Segmented Regression Analysis of Interrupted Time Series Design: Application in Health Science Research

Abstract

Currently, interrupted time series (ITS) design is increasingly applied in health science research since it is a strong statistical method for evaluating effects of interventions. The purpose of this article is to show how segmented regression analysis approach of ITS, which is
used to estimate intervention effects if data collection is longitudinal or time-series data by considering secular trend of time-series data after
the interventions. Time-series data are amount If those collected at equally spaced intervals, and need to be collected several times before
and after the interventions. Segmented regression  analysis is powerfully utilized for estimating intervention effects. This method can control
influence of other factors that may affect the change of time series data. The contents of this article include an introduction, ITS design, purpose and statistical method of ITS, analysis sample, and interpreting statistical results. Analysis samples are simulated for data analysis to clarify and make practical use of this method in health science research.