Feature selection and transformation by machine learning reduce variable numbers and improve prediction for heart failure readmission or death

Abstract

Background The prediction of readmission or death after a hospital discharge for heart failure (HF) remains a major challenge. Modern healthcare systems, electronic health records, and machine learning (ML) techniques allow us to mine data to select the most significant variables (allowing for reduction in the number of variables) without compromising the performance of models used for prediction of readmission and death. Moreover, ML methods based on transformation of variables may potentially further improve the performance. Objective To use ML techniques to determine the most relevant and also transform variables for the prediction of 30-day readmission or death in HF patients. Methods We identified all Western Australian patients aged 65 years and above admitted for HF between 2003–2008 in linked administrative data. We evaluated variables associated with HF readmission or death using standard statistical and ML based selection techniques. We also tested the new variables produced by transformation of the original variables. We developed multi-layer perceptron prediction models and compared their predictive performance using metrics such as Area Under the receiver operating characteristic Curve (AUC), sensitivity and specificity. Results Following hospital discharge, the proportion of 30-day readmissions or death was 23.7% in our cohort of 10,757 HF patients. The prediction model developed by us using a smaller set of variables ( n = 8) had comparable performance (AUC 0.62) to the traditional model ( n = 47, AUC 0.62). Transformation of the original 47 variables further improved (p<0.001) the performance of the predictive model (AUC 0.66). Conclusions A small set of variables selected using ML matched the performance of the model that used the full set of 47 variables for predicting 30-day readmission or death in HF patients. Model performance can be further significantly improved by transforming the original variables using ML methods.

Predicting mortality and hospitalization in heart failure using machine learning: A systematic literature review

“…of patients Setting Data source No. of features Primary outcome assessed Adler, E.D (2019) [10] 2006–2017 5 822 Inpatient and outpatient EHR and Trial 8 All-cause mortality Ahmad, T (2018) [30] 2000–2012 44 886 Inpatient and outpatient Registry 8 1-year all-cause mortality Allam, A (2019) [31] 2013 272 778 Inpatient Claims dataset 50 30-day all-cause readmission Angraal, S (2020) [13] 2006–2013 1 767 Inpatient Trial 26 All-cause mortality and HF hospitalization Ashfaq, A (2019) [32] 2012–2016 7 655 Inpatient and outpatient EHR 30-day all-cause readmission Awan, SE (2019) [33] 2003–2008 10 757 Inpatient and outpatient EHR 47 30-day HF-related readmission and mortality Chen, R (2019) [34] 2014–2017 98 Inpatient Prospective Clinical and MRI 32 Cardiac death, heart transplantation and HF-related hospitalization Chicco, D (2020) [11] 2015 299 Inpatient Medical records 13 One year survival Chirinos, J (2020) [35] 2006–2012 379 Inpatient Trial 48 Risk of all-cause death or heart failure-related hospital admission Desai, R.J (2020) [6] ...…”

“… First Author (year) Study Region No. of patients % Black Age % male % Hypertension % IHD Adler, E.D (2019) [10] USA and Europe 5 822 60.3 Ahmad, T (2018) [30] Europe 44 886 73.2 63 Allam, A (2019) [31] USA and Europe 272 778 73 ± 14 51 Angraal, S (2020) [13] USA, Canada, Brazil, Argentina, Russia, Georgia 1 767 72 (64–79) 50 Ashfaq, A (2019) [32] Europe 7 655 78.8 57 Awan, SE (2019) [33] Australia 10 757 82 ± 7.6 49 67 55 Chen, R (2019) [34] China 98 47 ± 14 79 23 Chicco, D (2020) [34] Pakistan 299 40–95* 65 Chirinos, J (2020) [35] USA, Canada, Russia 379 7.4 70 (62–77) 53.5 94.5 30.6 Desai, R.J (2020) [6] USA 9 502 5.1 78 ± 8 45 87.1 22 F...…”

Current Trends in Readmission Prediction: An Overview of Approaches

“…Even with the emergence of the ML algorithm, 29 out of 36 articles adopted traditional statistical methods. Among these studies, ~ 90% used LR either as a baseline [ 56 , 58 , 60 , 62 – 64 , 68 , 73 , 74 , 76 – 78 , 83 , 85 – 87 ] or the main model in prediction [ 60 , 69 , 71 , 82 , 88 – 90 ], and 3 studies derived their own risk scores on the basis of LR variable coefficients [ 61 , 66 , 84 ]. In the remaining 3 papers, the prognosis of readmission was carried out with Cox regression survival analysis.…”

Implementation of Artificial Intelligence-Based Clinical Decision Support to Reduce Hospital Readmissions at a Regional Hospital

“…[18][19][20][21][22][23][24][25][26] Although many studies have concluded that AI-based models are superior to traditional models for risk stratification, other studies have observed otherwise. 14,[27][28][29][30][31][32][33][34][35][36][37][38][39] While clinical factors have primarily been used to predict readmission, there has also been interest in incorporating sociodemographic factors into models to more accurately account for patients' sociopersonal contexts, which are increasingly recognized to affect health-related outcomes. [10][11][12][13]40 This is critical because health behaviors, social factors, and economic factors are estimated to account for 70% of a person's health.…”