Direct observation of weight counseling in primary care: alignment with clinical guidelines

Abstract

Primary care physicians provide care to a disproportionate number of overweight and obese patients and are uniquely positioned to help patients manage their weight in the context of a continuity relationship. The National Heart Lung and Blood Institute (NHLBI) developed evidence-based guidelines for the effective and efficient care of overweight/obese patients, but little is known about the use of these guidelines in practice. To determine the content of weight discussions and assess the elements of the NHLBI guidelines that were accomplished, office visits of 544 adult, overweight/obese patients to 28 primary care physicians were observed and audio-recorded. Associations between type of weight management discussion and patient, physician and visit characteristics were examined. Fifty percent (n=270) of visits included weight discussions; 47% and 38% included at least one NHLBI assessment or treatment elements, respectively. Only 36% (n=193) of discussions included an assessment and a treatment strategy; none included all NHLBI elements. Overall, adherence to guidelines was poor, particularly with regard to reporting BMI to the patient, measuring waist circumference, and setting realistic weight loss goals. Weight discussions did not clearly vary by the patient, physician or visit characteristics examined. These findings suggest opportunities to target resources for improved clinician training in communication and treatment techniques that are both consistent with current standards for effective, evidence-based practice and efficient enough for routine inclusion in the busy primary care visit.

Introduction

In the United States, addressing the rising prevalence of overweight and obesity in primary care is now recognized as a national health priority.(1) Primary care physicians provide care to a disproportionate number of overweight and obese patients(2) and are uniquely positioned to facilitate weight management discussions in the context of a continuity relationship.(3-5) The National Heart Lung and Blood Institute (NHLBI) developed evidence-based guidelines for the effective and efficient care of overweight/obese patients. These guidelines recommend primary care physicians address obesity as a chronic illness by first assessing and then treating overweight/obese patients. According to NHLBI guidelines, physicians should assess the patient’s body mass index (BMI), waist circumference, health risk status, and motivation for weight change. Subsequent treatment recommendations should include setting realistic weight loss goals as well as discussing diet, exercise, behavioral, pharmacological, or surgical therapies as appropriate.(6) Evidence suggests that this type of step-wise, multifaceted, weight management intervention is more comprehensive and effective than those focused on a single weight-loss method (e.g., diet advice only).(7)

Research on the content of weight counseling in primary care has focused on the frequency of one or two specific elements, such as discussions of diet or exercise,(4,8-15) referral to weight management experts(16,17), or BMI assessment.(18) No study has comprehensively examined the presence of all NHLBI assessment and treatment elements. Furthermore, most studies rely on patient or physician self-report through surveys or focus groups(4,8,10,11,19,20), methods prone to recall bias.(21) Accurate and comprehensive measurement of physician adherence to NHLBI guidelines can offer important insight about the quality of care received by overweight and obese patients.

The purpose of the current analysis was to document physician adherence to NHLBI guidelines for the assessment and treatment of overweight/obese patients using data obtained during directly observed primary care visits. Patient, physician, and visit characteristics were also examined for possible associations with the content of discussions about overweight/obesity during those visits.

Materials and Methods

Study Design, Setting, and Participants

We conducted a cross-sectional study within 16 outpatient primary care practices belonging to the Research Association of Practices (RAP), a practice-based research network in Northeast Ohio. Consecutively scheduled adult patients of 28 primary care physicians were mailed an invitation to participate in a study about physician-patient communication and were subsequently contacted by telephone. Participants completed a brief telephone-administered questionnaire prior to their office visit and their subsequent visit was directly observed and audio-recorded. Both physicians and patients were blinded to specific study hypotheses and were only informed that the study was about physician-patient communication. The University Hospital-Case Medical Centers Institutional Review Board approved the study protocol.

Physician characteristics.

Physicians’ gender, race/ethnicity, years since residency, and specialty type (internal or family medicine) were obtained from online physician directories for each physicians' affiliated hospital in Northeast Ohio.

Patient characteristics.

The telephone-administered survey assessed participant demographics (race, age, and education), health status (self-reported on a five point scale from excellent to poor), and diagnosis of a comorbid condition (diabetes, high cholesterol, high blood pressure, or heart disease). Comorbidities were categorized as none, one, or two or more. BMI (kg/m²) was calculated using self-reported height and weight and classified according to the World Health Organization’s BMI classification (22). Normal and underweight patients were excluded from the analyses. The study sample analyzed here consists of overweight patients (25 ≤ BMI < 30) with at least one comorbidity and obese patients (BMI ≥ 30) all of whom were eligible for weight management interventions during the clinical visit(6).

Visit Characteristics.

Duration of the office visit, defined as the time between the physician’s first entrance to the examination room and the physician’s exit at the end of the visit, was evaluated from the audio recording. Visits observed in practices predominately caring for the underserved were classified as safety net practices.

Content analysis and outcome variable definitions

Audio recordings of physician-patient encounters were transcribed and organized using ATLAS.ti Version 5.7.1 (Scientific Software Development GmbH, Berlin, Germany), a qualitative data management program. Each transcript was independently reviewed, and any text segment containing talk about weight, diet, or physical activity was identified. Next, analysts re-read transcripts and, using a standardized coding template, evaluated each text segment to determine the extent to which elements NHLBI guidelines were accomplished.

NHLBI weight assessment guidelines recommend that physicians assess a patient’s BMI, waist circumference, health risk status, and motivation to lose weight (6). If there was any mention of the patient’s BMI during the visit, the code BMI reported to patient was applied. Since no measurements of waist circumference occurred in any of the recorded visits, the code waist circumference measured was used for any talk referring to a patient’s pant size or inches lost around the waist. Health risk status was defined as any talk about how weight could contribute to a medical condition or when the physician recommended that a patient lose weight for health reasons. Assessed motivation to lose weight was coded whenever patients’ previous weight loss attempts or expressed desire to lose weight was discussed.

NHLBI treatment guidelines recommend setting a realistic weight loss goal and initiating a specific weight loss strategy, which should include a combination of low-calorie diets, increased physical activity, and behavior therapy (e.g., recommendation of self-monitoring, stimulus control or social support). Additionally, for eligible high-risk patients, NHLBI guidelines include options for pharmacotherapy and surgical procedures(6). For each transcript, reviewers coded discussions about setting a specific weight loss goal, exercise, or changes to diet. Talk in which physicians referred patients to commercial weight loss programs (e.g., Weight Watchers^™), nutritionists, weight loss clinics, or psychotherapists was coded, and together these codes defined a recommendation for behavioral therapy, since all incorporate behavioral strategies to increase patient adherence. Pharmacotherapy discussions included physician prescription or patient inquiry about weight loss medications. Talk about weight loss surgery, including referrals for bariatric surgery assessment, was also coded.

Finally, each case was categorized into one of three mutually exclusive categories: weight not discussed, assessment only, and treatment discussed. Weight not discussed was defined as having either no mention of weight or a single mention of weight that was ignored by the other person (physician or patient). Assessment only referred to discussion in which weight was identified as a problem, but there was no recommendation for change or treatment. Treatment discussed included physician- or patient-initiated talk about weight loss goals, diet, exercise or other therapies as methods for weight loss or maintenance.

Analyses

Descriptive statistics were used to summarize physician, patient, and visit characteristics, and the frequencies of each NHLBI guideline element coded. Generalized linear mixed models were used to examine association between these categories and patient, physician, and visit characteristics.

A multilevel multinomial analysis was conducted to determine the association of patient, physician, and visit characteristics with weight discussions. Weight not discussed was used as the reference group. Generalized linear mixed models were used to accommodate the multilevel structure and clustering of multiple patients seeing the same physician (Proc GLIMMIX). All associations were evaluated at p< 0.05. Analyses were performed with SAS Version 9.2 (SAS Institute Inc., Cary, NC).

Results

Among the 28 physician participants, roughly three-quarters (n=20) were internists, and one-quarter (n=8) were family physicians; half (n=14) were male; one-quarter (n=8) were black, and the average years in practice was 14. A total of 1108 patients enrolled in the study and completed the pre-visit survey, a 47% participation rate. Of these patients, 811 (73%) presented for their visit and were audio-recorded. Among these 811 patients, 67% (n=544) were obese or overweight with at least one chronic condition, and thus met the inclusion criteria for this analysis. These patients had a mean age of 53 ± 11years and were predominantly female (66%). Forty-one percent were black, and 32% had a high school diploma or less. Average BMI was 34 ± 7, and 36% reported very good or excellent health status. Mean time spent with the physician was 19 ± 10 minutes, and 33% of visits took place in a safety net practice.

Table 1 shows the frequency of each NHLBI guideline element during discussions between overweight/obese patients and their primary care physician. While all of these assessment strategies are recommended for each case, assessment of weight using at least one of the NHLBI guideline-recommended assessment strategies occurred in only 47% of discussions (n=255). At least one guideline-recommended treatment strategy was used in 38% of discussions (n=208). Discussions that included both an assessment strategy and a treatment strategy occurred in 36% (n=193) cases.

Table 1.

Frequency and percentage of discussions that included assessment and treatment recommendations from the NHLBI guidelines (n=544)

NHLBI Guidelines	Frequency (%)
Assessment (Any)	255 (47)
BMI reported to patient	10 (2)
Waist circumference measured*	4 (1)
Health risk status	163 (30)
Assessed patient’s motivation to lose weight	204 (38)
Treatment (Any)	208 (38)
Set specific weight loss goal(s)	30 (6)
Exercise discussed	169 (31)
Diet discussed	146 (27)
Behavioral therapy recommended	66 (12)
Commercial program	25 (5)
Nutritionist	33 (6)
Weight loss clinic	13 (2)
Psychotherapy	1 (1)
Pharmacotherapy	5 (1)
Surgery	3 (1)

NOTE: Discussions may have included implementation of more than one assessment or treatment guideline, therefore frequencies & percentages are not additive to the category total.

^*Determined by talk related to a patient’s pant size or inches lost around the waist.

The content of weight talk was categorized for each case as weight not discussed (50.4% n=274), assessment only (11.4% n=62), or treatment discussed (38.2% n=208). We examined the bivariate associations of physician, patient and visit characteristics with each weight talk group, while controlling for the clustering of patients seeing the same physician. Patients who were older, had fewer years of education, lower BMIs, or spent less time with the physician were less likely to be engaged in treatment discussion, as were patients seen at safety net practices.

In a multilevel, multivariable analysis (Table 2), independent of other variables in the model, patients with very good health status compared to fair/poor health status were almost three times more likely to engage in treatment discussion than no weight talk; however, this association was not observed for patients with excellent health status. Similarly, patients having only one comorbidity were more likely to engage in both assessment only and treatment discussed than no weight talk; however, this association was not observed for patients with two or more comorbidities. Younger age, higher BMI, and longer visit time were marginally associated with treatment discussion.

Table 2.

The association of characteristics with weight discussions during office visits using multinomial multilevel regression analysis (n=544)

	Assessment Only vs. Weight not Discussed OR (95% CI)	Treatment Discussed vs. Weight not Discussed OR (95% CI)
Physician Characteristics
Years since residency	1.01 (0.97, 1.05)	0.99 (0.95, 1.03)
Internal Medicine specialty	1.47 (0.62, 3.51)	1.08 (−2.08, 2.43)
Black	0.48 (0.12, 1.85)	0.82 (0.22, 3.04)
Male	1.09 (0.51, 2.34)	0.79 (0.37, 1.67)
Patient Demographic Characteristics
Age	1.00 (0.96, 1.03)	0.76 (0.95, 0.99)
Education
High school (Reference)	----	----
Some college	0.36 (0.16, 0.79)	0.97 (0.57, 1.65)
College degree or higher	0.81 (0.39, 1.70)	1.12 (0.64, 1.95)
Black	1.22 (0.53, 2.79)	1.17 (0.64, 2.10)
Male	0.82 (0.41, 1.64)	1.02 (0.62, 1.67)
Patient Health Characteristics
Body mass index	1.03 (0.99, 1.08)	1.05 (1.02, 1.08)
Self-reported health status
Fair/poor (Reference)	----	----
Good	1.65 (0.81, 3.38)	1.48 (0.86, 2.54)
Very good	1.30 (0.53, 3.17)	2.77 (1.53, 5.03)
Excellent	1.25 (0.39, 3.98)	1.88 (0.85, 4.15)
Co-morbidities*
None (Reference)	----	----
1	4.03 (1.24, 13.1)	1.92 (1.00, 3.66)
2 or more	2.86 (0.86, 9.53)	1.58 (0.80, 3.11)
Visit Characteristics
Time with physician	1.01 (0.98, 1.04)	1.03 (1.01, 1.05)
Safety net practice	1.78 (0.57, 5.54)	0.60 (0.20, 1.82)

NOTE: Multivariable multinomial model includes all potential predictor variables and adjusts for the clustering of multiple patients seeing the same physician. Two physicians had only one patient and therefore did not contribute to the multilevel analysis.

^*Comorbidities include diabetes, high cholesterol, high blood pressure, and heart disease.

Discussion

NHLBI guidelines incorporate key recommendations for both assessing and treating overweight/obesity in primary care and are designed to be implemented in a comprehensive, step-wise fashion. Given the prevalence of overweight/obesity and the important role primary care clinicians play in both the treatment and prevention of chronic diseases like diabetes, hypertension, and cardiovascular disease, assessments of clinician adherence to clinical guidelines can direct efforts to improve the effectiveness of brief advice for weight management in primary settings. This study used directly observed physician-patient discussions to document the content of weight talk and measured the degree to which that talk conforms to published guidelines.

Prior studies using direct observation of weight discussion among primary care physicians have shown considerable variation in the frequency with such talk occurs (17-56% of visits).(4,8,10) While nearly half of visits with overweight/obese patients in the current study included some discussion of weight, less than 36% of visits contained both an assessment and treatment discussion consistent with NHLBI recommendations. Perhaps more striking is the paucity of discussions that included the patient’s BMI (2%) or waist circumference (1%). Both BMI and waist circumference are practical indicators of the distribution of adipose tissue and are strongly correlated with increased risk for cardiovascular disease and mortality.(23,24) While it is possible that BMI or waist circumference were recorded in the patient’s medical record but not mentioned during the visit, explicit discussion of these indicators could potentially help patients understand the associated risk and offer a starting point for physicians to assist patients in establishing weight loss goals and a treatment plan.(25)

A more common assessment element was the identification of the patient’s health risk status by linking their weight to a presenting concern (30%). This strategy (i.e. framing the topic of weight by heightening its relevance as a problem) has been reported in other studies and may be a way for physicians to introduce the topic of weight management without negatively impacting the flow of the visit.(4,26,27) The most commonly observed assessment strategy was assessing a patient’s motivation to lose weight (38%), which is critically important as an accurate understanding of the patient’s current motivation to change behavior allows the clinician to tailor subsequent advice and offers of assistance.(28) However, despite being the most common form of assessment, nearly two-thirds of encounters with obese/overweight patients lacked any exploration of the patient’s level of motivation for weight loss.

Physicians and patients in our study rarely set weight loss goals during the visit (6%), despite NHLBI recommendations to do so. Achieving even modest weight loss (5-10% of body weight) is associated with decreased mortality and cardiovascular risk,(29,30) and research indicates that patients want physicians’ help in setting weight loss goals.(12) Furthermore, research has shown that those who meet initial goals for weight loss and physical activity are 1.5 to 3.0 times more likely to be successful in meeting long-term goals.(31) However, physicians report skepticism about their ability to help patients be successful with weight loss(32). Therefore, physicians and patients may be best served by focusing on incremental goals that are both medically meaningful and realistically achievable, which may have the dual benefit of increasing physician self-efficacy for providing effective weight loss advice as well as patient self-efficacy for losing weight. Finally, the continuity relationship developed in a primary care setting is ideal for setting modest incremental goals, monitoring of progress over time, and collaborating on strategy adjustments or additional goal-setting.

Diet modification and exercise were the most frequently discussed NHLBI-recommended treatment options, occurring in roughly one-third and one-quarter of discussions respectively. Other treatment options were rarely discussed. Evidence supports greater weight loss success when adjunctive behavioral therapies are used in addition to diet and exercise.(7,33) Given the overall low rate of treatment discussion, training clinicians to provide brief advice and assist patients with strategies and techniques for implementing new behaviors may be necessary in order to achieve acceptable levels of compliance with NHLBI guidelines.

Weight discussions in our study frequently lacked many NHLBI-recommended elements; however, weight discussions containing an assessment only or a discussion of treatment were, for the most part, equally likely to occur across most patient groups examined. While, previous studies have shown relationships between the frequency of weight discussions and patient BMI,(17,25,34-36) patient gender,(9,36,37) physician gender,(17) and the presence of chronic conditions.(9) Our analysis did show BMI associated with slightly increased odds of a treatment discussion taking place, but the difference was quite small. In the current study, observed associations between weight discussions and patient characteristics were marginal and lacked interpretable patterns of association. Furthermore, no associations between weight discussions and physician characteristics were found. Finally, the small association between both assessment and treatment discussions and the presence of one (but not multiple) comorbidity may suggest that while physicians use patients’ chronic concerns as opportunities to introduce or pursue the topic of weight management,(4,26) the complexity of treating multiple comorbidities in a brief office visit may constrain the time available to thoroughly address overweight/obesity.(38-40) Models for engaging in weight loss discussions that meet NHLBI guidelines should be developed and disseminated, but any new approaches to weight management talk will need to be both effective and efficient if they are to gain traction among the competing demands of primary care practice.

Several study limitations deserve mention. First, the cross-sectional design limits our ability to observe patients and primary care physicians as they navigate overweight/obesity in the context of a continuing relationship. Conceivably, some patients may have had their BMI or waist circumference assessed or had treatment options discussed during prior or subsequent visits to those observed in this study. Therefore, our findings may under-report the rate at which the NHLBI guidelines are followed over the span of clinical encounters that comprise the patient’s care. Second, our sample was limited to primary care physicians and patients from one region of the United States, so generalization of these findings to other practice settings or geographic areas should be done with caution. Lastly, this study focused only on the presence and content of weight management discussions captured in a single office visit; patient adherence to physician treatment plans or actual weight loss were not examined as outcomes. Despite these limitations, this study’s use of a rigorous method for analyzing the content of directly observed physician-patient communication provides good evidence for the degree to which primary care physicians’ assessment and treatment of overweight/obese patients conforms to NHLBI guidelines without the recall and reporting biases inherent in methods such as patient and physician self-report.

An improved understanding of physician’s adherence to NHLBI guidelines for assessing and treating overweight/obese patients provides opportunities to target resources for improved clinician training in communication and treatment techniques that are both consistent with current standards for effective, evidence-based practice and efficient enough for routine inclusion in the busy primary care visit. Equipping primary care clinicians to make effective use of available opportunities to provide weight loss treatment advice and tailored to the patient’s level of motivation for change is a necessary step in confronting the growing obesity epidemic.

Summary.

What is already known about this subject

• In the United States, addressing overweight and obesity in primary care is recognized as a national health priority.

• The National Heart Lung and Blood Institute (NHLBI) developed evidence-based guidelines which recommend that obesity be addressed as a chronic illness and that primary care physicians first assess and then treat their overweight/obese patients.

• Primary care physicians provide care to a disproportionate number of overweight and obese patients and are uniquely positioned to facilitate weight management.

What this study adds

• This study uses direct observation to describe primary care physicians’ assessment and treatment of overweight and obesity and evaluates physicians’ adherence to current evidence-based guidelines.

• Adherence to NHLBI guidelines was poor, particularly with regard to reporting BMI to the patient, measuring waist circumference, and setting realistic weight loss goals.

• The content of weight discussions did not show clear variation by patient, physician or visit characteristics.