Abstract
Background
The goal of this study was to examine referral and attendance patterns for diabetes self-management education and support (DSMES) among patients with both type 1 (T1D) and type 2 diabetes (T2D) at a regional medical center in Kentucky, and to identify demographic and clinical factors associated with these outcomes.
Methods
We analyzed electronic health records of adults with diabetes (n = 10,587; n = 817 with T1D and n = 9,770 with T2D) who received care from 1/1/2016-12/31/2019 at University of Kentucky HealthCare. We compared DSMES referral and attendance rates by various demographic and clinical factors and used logistic regression models to examine the association between these factors and DSMES referrals/attendance.
Results
DSMES referrals were made for 6.9% (n = 726) of our sample, and 40.2% (n = 292) of those referred attended DSMES. Referral rates were 11.6% for T1D and 6.5% for T2D. Attendance rates were 41.1% for T1D and 40.1% for T2D. Referrals was more common among females (OR 1.67, 95% CI 1.42–1.96), non-Hispanic Black and Hispanic individuals (OR 1.36, 95% CI 1.12–1.65 and OR 1.56, 95% CI 1.03–2.35), and less common among those aged 65+ (OR 0.31, 95% CI 0.20–0.48), those with public insurance (OR 0.75, 95% CI 0.63–0.88), and rural residents (OR 0.48, 95% CI 0.40–0.57). Patients with obesity (OR 1.42, 95% CI 1.18–1.71) and ≥ 9% A1C (OR 2.35, 95% CI 1.65–3.34) were also more likely to be referred. No factors were associated with DSMES attendance.
Conclusions
Despite clear guidelines recommending DSMES referrals for patients with diabetes, DSMES referral rates were low, and less than half of patients who were referred ultimately attended DSMES. Variation in referral rates across demographic and clinical characteristics highlights opportunities to improve and standardize referral processes.
Keywords: Diabetes self-management, Diabetes education, Referral, Attendance, Disparity
Prior presentation
Findings from this study were presented at the American Diabetes Association’s 82nd Scientific Sessions in New Orleans, LA on June 3–7, 2022, and at the University of Kentucky College of Public Health’s Public Health Showcase 2024 in Lexington, KY on March 25, 2024.
Introduction
In 2021, an estimated 38.4 million people (11.6% of the population) in the US had diabetes [1]. Diabetes remains the most expensive chronic condition in the nation and the economic costs of diabetes continue to grow. In 2022, the total estimated cost of diagnosed diabetes in the US was $412.9 billion, accounting for 1 in 4 health care dollars spent in the US [2]. Given the substantial burden that diabetes creates for the healthcare system, payers, patients and their caregivers, increasing uptake of strategies that improve diabetes management and reduce diabetes complications is crucial.
Diabetes self-management education and support (DSMES) is an example of one such strategy. DSMES is an evidence-based service delivered by a credentialed diabetes educator that helps people with diabetes and their families how to effectively manage their disease [3]. DSMES typically includes education on healthy eating, physical activity, medication use, monitoring, problem-solving, health coping and reducing risks. It is an integral part of diabetes management and is associated with a range of improvements in diabetes-related outcomes, including improved glycemic outcomes, a reduction in mortality, and improved quality of life, as well as a reduction in diabetes-related healthcare costs [3]. However, despite strong evidence supporting DSMES and clear guidelines from the American Diabetes Association (ADA) recommending its use -including referrals at four critical time points (at diagnosis, annually, when complicating factors arise, and during care transitions- DSMES is underutilized [3, 4]. Estimates suggest that utilization of DSMES among individuals with diabetes is < 10% in the first year following diagnosis [5, 6]. Studies that expand their scope beyond the first year following diagnosis also report low rates of DSMES attendance, with variability observed across certain demographic groups including race/ethnicity and age [7–11].
Studies examining barriers to DSMES have identified a variety of reasons for underutilization including barriers at the patient-level, provider-level and health system-level [3, 12]. Limited or insufficient access, ineffective referral processes, low reimbursement rates, confusion about when and how to make referrals, cost, timing and transportation challenges, low perceived seriousness of diabetes, and inadequate attention to cultural needs contribute to barriers [7, 11–15]. One of the fundamental barriers identified in prior studies is lack of a DSMES referral from a provider [16].
Kentucky is part of the ‘Diabetes Belt,’ a region of the southeastern United States with disproportionately high prevalence of diabetes and related complications. University of Kentucky HealthCare (UKHC), the largest academic medical center in the state, serves both urban and rural populations. In this study, we explore patterns of DSMES referral and attendance across key demographic and clinical characteristics of patients with diabetes receiving care at an academic medical center from 2016 to 2019. We also examined patterns of DSMES referral and attendance stratified by diabetes type, among individuals with type 1 diabetes (T1D) and type 2 diabetes (T2D).
Methods
Study design
We used electronic health record (EHR) data to construct a sample of adult patients (aged ≥ 18 years) diagnosed with diabetes (International Classification of Diseases 9th revision (ICD-9) codes 249.x, 250.x and ICD-10 codes E08-E13). Among adults with diabetes, we included all patients with at least 1 endocrinology or primary care visit at University of Kentucky Health Care (UKHC) during the study period between (1/1/2016 and 12/31/2019). UKHC is a regional academic medical center with a comprehensive diabetes center that serves Central and Eastern Kentucky. Each patient’s index date was defined as the date of their first diabetes diagnosis that occurred during the study period. To ensure sufficient follow-up, we extracted both the first and last encounter dates for each patient and restricted our sample to patients with at least 365 days of follow-up and at least two diabetes related diagnosis codes during the study period. We further excluded patients who had no recorded data for BMI or A1C during the study period. After implementing these criteria, our sample consisted of 10,587 unique patients.
DSMES referral and attendance
DSMES referrals were extracted from EHR data based on an internal workflow implemented by UKHC. If a patient had more than one referral, we only captured the first referral for a patient during the study period. DSMES attendance was documented by our internal DSMES program and only captured attendance at the UKHC DSMES program which offered both individual and group classes delivered in-person and via telehealth. The DSMES series at UKHC is structured as a comprehensive one-day course rather than separate sessions.
Covariate identification
Demographic variables included age, sex, race/ethnicity, insurance type, and rural or urban residence. Insurance types were categorized as commercial, public (Medicaid or Medicare coverage), or other based on the insurance type documented at their index visit. The classification of residence as rural or urban was determined by evaluating the five-digit zip code associated with the patients’ home address provided at their index visit. We used the 2010 Rural-Urban Commuting Area Codes (RUCA), a system developed at the University of Washington, to categorize patient zip codes as either urban or rural [17]. Clinical variables were extracted from the index visit as well and included obesity status, diabetes medication use, diabetes type and A1C values (< 6, 6–6.9, 7–7.9, 8–8.9, ≥ 9), with A1C defined as the first available measurement during the study period, all sourced from the EHR.
Statistical analysis
First, we calculated summary statistics for the entire sample, presenting frequencies and percentages for each variable. Next, we conducted comparisons of referral and attendance rates for DSMES across demographic and clinical characteristics, overall as well as stratified by diabetes type. Chi-squared tests were used to analyze differences in categorical variables.
Finally, to evaluate the association between demographic variables and DSMES referral and attendance during the study period, we fit logistic regression models adjusting simultaneously for all demographic and clinical variables. We also fit fully adjusted logistic regression models stratified by diabetes type. All analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). A two-sided p-value < 0.05 was considered statistically significant.
Results
Our sample included 10,587 adult patients (Table 1). 77.8% of them were Non-Hispanic White (NHW) individuals (n = 8,229), 79.9% were aged 45 or older (n = 8,467), 60.1% had public insurance (n = 6,367), 53.8% were living with obesity (n = 6,756), 55.9% were using insulin (n = 5,915). A1C levels were above 9% in 22.8% (n = 2,409) of the sample.
Table 1.
Demographic and clinical characteristics of patients with diabetes at UK Health Care
| Overall | DSMES Referrala | DSMES Attendanceb | |||
|---|---|---|---|---|---|
| Referred | Did not referred | Attended | Did not attend | ||
| n = 10,587 | n = 726 | n = 9,861 | n = 292 | n = 434 | |
| Demographic variables | |||||
| Sex | |||||
| Female | 5,326 (50.3%) | 450 (8.5%) | 4,876 (91.5%) | 177 (39.3%) | 273 (60.7%) |
| Male | 5,261 (49.7%) | 276 (5.3%) | 4,985 (94.8%) | 115 (41.7%) | 161 (58.3%) |
| Race/ethnicity | |||||
| non-Hispanic White | 8,229 (77.8%) | 490 (6.0%) | 7,739 (94.1%) | 190 (38.8%) | 300 (61.2%) |
| non-Hispanic Black | 1,884 (17.8%) | 187 (9.9%) | 1,697 (90.1%) | 77 (41.2%) | 110 (58.8%) |
| Hispanic | 328 (3.1%) | 36 (11.0%) | 292 (89.0%) | 18 (50.0%) | 18 (50.0%) |
| Other | 136 (1.3%) | 12 (8.8%) | 124 (91.2%) | 6 (50.0%) | 6 (50.0%) |
| Age Group | |||||
| 18–24 | 377 (3.5%) | 53 (14.1%) | 324 (85.9%) | 18 (34.0%) | 35 (66.0%) |
| 25–34 | 538 (5.1%) | 68 (12.6%) | 470 (87.4%) | 28 (41.2%) | 40 (58.8%) |
| 35–44 | 1,205 (11.4%) | 118 (9.8%) | 1,087 (90.2%) | 43 (36.4%) | 75 (63.6%) |
| 45–54 | 2,438 (23.0%) | 198 (8.1%) | 2,240 (91.9%) | 74 (37.4%) | 124 (62.6%) |
| 55–64 | 3,147 (29.7%) | 195 (6.2%) | 2,952 (93.8%) | 86 (44.1%) | 109 (55.9%) |
| 65+ | 2,882 (27.2%) | 94 (3.3%) | 2,788 (96.7%) | 43 (45.7%) | 51 (54.3%) |
| Insurance Type | |||||
| Commercial | 3,972 (37.5%) | 335 (8.4%) | 3,637 (91.6%) | 143 (42.7%) | 192 (57.3%) |
| Public | 6,367 (60.1%) | 370 (5.8%) | 5,997 (94.2%) | 139 (37.6%) | 231 (62.4%) |
| Other | 248 (2.3%) | 21 (8.5%) | 227 (91.5%) | 10 (47.6%) | 11 (52.4%) |
| Patient residence | |||||
| Urban areas | 5,537 (52.3%) | 500 (9.0%) | 5,037 (91.0%) | 215 (43.0%) | 149 (57.0%) |
| Rural areas | 5,050 (47.7%) | 226 (4.5%) | 4824 (95.5%) | 77 (34.1%) | 285 (65.9%) |
| Clinical variables | |||||
| Obesity | |||||
| With obesity | 6,756 (63.8%) | 518 (7.7%) | 6,238 (92.3%) | 211 (40.7%) | 307 (59.3%) |
| Without obesity | 3,831 (36.2%) | 208 (5.4%) | 3,623 (94.6%) | 81 (38.9%) | 127 (61.0%) |
| Medication | |||||
| Any Insulin | 5,915 (55.9%) | 539 (9.1%) | 5,376 (90.1%) | 210 (39.0%) | 329 (61.0%) |
| Oral Medication only | 2,157 (20.4%) | 131 (6.1%) | 2,026 (93.9%) | 55 (42.0%) | 76 (58.0%) |
| No Medication | 2,515 (23.8%) | 56 (2.2%) | 2,459 (97.8%) | 27 (48.2%) | 29 (51.8%) |
| Diabetes Type | |||||
| T1D | 817 (7.7%) | 95 (11.6%) | 722 (88.4%) | 39 (41.1%) | 56 (59.0%) |
| T2D | 9,770 (92.3%) | 631 (6.5%) | 9,139 (93.5%) | 253 (40.1%) | 378 (59.9%) |
| A1C categoryc | |||||
| < 6% | 1,392 (13.2%) | 40 (2.9%) | 1,352 (97.1%) | 16 (40.0%) | 24 (60.0%) |
| 6–6.9% | 3,103 (29.3%) | 160 (5.2%) | 2,943 (94.8%) | 72 (45.0%) | 88 (55.0%) |
| 7–7.9% | 2,258 (21.3%) | 159 (7.0%) | 2,099 (93.0%) | 72 (45.3%) | 87 (54.7%) |
| 8–8.9% | 1,425 (13.5%) | 102 (7.2%) | 1,323 (92.8%) | 39 (38.2%) | 63 (61.8%) |
| ≥ 9% | 2,409 (22.8%) | 265 (11.0%) | 2,144 (89.0%) | 93 (35.1%) | 172 (64.9%) |
aFor all comparisons p < 0.0001
bFor all comparisons p > 0.05; except patient residence p < 0.0001
cA1C levels represent baseline measurements, defined as the first available value during the study period
From the overall sample, 726 (6.9%) of patients with diabetes were referred to DSMES. Females (8.5% vs. 5.3% males; p < 0.0001), Non-Hispanic Black (NHB) and Hispanic patients (9.9% and 11.0% vs. 6.0% in non-Hispanic White; p < 0.0001), and younger patients (14.1% 18-24yo vs. 3.3% for 65+; p < 0.0001) were more likely to be referred. Patients with commercial insurance (8.4%) and those residing in urban areas (9.8%) were also more likely to be referred compared to their counterparts (p < 0.0001 for both). Patients living with obesity, those using insulin, and patients with T1D were more likely to be referred than their counterparts; who were not living with obesity, those on oral medication only or no medication, and those with T2D, respectively (p < 0.001 for all). Patients with higher A1C levels had a greater likelihood of referral compared to those with lower A1C levels (11.0% in the ≥ 9% group vs. 2.9% in the < 6% group; p < 0.0001). Among those referred, 40.2% attended DSMES; urban residency was the only demographic or clinical variable associated with DSMES attendance (43.0% urban vs. 34.1% rural; p < 0.0001).
In the results stratified by diabetes type (Table 2), 7.7% (n = 817) of participants had documented T1D and 92.3% (n = 9,770) had documented T2D. Among those with T2D, 6.5% of patients were referred to DSMES and 40.1% (n = 253) attended DSMES; in T1D, 11.6% (n = 95) were referred and, of those referred, 41.4% (n = 39) attended DSMES. The demographic and clinical characteristics of the T2D sample largely mirrored the overall sample, but T1D patients were younger (76.7% under 45), had higher rates of commercial insurance (53.9%), resided more in rural areas (53.1%), and had lower obesity prevalence (29.5%). Referral and attendance patterns in the T2D group were similar to the overall sample, with differences in referral patterns across all demographic and clinical variables examined but minimal differences in attendance patterns. Conversely, referral patterns within the T1D group only differed by sex, race/ethnicity and patient residence and attendance patterns only differed by age.
Table 2.
Demographic and clinical characteristics of patients with diabetes at UK HealthCare stratified by diabetes type
| Type 1 diabetes (n = 817) | Type 2 diabetes (n = 9,770) | |||||||
|---|---|---|---|---|---|---|---|---|
| DSMES Referrala | DSMES Attendanceb | DSMES Referralc | DSMES Attendanced | |||||
| Referred | Not Referred | Attended | Did not attend | Referred | Not Referred | Attended | Did not attend | |
| n = 95 | n = 722 | n = 39 | n = 56 | n = 631 | n = 9,139 | n = 253 | n = 378 | |
| Demographic variables | ||||||||
| Sex | ||||||||
| Female | 59 (14.2%) | 357 (85.8%) | 23 (39.0%) | 36 (61.0%) | 391 (4.9%) | 4,620 (95.1%) | 154 (39.4%) | 237 (60.6%) |
| Male | 36 (9.0%) | 365 (91.0%) | 16 (44.4%) | 20 (55.6%) | 240 (8.0%) | 4,519 (92.0%) | 99 (41.3%) | 141 (58.8%) |
| Race/ethnicity | ||||||||
| non-Hispanic White | 78 (10.4%) | 669 (89.6%) | 31 (39.7%) | 47 (60.3%) | 412 (5.5%) | 7,070 (94.5%) | 159 (38.6%) | 253 (61.4%) |
| non-Hispanic Black | 15 (26.8%) | 41 (73.2%) | 7 (46.7%) | 8 (53.3%) | 172 (9.4%) | 1,656 (90.6%) | 70 (40.7%) | 102 (59.3%) |
| Hispanic | 1 (11.1%) | 8 (88.9%) | 0 (0.0%) | 1 (100.0%) | 35 (11.0%) | 284 (89.0%) | 18 (51.4%) | 17 (48.6%) |
| Other | 0 (0.0%) | 4 (100.0%) | - | - | 12 (9.1%) | 120 (90.9%) | 6 (50.0%) | 6 (50.0%) |
| Age Group | ||||||||
| 18–24 | 43 (15.1%) | 242 (84.9%) | 14 (32.6%) | 29 (67.4%) | 10 (10.9%) | 82 (89.1%) | 4 (40.0%) | 6 (60.0%) |
| 25–34 | 21 (9.4%) | 203 (90.6%) | 8 (38.1%) | 13 (61.9%) | 47 (15.0%) | 267 (85.0%) | 20 (42.6%) | 27 (57.5%) |
| 35–44 | 9 (7.6%) | 109 (92.4%) | 5 (55.6%) | 4 (44.4%) | 9,109 (10.0%) | 978 (90.0%) | 38 (34.9%) | 71 (65.1%) |
| 45–54 | 13 (12.9%) | 88 (87.1%) | 4 (30.8%) | 9 (69.2%) | 185 (7.9%) | 2,152 (92.1%) | 70 (37.8%) | 115 (62.2%) |
| 55–64 | 6 (10.0%) | 54 (90.0%) | 5 (83.3%) | 1 (16.7%) | 189 (6.1%) | 2,898 (93.9%) | 81 (42.9%) | 108 (57.1%) |
| 65+ | 3 (10.3%) | 26 (89.7%) | 3 (100.0%) | 0 (0.0%) | 91 (3.2%) | 2,762 (96.8%) | 40 (44.0%) | 51 (56.0%) |
| Insurance Type | ||||||||
| Commercial | 53 (12.1%) | 387 (88.0%) | 24 (45.3%) | 29 (54.7%) | 292 (8.0%) | 3,250 (92.0%) | 119 (42.2%) | 163 (57.8%) |
| Public | 42 (11.4%) | 326 (88.6%) | 15 (35.7%) | 27 (64.3%) | 328 (5.5%) | 5,671 (94.5%) | 124 (37.8%) | 204 (62.2%) |
| Other | 0 (0.0%) | 9 (100.0%) | - | - | 21 (8.8%) | 218 (91.2%) | 10 (47.6%) | 11 (52.4%) |
| Patient residence | ||||||||
| Urban areas | 54 (14.1%) | 329 (85.9%) | 24 (44.4%) | 30 (55.6%) | 446 (8.7%) | 4,708 (91.4%) | 191 (42.8%) | 255 (57.2%) |
| Rural areas | 41 (9.5%) | 393 (90.6%) | 15 (36.6%) | 26 (63.4%) | 185 (4.0%) | 4,431 (96.0%) | 62 (33.5%) | 123 (66.5%) |
| Clinical variables | ||||||||
| Obesity | ||||||||
| With obesity | 35 (14.5%) | 206 (85.6%) | 11 (31.4%) | 24 (68.6%) | 483 (7.4%) | 6,032 (92.6%) | 200 (41.4%) | 283 (58.6%) |
| Without obesity | 60 (10.4%) | 516 (89.6%) | 28 (46.7%) | 32 (53.3%) | 148 (4.6%) | 3,107 (95.5%) | 53 (35.8%) | 95 (64.2%) |
| Medication | ||||||||
| Any Insulin | - | - | - | - | 444 (8.7%) | 4,679 (91.3%) | 171 (38.5%) | 273 (61.5%) |
| Oral Medication only | - | - | - | - | 131 (6.1%) | 2,025 (93.9%) | 55 (42.0%) | 76 (58.0%) |
| No Medication | - | - | - | - | 56 (2.3%) | 2,435 (97.8%) | 27 (48.2%) | 29 (51.8%) |
| A1C categorye | ||||||||
| < 6% | 1 (3.5%) | 28 (96.5%) | 1 (100.0%) | 0 (0.0%) | 39 (2.9%) | 1,324 (97.1%) | 15 (38.5%) | 24 (61.5%) |
| 6–6.9% | 7 (6.8%) | 96 (93.2%) | 4 (57.1%) | 3 (42.9%) | 153 (5.1%) | 2,847 (94.9%) | 68 (44.4%) | 85 (55.6%) |
| 7–7.9% | 22 (11.5%) | 169 (88.5%) | 8 (36.4%) | 14 (63.6%) | 137 (6.6%) | 1,930 (93.4%) | 64 (46.7%) | 73 (53.3%) |
| 8–8.9% | 24 (13.5%) | 154 (86.5%) | 11 (45.8%) | 13 (54.2%) | 78 (6.3%) | 1,169 (93.7%) | 28 (35.9%) | 50 (64.1%) |
| ≥ 9% | 41 (13.0%) | 275 (87.0%) | 15 (36.6%) | 26 (63.4%) | 224 (10.7%) | 1,869 (89.3%) | 78 (34.8%) | 146 (65.2%) |
aFor all comparisons p > 0.05; except sex, race/ethnicity and patient residence p < 0.05
bFor all comparisons p > 0.05; except age group p < 0.05
cFor all comparisons p < 0.0001
dFor all comparisons p > 0.05; except patient residence p < 0.0001
eA1C levels represent baseline measurements, defined as the first available value during the study period
In the logistic regression model presented in Fig. 1a, females (OR 1.67, 95% CI 1.42–1.96), NHBs (OR 1.36, 95% CI 1.12–1.65) and Hispanics (OR 1.56, 95% CI 1.03–2.35) were significantly more likely to be referred to DSMES than males and NHWs, respectively. Compared to the younger age groups, those aged 65 + were significantly less likely to be referred (OR 0.31, 95% CI 0.20–0.48). Individuals with public insurance (OR 0.75, 95% CI 0.63–0.88) and those residing in rural areas (OR 0.48, 95% CI 0.40–0.57) were less likely to be referred compared to their counterparts with commercial insurance and those residing in urban areas. No significant differences in odds of attendance were observed for any variables (Fig. 1b).
Fig. 1.
Fully adjusted multivariable logistic regression models examining the association between demographic and clinical variables, and the referral and attendance to DSMES 1a: Model 1 for DSMES Referral 1b: Model 2 for DSMES Attendance
Discussion
In this study of 10,587 adult patients with diabetes from a regional academic medical center, we found low rates of referral to DSMES (6.9%) and considerable variation in referrals across key demographic (sex, race/ethnicity, age, insurance type and patient residence) and clinical variables (obesity, diabetes medication use, diabetes type, and A1C category). Referral rates were higher in females compared with males, NHBs and Hispanics compared with NHWs, younger patients compared with older, patients with commercial versus public insurance and those residing in urban versus rural areas. Referral rates were also higher among patients living with obesity, those using insulin or any diabetes medication and those with higher A1C levels (A1C ≥ 9%) than their counterparts who were not living with obesity, used no medications and had lower A1C levels (A1C < 6%), respectively. In contrast to low referral rates, although attendance rates were higher than referral rates, fewer than half of referred patients attended DSMES (40.2%), with minimal differences in attendance across the same demographic and clinical variables. These findings suggest opportunities to improve DSMES referral rates for all eligible patients and highlight the importance of equitable referring practices for all patients who are eligible for DSMES.
To our knowledge, our study is among the first to comprehensively examine both referral and attendance rates of DSMES together across various demographic and clinical characteristics in a diverse population of patients with diabetes (overall as well as T1D and T2D) from an academic medical center located within the Diabetes Belt. This setting, which includes both urban and rural populations in a state with high diabetes burden, allowed us to examine subgroup differences in referral and attendance patterns that may not be visible in prior studies which have largely focused on patients with T2D, or relied on self-reported diabetes data from sources such as the Behavioral Risk Factor Surveillance System (BRFSS). Our findings were largely consistent with prior studies but additionally provide data on DSMES referral and attendance patterns in individuals with T1D. Findings from our overall and T2D patients were similar to estimates from Azam et al.‘s study, which found a 7.4% referral rate in a large and diverse sample including patients with T2D [11]. In contrast, Alsayed Hassan et al.‘s study, which examined referral and attendance rates among patients with T2D, reported higher rates at 53.5% and 55.4%, respectively [9]. The differences among the studies could be attributed to the differences between the healthcare systems and patient populations or methodologic differences in how DSMES eligible patients were ascertained for the denominator that contributed to the measured difference in referral rates. Earlier studies demonstrated a wide range of attendance rates from 6.5% to 53.7% [7, 9–12, 18, 19]. Although attendance rates were higher than referral, the DSMES completion remains unclear. Accredited programs require 10-hours of education, yet many studies only considered attendance without addressing persistence, and completion rates of the program might be lower [20]. To meet the Healthy People 2030 target of increasing the proportion of people with diabetes who receive formal diabetes education to 55.2%, efforts should be directed towards ensuring DSMES completion throughout the program [21].
At UKHC, DSMES is offered as a comprehensive one-day course rather than multiple sessions, a decision made by program leadership to address difficulties with patient retention. Although evidence on single-session DSMES formats is limited, one study found that patients who attended at least one DSMES visit had a 34% lower hospitalization rate compared to those with none [22]. Additionally, a systematic review reported that DSMES delivered over the short term (< 2.5 months) and the long term (>12 months) were equally effective in reducing A1C [23]. Further, prior studies examining DSMES referral and attendance have underscored the importance of accommodating patients’ circumstances and needs as a key strategy to reducing barriers to DSMES [9, 24]. These findings suggest that the important factor is engaging with DSMES at all, as even limited participation can lead to positive outcomes.
There may be various reasons for low referral and attendance rates. Primary among them are physician-related barriers, including a lack of knowledge about the referral process, providers’ apprehensions regarding patient refusal to attend, and a preference for patients to self-manage their disease rather than utilizing DSMES [9]. Physicians may tend to refer patients with worse clinical status, as indicated in the literature and our study, where higher referral rates were observed among patients with poorer clinical conditions (e.g., obesity, medication or insulin usage, higher A1C levels) [10, 11, 19]. On the other hand, duration of illness also can influence referral and attendance; for instance, in the initial year following diagnosis, approximately 10% of individuals engage in DSMES programs [25]. Moreover, individuals with a shorter duration of diabetes are less likely to participate in DSMES than those with a longer duration [19]. Access to DSMES in the early stages of the disease may prevent the worsening of their conditions.
Studies have demonstrated low DSMES participation rates among racial/ethnic minority groups, older adults and low-income patients [11, 18, 19, 26]. Patients in these groups may face challenges to participation in DSMES, including lower health literacy, difficulty navigating healthcare systems, financial limitations, transportation issues, and a lack of culturally or linguistically appropriate DSMES programs [3, 18]. In our study, public insurance was associated with lower referral rates compared to commercial insurance.
Race/ethnicity was also associated with DSMES referrals. In our study NHBs and Hispanics were more likely to be referred than NHWs; prior studies have reported higher referral rates in NHBs compared to NHWs but lower referral rates in Hispanics [8, 18, 27]. Prior studies have suggested that NHBs may be targeted for DSMES interventions due to higher prevalence of diabetes in this group, but further research is needed to understand the factors contributing to the higher rates of referrals among different race/ethnicity groups, especially for Hispanics [18].
Our finding that referral rates for DSMES declined with patient age, particularly for those above 65, aligns with previous studies [11, 18, 19, 26]. Older adults might be less familiar with or perceive DSMES programs as less relevant to their needs. Additionally, healthcare providers may hesitate to recommend DSMES to them due to concerns about whether they will adopt changes in their disease management behaviors [14].
Consistent with previous studies, rural residency emerged as a significant barrier in our study [18, 19, 28]. Interestingly, one study suggests it has a greater impact on access to DSMES programs than poverty level [29]. In our study, rurality was the only factor found to reduce both referral and attendance rates. To address this disparity, DSMES programs should expand their reach to include patients residing in rural areas, who encounter obstacles in accessing these programs for various reasons. Delivering DSMES through telehealth services can serve as an alternative approach to improving attendance.
Our study has many strengths. We utilized patient data from the electronic health record system of a large institution that serves individuals from various demographic and insurance groups including patients from both urban and rural areas in Kentucky. We also were able to differentiate between patients with T1D vs. T2D and present rates of DSMES referral and attendance in these two patient populations.
Our study has some limitations. First, the lack of clinician-level data prevents us from assessing the extent of inter-clinician variability in referrals for DSMES. Second, we were unable to ascertain whether referrals were verbally offered by clinicians during healthcare visits but declined by patients, leading to no referral documentation in the EHR. Therefore, it is possible that we are underestimating the proportion of patients offered DSMES. Third, our data did not allow us to align referrals with the four critical time points recommended by the ADA. Fourth, because our dataset did not capture the timing of diabetes diagnosis, we could not distinguish between newly diagnosed and long-standing patients. This may have influenced our overall referral rate estimates, as newly diagnosed patients are more likely to be referred to DSMES according to current guidelines. Fifth, the absence of information on age at diagnosis also prevented us from evaluating the impact of disease duration on referral and attendance. The sixth limitation was the low number of patients with T1D in our sample. Seventh, as this was an observational study, we cannot establish causality between patient characteristics and DSMES referrals or attendance. Eighth, DSMES attendance was only captured at the UKHC program; patients who may have attended DSMES elsewhere were not included, which could underestimate true attendance. Finally, the study data from a single center and single DSMES program limits the generalizability of our results.
Conclusions
Our study highlights low overall DSMES referral rates (6.9%) and variation in referral rates for DSMES across a number of demographic and clinical factors for patients with diabetes who received care at an academic medical center. Among patients referred to DSMES, however, there was little variation in terms of who attended DSMES. These findings highlight the importance of increasing patient referrals to DSMES overall and standardizing DSMES referral processes to ensure all eligible patients are referred to this service regardless of demographic or clinical characteristics. Efforts to improve DSMES referral rates may include integrating routine DSMES referrals into the diabetes care plan, particularly at diagnosis and during key management periods. Additionally, addressing disparities by utilizing telehealth for rural and older patients can enhance accessibility. Standardizing referral processes can further increase referral rates, leading to better diabetes management and outcomes. Future research should also assess outcomes such as A1C, complications, and healthcare utilization by referral and attendance status to better understand the real-world impact of DSMES on population.
Acknowledgements
Not applicable.
Abbreviations
- BRFSS
Behavioral Risk Factor Surveillance System
- DSMES
Diabetes self-management education and support
- NHB
Non-Hispanic black
- NHW
Non-Hispanic white
- RUCA
Rural-Urban Commuting Area Codes
- T1D
Type 1 diabetes
- T2D
Type 2 diabetes
- UKHC
University of Kentucky Health Care
Authors' contributions
MEL and JWK conceptualized the study. OA, GCB, WBB, KRH and KCD contributed to the study design. KRH performed statistical analyses and MEL advised on analyses. OA and MEL drafted the manuscript. GCB, WBB, KRH, JWK, and KCD revised the manuscript. All authors approved the version to be published.
Funding
This work was supported by the National Institutes of Health’s National Center for Advancing Translational Sciences [Grant numbers UL1TR001998 and KL2TR001996] and the University of Kentucky’s Igniting Research Collaboration Pilot Program and Priority Area in Obesity and Diabetes. The funder was not involved in the study design, data collection, data analysis, manuscript preparation, and/or publication decisions.
Data availability
The datasets generated and/or analyzed during the current study are not publicly available due to the data sharing restrictions of the home institution but are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
The Institutional Review Board at the University of Kentucky has approved a de-identified and limited dataset for UK Healthcare information for research use under protocol number 45668 and waived the need for consent to participate.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Centers for Disease Control and Prevention. National Diabetes Statistics Report. 2024. https://www.cdc.gov/diabetes/php/data-research/index.html. Accessed 23 Aug 2025.
- 2.Parker ED, Lin J, Mahoney T, Ume N, Yang G, Gabbay RA, et al. Economic costs of diabetes in the U.S. in 2022. Diabetes Care. 2024;47(1):26–43. 10.2337/dci23-0085. [DOI] [PubMed] [Google Scholar]
- 3.Powers MA, Bardsley JK, Cypress M, Funnell MM, Harms D, Hess-Fischl A, et al. Diabetes Self-management education and support in adults with type 2 diabetes: A consensus report of the American diabetes association, the association of diabetes care & education Specialists, the academy of nutrition and Dietetics, the American academy of family Physicians, the American academy of PAs, the American association of nurse Practitioners, and the American pharmacists association. Diabetes Care. 2020;43(7):1636–49. 10.2337/dci20-0023. [DOI] [PubMed] [Google Scholar]
- 4.Davis J, Fischl AH, Beck J, Browning L, Carter A, Condon JE, et al. 2022 national standards for diabetes self-management education and support. Diabetes Care. 2022(2). 10.2337/dc21-2396. [DOI] [PubMed]
- 5.Li R, Shrestha SS, Lipman R, Burrows NR, Kolb LE, Rutledge S. Diabetes self-management education and training among privately insured persons with newly diagnosed diabetes–United States, 2011–2012. MMWR. 2014;63(46):1045–9. [PMC free article] [PubMed] [Google Scholar]
- 6.Strawbridge LM, Lloyd JT, Meadow A, Riley GF, Howell BL. Use of Medicare’s diabetes self-management training benefit. Health Educ Behav. 2015;42(4):530–8. 10.1177/1090198114566271. [DOI] [PubMed] [Google Scholar]
- 7.Yoshida Y, Hong D, Nauman E, Price-Haywood EG, Bazzano AN, Stoecker C, et al. Patient-specific factors associated with use of diabetes self-management education and support programs in Louisiana. BMJ Open Diabetes Res Care. 2021. 10.1136/bmjdrc-2021-002136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Akyirem S, Choa E, Poghosyan H. Investigating racial and ethnic differences in diabetes self-management education among adults with diabetes. Sci Diabetes Self Manag Care. 2023;49(3):206–16. 10.1177/26350106231169693. [DOI] [PubMed] [Google Scholar]
- 9.Alsayed Hassan D, Curtis A, Kerver J, Vangsnes E. Diabetes self-management education and support: referral and attendance at a patient-centered medical home. J Prim Care Community Health. 2020;11:2150132720967232. 10.1177/2150132720967232. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bober T, Rothenberger S, Lin J, Ng JM, Zupa M. Factors associated with receipt of diabetes self-management education and support for type 2 diabetes: potential for a population health management approach. J Diabetes Sci Technol. 2023;17(5):1198–205. 10.1177/19322968231176303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Azam LS, Jackson TA, Knudson PE, Meurer JR, Tarima SS. Use of secondary clinical data for research related to diabetes self-management education. Res Social Adm Pharm. 2017;13(3):494–502. 10.1016/j.sapharm.2016.07.002. [DOI] [PubMed] [Google Scholar]
- 12.Peyrot M, Rubin RR, Funnell MM, Siminerio LM. Access to diabetes self-management education: results of National surveys of patients, educators, and physicians. Diabetes Educ. 2009;35(2):246–8. 10.1177/0145721708329546. 52 – 6. [DOI] [PubMed] [Google Scholar]
- 13.Carey ME, Agarwal S, Horne R, Davies M, Slevin M, Coates V. Exploring organizational support for the provision of structured self-management education for people with type 2 diabetes: findings from a qualitative study. Diabet Med. 2019;36(6):761–70. 10.1111/dme.13946. [DOI] [PubMed] [Google Scholar]
- 14.Winkley K, Stahl D, Chamley M, Stopford R, Boughdady M, Thomas S, et al. Low attendance at structured education for people with newly diagnosed type 2 diabetes: general practice characteristics and individual patient factors predict uptake. Patient Educ Couns. 2016;99(1):101–7. 10.1016/j.pec.2015.08.015. [DOI] [PubMed] [Google Scholar]
- 15.Horigan G, Davies M, Findlay-White F, Chaney D, Coates V. Reasons why patients referred to diabetes education programmes choose not to attend: a systematic review. Diabet Med. 2017;34(1):14–26. 10.1111/dme.13120. [DOI] [PubMed] [Google Scholar]
- 16.Arena L, Esquivel NS, Austin R, Millstein S, Kaelin-Kee J. 502-P: increasing provider referrals to diabetes Self-Management education and support services: recommendations from virtual focus groups with providers in new York state. Diabetes. 2022;71:502–P. 10.2337/db22-502-P. [Google Scholar]
- 17.University of Washington. RUCA Data 2010. https://depts.washington.edu/uwruca/ruca-uses.php. Accessed 8 Sept 2023.
- 18.Adjei Boakye E, Varble A, Rojek R, Peavler O, Trainer AK, Osazuwa-Peters N, et al. Sociodemographic factors associated with engagement in diabetes self-management education among people with diabetes in the united States. Public Health Rep. 2018;133(6):685–91. 10.1177/0033354918794935. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Luo H, Bell RA, Winterbauer NL, Xu L, Zeng X, Wu Q, et al. Trends and Rural-Urban differences in participation in diabetes Self-management education among adults in North carolina: 2012–2017. J Public Health Manag Pract. 2022;28(1):E178–84. 10.1097/PHH.0000000000001226. [DOI] [PubMed] [Google Scholar]
- 20.Centers for Medicare & Medicaid Services. Diabetes self-management training. https://www.medicare.gov/coverage/diabetes-self-management-training. Accessed 1 Mar 2024.
- 21.U.S. Department of Health and Human Services. Healthy People 2030: Increase the proportion of people with diabetes who get formal diabetes education — D–06 2020. https://health.gov/healthypeople/objectives-and-data/browse-objectives/diabetes/increase-proportion-people-diabetes-who-get-formal-diabetes-education-d-06. Accessed 10 Mar 2024.
- 22.Robbins JM, Thatcher GE, Webb DA, Valdmanis VG. Nutritionist visits, diabetes classes, and hospitalization rates and charges: the urban diabetes study. Diabetes Care. 2008;31(4):655–60. 10.2337/dc07-1871. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Chrvala CA, Sherr D, Lipman RD. Diabetes self-management education for adults with type 2 diabetes mellitus: a systematic review of the effect on glycemic control. Patient Educ Couns. 2016;99(6):926–43. 10.1016/j.pec.2015.11.003. [DOI] [PubMed] [Google Scholar]
- 24.Ufongene G. Implementing a Diabetes Self-Management Education Protocol in a Primary Care Clinic to Address Referral Rates [dissertation]. West Hartford (CT): University of Saint Joseph; 2018.
- 25.Baker KM, Nassar CM, Baral N, Magee MF. The current diabetes education experience: findings of a cross-sectional survey of adults with type 2 diabetes. Patient Educ Couns. 2023;108:107615. 10.1016/j.pec.2022.107615. [DOI] [PubMed] [Google Scholar]
- 26.White Perkins D, Milan P, Miazek K, Havstad S, Wegienka G. Identifying factors affecting diabetes education program participation within a metro Detroit integrated health system. Prev Med Rep. 2021;24:101646. 10.1016/j.pmedr.2021.101646. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Khan MM, Roberson S, Reid K, Jordan M, Odoi A. Geographic disparities and temporal changes of diabetes prevalence and diabetes self-management education program participation in Florida. PLoS ONE. 2021;16(7):e0254579. 10.1371/journal.pone.0254579. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Brown-Guion SY, Youngerman SM, Hernandez-Tejada MA, Dismuke CE, Egede LE. Racial/ethnic, regional, and rural/urban differences in receipt of diabetes education. Diabetes Educ. 2013;39(3):327–34. 10.1177/0145721713480002. [DOI] [PubMed] [Google Scholar]
- 29.Graves BA, Liu G, Kunberger J, Canaday L, Bambis B. Access to diabetes self-management education in a rural state: a GIS analysis. Online J Rural Nurs Health Care. 2019;19(2):98–126. 10.14574/ojrnhc.v19i2.575. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets generated and/or analyzed during the current study are not publicly available due to the data sharing restrictions of the home institution but are available from the corresponding author on reasonable request.