Hipertension Demand Forecasting Using Cross-Correlation and Lagged Multiple Linear Regression Models for Anticipating Health Resources Needs

Abstract

This article presents an algorithm that uses a combination of cross-correlation analysis and lagged multiple linear regression models to predict the time-series of future demand for clinical visits associated with a certain diagnosis, specifically hypertension, in the Catalan health-care system. The algorithm aims to provide a robust and explainable feature selection set of predictors. The study demonstrates that it is possible to predict demand associated with a diagnosis through the demand for previous clinical visits, and identifies important predictors for example case hypertension-related visits. The data used is from the primary care services of the Catalan Institute of Health, and the methodology can be applied to optimize resource allocation in the healthcare system.