Hospital Case Cost Estimates Modelling - Algorithm Comparison

TL;DR

This paper compares five mathematical models designed to estimate hospital case costs in Ontario by assigning dollar values to clinical cases, addressing data gaps in patient-specific financial information.

Contribution

It introduces and verifies five novel models for hospital cost estimation, classified into relative intensity and cost-per-diem based approaches.

Findings

Models successfully assigned cost estimates to clinical cases.

Verification conducted using real dataset.

Models offer different approaches for hospital cost modeling.

Abstract

Ontario (Canada) Health System stakeholders support the idea and necessity of the integrated source of data that would include both clinical (e.g. diagnosis, intervention, length of stay, case mix group) and financial (e.g. cost per weighted case, cost per diem) characteristics of the Ontario healthcare system activities at the patient-specific level. At present, the actual patient-level case costs in the explicit form are not available in the financial databases for all hospitals. The goal of this research effort is to develop financial models that will assign each clinical case in the patient-specific data warehouse a dollar value, representing the cost incurred by the Ontario health care facility which treated the patient. Five mathematical models have been developed and verified using real dataset. All models can be classified into two groups based on their underlying method: 1. Models based on using relative intensity weights of the cases, and 2. Models based on using cost per diem.