Improvement in Primary Care Provider Self-Efficacy and Use of Patient-Centered Counseling To Address Child Overweight and Obesity after Practice-Based Changes: Texas Childhood Obesity Research Demonstration Study

Background

Primary care pediatric providers confront the epidemic in childhood obesity daily. Over 18% of children aged 2–19 years have a BMI ≥95th percentile, and some practices serve populations with obesity prevalence >20%. ¹ Pediatricians have reported low self-efficacy in addressing childhood obesity.^2,3 In particular, providers perceive low confidence in counseling skills needed to address childhood obesity. Training has led to self-efficacy improvement,^4,5 but evaluation of training impact has been minimal and short term. Studies have also examined reported practice behavior. ⁶ However, such studies have focused on counseling topics (e.g., fruit and vegetable consumption or screen time) rather than the counseling techniques used by providers. In addition, little attention has been paid to the recommended follow-up steps after a visit, such as return visit or referral to a dietitian.

The Texas Childhood Obesity Research Demonstration (TX CORD) project evaluated a multisystem approach to childhood obesity prevention and intervention for children aged 2–12 years, which included training and support for primary care providers (PCPs) and practices. The multiyear project provided an opportunity to assess provider self-efficacy in obesity-related counseling techniques and reported practice behavior over several years.

The aim of this study was to evaluate the change in provider-reported self-efficacy and practice behaviors from baseline to 24 months during the TX CORD intervention to improve child overweight/obesity (OW/OB) care in primary care practices. In addition, the study examined the relationship of past training in nutrition, physical activity, or obesity care with these changes.

Methods

Results

Discussion

During a multisystem intervention in childhood obesity, medical providers in 12 primary care offices improved self-efficacy and practice behavior at 12 and 24 months compared with baseline measures. Specific areas of improved self-efficacy included counseling about sleep and use of behavior change strategies. Areas of evaluation and counseling practice behaviors that improved were patient-centered counseling approaches, which had low frequencies at baseline. Discussion of BMI, evaluation of medical conditions, and addressing lifestyle (all evaluation and counseling behaviors reported to be frequent at baseline) did not change significantly, nor did planned follow-up practice behaviors.

Prior training in physical activity education and child OW/OB education was associated with higher scores in both self-efficacy identification scale (confidence in identifying lifestyle factors associated with OW/OB) and self-efficacy counseling scale (confidence in counseling about lifestyle factors associated with OW/OB).

At baseline in this study, the providers reported high self-efficacy for the established foci of obesity management, including BMI assessment, eating behaviors, physical activity, and screen time. However, self-efficacy for addressing sleep behaviors, a behavior more recently recognized as associated with obesity, ¹³ as well as for parenting, goal setting, and use of motivational interviewing, were lower at baseline.

Previous studies have shown that confidence of pediatric providers has been higher for identifying OW/OB and managing medical comorbidities than for counseling skills used in addressing obesity.^2,3,14,15 The baseline ratings among the TX CORD providers were very similar to ratings of community providers in Perrin's study for the parallel questions of identification of and counseling about eating behaviors and physical activity behaviors. ⁵ Other studies have also found that providers rated confidence in assessment strategies for OW/OB higher than confidence in counseling and interventions.^6,16,17

The improvement in self-efficacy scores at 12 and 24 months, after correcting for multiple comparisons, is notable; the overall counseling score rose over the 2 years in contrast to the identification score, suggesting that the TX CORD training and intervention were especially beneficial for provider counseling for OW/OB. Other studies have demonstrated an effect of brief training on self-efficacy, but the longest time to reassessment was 6 months. In a study in a single practice (seven providers), 3 hours of training was associated with improved confidence at 6 months in assessment of readiness to change. ⁴ Another study with 2 hours of training, including the “5 A” provider steps to counsel patients (ask, advise, assess, assist, and arrange), reported improvement in self-efficacy immediately after training. ⁶ One-hour training in addressing obesity and motivational interviewing led to significant improvement in confidence at 4 months. ⁵

Our study demonstrated maintenance of higher self-efficacy scores between 12 and 24 months for most counseling topics. This suggests that the TX CORD clinic intervention process, which included training, office-based tools, and informal support from team, led to sustained impact on self-efficacy, with changes most notable in counseling areas (parenting, goal setting, and the five-item counseling scale).

In parallel with improved self-efficacy in counseling areas related to child OW/OB, reported frequency of patient-centered counseling practice behaviors rose between baseline and 12 months, and had no significant decrement between 12 and 24 months. The reported frequency at baseline for these counseling strategies was lower (means <3, which denoted “sometimes”) than the reported frequency of more conventional obesity care, which includes medical assessment and nutritional education, for example. The increase in reported frequency of patient-centered counseling practice behavior suggests that these techniques were unfamiliar at the start of the study but were adopted after training and support and were sustained. In contrast, the intervention had no effect on planned follow-up (return visit, referral to a dietitian, or referral to structured weight-management program), despite attention in the training to when and how to make referrals.

Several recent studies show high reported frequency of assessment of behaviors that are conventionally associated with obesity, both through provider report^6,18 and through chart audit. ¹⁹ Counseling about sleep, more recently recognized as associated with obesity risk, was reported in one other study, from 2016, with results similar to this study's baseline results. ¹⁹ Use of counseling methods, including patient-centered focus, is not always assessed, ¹⁹ or assessment has been narrow. One study assessed documentation of patient wellness goals at single time point, ¹⁸ and another audited charts of children with high BMI for documentation of motivational interviewing, which increased after training from 0% to 35%. ²⁰

Our study suggests that relatively brief training of providers from multiple offices improves reported use of patient-centered counseling. The increased use of motivational interviewing despite no self-efficacy improvement may result from the brevity of the training; providers may have used the technique without feeling fully confident in their skill level.

The lack of significant decreases in scores between interim (2013) and follow-up (2014) suggests that providers maintained the changes in self-efficacy and behaviors, despite less contact and check-ins by study staff. We speculate that improvement continued in the last year because providers found these tools useful and relevant, and these techniques were consistent with secular changes in the focus on childhood obesity, but confirmation would require a larger sample size, a controlled condition, or more focused qualitative work with the providers.

The lack of increase in use of follow-up appointments and referrals was disappointing because the TX CORD intervention highlighted available programs, the training included recommendations for follow-up after primary care evaluation, and the practices had materials to guide the referral process. Responders indicated that lack of reimbursement under Texas Medicaid for obesity was not a common reason. Although not assessed, the lack of availability of dietitians and long wait times for tertiary obesity program may have contributed to the low referrals. One study, carried out in a single practice, successfully increased return visits when EHR modifications were created to support standard obesity care practices, including follow-up visits. ²⁰

Prior training (defined as ≥10 hours) in child overweight and obesity and in physical activity education were positively associated with both self-efficacy identification and counseling scales although training in nutrition was not. This association suggests that brief training may have more effect when prior training has occurred; however, prior training may instead be a marker for interest and motivation. Further study with larger samples and detailed information about prior training are needed.

In contrast, training was not positively associated with reported practices (either evaluation and counseling or after-visit plan) with the exception of a baseline association between clinical nutrition training and evaluation and management scale. Further study should explore whether lack of after-visit planning reflects provider attitudes or external barriers.

Other research in this area is limited. Prior training was often not assessed, or was assessed but not examined for association with attitudes.^2,6 One study found an association between obesity training and higher self-efficacy and reported behaviors, but because the study was cross-sectional, it did not examine whether prior training enhanced effect of current training. ¹⁷ One study compared live and web-based training on knowledge and skills, assessed through simulation, among nursing students, and found no moderator effect of prior training immediately after training and 1 month later. However, only 15% of participants reported experience with prior weight-related training. ²¹

Strengths of this study include the 12- and 24-month evaluation after the training, longer than other reports, and the setting in busy, nonacademic pediatric practices. The clustering analysis and correction for multiple testing increase the statistical reliability of the findings. Limitations include the self-report of practice behavior, the small sample size, no tracking of individual providers over the intervention, and lack of a control group.

In conclusion, this study demonstrates that brief training and materials for childhood obesity care in primary care settings can result in improvement in PCP self-efficacy for behavior change skills and in reported practice behaviors, and these improvements are sustained for over a year. Prior training was associated with positive effects, suggesting that repeated, brief trainings of PCPs are beneficial for improving self-efficacy of provider OW/OB screening and counseling. Future studies should use a more rigorous, prospective design to document the impact of provider training and intervention on practice behavior, on family behavior, and ultimately on child weight. Other areas of investigation could include the association of provider interest in training with change in self-efficacy and practice behavior and also exploration of different behaviorally oriented approaches to the training.