Multiple State Analysis, a Multidimentional Approach to Multiple Time-to-Event Data and Life Course Health Trajectories: Application to Patients with Myocardial Infarction

TL;DR

This paper introduces a novel multiple state analysis method for studying complex life course health trajectories using multi-dimensional time-to-event data, demonstrated on myocardial infarction patient records.

Contribution

It presents a new empirical approach to analyze multiple health states over time, capturing complex trajectories and their associations with socio-demographics.

Findings

Different long-term condition patterns are linked to myocardial infarction onset.

Hypertension plays a recurrent role in health trajectories.

Health condition patterns vary by sex and ethnicity.

Abstract

Life course epidemiology of chronic diseases has been dominated so far by the environmental approach. Whether it focuses on early life exposures and events or later lifestyle behaviors, this approach assumes that previous life experiences interact at each stage of life and shape subsequent health trajectories. Inspired by the analysis of social trajectories, focusing on transitions between multiple states, in multiple dimensions of the social experience, we propose in this paper a novel empirical approach to multiple time-to-event health data, denoted as multiple state analysis. Alike the so-called state sequence analysis, the aim of multiple state analysis is to create typologies of the main life course trajectories. This approach is illustrated by the analysis of records from a south London general practice electronic health record from which multiple long term conditions associated with myocardial infarction were considered. Among expected results such as the recurrent role of hypertension, multiple state analysis shows that different patterns of long term conditions including physical and mental health conditions, are associated with the onset timing of myocardial infarction but also with socio-demographics such as sex and ethnicity.