Factors Influencing Patient Completion of Diabetes Self-Management Education Free

The IDEA study was a behavioral intervention to evaluate diabetes outcomes in patients randomized to group and individual DSME interventions.⁷  Potential subjects were recruited from adult patients diagnosed with diabetes from two health care systems in Minneapolis, Minn., and Albuquerque, N.M. Eligibility criteria included age 18–85 years, type 2 diabetes, suboptimal glycemic control (A1C ≥ 7%), and no recent participation in educational sessions.

Letters were mailed to 9,971 people who met initial eligibility criteria based on electronic health records. The study scheduled 760 enrollment visits and enrolled 623 subjects (82%). Informed consent was obtained using procedures approved by the participating medical groups' institutional review boards.

Subjects were randomly assigned to one of three study arms: group education (n = 243), individual education (n = 245), or usual care (no assigned education, n = 134). The usual care group was not included in this analysis because it had no active intervention.

All subjects received $50 gift cards for attending the baseline and consenting visit, and each of four mailed surveys during a year of follow-up included a $25 gift card. Subjects were not provided incentives for attending educational sessions.

All consenting subjects indicated a willingness to attend either the educational or group sessions and were scheduled into their assigned DSME sessions at a clinic choice within the care system using the routine clinic scheduling system. Subjects received phone calls to remind them of upcoming sessions.

The individual education arm consisted of three 1-hour DSME sessions—the same as available to patients in the care system as routine clinical practice. Content was based on the American Association of Diabetes Educators Seven Self-Care Behaviors program and delivered by certified diabetes educators (CDEs; dietitians or nurses) within the care system. Group education consisted of four 2-hour weekly sessions led by the same CDEs. Educators were trained by experts to use U.S. Diabetes Conversation Maps, an American Diabetes Association–sponsored program that uses laminated 3- by 5-foot drawings of situations familiar to people with diabetes to facilitate group interaction. Group sessions were held at a limited number of clinics, and there was less flexibility in scheduling these sessions than for scheduling individual visits.

Subjects completed baseline surveys on demographic, behavioral, and psychosocial measures. Psychosocial tests evaluated in the current analysis include: 1) the 12-Item Short-Form (SF-12) Health Survey,⁸  which measures perceived health function and well-being, and 2) the Patient Health Questionnaire (PHQ-9), which is used to diagnose depression and evaluate its severity.⁹  The PHQ-9 score was categorized as no depression (0–4), mild depression (5–9), or moderate to severe depression (≥ 10). Subjects were mailed evaluations asking them to rate their satisfaction with the content of DMSE, the facility, and, for group DSME subjects, the group experience.

Subjects who consented to the study but did not attend any sessions (for either intervention) were classified as Non-Starters. Because Non-Starters were not exposed to either group or individual DSME, they were pooled into a single group, and their characteristics were compared to those of all Completers (subjects who completed all assigned group or individual DSME sessions). Subjects who attended at least one but not all group DSME sessions were classified as Group DSME Partial Completers, and those who attended all sessions were classified as Group DSME Completers. Individual DSME Partial Completers and Completers were defined similarly.

Descriptive statistics included proportions, means, and standard deviations of independent variables, calculated according to study outcome. A two-sided α = 0.05 was considered statistically significant. For categorical variables, covariate-adjusted log odds were estimated by multiple logistic regression and presented as adjusted predicted probabilities. For continuous variables, mean differences between groups were estimated by multiple linear regression. All models adjusted for age category, sex, age-sex interaction, and study site. Non-Starter models, which included subjects randomized to both treatment groups, additionally adjusted for treatment group.

The Non-Starter group included 40 subjects who did not attend any of their assigned educational sessions (30 from group DSME and 10 from individual DSME; Table 1). The Group DSME Partial Completers included 38 subjects who attended at least one but not all four scheduled sessions. Individual DSME Partial Completers included 24 subjects who attended at least one but not all three scheduled sessions. Of subjects who attended at least one session, a greater proportion completed individual DMSE than group DMSE (89.8 vs. 82.2%; P = 0.02), but similar proportions attended one or two sessions (10.2 vs. 9.4%).

Non-starting of DSME was associated with age in men but not women, education, smoking status, and two of 10 SF-12 scales (General Health and Social Functioning) in adjusted analyses. The relationship between race/ethnicity and the Vitality scale were borderline significant. The probability of being a Non-Starter was higher in men < 60 years of age than for men who were older (17.1 vs. 4.7%; P = 0.005). Subjects with less education or current smoking were more likely to be Non-Starters. By education level, the probability of being a Non-Starter was 13.6, 5.0, and 6.2% for high school education or less, some college, and ≥ 4 years of college, respectively (P = 0.04). By smoking status, 14.3% of current smokers and 5.7% of nonsmokers were predicted to be Non-Starters (P = 0.006). Non-Starters scored lower than Completers on the SF-12 General Health (−9.1%, P = 0.005) and Social Functioning (−9.4%, P = 0.03) scales. Scores on the Vitality scale were nonsignificantly lower.

Being a Group DSME Partial Completer was strongly associated with age in men but not women and with smoking status in adjusted analyses (Table 2). Physical activity was associated with Partial Completer status at a borderline significance level. Group DSME Partial Completers scored lower than Group DSME Completers on 5 of the 10 SF-12 measures (Table 3). The probability of being a Partial Completer was substantially higher both in men < 60 years of age (44.8%) than those who were older (13.0%, P = 0.0009) and in current smokers (39.6%) than in nonsmokers (13.5%, P = 0.002). Group DMSE Partial Completers scored lower than Group DMSE Completers on General Health (−9.0%, P < 0.05) and Mental Health scales (−7.5%, P = 0.02). The Vitality score was nonsignificantly lower.

Subject satisfaction with the DSME content, group experience, and facility did not differ between Group Partial Completers and Group Completers in covariate-adjusted analyses (data not shown). Exploratory, unadjusted analyses suggested a single borderline association between Group DSME Partial Completer status and satisfaction ratings: Group DMSE Partial Completers tended to agree more strongly than did Group Completers with the statement, “I was disturbed by discussing my medical situation in front of others” (higher score indicating greater agreement; mean score 2.4 vs. 2.0, P = 0.05). Men and women responded similarly to the question (2.0 vs. 2.1, P = 0.48), whereas younger subjects agreed more strongly (2.3 vs. 2.0, P = 0.02). However, relative differences in the ratings to this question between Partial Completers and Completers appeared to be greater in older men (18.9%), older women (18.5%), and younger women (12.0%), than in younger men (7.7%), suggesting that younger men's discomfort with discussing medical situations in front of others was not a particularly strong explanation for high rates of Partial Completers in this group.

Individual DSME Partial Completer status was associated with age in men but not women, education, depression, and 7 of 10 SF-12 scores in adjusted analyses. The probability of being an Individual DMSE Partial Completer was higher in men < 60 years of age (11.1 vs. 2.1% in men ≥ 60 years of age) and higher in subjects with less education (12.6, 10.5, and 2.0% for subjects having high school or less, some college, and for ≥ 4 years of college, respectively; P = 0.01). The probability of being a Partial Completer was 3.4% in subjects with no depression, 8.9% for those with mild depression, and 18.8% for those with moderate to severe depression (P = 0.004). Individual DSME Partial Completers had lower, less advantageous mean SF-12 scores than Completers: Physical Composite −16.4% (P = 0.004), Physical Functioning −13.6% (P = 0.01), Role–Physical −25.0% (P = 0.003), Bodily Pain −13.1% (P = 0.02), General Health −14.8% (P = 0.001), Vitality −14.0% (P = 0.002), and Social Functioning −14.6% (P = 0.02). Individual DSME Partial Completers' mean Mental Composite and Mental Health scores were lower than those of Individual DSME Completers but at a border-line level of significance. Individual DSME Partial Completers and Individual DSME Completers did not differ significantly on any of the satisfaction measures.

The study evaluated personal characteristics influencing attendance at group and individual DSME in research subjects randomly assigned to their intervention group. Because the educational interventions were delivered within the patients' usual care delivery system and were quite similar to those available in routine clinical practice, the results are likely informative to clinical practice experience as well as to clinical trials.

DMSE Non-Starters were predicted to be younger men, have less education, be current smokers, and have poorer general health and lower social functioning. Group DSME Partial Completers were predicted to be younger men, be current smokers, and have poorer general and mental health. The lower probability of completion of education among younger men in all groups, but particularly group DMSE, and the relationships of partial completion of individual DMSE to depression and health status are notable. Individual DSME Partial Completers were predicted to be younger men, have less education, be depressed, and score lower on most SF-12 measures of health scales, including Physical Composite, Physical Functioning, Role–Physical, Bodily Pain, General Health, Vitality, and Social Functioning.

There are several possible explanations for why younger men were more likely to be Non-Starters or Partial Completers of DSME. Employment demands among younger, working-age subjects could make it difficult for them to attend scheduled DSME sessions. However, in our data, it did not appear that full-time employment was particularly associated with attendance. Younger men also may be uncomfortable discussing their diabetes in a group setting or may be skeptical of the benefits of learning in this setting. The study found limited evidence for differences between Partial Completers and Completers of group DSME in comfort with discussing medical conditions in front of others. However, this did not appear to be a particularly important factor among younger men, and we found no differences between Partial Completers and Completers for other satisfaction measures. Further research on the young male demographic may be warranted to identify educational needs, preferences, and issues related to attendance.

Thoolen et al.³  reported that DSME subjects who dropped out of a DSME trial reported doing so for practical reasons such as illness, traveling distance, other responsibilities, or time constraints. Subjects in this study assigned to the group DMSE intervention were more likely to miss sessions than those assigned to individual DMSE. The additional time requirements (8 vs. 3 hours), fewer clinic locations, and reduced scheduling flexibility of the group DSME sessions may have contributed to the lower completion rates in this group. These are practical logistical concerns that should be considered when recommending approaches to diabetes education.

The characteristics most robustly associated with completion of individual DSME were the presence and degree of depression and SF-12 scores indicating poorer physical and mental health. Among subjects without depression who attended at least one individual DMSE session, nearly 97% were predicted to complete all three sessions, compared to 81% of subjects with moderate or greater depression.

Depression is related to reduced adherence to medical treatment in general.¹⁰  Among people with diabetes, depression and depressive symptoms have been associated with poor adherence to various aspects of diabetes self-care,^11–14  including participation in education.

Our finding that subjects with a higher PHQ-9 score were less likely to complete individual treatment is consistent with meta-analysis results showing that diabetes patients with depression are less likely to keep medical appointments.¹⁵  From a clinical perspective, regular depression screening could be integrated more regularly into individual DSME assessment visits as a means of identifying an important risk factor for increased likelihood of missing future visits.

Engaging patients in DSME is challenging. Studies are needed to examine how attendance can be improved among groups differing in demographics, health status, and psychosocial functioning. Results suggest that screening instruments for depression, other measures of mental health, and physical health could help identify patients less likely to adhere to diabetes education so that necessary support can be provided.

Funding for the research described in this article was provided by Merck & Co. in North Wales, Pa.