Secondary Logo

Journal Logo

Institutional members access full text with Ovid®

Revealing Causal Heterogeneity Using Time Series Analysis of Ambulatory Assessments: Application to the Association Between Depression and Physical Activity After Myocardial Infarction

Rosmalen, Judith G.M. PhD; Wenting, Angela M.G. MSc; Roest, Annelieke M. MSc; de Jonge, Peter PhD; Bos, Elisabeth H. PhD

doi: 10.1097/PSY.0b013e3182545d47
Statistical Issues and Data Analysis

Objective Studies in psychosomatic medicine are characterized by analyses that typically compare groups. This nomothetic approach leads to conclusions that apply to the average group member but not necessarily to individual patients. Idiographic studies start at the individual patient and are suitable to study associations that differ between time points or between individuals. We illustrate the advantages of the idiographic approach in analyzing ambulatory assessments, taking the association between depression and physical activity after myocardial infarction as an example.

Methods Five middle-aged men who had myocardial infarction with mild to moderate symptoms of depression were included in this study. Four of these participants monitored their physical activity and depressive symptoms during a period of 2 to 3 months using a daily self-registration form. The time series of each individual participant were investigated using vector autoregressive modeling, which enables the analysis of temporal dynamics between physical activity and depression.

Results We found causal heterogeneity in the association between depression and physical activity. Participants differed in the predominant direction of effect, which was either from physical activity to depression (n = 1, 85 observations, unstandardized effect size = −0.183, p = .03) or from depression to physical activity (n = 2, 65 and 59 observations, unstandardized effect sizes = −0.038 and −0.381, p < .001 and p = .04). Also, the persistency of effects differed among individuals.

Conclusions Vector autoregressive models are suitable in revealing causal heterogeneity and can be easily used to analyze ambulatory assessments. We suggest that these models might bridge the gap between science and clinical practice by translating epidemiological results to individual patients.

Abbreviations PEP = Psycho-Educational Prevention Module

BDI = Beck Depression Inventory

PCI = percutaneous coronary intervention

CABG = coronary artery bypass graft

LVEF = left ventricular ejection fraction

BMI = body mass index

VAR = vector autoregressive modeling

From the University of Groningen (J.G.M.R., P.d.J., E.H.B.), Groningen; and Tilburg University (A.M.G.W., A.M.R.), Tilburg, the Netherlands.

Address correspondence and reprint requests to Peter de Jonge, PhD, Department of Psychiatry, University of Groningen, Groningen, the Netherlands. E-mail:

Received for publication April 28, 2011; revision received November 2, 2011.

Copyright © 2012 by American Psychosomatic Society
You currently do not have access to this article

To access this article:

Note: If your society membership provides full-access, you may need to login on your society website