Send to

Choose Destination
Am J Epidemiol. 2019 Oct 8. pii: kwz218. doi: 10.1093/aje/kwz218. [Epub ahead of print]

Heterogeneous Exposure Associations in Observational Cohort Studies: The Example of Blood Pressure in Older Adults.

Author information

Department of Health Research and Policy, Stanford University, Stanford, CA, USA.
School of Biological and Population Health Sciences, Oregon State University, Corvallis, USA.
School of Electrical Engineering and Computer Science, Oregon State University, Corvallis, USA.
Department of Medicine, San Francisco VA Medical Center, San Francisco, CA, USA.
Kidney Health Research Collaborative, University of California, San Francisco, USA.
Department of Medicine, University of California, San Francisco, USA.
Department of Epidemiology and Biostatistics, University of California, San Francisco, USA.
Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA, USA.
Cardiovascular Health Research Unit, Departments of Medicine, Epidemiology, and Health Services, University of Washington, Seattle, USA.
Kaiser Permanente Washington Health Research Institute, Seattle, USA.


Heterogeneous exposure associations (HEAs) can be defined as differences in the association of a exposure with an outcome among subgroups that differ by a set of characteristics. This manuscript intends to foster discussion of HEAs in the epidemiological literature, and present a variant of the random forest algorithm that can be used to identify HEAs. We demonstrate the use of this algorithm in the setting of the association of systolic blood pressure and death in older adults. The training set included pooled data from the baseline examination of the Cardiovascular Health Study (1989-1993), the Health, Aging, and Body Composition study (1997-1998), and the Sacramento Area Latino Study on Aging (1998-1999). The test set included data from the National Health and Nutrition Examination Survey (1999-2002). The hazard ratios ranged from 1.25 (95% CI: 1.13, 1.37) per 10 mmHg higher systolic blood pressure in men aged ≤67 years with diastolic blood pressure >80 mmHg, to 1.00 (0.96, 1.03) in women with creatinine <0.7 mg/dL and a history of hypertension. HEAs have the potential to improve our understanding of disease mechanisms in diverse populations, and guide the design of randomized controlled trials to control exposures in heterogeneous populations.


Methods; blood pressure; effect modification; random forest


Supplemental Content

Full text links

Icon for Silverchair Information Systems
Loading ...
Support Center