PT  - JOURNAL ARTICLE
AU  - Liaw, Winston
AU  - Adepoju, Omolola
AU  - Womack, Pete
AU  - King, Ben
AU  - Prewitt, Todd
AU  - Dobbins, Jessica
AU  - Glasheen, William
AU  - Woodard, LeChauncy
AU  - Kakadiaris, Ioannis
TI  - Factors Associated with Accumulating Diabetes Complications in a Medicare Advantage Cohort to Inform a Prediction Tool
AID  - 10.1370/afm.22.s1.5940
DP  - 2024 Nov 20
TA  - The Annals of Family Medicine
PG  - 5940
VI  - 22
IP  - Supplement 1
4099  - http://www.annfammed.org/content/22/Supplement_1/5940.short
4100  - http://www.annfammed.org/content/22/Supplement_1/5940.full
SO  - Ann Fam Med2024 Nov 20; 22
AB  - Context: Complications contribute to significant type 2 diabetes mellitus (T2DM) morbidity. While preventable, many fail to receive timely treatment in primary care, leading the most vulnerable to accumulate complications.Objective: To identify the factors associated with the accumulation of T2DM complicationsStudy Design and Analysis: In this cohort study, we assessed the factors associated with the Diabetes Complications Severity Index (DCSI; which captures the number and severity of complications).Population Studied: Adults aged 65 years and older with T2DM enrolled in a national Medicare Advantage planSetting or Dataset: Enrollment and claims data (2016-2020)Measures: The dependent variable was DCSI. Independent variables included year, demographics (age, sex, race/ethnicity, language, dual eligibility), geography (rural vs. non-rural), comorbidities (Charlson Comorbidity Index (CCI) and Functional Comorbidity Index (FCI)), utilization, and quality measures. We used multilevel mixed effects models with a progressively expanding set of variables, including demographics (model 1), comorbidities (model 2), utilization (model 3), and quality measures (model 4).Results: We included 49,843 individuals in model 1. All four models showed a consistent relationship between year (incidence rate ratio (IRR)=1.30, p<0.001, 2020 vs. 2016, model 1), sex (IRR=0.85, p<0.001, female vs. male, model 1), race / ethnicity (IRR=1.05, p<0.001, Black vs. white, model 1), dual eligibility (IRR=1.26, p<0.001 yes vs. no), and rurality (IRR=0.90, p<0.001, yes vs. no, model 1). In model 2, CCI (IRR=1.18, p<0.001) and FCI (IRR=1.08, p<0.001) were associated with higher DCSI. In model 3, emergency department visits (IRR=1.002, p=0.01) were associated with higher DCSI while physician visits (IRR=0.998, p<0.001) were associated with lower DCSI. In model 4, not meeting the quality measure for blood pressure (a target blood pressure that is less than 140/90) was associated with higher DCSI (IRR=1.08, p<0.001), while meeting the quality measures for LDL cholesterol and hemoglobin A1c were not associated with DCSI.Conclusions: Year, male sex, race / ethnicity, non-rural status, comorbidities, emergency department visits, and not meeting the quality measure for blood pressure were associated with higher DCSI. These results will inform a primary care tool that uses artificial intelligence/machine learning to predict those individuals at high risk for diabetes progression.