Health Care Provider Prescribing Habits and Barriers to Use of New Type 2 Diabetes Medications: A Single-System Survey Study Free

This study was conducted at an academic health care system that includes >1,200 physicians and 800 residents and fellows who practice at urban and suburban hospitals and ambulatory medical centers across southeast Michigan. The health system’s staff physicians, trainees, and advanced practice providers in cardiology, endocrinology, nephrology, and primary care (internal medicine and family medicine) were asked to complete a survey between 9 February and 14 March 2022. The survey was sent to 515 HCPs (113 cardiologists, 19 endocrinologists, 26 nephrologists, and 357 primary care physicians). We chose clinicians in these specialties because they would be most likely to have familiarity with and prescribe medications for patients with type 2 diabetes, ASCVD, CKD, heart failure, and multiple cardiovascular risk factors.

We developed an anonymous, online, multiple-choice questionnaire (SurveyMonkey, San Mateo, CA) that was based on review of published literature (23,24,28,31,32) and feedback from institution-based clinicians and social scientists. The full survey instrument is provided in Supplementary Table S1. A pilot was administered to 10 HCPs (seven in internal medicine, two in endocrinology, and one in cardiology) in addition to an epidemiologist, a biostatistician, and a medical writer experienced in survey development to validate the survey. The pilot found that respondents required on average <5 minutes to complete the survey. In addition to demographic data, participants were asked how frequently they prescribed drugs in each class of medication, which factors influenced their decision to prescribe a medication, how familiar they were with type 2 diabetes medications, and what they perceived as benefits of and barriers to prescribing GLP-1 receptor agonists and SGLT2 inhibitors. Respondents were informed that participation was voluntary and anonymous with no personal identifying information collected by the survey instrument. No incentive was offered to complete the survey. The link to the survey was sent via e-mail on 9 February 2022, with a reminder e-mail sent on 27 February 2022. The survey remained open until 4 March 2022.

We asked respondents to self-report their prescribing activities because we aimed to assess HCPs’ perceptions not only about their actual prescribing behavior, but also about their personal views on the medications, understanding of diabetes management, and self- perceived barriers to use of the medications. Thus, we aimed to obtain a generalized estimate to reflect the practitioners’ attitudes and to better understand the real world, lived experiences of HCPs in different disciplines and with different professional degrees or levels of training. The study was approved by the institutional review board of Henry Ford Health.

We sent the survey to 515 eligible HCPs. We expected a response rate between 25 and 50%. Therefore, the expected sample size was 129–258 HCPs. The demographic characteristics of HCPs in the five different specialties were compared using Fisher exact tests. All other survey responses were summarized using descriptive statistical analysis. A two-tailed P value of 0.05 was used to determine statistical significance.

The survey was sent to 515 HCPs, and 125 (24%) responded, including 22 (17.6%) in cardiology, 13 (10.4%) in endocrinology, 27 (21.6%) in family medicine, 54 (43.2%) in internal medicine, and 9 (7.2%) in nephrology. Demographic and descriptive characteristics of respondents are shown in Table 1. The group included seven (5.6%) certified nurse practitioners (CNPs) and 116 (92.8%) physicians, of whom seven (5.6%) were fellows and 30 (24%) were residents. Almost half of the respondents (60 [48%]) had been in practice for ≥11 years, 50 (40%) for <5 years, and 15 (12%) for 6–10 years. Most respondents (72 [57.6%]) practiced primarily at the health system’s flagship hospital, 28 (22.4%) at a satellite medical center associated with the health system, and 20 (16%) at other regional hospitals that are part of the health system.

All 94 respondents in endocrinology, family medicine, and internal medicine; 10 (45.5%) in cardiology; and six (66.7%) in nephrology prescribed medications to treat type 2 diabetes. Almost all respondents (95.2%) had some type of training in diabetes management. Nearly all (96.3%) of the respondents in endocrinology, nephrology, family medicine, and internal medicine and more than half (68.2%) of those in cardiology were aware of guidelines for the pharmacological management of type 2 diabetes. There were differences by specialty in respondents’ professional designations and main practice sites, as well as whether type 2 diabetes medications were prescribed and whether respondents were aware of guidelines (Table 1).

Prescribing data are summarized in detail in Supplementary Figure S1. Most respondents from all specialties and of all professional designations frequently prescribed (6 to >20 times per year) SGLT2 inhibitors, metformin, insulin, and GLP-1 receptor agonists (78.4, 76.8, 73.4, and 66.9% of respondents, respectively), followed by dipeptidyl peptidase 4 (DPP-4) inhibitors (43.9%), sulfonylureas (35.5%), and thiazolidinediones (TZDs) (8.9%). There were differences among the specialties in prescribing frequency of each medication class. Most respondents (54–100%) in endocrinology and primary care (internal medicine and family medicine) frequently prescribed metformin, DPP-4 inhibitors, GLP-1 receptor agonists, SGLT2 inhibitors, and insulin. Few respondents (<33%) in cardiology and nephrology frequently prescribed any of the diabetes medications except for SGLT2 inhibitors, which were frequently prescribed by 64–67% of respondents in these two specialties. More respondents in family medicine (70%) than in cardiology, endocrinology, nephrology, and internal medicine (9–38%) frequently prescribed sulfonylureas. More respondents in family medicine (25.9%) than in the other four specialties (<9%) also frequently prescribed TZDs.

There were differences among specialties with regard to prevalence of never prescribing different classes of diabetes medications. Most (59–95.2%) of cardiology HCPs never prescribed any of the diabetes medications, whereas only 13.6% of cardiology HCPs never prescribed SGLT2 inhibitors. Most nephrology HCPs (55.6–87.5%) never prescribed sulfonylureas, TZDs, DPP-4 inhibitors, GLP-1 receptor agonists, or insulin, and 33.3% never prescribed metformin; however, no respondents chose “never” for frequency of prescribing SGLT2 inhibitors. No respondents from endocrinology or primary care chose “never” for frequency of prescribing metformin or insulin. Almost no respondents (0–3.7%) in endocrinology or primary care chose “never” for prescribing GLP-1 receptor agonists or SGLT2 inhibitors. Some (<35%) of the respondents from endocrinology and primary care chose “never” for prescribing sulfonylureas or DPP-4 inhibitors, while more (29.6–63%) chose “never” for prescribing TZDs.

Prescribing frequency for most of the different diabetes medications was similar among HCPs of different professional designations. More than 50% of HCPs of all professional designations frequently prescribed metformin, GLP-1 receptor agonists, SGLT2 inhibitors, and insulin, while fewer (32–50%) frequently prescribed DPP-4 inhibitors. Differences in prescribing of sulfonylureas and TZDs were seen among the groups; more attending physician respondents frequently prescribed sulfonylureas (50%) and TZDs (14.3%) compared with <14.3% of trainees (interns, residents, or fellows) or CNPs for both medications.

Most respondents were comfortable to very comfortable prescribing metformin (90.4%), SGLT2 inhibitors (83.2%), insulin (75%), GLP-1 receptor agonists (72%), and DPP-4 inhibitors (58.4%), while fewer felt comfortable or very comfortable prescribing sulfonylureas (52.4%) and TZDs (29.6%). There were differences between specialties regarding comfort levels in prescribing type 2 diabetes therapies. Respondents in cardiology felt most comfortable prescribing SGLT2 inhibitors (63.7%) and metformin (54.5%); fewer (4.6–23.8%) felt comfortable prescribing insulin, GLP-1 receptor agonists, sulfonylureas, DPP-4 inhibitors, and TZDs. Nephrology respondents similarly had the highest comfort level prescribing SGLT2 inhibitors (100%) and metformin (88.9%), followed by insulin (55.5%) and sulfonylureas (52%); fewer (11–33.3%) felt comfortable prescribing GLP-1 receptor agonists, TZDs, and DPP-4 inhibitors. Respondents in endocrinology were comfortable prescribing all classes of type 2 diabetes medications (77–100%). Family medicine practitioners also had a high comfort level with prescribing all classes of type 2 diabetes medications (74–100%) except for TZDs (37%). Internal medicine practitioners had a high comfort level prescribing all medications (68–98%) except for TZDs (27.8%) and sulfonylureas (40.7%).

Comfort levels in prescribing most type 2 diabetes therapies were similar among HCPs of different professional designations, with >50% of respondents in all groups answering that they were comfortable or very comfortable with prescribing metformin, GLP-1 receptor agonists, SGLT2 inhibitors, and insulin. More than half of respondents who were attending physicians or CNPs also had a high comfort level with prescribing sulfonylureas and DPP-4 inhibitors, while slightly fewer (40–43%) in these two groups felt comfortable prescribing TZDs. Few (5.4–37.8%) trainees felt comfortable prescribing sulfonylureas, TZDs, or DPP-4 inhibitors.

Data about HCPs’ preferences regarding an add-on therapy to metformin are summarized in Supplementary Figure S2. The most frequently chosen second type 2 diabetes medications to be added to metformin were SGLT2 inhibitors (57.6%) and GLP-1 receptor agonists (55.2%); few (10.4–20%) chose insulin, sulfonylureas, or DPP-4 inhibitors, and no respondents chose TZDs as a second agent. Breakdown by specialty indicated a preference for SGLT2 inhibitors by respondents in cardiology (59.1%) and nephrology (66.7%), a preference for GLP-1 receptor agonists by respondents in endocrinology (100%) and family medicine (63%), and about equal preference for GLP-1 receptor agonists and SGLT2 inhibitors by internal medicine (61.1 and 63%, respectively) as additional agents. Analysis by professional designation revealed similar preference by respondents in all professional designations (≥50%) for GLP-1 receptor agonists or SGLT2 inhibitors as second agents to be added to metformin.

More than half of all respondents took the following factors into consideration when choosing a second type 2 diabetes medication to be added to metformin: reduction of comorbidities (82.4%), improvement in A1C (68.8%), favorable weight profile (67.2%), ease of taking medication (64%), lower cost (56.8%), minimization of hypoglycemia (54.4%), and mortality benefit (52.8%). Breakdown by specialty revealed that prescribing of type 2 diabetes medications by cardiologists and nephrologists was unlikely to be influenced by lower cost (13.6 and 22.2%, respectively), improvement in A1C (18.2 and 33.3%, respectively), favorable weight profile (18.2 and 22.2%, respectively), minimization of hypoglycemia (9.1 and 22.2%, respectively), or ease of taking medication (13.6 and 22.2%, respectively). Breakdown by professional designation did not reveal any clear differences, as most (>50%) of the respondents in each professional designation took all of the factors listed into consideration.

HCPs’ perceptions about benefits of and barriers to prescribing GLP-1 receptor agonists and SGLT2 inhibitors are summarized in Supplementary Figure S3. Respondents felt that the most important benefits of using GLP-1 receptor agonists and SGLT2 inhibitors were improvement of A1C (88.8 and 83.2%, respectively) and decrease in risk of comorbidities (75.2 and 94.4%, respectively). Most respondents also perceived favorable weight profile (88.8%) to be a benefit of GLP-1 receptor agonist use and decrease in mortality (69.6%) and ease of taking medication (62.4%) to be benefits of SGLT2 inhibitor use. The main differences among the specialties in perceptions of benefits to using GLP-1 receptor agonists were that few (<33%) respondents in cardiology and nephrology perceived minimization of hypoglycemia and ease of taking medication to be benefits compared with >50% of respondents in the other specialties. Also, 84.6% of endocrinology respondents felt that reduction in mortality was a benefit of using GLP-1 receptor agonists compared with 40–48% of respondents in all other specialties. The main differences among the specialties in perceptions of benefits to using SGLT2 inhibitors were that few (27–31%) of the respondents in cardiology perceived minimization of hypoglycemia and minimization of weight gain to be benefits compared with ∼50% or more respondents in the other specialties. Analysis by professional designation demonstrated that slightly fewer trainee respondents than attending physicians and CNPs perceived decrease in mortality (32.4 vs. 57.5%) and minimization of hypoglycemia (40.5 vs. 57%) as being benefits of prescribing GLP-1 receptor agonists. Fewer trainee respondents than attending physicians and CNPs perceived weight benefits (32.4 vs. 57.5%), minimization of hypoglycemia (35.1 vs. 57%), and ease of taking medication (46 vs. 76%) to be benefits of prescribing SGLT2 inhibitors.

More than 80% of respondents perceived high cost and problems with prior authorization to be the primary barriers to prescribing GLP-1 receptor agonists and SGLT2 inhibitors. More than half (58.4%) of respondents felt that the need to inject was a barrier to GLP-1 receptor agonist use, and 52.8% perceived adverse reactions to be a barrier to SGLT2 inhibitor use. Breakdown by specialty revealed that respondents in cardiology and nephrology (50 and 44.4%, respectively) felt that no or limited experience was a barrier to the use of GLP-1 receptor agonists compared with 0–7.4% of respondents in endocrinology and primary care. Limited experience was not perceived to be a barrier to SGLT2 inhibitor use by most respondents in all specialties. Analysis by professional designation demonstrated that fewer trainees than attending physicians or CNPs had concerns about adverse reactions with GLP-1 receptor agonists (8.1 vs. 25.5%) and SGLT2 inhibitors (40.5 vs. 57.5%). For GLP-1 receptor agonists, the need for injections was rarely perceived to be a barrier by CNPs (14%) compared with trainees or attending physicians (54–63%). Very few (0–18%) of the respondents of any professional designation perceived limited experience to be a barrier to using either drug.

Data on perceptions of which specialties should be responsible for GLP-1 receptor agonist and SGLT2 inhibitor prescribing are summarized in Supplementary Figure S4. Most respondents felt that the main prescribers of GLP-1 receptor agonists and SGLT2 inhibitors should be primary care physicians (85.6 and 90.4%, respectively) and endocrinologists (69.6% for both drug classes). Breakdown by specialty revealed that 60–80% of respondents in cardiology, endocrinology, and nephrology also felt that cardiologists should be a main prescriber of SGLT2 inhibitors. Most respondents in endocrinology and nephrology (78–84%) felt that SGLT2 inhibitor prescribing should also be a responsibility of nephrologists.

Data on which staff members were responsible for securing prior authorizations are summarized in Supplementary Figure S5. Prior authorizations were reported to be completed by the department secretary (35.2%), a specialty pharmacy (20.8%), a regular pharmacy (7.2%), or other (21.6%), which included clinical pharmacists, medical assistants, nurses, clinic nurse managers, attending physicians, or office staff. Fourteen (11.2%) of the respondents did not have help with prior authorizations. Analysis by specialty showed that 76.9% of respondents in endocrinology reported that a specialty pharmacy handled prior authorizations, and 50% of respondents in internal medicine had help from a department secretary. Fewer than 33% of respondents in cardiology, family medicine, and nephrology reported having a department secretary, a pharmacy, or another resource person available as an option for help with prior authorizations. Among the respondents who perceived high cost and/or prior authorizations to be burdensome for GLP-1 receptor agonists and SGLT2 inhibitors, 44.7–48.1% had a department secretary, and 21.5–28.2% had a specialty pharmacy to help with prior authorizations; few (14.1–17.7%) did not have anyone to help with prior authorizations (Table 2).

Our health care system–wide survey of HCPs who treat patients with type 2 diabetes showed that, although most practitioners follow current practice guidelines for early use of GLP-1 receptor agonists and SGLT2 inhibitors, barriers to prescribing these therapies still exist. Among our main findings, we observed that agents in these classes were frequently prescribed and chosen as early therapy for diabetes and that most HCPs were aware of their benefits. Notably, high cost and problems with prior authorization were perceived as barriers to prescribing these newer agents, and cardiology and nephrology practitioners rarely prescribed GLP-1 receptor agonists, feeling that limited knowledge was a barrier. Analysis by professional designation showed similar prescribing practices of most diabetes medications.

Few studies have looked at prescribing patterns for type 2 diabetes medications, the factors influencing prescribing, and, specifically, the barriers to prescribing GLP-1 receptor agonists and SGLT2 inhibitors. Cross-sectional studies evaluating trends in prescription patterns for type 2 diabetes medications from 2003 to 2019 in the United States have shown persistently high prescribing of metformin, declining prescriptions for sulfonylureas and TZDs, a steady increase in prescribing of newer insulin analogs, and an increase in prescriptions for DPP-4 inhibitors, GLP-1 receptor agonists, and SGLT2 inhibitors, although overall prescriptions for the latter two medications were still low (31–33). The most recent reports of diabetes medication prescribing practices in the United States from 2015 to 2019 showed that most prescriptions were for metformin (72–83%), followed by insulin (10–50%), sulfonylureas (5–50%), DPP-4 inhibitors (2–20%), GLP-1 receptor agonists and SGLT2 inhibitors (<10%), and TZDs (<5%) (31–33). A 2018 survey of HCPs who treat adults with type 2 diabetes demonstrated similar findings, with most HCPs reporting that they most frequently prescribed metformin (85%), followed by insulin (77.8%), sulfonylureas (65.8%), DPP-4 inhibitors (42.4%), GLP-1 receptor agonists (23.3%), SGLT2 inhibitors (8.4%), and TZDs (4.1%) (23).

Few studies have compared the prescribing patterns of type 2 diabetes medications between practitioners of different specialties, particularly HCPs who may frequently encounter patients with type 2 diabetes and concurrent cardiorenal comorbidities. Cross-sectional studies analyzing data through 2018 and 2019 in adult patients with type 2 diabetes have shown that most diabetes medications, including GLP-1 receptor agonists and SGLT2 inhibitors, were prescribed by an internal medicine physician (55–69%), while fewer were prescribed by an endocrinologist (2.4–19%) or cardiologist (0.3–10%) (31,34). Surveys of clinicians in primary care, endocrinology, and cardiology have shown that GLP-1 receptor agonists and SGLT2 inhibitors were perceived to be most frequently prescribed by endocrinologists and primary care providers (>50%) relative to cardiologists (<15%) (24,30,35). Interestingly, large U.S. studies of patients with type 2 diabetes with and without ASCVD or other cardiorenal comorbidities who were being treated in cardiology, endocrinology, and primary care revealed that, although 26–48% of patients were eligible for treatment with a GLP-1 receptor agonist or SGLT2 inhibitor, only 2.5–11% were actually prescribed these medications, and prevalence of use was lower in the subset of patients with ASCVD (26–29). One retrospective analysis evaluated trends in prescribing GLP-1 receptor agonists and SGLT2 inhibitors relative to DPP-4 inhibitors for patients with cardiorenal comorbidities by HCPs not only in primary care, cardiology, and endocrinology, but also in nephrology (36). Compared with endocrinology HCPs, primary care and nephrology HCPs were significantly less likely to initiate GLP-1 receptor agonists or SGLT2 inhibitors as opposed to DPP-4 inhibitors; however, cardiologists were more likely to initiate SGLT2 inhibitors, but less likely to initiate GLP-1 receptor agonists than DPP-4 inhibitors (36). In this study, patients with ASCVD, nephropathy, and/or heart failure were also less likely to be started on a GLP-1 receptor agonist and/or an SGLT2 inhibitor than those without these conditions (36).

Our study similarly revealed that endocrinology and primary care HCPs predominantly prescribed all diabetes medications, and most practitioners frequently prescribed metformin and insulin, while few frequently prescribed sulfonylureas and TZDs. In contrast to previous studies, most HCPs in our study frequently prescribed SGLT2 inhibitors and GLP-1 receptor agonists. Our study also demonstrated frequent prescribing of SGLT2 inhibitors by respondents in cardiology and nephrology. Reasons proposed for the low use of GLP-1 receptor agonists and SGLT2 inhibitors in prior studies of diabetes medication prescribing patterns include lack of HCP familiarity with these newer medications and clinical guidelines at the time the studies were conducted, high cost of these agents compared with other medications, and formulary restrictions (31). Another reason may be that HCPs need a certain amount of time to translate research evidence into clinical practice and adopt clinical practice guidelines. Research on implementation of clinical practice guidelines has suggested that up to 17 years, on average, is required for only 14% of published evidence to be incorporated into clinical practice (37). Many of the published studies on diabetes medication prescribing patterns evaluated data up to 2015–2019, before or only a few years after the first SGLT2 inhibitor, empagliflozin, was approved by the FDA. Our study was conducted 3–7 years later than these published studies, thereby allowing more time for HCPs to have become familiar with the cardiovascular outcome trials of GLP-1 receptor agonists and SGLT2 inhibitors, clinical practice recommendations for early use of these agents, and prescribing for these medications. Respondents in our study practiced in an academic environment and were required to teach and take part in continuing medical education, further increasing the likelihood of awareness of new medications and guidelines. Furthermore, previous studies on type 2 diabetes medication prescribing patterns (24,30,31,34,35) were conducted before empagliflozin and dapagliflozin were approved by the FDA for heart failure (and also for CKD in the case of dapagliflozin), in the absence of type 2 diabetes (16–19). The greater proportion of cardiologists and nephrologists who reported frequently prescribing SGLT2 inhibitors in our study may have been a result of these clinicians using these agents to treat heart failure and CKD rather than type 2 diabetes.

We found only one retrospective analysis on prescribing patterns of SGLT2 inhibitors and GLP-1 receptor agonists that included nephrologists (36). Our study is the first to evaluate nephrologists’ perceptions of their diabetes medication–prescribing behaviors; however, the sample size for this group was small, and our findings need to be confirmed by larger studies. Overall, the current available data on diabetes medication prescribing patterns and our study results suggest that interprofessional dialogue is needed to clarify the expectations of cardiologists and nephrologists with regard to prescribing GLP-1 receptor agonists and SGLT2 inhibitors for type 2 diabetes and specifically for cardiorenal benefits outside of the context of diabetes.

Although HCPs in our study reported frequently prescribing the newer diabetes medications, respondents still perceived barriers to prescribing these drugs. Reasons for their underutilization identified by earlier studies include high out-of-pocket cost, problems with insurance coverage, and knowledge gaps (23,24,31). In our study, the barriers to prescribing these agents most frequently chosen by respondents were also high cost and problems with prior authorizations. Less than half of our respondents identified staff who helped with prior authorizations. One reason why high cost and need for prior authorizations are major concerns in our health system is the high percentage of patients covered by Medicare (42%) and Medicaid (17%) in our patient population. The patients covered by Medicare generally encounter high out-of-pocket costs, whereas those covered by Medicaid have lower out-of-pocket costs but may be unable to obtain newer medications because of formulary restrictions. The annual out-of-pocket costs for an SGLT2 inhibitor and a GLP-1 receptor agonist in 2019 was reported to be $1,000 and >$1,500, respectively, for the average beneficiary covered by a Medicare Part D plan; this cost is similar to or even higher than the cost of insulin (38). With regard to Medicare coverage for GLP-1 receptor agonists and SGLT2 inhibitors in 2019, coverage without prior authorization and without step therapy varied from 64 to 95% depending on the specific medication (38). Patients covered by Medicaid, on the other hand, were found to have low out-of-pocket costs for SGLT2 inhibitors, but they faced stringent formulary restrictions, with only 30–35% of SGLT2 inhibitors being covered without prior authorization or step therapy in 2021 (39). Of note, even patients with commercial insurance seen at our health system face formulary restrictions. Commercial or employer-based health insurance was found to cover only 25–60% of SGLT2 inhibitors without prior authorization or step therapy in 2021 (39). High out-of-pocket costs, formulary restrictions with requirements for step therapy, and the need for prior authorization indicate a need to educate HCPs and clinic staff regarding cost-saving resources such as manufacturer patient assistance programs and copay cards for GLP-1 receptor agonists and SGLT2 inhibitors. Furthermore, greater administrative assistance is needed to help HCPs with prior authorizations, formulary rules, and understanding of which medications should be prescribed first when step therapy is required by insurance providers.

A barrier to prescribing GLP-1 receptor agonists identified in our study by respondents in cardiology and nephrology was lack of comfort and experience with prescribing these medications. A survey study at an academic health care system also found that cardiologists reported a lack of knowledge and discomfort with prescribing type 2 diabetes medications as top barriers to the use of GLP-1 receptor agonists and SGLT2 inhibitors (24). The discomfort of these specialists may stem from their perception that prescribing type 2 diabetes medications does not fall within the scope of their discipline. A survey of cardiologists in the United Kingdom found that only 11% felt responsible for delivering diabetes care for their patients who had concurrent acute coronary syndrome, with most deferring to a diabetes specialist (30).

All of the barriers identified in our study contribute to therapeutic or clinical inertia, which refers to the failure to advance or intensify treatment regimens when indicated. This phenomenon has been described as a key reason why people with diabetes may not reach their glycemic targets (40) and why eligible patients with diabetes and cardiorenal risk are not started on GLP-1 receptor agonists and SGLT2 inhibitors (25). This inertia is felt to be driven in large part by HCP barriers such as time constraints, the need to address other pressing medical issues during visits, preferential use of familiar type 2 diabetes agents, lack of knowledge and experience with newer therapies, and perceived adverse effects of medications (25,40). Patient barriers (e.g., concerns about side effects, lack of knowledge regarding treatment options, difficulty with adherence, and fear of injections) and system barriers (e.g., cost and availability of new medications) also contribute, but to a lesser degree than HCP barriers (40). Disparities related to certain patient demographics have also been described, with certain races/ethnicities (Black, Hispanic, and Asian), female sex, and older age being associated with decreased likelihood of starting a GLP-1 receptor agonist or an SGLT2 inhibitor (36). We did not study the effect of different patient races/ethnicities on prescribing, but despite our health system being located in Detroit, MI, where 77.1% of the population is Black/African American as of the 2021 census, prescribing of these newer agents was perceived to be frequent by HCPs, indicating that race/ethnicity was probably not a major determinant of prescribing practices of diabetes medications in our health system. We did not specifically ask about therapeutic inertia, but HCPs’ responses perceiving a lack of experience or knowledge, adverse effects, and costs as barriers to prescribing newer medications suggest that HCP and system barriers are present and contribute to therapeutic inertia.

Our analysis of prescribing patterns by different professional designations did not reveal major differences in prescribing behavior and attitudes among trainees, attending physicians, and CNPs, with all types of HCPs having high prescribing frequency of GLP-1 receptor agonists and SGLT2 inhibitors, good awareness of the benefits of using these medications, and similar perceived barriers of high cost and prior authorization for these medications. Trainees did have lower prescribing frequency and comfort level with older medications (sulfonylureas and TZDs), compared with attending physicians and CNPs. Trainees in our health care system are taught current practice recommendations for management of diabetes in the form of lectures and informal teaching by attending endocrinologists, endocrinology fellows, and supervising attending physicians during their clinical rotations. Trainees likely model the prescribing behaviors of their supervising attending physicians, which would explain the similarities in prescribing habits between trainees and attending physicians. We only found one study that evaluated differences in prescribing practices of diabetes medications among providers based on years of practice; this study also did not find any differences (23).

Our single-system sample population may not be generally representative of all HCPs, particularly those who practice in communities rather than in academic institutions with access to subspecialty care and regular continuing medical education. Respondents also may have had recall bias.

The number of nephrologists in our study was small and may not be representative of the prescribing behavior and attitudes of this specialty. The number of CNPs was also small and likely does not represent the prescribing behavior and attitudes of this group. No physician assistants responded to the survey, so this group was not included in the analysis. Because of the small number of fellows, this group was combined with interns and residents and analyzed as a whole.

Our 24% survey response rate was typical of response rates for online surveys of HCPs, which have been reported in other studies as 11–35% (41) or 31–36% (23,24). Our survey’s length and the short time the survey was open with only one reminder e-mail may have limited participation. However, the number of respondents (n = 125) was comparable to or larger than that reported in recent studies (23,24,30).

We kept the survey open for 33 days based on data from SurveyMonkey showing that 80% of responses were collected within 7 days, whereas only 11 and 4% of responses were collected in the second and third weeks of the survey period, respectively (42). Published survey studies on this topic similarly kept their surveys open for 25 days (24) or 31 days (23). A survey of cardiologists was kept open for 5 months, yet the number of respondents was 103—less than our survey (30). One reminder e-mail was sent in our study because studies have shown that the majority of responses (17.8%) occur after the initial invitation, with only 3.4% additional responses after first and subsequent e-mail reminders (43).

We assessed self-reported prescribing behavior, which may not align with actual prescription activity; therefore, our findings represent estimates. However, our aim was to evaluate HCPs’ perceptions of type 2 diabetes medications and attitudes regarding prescribing, especially for GLP-1 receptor agonists and SGLT2 inhibitors.

Clinicians in different specialties at our institution reported prescribing type 2 diabetes medications in accordance with published clinical practice recommendations that advise early use of GLP-1 receptor agonists and SGLT2 inhibitors for cardiovascular and renal benefits in patients with type 2 diabetes. Most HCPs in all specialties recognized the glycemic, cardiovascular, and renal benefits of agents in these classes. Primary care and endocrinology clinicians reported prescribing type 2 diabetes medications most frequently, whereas cardiology and nephrology specialists reported low prescribing activity for GLP-1 receptor agonists but frequent prescribing of SGLT2 inhibitors. High cost and need for prior authorization were the main perceived barriers to prescribing these newer agents. Cardiologists and nephrologists cited limited experience with GLP-1 receptor agonists as a barrier to using these medications.

Our study suggests that more education is needed for HCPs in all specialties on manufacturers’ patient assistance programs, copay cards, and insurance formularies. Our findings suggest that greater administrative assistance is needed to simplify and expedite the prior authorization process and that algorithms for prescribing GLP-1 receptor agonists and SGLT2 inhibitors when step therapy is required by insurance companies would be helpful. More education targeted at specialists in cardiology and nephrology on how to identify candidates for GLP-1 receptor agonists, how to prescribe these medications, and how to monitor patients on these therapies for adverse effects and efficacy would help HCPs in these specialties incorporate GLP-1 receptor agonists into their practices. Discussion among specialties would be beneficial in clarifying the role of cardiologists and nephrologists in prescribing medications for type 2 diabetes and diabetes medications that may be used for cardiorenal benefits outside the context of type 2 diabetes.

The authors thank Karla Passalacqua, PhD, of Henry Ford Hospital, for her editorial assistance with this article.

No potential conflicts of interest relevant to this article were reported.

A.Y. researched the data, contributed to writing, and reviewed the manuscript. S.W.L. researched the data, wrote the manuscript, and reviewed and edited the manuscript. S.W.L. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.