Estimation of optimal treatment regimes with electronic medical record data using the residual life value estimator.

Clinicians and patients must make treatment decisions at a series of key decision points throughout disease progression. A dynamic treatment regime is a set of sequential decision rules that return treatment decisions based on accumulating patient information, like that commonly found in electronic medical record (EMR) data. When applied to a patient population, an optimal treatment regime leads to the most favorable outcome on average. Identifying optimal treatment regimes that maximize residual life is especially desirable for patients with life-threatening diseases such as sepsis, a complex medical condition that involves severe infections with organ dysfunction. We introduce the residual life value estimator (ReLiVE), an estimator for the expected value of cumulative restricted residual life under a fixed treatment regime. Building on ReLiVE, we present a method for estimating an optimal treatment regime that maximizes expected cumulative restricted residual life. Our proposed method, ReLiVE-Q, conducts estimation via the backward induction algorithm Q-learning. We illustrate the utility of ReLiVE-Q in simulation studies, and we apply ReLiVE-Q to estimate an optimal treatment regime for septic patients in the intensive care unit using EMR data from the Multiparameter Intelligent Monitoring Intensive Care database. Ultimately, we demonstrate that ReLiVE-Q leverages accumulating patient information to estimate personalized treatment regimes that optimize a clinically meaningful function of residual life.

DYNAMIC PREDICTION OF RESIDUAL LIFE WITH LONGITUDINAL COVARIATES USING LONG SHORT-TERM MEMORY NETWORKS.

Sepsis, a complex medical condition that involves severe infections with life-threatening organ dysfunction, is a leading cause of death worldwide. Treatment of sepsis is highly challenging. When making treatment decisions, clinicians and patients desire accurate predictions of mean residual life (MRL) that leverage all available patient information, including longitudinal biomarker data. Biomarkers are biological, clinical, and other variables reflecting disease progression that are often measured repeatedly on patients in the clinical setting. Dynamic prediction methods leverage accruing biomarker measurements to improve performance, providing updated predictions as new measurements become available. We introduce two methods for dynamic prediction of MRL using longitudinal biomarkers. in both methods, we begin by using long short-term memory networks (LSTMs) to construct encoded representations of the biomarker trajectories, referred to as "context vectors." In our first method, the LSTM-GLM, we dynamically predict MRL via a transformed MRL model that includes the context vectors as covariates. In our second method, the LSTM-NN, we dynamically predict MRL from the context vectors using a feed-forward neural network. We demonstrate the improved performance of both proposed methods relative to competing methods in simulation studies. We apply the proposed methods to dynamically predict the restricted mean residual life (RMRL) of septic patients in the intensive care unit using electronic medical record data. We demonstrate that the LSTM-GLM and the LSTM-NN are useful tools for producing individualized, real-time predictions of RMRL that can help inform the treatment decisions of septic patients.

Complete Revascularization and Angina-Related Health Status in the ISCHEMIA Trial.

BACKGROUND: The impact of complete revascularization (CR) on angina-related health status (symptoms, function, quality of life) in chronic coronary disease (CCD) has not been well studied. OBJECTIVES: Among patients with CCD randomized to invasive (INV) vs conservative (CON) management in ISCHEMIA (International Study of Comparative Health Effectiveness with Medical and Invasive Approaches), we compared the following: 1) the impact of anatomic and functional CR on health status compared with incomplete revascularization (ICR); and 2) the predicted impact of achieving CR in all INV patients compared with CON. METHODS: Multivariable regression adjusting for patient characteristics was used to compare 12-month health status after independent core laboratory-defined CR vs ICR in INV patients who underwent revascularization. Propensity-weighted modeling was then performed to estimate the treatment effect had CR or ICR been achieved in all INV patients, compared with CON. RESULTS: Anatomic and functional CR were achieved in 43.3% and 57.8% of 1,641 INV patients, respectively. Among revascularized patients, CR was associated with improved Seattle Angina Questionnaire Angina Frequency compared with ICR after adjustment for baseline differences. After modeling CR and ICR in all INV patients, patients with CR and ICR each had greater improvements in health status than CON, with better health status with CR than ICR. The projected benefits of CR were most pronounced in patients with baseline daily/weekly angina and not seen in those with no angina. CONCLUSIONS: Among patients with CCD in ISCHEMIA, health status improved more with CR compared with ICR or CON, particularly in those with frequent angina. Anatomic and functional CR provided comparable improvements in quality of life. (International Study of Comparative Health Effectiveness With Medical and Invasive Approaches [ISCHEMIA]; NCT01471522).

Impact of Complete Revascularization in the ISCHEMIA Trial.

BACKGROUND: Anatomic complete revascularization (ACR) and functional complete revascularization (FCR) have been associated with reduced death and myocardial infarction (MI) in some prior studies. The impact of complete revascularization (CR) in patients undergoing an invasive (INV) compared with a conservative (CON) management strategy has not been reported. OBJECTIVES: Among patients with chronic coronary disease without prior coronary artery bypass grafting randomized to INV vs CON management in the ISCHEMIA (International Study of Comparative Health Effectiveness with Medical and Invasive Approaches) trial, we examined the following: 1) the outcomes of ACR and FCR compared with incomplete revascularization; and 2) the potential impact of achieving CR in all INV patients compared with CON management. METHODS: ACR and FCR in the INV group were assessed at an independent core laboratory. Multivariable-adjusted outcomes of CR were examined in INV patients. Inverse probability weighted modeling was then performed to estimate the treatment effect had CR been achieved in all INV patients compared with CON management. RESULTS: ACR and FCR were achieved in 43.4% and 58.4% of 1,824 INV patients. ACR was associated with reduced 4-year rates of cardiovascular death or MI compared with incomplete revascularization. By inverse probability weighted modeling, ACR in all 2,296 INV patients compared with 2,498 CON patients was associated with a lower 4-year rate of cardiovascular death or MI (difference -3.5; 95% CI: -7.2% to 0.0%). In comparison, the event rate difference of cardiovascular death or MI for INV minus CON in the overall ISCHEMIA trial was -2.4%. Results were similar but less pronounced with FCR. CONCLUSIONS: The outcomes of an INV strategy may be improved if CR (especially ACR) is achieved. (International Study of Comparative Health Effectiveness With Medical and Invasive Approaches [ISCHEMIA]; NCT01471522).

Impact of complete revascularization in the IschemiaTrial

BACKGROUND: Anatomic complete revascularization (ACR) and functional complete revascularization (FCR) have been associated with reduced death and myocardial infarction (MI) in some prior studies. The impact of complete revascularization (CR) in patients undergoing an invasive (INV) compared with a conservative (CON) management strategy has not been reported. OBJECTIVES: Among patients with chronic coronary disease without prior coronary artery bypass grafting randomized to INV vs CON management in the ISCHEMIA (International Study of Comparative Health Effectiveness with Medical and Invasive Approaches) trial, we examined the following: 1) the outcomes of ACR and FCR compared with incomplete revascularization; and 2) the potential impact of achieving CR in all INV patients compared with CON management. METHODS: ACR and FCR in the INV group were assessed at an independent core laboratory. Multivariable-adjusted outcomes of CR were examined in INV patients. Inverse probability weighted modeling was then performed to estimate the treatment effect had CR been achieved in all INV patients compared with CON management. RESULTS: ACR and FCR were achieved in 43.4% and 58.4% of 1,824 INV patients. ACR was associated with reduced 4-year rates of cardiovascular death or MI compared with incomplete revascularization. By inverse probability weighted modeling, ACR in all 2,296 INV patients compared with 2,498 CON patients was associated with a lower 4-year rate of cardiovascular death or MI (difference -3.5; 95% CI: -7.2% to 0.0%). In comparison, the event rate difference of cardiovascular death or MI for INV minus CON in the overall ISCHEMIA trial was -2.4%. Results were similar but less pronounced with FCR. CONCLUSIONS: The outcomes of an INV strategy may be improved if CR (especially ACR) is achieved. (International Study of Comparative Health Effectiveness With Medical and Invasive Approaches [ISCHEMIA]; NCT01471522).