IPROC iproc Iproceedings 2369-6893 JMIR Publications Toronto, Canada v4i2e11897 10.2196/11897 Abstract Abstract Prospective Real-World Performance Evaluation of a Machine Learning Algorithm to Predict 30-Day Readmissions in Patients with Heart Failure Using Electronic Medical Record Data Hale Timothy Kakarmath Sujay S MBBS, MS 1 2 3 Derakhshani Neda MSc 1
Connected Health Innovation Partners HealthCare Boston, MA, United States snderakhshani@partners.org
Golas Sara B MA 1 Felsted Jennifer PhD 1 3 Shibahara Takuma PhD 4 Aoki Hideo 4 Takata Mika 5 Naono Ken PhD 4 Kvedar Joseph MD 1 2 3 Jethwani Kamal MD, MPH 1 2 3 Agboola Stephen MD, MPH 1 2 3 Connected Health Innovation Partners HealthCare Boston, MA United States Massachusetts General Hospital Boston, MA United States Harvard Medical School Boston, MA United States Research and Development Group Hitachi, Ltd Tokyo Japan Big Data Laboratory Hitachi America Ltd Santa Clara, CA United States Corresponding Author: Neda Derakhshani snderakhshani@partners.org Jul-Dec2018 17 09 2018 4 2 e11897 9 8 2018 29 8 2018 ©Sujay S Kakarmath, Neda Derakhshani, Sara B Golas, Jennifer Felsted, Takuma Shibahara, Hideo Aoki, Mika Takata, Ken Naono, Joseph Kvedar, Kamal Jethwani, Stephen Agboola. Originally published in Iproceedings (http://www.iproc.org), 17.09.2018. 2018

This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in Iproceedings, is properly cited. The complete bibliographic information, a link to the original publication on http://www.iproc.org/, as well as this copyright and license information must be included.

 Background

Heart failure (HF) patients have a high readmission rate with approximately 20% of patients being readmitted within 30-days after discharge. Hospital interventions to reduce HF readmissions are resource- and effort-intensive. Widespread availability of electronic medical record data has spurred interest in using machine learning-based techniques for risk stratification of heart failure patients. The predictive performance of machine learning-based predictive models is often evaluated solely using the Area Under the Receiver Operating Characteristic (AUROC) curve. However, the AUROC is independent of prevalence therefore predictive models with the same AUROC can have differential clinical utility. Furthermore, the AUROC does not provide any insight about the presence of overfitting or decay in predictive performance of a model over time, both of which can affect its real-world performance.

 Objective

Our primary objective is to assess real-world performance of a 30-day readmission risk prediction model for HF patients, which had an AUROC of 0.71 in the training dataset.

 Methods

Predictions for risk of 30-day readmissions in HF patients in the Partners Healthcare System were prospectively obtained from the model. We assessed the positive (PPV) and negative predictive value (NPV), in addition to sensitivity, specificity, accuracy, model calibration and Brier score.

 Results

Four hundred twenty index admissions that were not part of the training dataset were included in this prospective evaluation. Readmission rate was 24% (101 30-day readmissions). The AUROC of the predictive model was 0.57. At a discrimination threshold of 0.2 for flagging high-risk index admissions, the sensitivity and specificity of the model were 53.46% and 63.32%, respectively. The PPV and NPV were 31.57% and 81.12%, respectively. The Brier score was 0.19.

 Conclusions

Our analysis offers important insights about the real-world performance of this predictive model. The NPV suggests that the model’s prediction about patients at low risk for readmission are reliable. This insight can be useful in optimizing resource allocation for patients with heart failure.

 accuracy machine learning positive predictive value validation