Early detection of heart failure with varying prediction windows by structured and unstructured data in electronic health records

Abstract

Heart failure (HF) prevalence is increasing and is among the most costly diseases to society. Early detection of HF would provide the means to test lifestyle and pharmacologic interventions that may slow disease progression and improve patient outcomes. This study used structured and unstructured data from electronic health records (EHR) to predict onset of HF with a particular focus on how prediction accuracy varied in relation to time before diagnosis. EHR data were extracted from a single health care system and used to identify incident HF among primary care patients who received care between 2001 and 2010. A total of 1,684 incident HF cases were identified and 13,525 controls were selected from the same primary care practices. Models were compared by varying the beginning of the prediction window from 60 to 720 days before HF diagnosis. As the prediction window decreased, the performance [AUC (95% CIs)] of the predictive HF models increased from 65% (63%-66%) to 74% (73%-75%) for the unstructured, from 73% (72%-75%) to 81% (80%-83%) for the structured, and from 76% (74%-77%) to 83% (77%-85%) for the combined data.

Figure 1

Illustration of the timeline for predictive modeling.

Figure 2

Predictive modeling performance (AUC) with different feature types and predictive window sizes.

Figure 3

Heat maps of the top m features selected by any fold in 10-fold cross validation (CV) for unstructured feature types (a; m=15), structured feature types (b; m=100) and combined feature types (c; m=100). Each row along the vertical axis is a feature. The horizontal axis indicates the prediction window size from 720 days to 60 days prior to the diagnosis date. The dark red indicates that the feature is selected in all of the 10 folds in 10-fold CV; the dark blue indicates that the feature is not selected in any of the 10 CV folds. Examples of clinical variables are labeled to highlight the different patterns of predictive ability as a function of predictive window size.