Development of Predictive Models for Health Care Outcomes of Patients with Chronic Kidney Disease and Type 2 Diabetes

Progression of chronic kidney disease (CKD) is associated with increased risk of cardiovascular or renal events that contribute to increasing healthcare costs. The purpose of this study was to develop predictive models for rapid progression of CKD, days out of the home, and high costs for patients with both CKD and type 2 diabetes (T2D) . This retrospective observational cohort study used administrative claims data for patients with CKD (stage 3-4) and T2D aged 65-89 years enrolled in a Humana Medicare Advantage Prescription Drug plan 1/1/2012-12/31/2017. Patients were enrolled ≥1 year pre-index and followed up to one year post-index. Outcomes included rapid progression of CKD (>1 decrease in estimated glomerular filtration rate [eGFR] >5 mL/min/1.73m2 per year) , days out of the home (days spent out of their residence [e.g., hospital]; ≥2% days out of the home) and high costs (75-90th and >90th percentiles) post-index. Pre-index demographic and clinical characteristics, selected based on the LASSO technique, were included in logistic regression models to generate parameter estimates and model performance statistics. We identified 169,876 patients with CKD stage 3-4 and T2D. The C-statistics for the models ranged from 0.694 to 0.745. Lower initial eGFR led to incrementally higher risk for more days out of the home and high costs. Patients with urinary albumin to creatinine ratio (UACR) ≥300 mg/g had three times increased risk for rapid progression of CKD, 57% higher risk of more days out of the home, and, among those in the 75-90th cost percentile, had 38% higher costs than patients with UACR<30 mg/g. The presence of congestive heart failure, anemia, or higher pre-index utilization (≥5 physician visits) increased the risk for all outcomes. The predictive models developed in this study can be potentially used as decision support tools for clinicians and payers, and the risk scores from these models can be applied to future outcomes studies focused on patients with T2D and CKD. Disclosure R. Nair: None. M.K. Pasquale: Consultant; Bayer AG. T. Evers: Employee; Bayer AG. M.M. Cockrell: Employee; Humana. Stock/Shareholder; Humana. A. Gay: Employee; Bayer AG. R. Singh: None. N. Schmedt: Employee; Bayer AG. Funding This study was funded by Bayer.