Format

Send to

Choose Destination
Stat Methods Med Res. 2018 Oct 31:962280218808817. doi: 10.1177/0962280218808817. [Epub ahead of print]

Causal inference with multiple concurrent medications: A comparison of methods and an application in multidrug-resistant tuberculosis.

Author information

1
1 Department of Statistics, McMaster University, Hamilton, Canada.
2
2 Faculty of Pharmacy, Université de Montréal, Montreal, Canada.
3
3 Department of Epidemiology, Biostatistics & Occupational Health, McGill University, Montreal, Canada.
4
4 Division of Global HIV and TB, Centers for Disease Control and Prevention, New Delhi, India.
5
5 World Health Organization Collaborating Centre for Tuberculosis and Lung Diseases, Fondazione S. Maugeri, Tradate, Italy.
6
6 Clinical Epidemiology and Medical Statistics Unit, Department of Medical, Surgical and Experimental Sciences, University of Sassari, Sassari, Italy.
7
7 Rollins School of Public Health and Emory School of Medicine, Emory University, Atlanta, USA.
8
8 Departamento de Investigación en Hiperreactividad Bronquial, Instituto Nacional de Enfermedades Respiratorias, Mexico City, Mexico.
9
9 Unidad de Investigación Médica en Enfermedades Respiratorias, Instituto Mexicano del Seguro Social, Mexico City, Mexico.
10
10 Respiratory Epidemiology and Clinical Research Institute, McGill University Health Centre, Montreal, Canada.
11
11 Department of Medicine, McGill University, Montreal, Canada.

Abstract

This paper investigates different approaches for causal estimation under multiple concurrent medications. Our parameter of interest is the marginal mean counterfactual outcome under different combinations of medications. We explore parametric and non-parametric methods to estimate the generalized propensity score. We then apply three causal estimation approaches (inverse probability of treatment weighting, propensity score adjustment, and targeted maximum likelihood estimation) to estimate the causal parameter of interest. Focusing on the estimation of the expected outcome under the most prevalent regimens, we compare the results obtained using these methods in a simulation study with four potentially concurrent medications. We perform a second simulation study in which some combinations of medications may occur rarely or not occur at all in the dataset. Finally, we apply the methods explored to contrast the probability of patient treatment success for the most prevalent regimens of antimicrobial agents for patients with multidrug-resistant pulmonary tuberculosis.

KEYWORDS:

Causal inference; concurrent medications; generalized propensity score; machine learning; multidrug-resistant tuberculosis; targeted maximum likelihood estimation

PMID:
30381005
DOI:
10.1177/0962280218808817

Supplemental Content

Full text links

Icon for Atypon
Loading ...
Support Center