Endocrinology/Primary Care Pharmacy Collaboration vs Endocrinology Care Alone in Patients With Type 2 Diabetes and A1c >9%

Abstract

Background: Type 2 diabetes (T2D) requires close collaboration between patients and their care management team, often including endocrinology. Primary care pharmacist impact on diabetes management in collaboration with endocrinology is not well established. Objective: To assess if pharmacy and endocrinology collaboration results in a greater A1c reduction in patients with T2D vs endocrinology alone. Methods: This retrospective, observational cohort study was conducted in adult outpatients with T2D and baseline A1c >9% who saw endocrinology within 1 year preceding the study period (January 1, 2021 to January 1, 2022). Patients were included if they had a follow-up A1c 6 months (±90 days) from index date and completed at least 1 endocrinology visit during the study period. Patients managed by endocrinology/primary care pharmacist collaboration (Endo/PharmD) were compared with those who received endocrinology care alone (Endo). Primary outcome was change in A1c from baseline to 6 months. Secondary outcomes included total number of completed visits and percentage of patients achieving A1c <6.5%, <7%, <8%, and <9% between groups at 6 months. Results: A total of 418 patients were included (22 Endo/PharmD, 396 Endo). The change in follow-up A1c was not significantly different between groups, −0.481% (standard error [SE] = 0.396); P = 0.6179. Endo/PharmD patients had significantly more provider visits during the study period (5.3 ± 2.3 vs 2.3 ± 1.2; P < 0.001). No significant difference was observed in odds of A1c goal attainment between groups at 6 months. Conclusion and Relevance: Endocrinology/primary care pharmacist collaboration occurred infrequently but was associated with a trend toward greater A1c reduction in patients with T2D and A1c >9%.

Introduction

Type 2 diabetes (T2D) is a complex, chronic disease that requires close collaboration between patients and their care management team. Collaboration between health care providers including primary care and subspecialty clinicians, nurses, dietitians, pharmacists, and other caregivers is beneficial for patients living with diabetes.^1,2 People with uncontrolled T2D may need to see an endocrinology specialist for treatment.

The Center for Disease Control National Diabetes Statistics Report in 2020 reported a total of 37.3 million people (11.3% of the US population) living with diabetes. ³ The percentage of the population diagnosed with diabetes is increasing, with 1 study projecting as many as 1 in 3 adults could have diabetes by 2050. ⁴ Increased prevalence of diabetes in the United States has augmented the demand for specialized providers to care for these patients.

Projections from the Endocrine Society predict a gap of approximately 1484 full-time endocrinologists which is anticipated to persist through 2025. ⁵ Continued shortage of endocrinologists means other resources are needed to fill this void, improve access, and optimize health outcomes and patient satisfaction. A systematic review and meta-analysis from 2021 found that in patients with T2D, the most effective approach for A1c reduction was involvement of nonphysician providers initiating and intensifying treatment. ²

The Northeast Ohio region of Cleveland Clinic provides multiple ambulatory pharmacy services including primary care and endocrinology services. There are 24 primary care pharmacists embedded in 17 primary care clinics who work collaboratively with primary care providers, whereas the 2 endocrinology pharmacists work in 2 endocrinology clinics and work collaboratively with endocrinology providers. Primary care and endocrinology pharmacists work under collaborative practice agreements (CPAs) with physicians and advance practice providers. The CPA allows pharmacists to provide medication management for various disease states including, but not limited to, diabetes. Primary care pharmacists at Cleveland Clinic operate at the top of their licenses through medication management by initiating, discontinuing, and adjusting medications. In addition, the pharmacists order and evaluate laboratory values to ensure safe and effective therapy, identify and address barriers to patient adherence, and provide extensive medication education to help patients reach therapeutic goals.

Literature has repeatedly demonstrated positive clinical outcomes when pharmacists in primary care settings are involved in diabetes management.^{2,6 -8} In addition, pharmacists are an accessible resource who provide care between endocrinology visits. Although many studies have evaluated the impact of pharmacists on diabetes outcomes in primary care settings, there are no studies evaluating the collaboration between primary care pharmacists and endocrinologists in the management of T2D.

Objectives

This study aimed to assess if primary care pharmacists in conjunction with endocrinology improved reduction of A1c in patients with T2D and A1c >9% as compared with endocrinology alone. The primary objective compared the change in A1c between groups 6 months following the initial visit. Secondary objectives compared the total number of completed visits between groups 6 months after the initial visit and the percentage of patients who reached A1c goals of <6.5%, <7%, <8%, and <9% at 6 months.

Methods

This Institutional Review Board exempt, retrospective, observational cohort study was conducted within an integrated health system from January 1, 2021 through January 1, 2022. Patients were included if they had a diagnosis of T2D, baseline A1c >9%, endocrinology provider visit within 1 year prior to the study period, follow-up A1c 6 months (±90 days) from index date (date of first visit during the study period) and had at least 1 follow-up visit with an endocrinology provider during the study period. Patients were excluded if they had type 1 diabetes, gestational diabetes, contact with a pharmacist making clinical decisions in T2D management prior to the study period, or used insulin pumps and/or insulin pods at baseline.

Patients were divided into 2 cohorts: endocrinology/primary care pharmacist collaboration (Endo/PharmD) and usual endocrinology care alone (Endo). Endocrinology/primary care pharmacist collaboration cohort patients were those who completed visits with an endocrinology provider within 1 year prior to the index date and an initial visit with a primary care pharmacist during the study period. Primary care pharmacists work under a CPA to initiate, adjust, and stop medications, order drug therapy monitoring labs, and provide education and counseling. Endocrinology care alone cohort patients were those who completed an endocrinology provider visit and had no primary care pharmacist visits during the study period. Patients in the Endo cohort did not have any pharmacist, including the endocrinology pharmacists, making decisions in the management of T2D during the study period. In this study, endocrinology providers included both endocrinologists and advanced practice practitioners. Advanced practice practitioners were defined as nurse practitioners and physician assistants. The Endo/PharmD cohort index date was the date of the initial pharmacist visit; the Endo cohort index date was the date of the first follow-up visit with the endocrinology provider within the study period. The patient’s study period was defined as the index date to 6 months.

Baseline demographics include age, gender, race, body mass index (BMI), estimated glomerular filtration rate (eGFR), insurance type, and number and type of comorbid conditions. Body mass index and eGFR were collected at the time closest to the index date. Comorbid conditions collected via the International Classification of Diseases 10th Revision (ICD-10) codes included heart failure, hypertension, chronic kidney disease (CKD), and coronary artery disease (CAD). Coronary artery disease diagnosis included a history of myocardial infarction (MI), stroke, peripheral artery disease (PAD), or unstable angina (UA). Chronic kidney disease diagnosis was based on eGFR categories (stage 3a eGFR 45-59, stage 3b eGFR 30-44, stage 4 eGFR 15-29, stage 5 eGFR <15) according to the Kidney disease: Improving Global Outcomes (KDIGO) guidelines. ⁹ Other data collected included consulting prescriber specialty and type and number of diabetes medications at baseline and follow-up (at end of the 6-month study period). Outcome data included baseline A1c, follow-up A1c, and the total number of visits with providers. The baseline A1c was the A1c closest to the index date, up to 3 months prior to, or up to 30 days after, the index date. The follow-up A1c was the repeat A1c drawn 6 months (±90 days) after the index date. Eligible visits included both in-person and distance health visits where the health care provider was making clinical decisions in the management of T2D. Pharmacists conducted 30-minute established visits and 60-minute new patient visits, whereas endocrinologists conducted 20-minute and 40-minute visits, respectively. Total visit count included the initial visit with the endocrinology provider or pharmacist plus all other follow-up visits during the 6-month study period.

Categorical data were analyzed using Pearson’s chi-square test or Fisher’s exact test, and continuous data were analyzed using analysis of variance (ANOVA) or Kruskal-Wallis. Due to the large imbalance between cohorts, this did not allow for matching cases and controls across potential confounders. Three sets of propensity scores were constructed to allow for comparison between groups and account for potential confounders. Potential confounders included for consideration in constructing the propensity scores were baseline A1c, age, gender, race, insurance type, BMI, eGFR, number of comorbid conditions, and number of diabetes medications at follow-up. Stepwise logistic regression methods were used to construct the 3 sets of scores. For each set, the final propensity scores were based only on those variables which were statistically significant (P < 0.05). Change in A1c was analyzed by multivariable linear regression with repeated measures to account for differences in baseline A1c. Propensity scores balanced the presence of significant confounders. Their inclusion in the model allowed for an unbiased comparison between groups. The list of variables used to construct propensity scores for the change in A1c included all potential confounding factors except for the baseline A1c. Secondary outcomes were analyzed by multivariable logistic regression for A1c goals achieved and multivariable linear regression for total number of visits between groups at 6 months. The derivation of propensity scores for A1c goals achieved used all potential confounding factors listed previously, and for total number of visits, propensity score calculations were based on all factors except for the total number of visits. All tests of significance were 2-tailed with an a priori significance level of P < 0.05.

Results

A total of 418 eligible patients were identified, 396 in the Endo cohort and 22 patients in the Endo/PharmD cohort. Table 1 summarizes patient demographics and clinical characteristics. Patients were similar between groups. Most patients were white (74%), female (52%), with a mean age (standard deviation [SD]) of 59 years (±13.2). Patients in the Endo/PharmD group were more often referred to the primary care pharmacist (77%) from primary care providers than by endocrinology providers (18%). Supplemental Table S1 summarizes the baseline diabetes medication use; there was no significant difference between groups as both used an average of 3 diabetes medications (P = 0.85). There was a significant difference in the use of glucagon-like peptide-1 receptor agonists (GLP-1 RAs), with more frequent use in the Endo/PharmD cohort than the Endo cohort (59% vs 32.6%; P = 0.011). Although metformin was more frequently used in the Endo/PharmD cohort (77% vs 61%; P = 0.12), the difference between groups was not statistically significant. The distribution of diabetes medication use was otherwise similar between groups. Overall, the most common diabetes medications used at baseline were insulin (72%), metformin (62%), and GLP-1 RAs (34%).

Table 1.

Baseline Characteristics.

	Total (N = 418)	Endo/PharmD (N = 22)	Endo (N = 396)	P
Age, y	58.8 ± 13.2	54.3 ± 13.4	59.0 ± 13.2	0.10
Female gender	216 (51.7)	13 (59.1)	203 (51.3)	0.47
Race
White	308 (73.7)	14 (63.6)	294 (74.2)	0.29
Black	61 (14.6)	6 (27.3)	55 (13.9)
Multiracial	23 (5.5)	2 (9.1)	21 (5.3)
Asian	5 (1.2)	0 (0)	5 (1.3)
Other	21 (5.0)	0 (0)	21 (5.3)
BMI, kg/m2	35.4 ± 8.2	36.0 ± 8.6	35.4 ± 8.2	0.76
eGFR, mL/min/1.73m2	85.5 ± 27.3	93.6 ± 25.3	85.1 ± 27.3	0.15
Comorbid conditions
Heart failure	75 (17.9)	3 (13.6)	72 (18.2)	0.59
Hypertension	365 (87.3)	19 (86.4)	346 (87.4)	0.89
Cardiovascular disease
MI	27 (6.5)	3 (13.6)	24 (6.1)	0.16
Stroke	55 (13.2)	4 (18.2)	51 (12.9)	0.47
PAD	48 (11.5)	4 (18.2)	44 (11.1)	0.31
CAD	108 (25.8)	7 (31.8)	101 (25.5)	0.51
UA	17 (4.1)	1 (4.5)	16 (4.0)	0.91
Chronic kidney disease
Stage 5	5 (1.2)	0 (0)	5 (1.3)	0.99
Stage 4	12 (2.9)	1 (4.5)	11 (2.8)	0.63
Stage 3b	31 (7.4)	0 (0.0)	31 (7.8)	0.17
Stage 3a	31 (7.4)	2 (9.1)	29 (7.3)	0.76
Total number of comorbid conditions	1.9 ± 1.5	2.0 ± 1.5	1.8 ± 1.5	0.63
Consulting prescriber specialty
Internal/family medicine	17 (4.1)	17 (77.3)	0 (0)	<0.001
Endocrinology	400 (95.7)	4 (18.2)	396 (100)
Other	1 (0.24)	1 (4.5)	0 (0)
Insurance type
Commercial	126 (30)	9 (41)	117 (29)	0.59
Medicare	192 (46)	10 (45)	182 (46)
Medicaid	64 (15)	2 (9)	62 (15)
Other	36 (9)	1 (5)	35 (9)

Abbreviations: BMI, body mass index; CAD, coronary artery disease; eGFR, estimated glomerular filtration rate; MI, myocardial infarction; PAD, peripheral artery disease; UA, unstable angina.

Statistics presented as mean ± SD or N (column %).

The primary and secondary outcomes and multivariable analyses are provided in Table 2. The primary outcome assessed the change in A1c between groups at 6 months. Baseline A1c values were significantly higher in the Endo/PharmD cohort (11.5% ±2.0) vs Endo cohort (10.3% ±1.3), P < 0.001. Follow-up A1c values were not significantly different between Endo/PharmD (8.1% ±2.0) and Endo (8.9% ±1.8) cohorts, P = 0.059. Patients in the Endo/PharmD cohort had significantly greater unadjusted A1c reduction compared with Endo cohort (−3.3% ±2.4 vs −1.5% ±2.0; respectively, P < 0.001). There was a clinically meaningful trend toward improved A1c reduction in the Endo/PharmD cohort; however, after models were adjusted based on propensity scores, no significant difference in the change in A1c between groups was observed (−0.481 [standard error (SE) = 0.396]; P = 0.6179).

Table 2.

Clinical Outcome Results.

Primary outcome
	Total (N = 418)	Endo/PharmD (N = 22)	Endo (N = 396)	P a
Baseline A1c	10.4 ± 1.4	11.5 ± 2.0	10.3 ± 1.3	<0.001
Follow-up A1c	8.8 ± 1.8	8.1 ± 2.0	8.9 ± 1.8	0.059
Change in A1c	−1.6 ± 2.0	−3.3 ± 2.4	−1.5 ± 2.0	<0.001
Change in A1c between groups	Mean change in A1c (%)		Standard error	P b,c
Follow-up A1c	−0.481		0.396	0.6179
Secondary outcomes
A1c goals achieved at 6 months between groups
A1c	Total (N = 418)	Endo/PharmD (N = 22)	Endo (N = 396)	Odds ratio b	P (95% CI) b
<6.5%	29 (6.9)	5 (22.7)	24 (6.1)	0.916	0.922 (0.16-5.233)
<7%	59 (14.1)	7 (31.8)	52 (13.1)	1.355	0.648 (0.367-4.993)
<8%	136 (32.5)	12 (54.5)	124 (31.3)	1.79	0.287 (0.614-5.222)
<9%	246 (58.9)	16 (72.7)	230 (58.1)	1.34	0.616 (4.196-0.616)
Differences in patient visits by patient group
	Total (N = 418)	Endo/PharmD (N = 22)	Endo (N = 396)	P b
Endo visit count	2.3 ± 1.2	1.7 ± 0.78	2.3 ± 1.2	<0.001
PharmD visit count	0.19 ± 0.94	3.6 ± 2.2	0 ± 0	<0.001
Total visits	2.5 ± 1.4	5.3 ± 2.3	2.3 ± 1.2	0.013
Mean difference
Endo/PharmD	Endo	Difference of means	Standard error	P b
		2.9	0.285	<0.0001

Abbreviation: CI, confidence interval.

Statistics presented as mean ± SD or N (column %).

^aANOVA.

^bModels adjusted based on propensity scores.

^cP values adjusted for multiple comparisons using Tukey-Kramer methods.

Secondary outcomes included the proportion of patients achieving A1c levels and the total number of visits. The proportion of patients achieving A1c goals were greater in patients in the Endo/PharmD vs Endo: <6.5% (22.7% vs 6.9%), <7% (31.8% vs 14.1%), <8% (54.5% vs 32.5%), and <9% (72.7% vs 58.9%, between Endo/PharmD and Endo cohorts, respectively). When adjusted based on propensity scores, the odds of A1c goal attainment were not significantly different between groups for any A1c category. Patients in the Endo/PharmD cohort had significantly more visits than the Endo cohort (5.3 ± 2.3 vs 2.3 ± 1.2; P < 0.001). Patients in the Endo/PharmD group had a mean of 3.6 ± 2.2 primary care pharmacist visits and 1.7 ± 0.8 endocrinology visits. Patients in the Endo cohort had no pharmacist visits and an average of 2.3 ± 1.2 endocrinology visits. After adjustment, the mean difference of visits between groups was significantly different (2.9 [SE = 0.285]; P < 0.0001).

Discussion

This analysis identified a clinically meaningful trend toward improved A1c reduction and A1c goal attainment among patients whose T2D was managed through a collaboration between primary care pharmacists and endocrinology providers. Although the small number of patients who met eligibility criteria for the Endo/PharmD cohort is a limitation of this report, the finding highlights that interdisciplinary collaboration is being underutilized.

One hypothesis for the underutilization of primary care pharmacists could be that endocrinology providers are referring patients to endocrinology specialty pharmacists or their own mid-level providers instead of primary care pharmacists. It was observed that patients in the Endo/PharmD group were most often referred to the primary care pharmacist by the primary care provider instead of the already-established endocrinology provider. The integrated delivery network the study was conducted at has a large team of primary care pharmacists, although only having a total of 2 pharmacists at 2 locations within the endocrinology department. Thus, endocrinologists at these 2 locations are more likely referring patients to their endocrinology specialty pharmacist. This leaves an opportunity for endocrinology providers at other sites to refer to their local primary care pharmacists.

One noted difference between groups was the more frequent use of GLP-1 RAs in the Endo/PharmD cohort compared with the Endo only cohort. This difference could be due to the increased time for education regarding medication adverse effects, injection training, and medication procurement during pharmacist visits (30-60 minute pharmacist visits vs 20-40 minute endocrinology visits). In addition, Endo/PharmD also had a significantly higher number of total visits, which could have provided more opportunities to reinforce medication or diet adherence, as well as the recognition and management of adverse effects that could potentially lead to medication discontinuation.

Another key difference between cohorts included the baseline A1c. The baseline A1c in the Endo/PharmD cohort (11.5% ±2.0) was significantly higher than the Endo group (10.3% ±1.3), which may suggest a fundamental difference in the patients who are being referred to primary care pharmacists in the Endo/PharmD cohort. Patients in the Endo/PharmD cohort had more poorly controlled diabetes and had a nonsignificant higher number of comorbidities suggesting they may be more medically complex than patients in the Endo cohort. Although we traditionally think of medically complex patients as having difficult to control diabetes, this difference between populations may confound the primary outcome results as patients with higher baseline A1c could have more A1c reduction with appropriate interventions.

The findings of this study support previously published analyses and guidelines, that a multidisciplinary approach to patient management is key in patients with diabetes. The American Diabetes Association guidelines specifically mention that patients with diabetes should receive care from a coordinated multidisciplinary team, including pharmacists. ¹ A 2021 systemic review and meta-analysis investigated strategies for overcoming therapeutic inertia on glycemic control in patients with T2D. ² The study found that involvement of nonphysician providers in management of T2D was the most effective approach to mitigate therapeutic inertia and improve A1c. Compared with physician interventions, pharmacist-interventions over a median duration of 1 year (0.9-36 months) were significantly more effective in reducing A1c (−0.90% to −0.60% vs −0.40% to 0.26%). In this study, pharmacy collaboration led to improvements in A1c similar to the pharmacy-based intervention studies included in the analysis by Powell et al. ²

Previous studies have evaluated the clinical outcomes of pharmacists practicing within primary care settings. A study by Schultz et al ⁶ compared the change of A1c between patients managed in a pharmacist-managed diabetes clinic vs usual care, which was defined as management by a primary care provider. Patients in a pharmacist-managed diabetes clinic had significantly greater reduction in A1c compared with usual care (absolute A1c change of −1.63% [95% confidence interval (CI) = −1.28 to −1.97] vs 1.53% [95% CI = 1.10 to 1.96]; P < 0.001). ⁶ Benedict et al ⁷ found that patients who had a primary care pharmacist on their diabetes management care team reached A1c goal <8% faster compared with usual primary care provider care (3.4 [2.7] vs 4.6 [2.7] months; P < 0.0001) and had a greater reduction in A1c at 6 months (−1.19% [1.65] vs −0.99% [1.59]; P < 0.009, between groups, respectively). Similar to this study, the patients in the pharmacist group had greater A1c reduction at 6 months compared with primary care provider care alone. The length of this study (6 months) could be seen as a limitation, as it may take medically complex patients especially longer to see changes in A1c. Despite the short time frame of the study, we were still able to observe a nonsignificant trend toward improved outcomes in patients in the Endo/PharmD group compared with the Endo group by comparing change in A1c at 6 months.

Although this study did not find a statistically significant difference in A1c between groups, and the results are limited by a low sample size, it did reveal a trend toward clinical benefit of interdisciplinary collaboration for the management of T2D. It highlights an opportunity to increase primary care pharmacists’ involvement in the management of patients with uncontrolled T2D seen by endocrine providers. Patients in the Endo/PharmD cohort had significantly more visits than patients in Endo cohort and a higher baseline A1c. This highlights the value of pharmacists serving as providers to initiate or intensify treatments, provide education, and reinforce medication and lifestyle regimens in between endocrinology visits. With pharmacists acting as providers between endocrinology visits, this may increase access to endocrinology care so that these providers may be able to see more unique patients.

Limitations of this study include the retrospective design, specifically regarding documented medication adherence. As medications were collected from the patients’ medical record, there was no way to account for adherence to the medication. We did observe a difference in the change in A1c between groups, but the low sample size of patients who received Endo/PharmD collaborative care limits the strength and generalizability of our results. In addition, the small sample size in the Endo/PharmD cohort did not allow for 1:1 matching of controls and cases, which is a more traditional approach to propensity score matching. However, to control for potential confounding factors that could influence the change in A1c, 3 separate propensity scores were constructed to allow for comparison between groups. An additional limitation of this study is the small number of nonwhite patients, which may limit generalizability of the study results. This study did not detect a difference in the change in A1c between groups at 6 months likely due to a small sample size.

Conclusions

In conclusion, no statistical significance was observed in the change in A1c between groups, although the results showed a clinically meaningful trend toward improved A1c reduction in patients receiving primary care pharmacist and endocrinology provider collaboration to manage T2D. Collaboration between primary care pharmacists and endocrinology providers was observed to occur infrequently, which highlights the opportunity to increase collaboration of endocrinology with primary care pharmacists in the management of T2D. Further investigation is warranted to determine the impact of such collaboration on the change in A1c as well as A1c goal attainment.