TY - JOUR
T1 - Trajectories of Glycemic Change in a National Cohort of Adults with Previously Controlled Type 2 Diabetes
AU - McCoy, Rozalina G.
AU - Ngufor, Che
AU - Van Houten, Holly K.
AU - Caffo, Brian
AU - Shah, Nilay D.
N1 - Publisher Copyright:
© 2017 Wolters Kluwer Health, Inc. All rights reserved.
PY - 2017
Y1 - 2017
N2 - Background: Individualized diabetes management would benefit from prospectively identifying well-controlled patients at risk of losing glycemic control. Objectives: To identify patterns of hemoglobin A 1c (HbA 1c) change among patients with stable controlled diabetes. Research Design: Cohort study using OptumLabs Data Warehouse, 2001-2013. We develop and apply a machine learning framework that uses a Bayesian estimation of the mixture of generalized linear mixed effect models to discover glycemic trajectories, and a random forest feature contribution method to identify patient characteristics predictive of their future glycemic trajectories. Subjects: The study cohort consisted of 27,005 US adults with type 2 diabetes, age 18 years and older, and stable index HbA 1c <7.0%. Measures: HbA 1c values during 24 months of observation. Results: We compared models with k=1, 2, 3, 4, 5 trajectories and baseline variables including patient age, sex, race/ethnicity, comorbidities, medications, and HbA 1c. The k=3 model had the best fit, reflecting 3 distinct trajectories of glycemic change: (T1) rapidly deteriorating HbA 1c among 302 (1.1%) youngest (mean, 55.2 y) patients with lowest mean baseline HbA 1c, 6.05%; (T2) gradually deteriorating HbA 1c among 902 (3.3%) patients (mean, 56.5 y) with highest mean baseline HbA 1c, 6.53%; and (T3) stable glycemic control among 25,800 (95.5%) oldest (mean, 58.5 y) patients with mean baseline HbA 1c 6.21%. After 24 months, HbA 1c rose to 8.75% in T1 and 8.40% in T2, but remained stable at 6.56% in T3. Conclusions: Patients with controlled type 2 diabetes follow 3 distinct trajectories of glycemic control. This novel application of advanced analytic methods can facilitate individualized and population diabetes care by proactively identifying high risk patients.
AB - Background: Individualized diabetes management would benefit from prospectively identifying well-controlled patients at risk of losing glycemic control. Objectives: To identify patterns of hemoglobin A 1c (HbA 1c) change among patients with stable controlled diabetes. Research Design: Cohort study using OptumLabs Data Warehouse, 2001-2013. We develop and apply a machine learning framework that uses a Bayesian estimation of the mixture of generalized linear mixed effect models to discover glycemic trajectories, and a random forest feature contribution method to identify patient characteristics predictive of their future glycemic trajectories. Subjects: The study cohort consisted of 27,005 US adults with type 2 diabetes, age 18 years and older, and stable index HbA 1c <7.0%. Measures: HbA 1c values during 24 months of observation. Results: We compared models with k=1, 2, 3, 4, 5 trajectories and baseline variables including patient age, sex, race/ethnicity, comorbidities, medications, and HbA 1c. The k=3 model had the best fit, reflecting 3 distinct trajectories of glycemic change: (T1) rapidly deteriorating HbA 1c among 302 (1.1%) youngest (mean, 55.2 y) patients with lowest mean baseline HbA 1c, 6.05%; (T2) gradually deteriorating HbA 1c among 902 (3.3%) patients (mean, 56.5 y) with highest mean baseline HbA 1c, 6.53%; and (T3) stable glycemic control among 25,800 (95.5%) oldest (mean, 58.5 y) patients with mean baseline HbA 1c 6.21%. After 24 months, HbA 1c rose to 8.75% in T1 and 8.40% in T2, but remained stable at 6.56% in T3. Conclusions: Patients with controlled type 2 diabetes follow 3 distinct trajectories of glycemic control. This novel application of advanced analytic methods can facilitate individualized and population diabetes care by proactively identifying high risk patients.
KW - diabetes mellitus type 2
KW - glycosylated hemoglobin
KW - machine learning
KW - mixture of generalized linear mixed effects model (MGLMM)
KW - patient-centered medicine
KW - random forest feature contribution (rfFC) method
UR - http://www.scopus.com/inward/record.url?scp=85031817336&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85031817336&partnerID=8YFLogxK
U2 - 10.1097/MLR.0000000000000807
DO - 10.1097/MLR.0000000000000807
M3 - Article
C2 - 28922296
AN - SCOPUS:85031817336
SN - 0025-7079
VL - 55
SP - 956
EP - 964
JO - Medical care
JF - Medical care
IS - 11
ER -