Primary Care Perspectives on Prescribing Anti-Obesity Medications for Adolescents

Structured Abstract

Background:

Recent pediatric guidelines recommend clinicians offer anti-obesity medication (AOM) as an adjunct to intensive lifestyle intervention.

Objective:

To investigate pediatricians’ perspectives about prescribing AOM, including barriers and facilitators.

Methods:

An investigator-developed survey was emailed to primary care pediatric physicians (n=187) and advanced practice providers (n=190) within an academic-affiliated network. The survey evaluated willingness to prescribe AOM and agreement with 25 statements about barriers and facilitators. Three vignettes explored AOM decision-making. Multinomial logistic regression was used to determine relative risk ratios for willingness to prescribe by agreement with each statement.

Results:

Among 74 respondents (20% response rate), 24% were willing, 42% uncertain, and 34% unwilling to prescribe. Most (64%) agreed that AOM should be managed only by specialists. Willingness to prescribe was associated with clinician motivation and belief in guideline practicality and applicability. Unwillingness was associated with beliefs that patients would not continue AOM long enough for benefit and that there was insufficient time or resources to implement. In vignettes, 52% were willing to prescribe AOM for a patient with severe obesity and metabolic complications, versus 11% for a patient with obesity and possible disordered eating.

Conclusions:

Willingness to prescribe AOM was low and was associated with perceived practicality and appropriateness for patients.

Introduction

Pediatric obesity affects nearly one-quarter of American adolescents,¹ is associated with comorbidities that adversely impact long-term health,² and disproportionately impacts youth who are marginalized by race, ethnicity, and socioeconomic status.³ Further, the rate of body mass index (BMI) increase among youth rose significantly during the COVID-19 pandemic, particularly among those with pre-existing overweight or obesity.⁴ In January 2023, to address the top public health priority of pediatric obesity,⁵ the American Academy of Pediatrics (AAP) released a Clinical Practice Guideline (CPG) for the Evaluation and Treatment of Children and Adolescents with Obesity.⁶ This CPG was the first to recommend that pediatricians should offer anti-obesity medication (AOM) to adolescents with obesity as an adjunct to health behavior and lifestyle treatment.

Although AOM can induce clinically meaningful weight loss,^7,8 concerns have been raised about accessibility, costs, long-term efficacy and safety.⁹ With over 50% of children in the U.S. covered by Medicaid/CHIP,¹⁰ cost is a significant factor for access to obesity treatment. Notably, only four medications have U.S. Food and Drug Administration approval for long-term use in children for weight loss (orlistat, phentermine/topiramate, liraglutide, and semaglutide), three of which were approved in 2020 or later.¹¹ Semaglutide, the most effective anti-obesity medication, with a 16.1% decrease in BMI, was approved in the month preceding the AAP’s CPG release.^7,12 However, the recency of these medication approvals has raised questions about the appropriate duration of use, particularly given the anticipated weight re-gain upon discontinuing the medication,⁷ which can be inevitable amidst supply shortages and changing insurance coverage. Another significant safety concern has been the theoretical onset/exacerbation of disordered eating behaviors in adolescents who are already at elevated risk due to higher BMI.¹³ Additionally, although obesity is discussed regularly during pediatric well-child care in the general pediatric setting and general pediatricians are often the ones supporting families with weight management, prescription of AOM is not currently standard of care for pediatricians. Families with marginalized or multiply marginalized identities may not have access to subspecialist prescribing AOM, further demonstrating the importance of general pediatricians in AOM prescription and management. Scant work has examined pediatricians’ perspectives about AOM prescriptions, which will be critical to equitable provision of these medications. Therefore, because primary care pediatricians are central to obesity management and are potential prescribers of AOM, we examined their perspectives about prescribing AOM to adolescents with obesity using a knowledge, attitudes, and behaviors framework for adopting new practice guidelines.¹⁴

Methods

An investigator-developed Qualtrics-based survey was sent by email to primary care pediatric physicians (n=187) and advanced practice providers (n=190) within an academic-affiliated primary care pediatric network in Pittsburgh, PA, in June 2023. No incentives were provided for survey completion, and two reminders were sent 1- and 2-weeks before the survey closed. The primary survey question of interest was: “To what degree are you willing to implement this guideline by prescribing weight-loss pharmacotherapy to adolescents 12 years and older with obesity (BMI ≥95th percentile) in your practice as it is now?” Of note, the wording of the survey questions included “weight-loss pharmacotherapy” rather than “anti-obesity medication” in order to maintain consistency with the AAP guideline. A 5-point Likert scale of very unwilling, somewhat unwilling, uncertain, somewhat willing, and very willing was used describe categories of willingness; however, due to the very small number of respondents, responses were collapsed to 3 categories of: unwilling (composed of very unwilling and somewhat unwilling), uncertain, or willing (composed of somewhat willing and very willing). The survey also included 25 items assessing attitudes about barriers and facilitators of use of AOM in accordance with the AAP CPG, as described in Table 2. The survey concluded with 3 clinical vignettes to explore clinical decision-making about prescribing AOM for adolescents with obesity. The vignettes were initially designed by a pediatric endocrinologist and medical student, then refined by a clinical psychologist who specializes in disordered eating in the setting of obesity and a pediatric obesity medicine specialist. The survey was pilot tested and further refined based on feedback by two primary care pediatricians. For each vignette, respondents were asked to rate how likely they were to prescribe AOM for the patient based on the description, as well as to rank the most important factors influencing their decision.

Table 2.

Full survey statements assessing attitudes about barriers and facilitators to anti-obesity medication prescribing.

Survey Statement	Associated Element in Knowledge-Attitudes-Behavior Framework14
Rate your agreement with each of these statements regarding weight loss pharmacotherapy in adolescents 12 y and older with obesity (BMI ≥95th percentile): (Survey displayed 5 scaled options for each statement: Strongly agree, agree, neither agree nor disagree, disagree, strongly disagree)
It will significantly reduce obesity and/or its consequences.	Outcome Expectancy
Pharmacotherapy for weight loss should be managed by specialist physicians only.
Adherence to these guidelines will cause increased eating disorders.
I am motivated to implement these guidelines into my clinical practice.	Motivation
I will be able to implement these guidelines into my clinical practice.	Self-Efficacy
I feel that these guidelines were developed in the best interest of adolescents with obesity.	Agreement with Guidelines (Specific and General)
It is safe.
It is practical.
It is cost-beneficial.
These guidelines are applicable to my patients.
I am confident in the people that developed these guidelines.
The evidence is strong.
Diet and exercise should be the focus for weight loss.
Weight loss is not linked to health outcomes.
Rate your agreement with these statements about potential external barriers that may prevent you from implementing the guidelines on weight-loss pharmacotherapy in adolescents 12 y and older with obesity (BMI ≥95th percentile) in your practice: (Survey displayed 5 scaled options for each statement: Strongly agree, agree, neither agree nor disagree, disagree, strongly disagree)
These guidelines offering weight-loss pharmacotherapy will increase my malpractice liability.	Environmental Factors
I will not be adequately reimbursed for implementing these guidelines offering weight-loss pharmacotherapy.
My organization has constraints on implementing these guidelines offering weight-loss pharmacotherapy.
There aren’t enough resources to implement these guidelines offering weight-loss pharmacotherapy.
There’s not enough time to implement these guidelines on offering weight-loss pharmacotherapy.
The pharmacotherapy guidelines are too complex or hard to follow.	Guideline Factors
There are contradictory medication guidelines influencing my practice.
My patients’ insurance won’t cover weight-loss pharmacotherapy.	Patient Factors
My patients’ families won’t want weight-loss pharmacotherapy.
My patients won’t want weight-loss pharmacotherapy.
My patients may not continue weight-loss pharmacotherapy long enough for sustained benefits

Note: “Weight-loss pharmacotherapy” rather than “anti-obesity medication” was the terminology used in survey questions to reflect exact wording in the AAP guideline.

Respondent characteristics were summarized using descriptive statistics. Differences in distribution of willingness to prescribe by respondent demographics were assessed using the chi-square test. Univariable multinomial logistic regression models were used to determine the relative risk ratio (RRR) for willingness to prescribe by strength of agreement for each CPG-related statement (strong disagreement to strong agreement; reference: neutral). To facilitate interpretation of findings, the reference category for the regression models was varied by the tone of statements: for positive statements about the CPG, the reference category was unwillingness to prescribe (i.e., RRR>1 indicate greater willingness to prescribe), while for negative statements, the reference category was willingness (i.e., RRR>1 indicate greater unwillingness to prescribe). For each vignette, the respondents’ likelihood to prescribe AOM and the most important factor driving the decision were tabulated. Analyses were completed in Stata 17, and a two-sided alpha of 0.05 was used to determine statistical significance. The University of Pittsburgh Institutional Review Board deemed the study as exempt.

Results

Eighty-two pediatric clinicians participated in the survey, 74 of whom reported their willingness to prescribe AOM (20% response rate). The vast majority self-identified as female (82%) and non-Hispanic White (91% White, 5% Asian, 1% Black/African American, 3% preferred not to say; 95% non-Hispanic/Latino/a, 3% Hispanic/Latino/a, 3% preferred not to say). Additional demographic characteristics are summarized in Table 1. Most reported at least some familiarity with the AAP guidance to offer AOM (36% aware of guidelines but had not read/reviewed, 27% reviewed guidelines but were unfamiliar with specific recommendations, and 12% read and understood the guidance, while 24% were unaware of guidelines). When presented with the verbatim statement from the AAP guidelines about use of AOM, approximately one-quarter (24%) were willing to prescribe AOM, 42% uncertain, and 34% unwilling. Most (64%) agreed that AOM should be managed by specialist physicians only, but less than half agreed that referral to obesity medicine specialists or endocrinologists was effective for management of pediatric obesity (48% and 41%, respectively). Effectiveness of referral to nutritionists was rated similarly to that of obesity medicine specialists and endocrinologists (47%).

Table 1.

Demographics of Respondents and Willingness to Prescribe AOM

	Overall, n (%)	Willing, n (%)	Uncertain, n (%)	Unwilling, n (%)	p
Gender
F	61 (82%)	13 (21%)	26 (43%)	22 (36%)	0.4
M	13 (18%)	5 (39%)	5 (39%)	3 (23%)
Clinician type
Attending Physician	18 (24%)	3 (17%)	10 (56%)	5 (28%)	0.7
Nurse Practitioner	44 (60%)	12 (27%)	16 (36%)	16 (36%)
Physician Assistant	12 (16%)	3 (25%)	5 (42%)	4 (33%)
Years in Practice
0–5	12 (16%)	2 (17%)	6 (50%)	4 (33%)	0.5
6–10	14 (19%)	3 (21%)	4 (29%)	7 (50%)
11–15	17 (23%)	5 (29%)	5 (29%)	7 (41%)
16+	31 (42%)	8 (26%)	16 (52%)	7 (23%)
Number of Patients per Week (n=1 missing)
< 100	44 (60%)	11 (25%)	18 (41%)	15 (34%)	1.0
≥ 100	29 (40%)	7 (24%)	12 (41%)	10 (34%)
Practice Setting^
Urban	22 (30%)	6 (27%)	8 (36%)	8 (36%)	0.9
Suburban	39 (53%)	10 (26%)	16 (41%)	13 (33%)
Rural	13 (18%)	2 (15%)	7 (54%)	4 (31%)
Proportion of Patients Using Medicaid
0–25%	19 (26%)	6 (32%)	6 (32%)	7 (37%)	0.4
26–50%	34 (46%)	6 (18%)	16 (47%)	12 (35%)
51–75%	15 (20%)	3 (20%)	6 (40%)	6 (40%)
76–100%	6 (8%)	3 (50%)	3 (50%)	0

^{^}Multiple clinicians provided care in multiple settings; categorized as urban if any practice was in urban setting, rural if any in rural, and suburban if only suburban; none were rural and urban

Figure 1 depicts agreement with each statement about the CPG guidance about prescribing AOM. Respondents were more likely to be willing to prescribe if they agreed that the guidance was applicable to their patients (RRR 24 [95% CI 2–331] for strongly agree), was practical (RRR 6.0 [95% CI 1.0–35.4] for somewhat agree), was in the best interest of their patients (RRR 11.7 [95% CI 1.5-.89.1] for strongly agree and willing; RRR 8.7 [95% CI 1.7–45.2] for strongly agree and unsure), and if they were motivated to implement it (RRR 22.7 [95% CI 2.3–221.0] for somewhat agree).

Figure 1.

Ratings of agreement with statements related to the CPG and AOM use in adolescents with obesity, separated by agreement with positive and negative statements. For each statement, responses were stratified by willingness to prescribe AOM: unwilling (U), neither (N), or willing (W). Statements marked with * indicate significant differences in agreement across willingness categories (p<0.05) in multinomial logistic regression. CPG: clinical practice guideline.

Respondents who disagreed that they would be able to implement the recommendation were more likely to be unwilling than unsure (using reference category of unwilling, unsure RRR 0.2 [95% CI 0.03–0.8] and 0.04 [95% CI 0.004–0.4] for strongly and somewhat disagree). Unwillingness to prescribe was more common in respondents who agreed that patients would not continue AOM long enough to see a benefit (RRR 30.0 [95% CI 2.9–315.6] for strongly agree and 8.3 [95% CI 1.5–46.9] for somewhat agree), that AOM should be managed by specialists (RRR 18.7 [95% CI 1.5–232.3] for strongly agree), and that there was insufficient time or resources to implement the guidelines (RRR 12.0 [95% CI 1.2–122.3] and RRR 30.2 [95% CI 2.9–315.6] for strongly agree with not enough time and resources, respectively).

Concerns about cost-effectiveness, patient/family desire for AOM, insurance, reimbursement, and liability were not associated with willingness to prescribe. Only 6.8% (5/73) agreed that AOM would increase the risk of eating disorders, and this belief was not associated with willingness to prescribe. Although 79% (58/73) agreed that diet and exercise should be the focus for supporting weight loss, this belief was also not associated with willingness to prescribe. Knowledge of the guidelines prior to the survey was not associated with willingness to prescribe.

When presented with clinical vignettes (Figure 2), reported likeliness of prescribing AOM varied for each case, with the highest reported likeliness (52%) to prescribe for a 12-year-old adolescent who was new to the provider’s practice, with history of depression, anxiety, irritable bowel syndrome, and weight of 300 pounds and metabolic complications including elevated blood pressure, prediabetes-range hemoglobin A1c, dyslipidemia, and elevated ALT. Presence of obesity-related complications was cited as the most common factor influencing higher likeliness to prescribe, while the short duration of the provider/patient relationship was the top factor influencing unlikeliness to prescribe AOM. The lowest likeliness (11%) to prescribe was reported for a case presenting a 16-year-old male with BMI at the 98^th percentile in the setting of a restrictive diet, low heart rate, and dizziness when standing, which was intended to reflect possible disordered eating and physiologic sequelae. Notably, “current weight loss attempts” was reported to be the most influential factor in decision-making for all categories of respondents (i.e., likely, uncertain, and unlikely to prescribe).

Figure 2.

Clinical vignettes and reported likelihood of prescribing anti-obesity medications and top factors influencing reported decision to prescribe.

Discussion

In this study examining attitudes related to recent AAP guidance about obesity management, a minority of primary care pediatricians were willing to prescribe AOM, while a majority agreed that AOM should be managed by specialist physicians. Respondents who reported personal motivation to prescribe and who agreed that the guidelines were applicable to their patients had greater willingness to prescribe. However, insufficient time and resources were identified as potential barriers to willingness to prescribe AOM.

Our findings underscore the roles of both attitudes and external barriers as factors influencing uptake of the recent AAP obesity guidelines. This is in alignment with a widely used framework developed by Cabana and colleagues that provides a “differential diagnosis” for why clinicians do or do not adhere to practice guidelines.¹⁴ This framework highlights knowledge, attitudes, and behavior as sources of clinical practice-related behavior change. In our study, we found that knowledge of the guidelines was not associated with willingness to prescribe, likely because awareness and knowledge of the guidelines were high overall in this sample. Consistent with Cabana et al.’s framework, we found that attitudes reflecting beliefs about expected outcomes of the therapy (e.g., patients will not continue AOM long enough to see benefits), self-efficacy (e.g., AOM should be managed by specialist physicians only), and applicability to patients were strongly associated with willingness to prescribe. We also found that external barriers were important factors in willingness to prescribe, including constraints on time and resources. These findings suggest that in order to modify pediatric clinicians’ behaviors related to the AAP guidance for use of AOM, both clinician attitudes and barriers must be addressed. This is also consistent with the Consolidated Framework for Implementation Research (CFIR), in which implementation depends not only on characteristics of individuals, but also on inner and outer settings, as well as processes and characteristics of an intervention or new practice.¹⁵

Our use of clinical vignettes provides insights into factors driving clinical decision-making about AOM. Each of these vignettes included adolescents meeting the age and BMI threshold for use of AOM according to the AAP guideline, yet reported likeliness of prescribing AOM was variable and informative. Our findings suggest that obesity-related complications are an important factor supporting willingness to prescribe AOM even for patients new to a provider’s practice. Importantly, we found that reported weight-loss attempts may be considered differently by different providers, with “current weight-loss attempts” cited as the most important factor driving decisions to prescribe and to not prescribe AOM for an adolescent with possible disordered eating. This finding likely reflects the pervasive agreement among respondents that diet and exercise should be the focus when supporting weight-loss. This case also illustrated the importance of clinical context when evaluating decision-making about AOM: although only 6.8% of respondents agreed with the statement that AOM would increase risk of eating disorders, in the vignette with the patient with possible disordered eating, 27% cited concern for the patient’s mental health as the top factor in unwillingness to prescribe AOM. Notably, in a systematic review and meta-analysis, structured obesity treatment has been found to be associated with lower risk for eating disorders among children and adolescents,¹⁶ but studies are needed to determine outcomes with real-world use of AOM for adolescents who do not take part in such structured, professionally-managed programs.

Our study included a timely and comprehensive assessment of attitudes and clinical decision-making about AOM use in the context of an AAP guideline that has significant implications for the approach to obesity management in adolescents. The use of clinical vignettes added context to clinicians’ reported agreement with statements about the CPG. These findings may guide primary care pediatricians and pediatric specialists involved in obesity management in understanding facilitators and barriers to uptake of AOM for adolescents with obesity. Limitations include the focus on a primary care network associated with a single academic center, the 20% response rate, and inability to examine actual prescribing.

In summary, within the first few months after the AAP released guidelines in support of AOM use for adolescents with obesity, primary care pediatric clinicians remained largely unwilling to prescribe AOM. While insufficient time and resources were associated with unwillingness to prescribe, individual motivation and belief in the guideline’s practicality and applicability were associated with greater willingness. As subspecialist access is limited and thus primary care pediatric clinicians provide frontline obesity management for most adolescents, their beliefs on barriers and facilitators can be viewed as expert opinion to inform practices going forward. For context, the network in which this study was conducted has limited availability of interdisciplinary specialty care for adolescents with obesity, emphasizing the importance of primary care findings in this and similar networks. Concerns about cost or insurance coverage were not associated with willingness to prescribe AOM, but these factors do impact access and are important considerations to achieve pharmacoequity^17,18 for youth with obesity. In addition, as disordered eating behaviors are more common among youth with higher BMI,¹³ the impact of AOM use, particularly outside of a structured obesity treatment program, should be the focus of future research. As more real-world data emerges, attitudes and practices of primary care pediatricians should be reevaluated to guide best practices about appropriate use and policies for AOM in adolescents.

Abbreviations:

CPG

Clinical Practice Guideline

AOM

Anti-obesity medication