RESUMEN
We present a general framework for developing a machine learning (ML) tool that supports clinician assessment of patient risk using electronic health record-derived real-world data and apply the framework to a quality improvement use case in an oncology setting to identify patients at risk for a near-term (60 day) emergency department (ED) visit who could potentially be eligible for a home-based acute care program. Framework steps include defining clinical quality improvement goals, model development and validation, bias assessment, retrospective and prospective validation, and deployment in clinical workflow. In the retrospective analysis for the use case, 8% of patient encounters were associated with a high risk (pre-defined as predicted probability ≥20%) for a near-term ED visit by the patient. Positive predictive value (PPV) and negative predictive value (NPV) for future ED events was 26% and 91%, respectively. Odds ratio (OR) of ED visit (high- vs. low-risk) was 3.5 (95% CI: 3.4-3.5). The model appeared to be calibrated across racial, gender, and ethnic groups. In the prospective analysis, 10% of patients were classified as high risk, 76% of whom were confirmed by clinicians as eligible for home-based acute care. PPV and NPV for future ED events was 22% and 95%, respectively. OR of ED visit (high- vs. low-risk) was 5.4 (95% CI: 2.6-11.0). The proposed framework for an ML-based tool that supports clinician assessment of patient risk is a stepwise development approach; we successfully applied the framework to an ED visit risk prediction use case.
RESUMEN
PURPOSE: As immune checkpoint inhibitors (ICIs) have transformed the care of patients with cancer, it is unclear whether treatment at the end of life (EOL) has changed. Because aggressive therapy at the EOL is associated with increased costs and patient distress, we explored the association between the Food and Drug Administration (FDA) approvals of ICIs and treatment patterns at the EOL. METHODS: We conducted a retrospective, observational study using patient-level data from a nationwide electronic health record-derived database. Patients had advanced melanoma, non-small-cell lung cancer (NSCLC; cancer types with an ICI indication), or microsatellite stable (MSS) colon cancer (a cancer type without an ICI indication) and died between 2013 and 2017. We calculated annual proportions of decedents who received systemic cancer therapy in the final 30 days of life, using logistic regression to model the association between the post-ICI FDA approval time and use of systemic therapy at the EOL, adjusting for patient characteristics. We assessed the use of chemotherapy or targeted/biologic therapies at the EOL, before and after FDA approval of ICIs using Pearson chi-square test. RESULTS: There was an increase in use of EOL systemic cancer therapy in the post-ICI approval period for both melanoma (33.9% to 43.2%; P < .001) and NSCLC (37.4% to 40.3%; P < .001), with no significant change in use of systemic therapy in MSS colon cancer. After FDA approval of ICIs, patients with NSCLC and melanoma had a decrease in the use of chemotherapy, with a concomitant increase in use of ICIs at the EOL. CONCLUSION: The adoption of ICIs was associated with a substantive increase in the use of systemic therapy at the EOL in melanoma and a smaller yet significant increase in NSCLC.