Project

Back to overview

Pathways from mental disorders to burden of disease: Causal mediation analysis of big data from South Africa

Applicant Haas Andreas
Number 193381
Funding scheme Ambizione
Research institution Institut für Sozial- und Präventivmedizin Universität Bern
Institution of higher education University of Berne - BE
Main discipline Mental Disorders, Psychosomatic Diseases
Start/End 01.07.2021 - 30.06.2025
Approved amount 924'952.00
Show all

All Disciplines (5)

Discipline
Mental Disorders, Psychosomatic Diseases
Accidents
Medical Statistics
Addictive Diseases
Methods of Epidemiology and Preventive Medicine

Keywords (6)

Causal inference; Global mental health; Mental illness; Physical comorbidity; South Africa; Mortality

Lay Summary (German)

Lead
Menschen, die an schweren psychischen Erkrankungen leiden, haben eine kürzere Lebenserwartung als psychisch Gesunde. In Ländern mit niedrigem und mittlerem Einkommensniveau sind die Ursachen für die verkürzte Lebenserwartung von Menschen mit psychischen Erkrankungen wenig erforscht. In diesem Projekt werden mögliche Ursachen für die erhöhte Sterblichkeit von Menschen mit psychischen Erkrankungen in Südafrika untersucht.
Lay summary

Studien aus Industrieländern haben gezeigt, dass Menschen, die an psychischen Erkrankungen leiden, auch häufiger schwer körperlich erkranken als psychisch Gesunde. Das häufigere Auftreten von bestimmten körperlichen Erkrankungen, wie beispielsweise Herz-Kreislauferkrankungen, trägt massgelblich zur geringeren Lebenserwartung von Menschen mit psychischen Erkrankungen bei. In Ländern mit niedrigem und mittlerem Einkommensniveau sind die Ursachen für die verkürzte Lebenserwartung von Menschen mit psychischen Erkrankungen wenig erforscht. In diesem Projekt werden mögliche Ursachen für die erhöhte Sterblichkeit von Menschen mit psychischen Erkrankungen in Südafrika untersucht. Im Speziellen wird der ursächliche Einfluss von Herz-Kreislaufkrankheiten, Stoffwechselkrankheiten, Infektionskrankheiten und unnatürlichen Todesarten auf die erhöhte Sterblichkeit von Menschen mit psychischen Erkrankungen berechnet. Hierzu werden routinemässig erhobene Gesundheitsdaten mit statistischen Verfahren ausgewertet.

Direct link to Lay Summary Last update: 20.05.2021

Responsible applicant and co-applicants

Abstract

Background: People with mental illness have sharply increased mortality compared with the general population. Lower life expectancy in people with mental illness reflects a variety of reasons including higher rates of physical illness, unnatural deaths, and health care disparities. The relative contribution of the various causal pathways to mortality in people with mental illness is largely unknown, especially for low- and middle-income countries. Recent methodological advances in causal mediation analysis make it possible to quantify the contribution of mental disorders to mortality through indirect causal pathways. Previous studies, for example the highly influential Global Burden of Disease (GBD) study are limited in their ability to capture mortality attributable to mental illness through indirect causal pathways. Aim: I aim to analyze big data from South Africa with state-of-the-art causal inference methods to quantify the effect of mental disorders on adverse health outcomes and mortality. Objectives:1) To estimate the difference in life expectancy between people with and without mental illness. 2) To study mental illness as risk factors for adverse health outcomes including (a) physical illnesses, (b) unnatural death, and (c) poor medication adherence. 3) To estimate the causal effect of mental illness on mortality through indirect causal pathways. 4) To collaborate with the GBD study to improve estimates of the burden of mental disorders in low- and middle-income countries.Hypotheses:1) Mental disorders are etiological risk factors for adverse health outcomes including physical illnesses, unnatural death, and poor medication adherence. 2) Mental disorders have a large indirect causal effect on mortality. Data sources: Large observational databases integrating clinical, laboratory, pharmacy, and mortality surveillance data of 9 million people. Data come from a) public health facilities in the Western Cape province of South Africa, b) a large South African private-sector disease management program, and c) the world’s largest international HIV cohort collaboration. Methods: Adults, 18 years or older, who were living in South Africa, on January 1, 2011 will be included. Mental disorders and physical comorbidities will be ascertained based on ICD-10 diagnoses from hospital discharge data and pharmacy records. Medication adherence will be measured using pharmacy claims data and mortality through linkage of medical records with death certificates from the population registry. Survival models will be used to examine differences in life expectancy between people with and without mental disorders (objective 1). The effect of mental disorders on adverse health outcomes will be estimated using causal inference models (targeted maximum likelihood estimation) (objective 2). The indirect effect of mental disorders on mortality through intermediate causes will be estimated using causal mediation analysis (g-formula) (objective 3). Estimates and data will be shared with the GBD study (objective 4). Significance: Excess mortality in people with mental illness is a neglected public health problem. This study will harvest the power of big data from heath information systems using state-of-the-art causal inference methods to increase our understanding of the causal pathways that lead to the much higher death rates among people with mental illness. Results are relevant for public mental health policy and program, and the estimation of the burden of mental disorders in low- and middle-income countries.
-