 However, the focus of dietary advice was generally about the contents of a healthy infant diet and less about infant feeding practices (e.g., responsive feeding). 25,47 Further, PCPs tended to provide "blanket" nutritional advice and not discuss specific diet and nutrition topics; also, they were more likely to discuss fruit and vegetable consumption than consumption of sugary drinks, fast foods, and energy dense foods. 34,42,44,[64][65][66]70 PCPs lacked awareness of the importance of physical activity in young children 25,31,50,65,67,68 and placed low priority on raising the topic. 25

| Parent-level factors
A positive relationship between the PCP and the family 25,39,40,47,66,74 and parental concern about childhood overweight 26,49 were identified as facilitators. PCPs believed that when parents themselves raised concerns about their child's weight, they were more likely to engage with PCPs and comply with recommendations.

| Organization-level factors
Implementation was enabled by perception of role support from the organization 23,33,64,73 and access to training opportunities. [32][33][34]45,53,62,70,72 PCPs identified availability of sufficient time and support staff, 24,25,32,55 access to specialist staff (dieticians and breastfeeding support staff) and local community based familycentered obesity prevention programs 25,45,70 as potential facilitators of implementation. Some PCPs believed that a uniform coherent approach to obesity prevention 23,26,41,74 and closer working between physicians and nurses 26,41,53,62 can help improve the quality of preventive care.
Due to the potential of this attributional bias, caution must be exercised when interpreting the findings of the barriers and facilitators.

| Gaps in literature
Several gaps emerged from the data. Firstly, there was lack of information on collaborative working and team-based approach to

| Strengths and limitations
To our knowledge, this is the first systematic review to report on childhood-obesity prevention practice behaviors of key PCP groups and conduct a theoretical analysis of the barriers and facilitators. This review was led by DR as part of a doctoral research project and is "restricted" because certain elements that are required in a "full" review were simplified. 81 A single reviewer screening of abstracts may limit the methodological standard of a review. 82 However, the conduct of the review was closely supervised by DR's supervisors who were also members of the review team. As second reviewers, review team members verified the lead reviewer's work on quality assessment and data extraction on a randomly selected proportion of the papers. A well conducted restricted review with minimum 20% checking by a second reviewer is considered an appropriate strategy in situations where a "full" SR process cannot be implemented. 82

CONFLICT OF INTEREST
There are no other relationships/conditions/circumstances that present a potential conflict of interest.