Feature importance analysis for patient management decisions

TL;DR

This paper analyzes clinical data to identify key features influencing physicians' decisions on lab tests and medications, demonstrating that a small subset of features can predict these decisions effectively.

Contribution

It provides insights into feature importance in clinical decision-making and shows that a limited number of features can predict physician choices accurately.

Findings

Physician decisions can be predicted from a small subset of features.

Summary statistics for 335 lab and 407 medication decisions are reported.

Analysis based on electronic health records of 4486 post-surgical cardiac patients.

Abstract

The objective of this paper is to understand what characteristics and features of clinical data influence physician's decision about ordering laboratory tests or prescribing medications the most. We conduct our analysis on data and decisions extracted from electronic health records of 4486 post-surgical cardiac patients. The summary statistics for 335 different lab order decisions and 407 medication decisions are reported. We show that in many cases, physician's lab-order and medication decisions can be well predicted from a small subset of all features.