Development and Evaluation of EHR-enabled Population Health Outreach Strategies to Improve Diabetes Screening in a Safety-net Health System: a Pragmatic Randomized Controlled Trial

Modified Abstract Section
ABSTRACT
Type 2 diabetes (T2D) screening remains suboptimal in spite of well-recognized, national screening guidelines. In the US, 7.3 million adults with T2D and 74.5 with prediabetes (PDM) remain undiagnosed. In spite of opportunistic screening in clinical practice, nearly one-third of primary care patients have undiagnosed dysglycemia (PDM + T2D). To close screening gaps, new strategies are needed. We adapt evidence-based approaches from cancer screening to conceptualize T2D screening as a multi-step process (risk assessment, screening invitation, test ordering, and test completion) requiring coordination across patient, provider, and health system interfaces. We previously developed the Parkland Dysglycemia Detection Program (PDDP) – an EHR-based, multicomponent population health T2D screening intervention that automates risk assessment, bulk orders screening tests, and facilitates bulk patient outreach via screening invitations. The PDDP closes multiple gaps in the screening process and supplements opportunistic screening in clinical practice. In our PDDP pilot study, a single, generic ‘overdue for screening’ invitation had a 41% response rate vs. 13% in usual care alone. Of those completing screening, 37% had PDM and 5% had T2D, representing cases ‘missed’ by opportunistic screening alone. Although the PDDP helped close overall screening gaps and detected cases of undiagnosed dysglycemia, response rates to generic invitations were similar across racial/ethnic subgroups (Hispanics 42%; NH Blacks 41%; NH whites 39%) and those with known PDM vs. unknown glycemic status (38% vs. 41%). To address known screening and outcome differences in racial/ethnic minorities and those with PDM, improved screening is needed. In this proposal, we seek to improve the PDDP response in racial/ethnic minorities and those with known PDM to close screening gaps. To accomplish this, we will develop Targeted (by race/ethnicity), Tailored (by known PDM vs. unknown glycemic state) (TT) screening invitations (Aim 1) to increase engagement of high risk subgroups. We will then conduct a 3-arm split-cluster RCT (Aim 2) to evaluate the efficacy of PDDP-delivered TT screening outreach + navigation of non-responders vs. PDDP-delivered generic invitations to improve screening completion in high risk patients and evaluate the effectiveness of the TT PDDP and Generic PDDP to improve screening completion vs. usual care, opportunistic screening. Lastly, we will conduct cost-effectiveness analyses (Aim 3) to compare direct costs and the cost per patient screened and case found across the three study arms. Together, these findings will provide actionable evidence on clinical and cost-effective ways to close screening gaps in high-risk patients. Because the PDDP is highly automated and scalable using a common EHR, our findings can be practically implemented in most health systems. Our findings will have important implications for clinics and health systems seeking to close T2D screening gaps and decrease screening differences through scalable, population-health T2D screening strategies to supplement opportunistic screening in usual care.

Grant Number: 5R01MD016101-04
NIH Institute/Center: NIH
Principal Investigator: Michael Bowen