An electronic health record based model predicts statin adherence, LDL cholesterol, and cardiovascular disease in the United States Military Health System

Abstract

HMG-CoA reductase inhibitors (or “statins”) are important and commonly used medications to lower cholesterol and prevent cardiovascular disease. Nearly half of patients stop taking statin medications one year after they are prescribed leading to higher cholesterol, increased cardiovascular risk, and costs due to excess hospitalizations. Identifying which patients are at highest risk for not adhering to long-term statin therapy is an important step towards individualizing interventions to improve adherence. Electronic health records (EHR) are an increasingly common source of data that are challenging to analyze but have potential for generating more accurate predictions of disease risk. The aim of this study was to build an EHR based model for statin adherence and link this model to biologic and clinical outcomes in patients receiving statin therapy. We gathered EHR data from the Military Health System which maintains administrative data for active duty, retirees, and dependents of the United States armed forces military that receive health care benefits. Data were gathered from patients prescribed their first statin prescription in 2005 and 2006. Baseline billing, laboratory, and pharmacy claims data were collected from the two years leading up to the first statin prescription and summarized using non-negative matrix factorization. Follow up statin prescription refill data was used to define the adherence outcome (> 80 percent days covered). The subsequent factors to emerge from this model were then used to build cross-validated, predictive models of 1) overall disease risk using coalescent regression and 2) statin adherence (using random forest regression). The predicted statin adherence for each patient was subsequently used to correlate with cholesterol lowering and hospitalizations for cardiovascular disease during the 5 year follow up period using Cox regression. The analytical dataset included 138 731 individuals and 1840 potential baseline predictors that were reduced to 30 independent EHR “factors”. A random forest predictive model taking patient, statin prescription, predicted disease risk, and the EHR factors as potential inputs produced a cross-validated c-statistic of 0.736 for classifying statin non-adherence. The addition of the first refill to the model increased the c-statistic to 0.81. The predicted statin adherence was independently associated with greater cholesterol lowering (correlation = 0.14, p < 1e-20) and lower hospitalization for myocardial infarction, coronary artery disease, and stroke (hazard ratio = 0.84, p = 1.87E-06). Electronic health records data can be used to build a predictive model of statin adherence that also correlates with statins’ cardiovascular benefits.

Prevalence, treatment, and control of severe hyperlipidemia

“…Our study demonstrates the power of utilizing EHR to capture a high-risk cohort of patients with severe hypercholesterolemia, whose treatment can be optimized with affordable and low-risk medications in order to prevent future adverse outcomes. The potential implications for providing a pragmatic approach to identify both patient-specific and system-wide care gaps are tremendous [ [34] , [35] , [36] , [37] , [38] , [39] ]. With an enhanced ability to pinpoint these patients, we will also have opportunities to develop and capitalize on novel implementation strategies to actually improve patient care.…”

Predicting Lipid-Lowering Medication Persistence after the First Cardiovascular Disease Hospitalization

“…Consistent with international evidence, 98,103,104 we found obesity to be a positive factor in medication persistence, which may relate to patients' perception of their obesity putting them at increased risk of a subsequent CVD event, thus, motivating their greater general interest in their health including medication adherence. Previous studies have shown that the initial refill history has the potential to improve subsequent medication adherence prediction for new therapy users 40,48 and past refill history was suggested as a significant factor for medication adherence prediction in primary prevention. 105 Our BRT, SVM, and ANN models using both forward and backward variable selection processes, demonstrated that variables related to past adherence behavior (e.g., past refill history, days between the latest prehospitalization dispensing date, and the hospital admission date) improved the prediction performance.…”

“…Among the few relevant studies, random forest (RF) and boosted regression tree (BRT) have been applied to predict medication adherence among patients who were new to statin or statin combination drugs achieving cross-validated c-statistics of 0.81 and 0.842, respectively. 40,48 Although not a cross-validated estimation of prediction accuracy, Bourdès et al observed that a one hidden layer neural network performed better than LR for predicting medication persistence for patients with acute coronary syndrome after the first hospital discharge. 49 Prediction of high risk for readmission has been used successfully as the basis for CDSS in hospital discharge planning.…”

A framework to build similarity-based cohorts for personalized treatment advice – a standardized, but flexible workflow with the R package SimBaCo

“…Hospital EHR, on the contrary, provide a more complete picture of the patient [26] as they can be considered as a quasi-electronic form of the patient record [27]. Those diagnosis codes are presumably more accurate and are complemented by additional data sources such as laboratory values [28]. However, this information is not standardized and therefore complicates subsequent analysis steps.…”