Forecasting Patient Demand at Urgent Care Clinics using Machine Learning

TL;DR

This paper evaluates machine learning models for predicting patient demand at urgent care clinics, demonstrating that ensemble methods significantly outperform existing approaches, especially amid COVID-19 disruptions.

Contribution

It introduces the application of ensemble machine learning techniques for demand forecasting in urgent care, with analysis of feature importance and adaptability during pandemic-related demand fluctuations.

Findings

Ensemble methods improved demand prediction accuracy by 23-27%.

Models effectively adapted to COVID-19 demand volatility.

Feature analysis identified key predictors for demand forecasting.

Abstract

Urgent care clinics and emergency departments around the world periodically suffer from extended wait times beyond patient expectations due to inadequate staffing levels. These delays have been linked with adverse clinical outcomes. Previous research into forecasting demand this domain has mostly used a collection of statistical techniques, with machine learning approaches only now beginning to emerge in recent literature. The forecasting problem for this domain is difficult and has also been complicated by the COVID-19 pandemic which has introduced an additional complexity to this estimation due to typical demand patterns being disrupted. This study explores the ability of machine learning methods to generate accurate patient presentations at two large urgent care clinics located in Auckland, New Zealand. A number of machine learning algorithms were explored in order to determine the most effective technique for this problem domain, with the task of making forecasts of daily patient demand three months in advance. The study also performed an in-depth analysis into the model behaviour in respect to the exploration of which features are most effective at predicting demand and which features are capable of adaptation to the volatility caused by the COVID-19 pandemic lockdowns. The results showed that ensemble-based methods delivered the most accurate and consistent solutions on average, generating improvements in the range of 23%-27% over the existing in-house methods for estimating the daily demand.