Primary Care Residents' Knowledge, Attitudes, Self-Efficacy, and Perceived Professional Norms Regarding Obesity, Nutrition, and Physical Activity Counseling

Being overweight or obese is a risk factor for cardiovascular disease, type 2 diabetes mellitus, dyslipidemia, hypertension, osteoarthritis, and several cancers^1–3  and remains common in the United States. Today, more than two-thirds of American adults are overweight, with approximately 35% considered obese (body mass index ≥ 30).^4,5  Primary care providers have been called on to help address this epidemic by screening for obesity during ambulatory visits and providing weight reduction, nutrition, and physical activity counseling to their overweight and obese patients.^6–11  Although several guidelines exist for this counseling,^8,9,12–14  and mounting evidence shows physician advice can encourage weight loss among patients,^15–17  weight-related counseling during ambulatory visits occurs at modest rates and is infrequently aligned with recommended methods shown to support behavior change.^18–23  Studies examining barriers to weight counseling in primary care practice suggest physicians feel inadequately trained, and report a lack of time and poor self-efficacy to counsel on weight management.^24–28  Additionally, physicians may hold negative attitudes toward overweight and obese patients and their ability to sustain change, which may inhibit counseling.^29–31  Given this context, research assessing the extent to which residency training prepares primary care physicians to address patient obesity is warranted.

Previous reports show that only half of graduating primary care residents rate themselves as “very prepared” to counsel patients about diet and exercise.^32,33  Few studies have looked into explaining this issue or identifying specific areas of need, such as knowledge deficit, poor attitudes, or low confidence.^25,34–37  In addition, the literature is predominately focused on internal medicine,^34–37  and most studies include only a small number of programs,^25,34–37  raising concern about the ability to generalize from the findings. The purpose of this study is to assess the degree to which a statewide sample of senior family medicine (FM), internal medicine (IM), and obstetrics and gynecology (ob-gyn) residents are prepared to provide counseling on the topics of obesity, nutrition, and physical activity (ONPA). We assessed resident knowledge, attitudes, self-efficacy, and perceived professional role. We also examined demographic and training program characteristics potentially associated with variation in levels of preparedness.

This study represents a cross-sectional survey of senior primary care residents across Ohio, encompassing FM and IM postgraduate year (PGY)-3 and ob-gyn PGY-3 and PGY-4 residents.

Using maximum variation sampling, 30 of the 54 FM, IM, and ob-gyn programs in Ohio were selected for recruitment based on diversity of specialty, program size, location (north coast, northeast, northwest, and central/southwest), and hospital affiliation. Programs were invited to participate via a letter mailed to the program director describing the purpose and scope of the study. Program enrollment and subsequent data collection scheduling were completed by the study coordinator through follow-up contacts by phone and/or e-mail. Program participation was incentivized by offering an educational presentation about evidence-based strategies for ONPA counseling.

Surveys were collected in person as part of the educational presentation, which was scheduled during a didactic session for residents at each residency site. Resident attendance was incentivized by offering a light lunch or snack during the session. The first 15 to 20 minutes of each session were allotted for survey completion. The study and participation were described, and an information sheet and a paper survey were distributed to each resident in attendance. Surveys were collected as completed, and the 45-minute presentation followed.

Resident participation was voluntary, and no identifying data were collected on the survey. However, a sign-in sheet was used so that senior residents unable to attend the session could be identified and offered the opportunity to participate. For these individuals, survey administration was completed by e-mailing a link to an electronic version of the survey. Residents who completed the electronic survey received a $5 Starbucks e-gift card. Additional recruitment strategies were used when resident participation rates were below 70%, including a $100 Visa gift card raffle, and the mailing of a paper survey with a postage-paid return envelope.

Demographic characteristics on the survey included sex, age, ethnicity, race, US citizenship, medical school location, and chief resident status. Residency program characteristics that were assessed included specialty, self-reported time spent in ambulatory practice, and exposure to an elective rotation emphasizing ONPA counseling. Program location and number of first-year positions available (used as an indicator of program size) were both obtained from an online directory.³⁸ 

The indicators of counseling preparedness assessed included ONPA counseling knowledge and attitudes, self-efficacy, and perceived professional norms toward ONPA counseling. ONPA counseling knowledge was assessed by presenting 2 case studies with associated questions (15 items total), which is similar to the format of the United States Medical Licensing Examination and board examinations.³⁹  Items focused on obesity risks and recommended assessment and treatment.¹²  A knowledge summary score was computed for each respondent by assigning a value of 1 to each correct response, totaling, and transforming to a 0 to 100 scale. Resident attitudes, self-efficacy, and perceived professional norms regarding ONPA counseling were assessed using 12-item, 9-item, and 10-item measures, respectively. Items for the attitudes and self-efficacy measures were adapted from scales identified in the literature.^25,30,40  The professional norms measure was developed by the study team, guided by relevant theory and research.^41–43  All items used a 5-point Likert response format (1, strongly disagree, to 5, strongly agree). Summary scores were created by totaling responses and dividing by the number of items completed, then transforming into 0 to 100 scales. Higher scores represented greater or more positive responses. All survey items were pilot tested with a sample of 25 FM, IM, and ob-gyn fellows. The final measures are provided as online supplemental material.

The University Hospitals Case Medical Center Institutional Review Board approved the study protocol.

Data were stored using REDCap⁴⁴  and analyzed using SPSS Statistics for Windows 2013 (IBM Corp). Descriptive statistics were used to summarize resident demographic and training program characteristics and ONPA summary scores; associations were examined using bivariate linear regressions. Generalized linear mixed models were used to compute adjusted P values for resident respondents clustered within programs. The study was powered with a target sample size of 24 programs, with an average of 8 residents per program, providing adequate power to detect differences between groups in the magnitude of 0.35 of a standard deviation. Statistical significance was evaluated at P < .05.

Of the 30 residency programs invited to participate, 25 (83%) completed data collection. A total of 219 senior residents completed the survey, representing a response rate of 62%. TABLE 1 shows the demographic and program characteristics of participating residents. Just over half of respondents were female (58%, 113 of 194), white (53%, 101 of 190), and training in IM (58%, 127 of 219). Few (14%, 27 of 192) reported completing an elective rotation involving significant instruction in ONPA counseling. As some participants did not complete every item, valid percentages are reported for all frequencies.

Statistics for each ONPA counseling domain, including mean, minimum, and maximum scores and person and item reliability,⁴⁵  are shown in TABLE 2. Mean ONPA counseling knowledge scores were low (50.8, SD = 15.6), with senior residents answering just over half of the items correctly. Scores for attitudes, self-efficacy, and professional norms toward ONPA counseling also were low, on average, with a wide range of scores observed across the sample.

TABLE 3 shows associations between residents' ONPA scores and demographic and training program characteristics. ONPA counseling knowledge significantly varied by resident characteristics, with US citizens (B = 7.52, P = .01) and US medical school graduates (B = 6.48, P = .02) having higher scores with a magnitude of plus one-half standard deviation. However, these same characteristics were associated with significantly less positive attitudes and professional norms toward ONPA counseling. ONPA counseling self-efficacy varied by sex and location of medical school training, with female and US-trained residents reporting significantly lower self-efficacy. Scores for the 4 ONPA domains varied by race, but with no consistent trend. For program characteristics, specialty was associated with self-efficacy to provide counseling, with IM and ob-gyn residents significantly less confident providing ONPA counseling (B = −8.28 and B = −17.03, respectively) compared to FM residents. Residents training in ob-gyn also reported lower professional norms toward ONPA counseling compared to FM residents (B = −11.31, P = .002). Spending more time in an ambulatory practice location was associated with higher self-efficacy (B = 2.04, P = .01), but not knowledge, attitudes, or perceived norms toward counseling. Finally, residents reporting participation in an elective rotation that focused on ONPA had higher self-efficacy and more positive attitudes and professional norms scores (P < .001) with a magnitude of plus 1 standard deviation.

Our results show that senior FM, IM, and ob-gyn residents in Ohio currently are ill-equipped to provide ONPA counseling. On average, residents answered half of the knowledge items incorrectly, indicating that knowledge of recommended obesity counseling guidelines and techniques is limited. Likewise, mean counseling self-efficacy was modest (52 out of 100), suggesting that residents do not feel confident in their ability to perform basic ONPA counseling tasks. These findings are consistent with previous studies showing low self-perceived ONPA counseling proficiency and a poor grasp of strategies to address obesity among IM residents.^25,34–37  Studies examining barriers to weight management counseling in primary care practice also report inadequate training as a common barrier.^24,26,28 

We also assessed attitudes and professional norms toward ONPA counseling to explore the importance residents place on ONPA counseling and their perceptions regarding expectations and the importance placed on this area by their peers, preceptors, and professional societies. While no differences were observed in attitudes toward ONPA counseling across FM, IM, and ob-gyn, perception of professional norms were significantly lower in ob-gyn. This suggests that perceived barriers and challenges to addressing ONPA are similar across subspecialties; however, the perception of professional roles differ.

If primary care physicians are expected to address ONPA with their patients, training programs need to change. Examination of current curricula and critical reflection on the culture and attitudes maintained by program faculty and preceptors both are warranted. Efforts to increase instruction and experiential learning opportunities around ONPA and/or embed novelties of non-US predoctoral programs into US training should be explored to improve resident preparedness regarding ONPA counseling. This is considering the fact that international medical school graduates and residents who reported engagement in an elective rotation emphasizing ONPA showed greater counseling preparedness across domains.

Several study limitations deserve to be noted. First, we surveyed residents in a single state, and the generalizability of these findings may be limited to states and programs with similar characteristics. Second, despite multiple recruitment strategies and additional incentives, the response rate was modest. Residents with a greater interest in obesity care might have been more likely to participate, introducing the potential for participation bias and inflated scores. Third, the measures used in this study are new and have limited validity evidence. However, they were developed by drawing from the literature and were pilot tested. Finally, pediatrics residency programs were excluded from this study because we decided to focus on adult primary care trainees.

Primary care physicians see the majority of the population annually,⁵  and therefore have the opportunity to play a key role in addressing the American obesity epidemic by routinely counseling about obesity, nutrition, and physical activity during visits. Our findings show that significant gaps exist in primary care residents' preparedness to provide such counseling. Added training or an increased emphasis on ONPA is needed to enhance primary care physicians' skills to provide effective ONPA counseling and to prevent the health risks associated with obesity.