Influence of Framing: Recruitment to a Diabetes Disease Management Program From an Emergency Department Improves Enrollment and Outcomes

Introduction Disease management programs (DMPs) provide education, self-management skills, care coordination, and frequent clinical assessment and medication adjustment. Our health system’s diabetes mellitus (DM) DMP recruited patients from an emergency department (ED) and outpatient settings (primary care physicians’ [PCP] and endocrinologists’ offices; cold calling patients with poorly-controlled diabetes). We investigated whether recruitment to a DMP from an ED is feasible and effective, hypothesizing such patients would have better enrollment rates, future A1c control, and ED utilization because their receptiveness to change was “framed” by their ED visit. “Framing” is the notion that the same problem presented using a different context impacts response to the information. Being told in an acute-care ED setting one has newly-diagnosed or poorly-controlled DM, or DM-related complications may influence desire/commitment to enroll in the DMP and make lifestyle/medication changes. That is, acute illness or acute setting may influence/”frame” willingness to enroll and improve glycemic control. Methods We captured all DMP recruitees’ demographic, medical, insurance, A1c, and recruitment venue characteristics and evaluated future enrollment rates, A1c, and ED utilization from any ED in our health system. We analyzed pre- vs. post-recruitment changes in A1c and ED visit rates, comparing patients recruited from the ED who enrolled, patients recruited from the ED who did not enroll, patients recruited from outpatient settings who enrolled, and patients recruited from outpatient settings who did not enroll. Continuous enrollment predictor and outcome variables were compared using the Mann-Whitney test; categorical outcome variables were compared using Fisher’s exact test. Results There were no statistically significant differences in characteristics (including mean baseline A1c [~11.4%]) among patients recruited from the ED, clinics, or cold calling. Twenty-five percent of all ED-recruited patients enrolled vs. 35% from outpatient settings. When a recruiter familiar with the DMP was in the ED, 41% of ED patients enrolled vs. 12% at other times (p=0.0001). Nearly 84% of ED visits were for direct DM-related causes (eg, diabetic ketoacidosis, hyperosmolar hyperglycemic state) or complications with a well-established link to diabetes (eg, acute coronary syndrome, stroke, wound infection); there was no statistically-significant difference in enrollment rates between patients whose ED visit was vs. was not for a DM-related complaint (53.8% vs. 60.0%, p=0.8018). No other variables, including whether the patient had newly diagnosed DM, were associated with enrollment. Enrollees with worse baseline glycemic control (A1c ≥11%) had a greater median A1c decrease (3.5% vs. 1.9%) vs. those with less-poor baseline glycemic control (A1c <11%) or those declining the program (p=0.05). Post-recruitment ED visits-per-patient-per-month decreased among patients recruited from the ED (-0.08), but not among those recruited from outpatient settings. (+0.08), p<0.0001). Conclusion ED recruitment to a diabetes DMP is feasible and effective. An ED-based diabetes DMP recruiter had enrollment rates substantially greater than a cold-calling DMP recruiter, comparable to enrollment rates from PCPs and endocrinologists, suggesting the importance of the recruitment framing/context. ED-recruited patients achieved substantial improvements in A1c and future ED visit rates.