zbMATH — the first resource for mathematics

The intensity-score approach to adjusting for confounding. (English) Zbl 1210.62146
Summary: In a recent article on the efficacy of antihypertensive therapy, Berlowitz et al. [New England J. Med. 339, 1957–1963 (1998)] introduced an ad hoc method of adjusting for serial confounding assessed via an intensity score, which records cumulative differences over time between therapy actually received and therapy predicted by prior medical history. Outcomes are subsequently regressed on the intensity score and baseline covariates to determine whether intense treatment or exposure predicts a favorable response. We use a structural nested mean model to derive conditions sufficient for interpreting the Berlowitz results causally. We also consider a modified approach that scales the intensity at each time by the inverse expected treatment given prior medical history. This leads to a simple, two-step implementation of G-estimation if we assume a nonstandard but useful structural nested mean model in which subjects less likely to receive treatment are more likely to benefit from it. These modeling assumptions apply, for example, to health services research contexts in which differential access to care is a primary concern. They are also plausible in our analysis of the causal effect of potent antiretroviral therapy on change in CD4 cell count, because men in the sample who are less likely to initiate treatment when baseline CD4 counts are high are more likely to experience large positive changes. We further extend the methods to accomodate repeated outcomes and time-varying effects of time-varying exposures.

62P10 Applications of statistics to biology and medical sciences; meta analysis
62N02 Estimation in survival analysis and censored data
Full Text: DOI
[1] Ahdieh, Selection by indication of potent antiretroviral therapy use in a large cohort of women infected with Human Immunodeficiency Virus, American Journal of Epidemiology 152 pp 923– (2000) · doi:10.1093/aje/152.10.923
[2] Berlowitz, Inadequate management of blood pressure in a hypertensive population, New England Journal of Medicine 339 pp 1957– (1998) · doi:10.1056/NEJM199812313392701
[3] Bickel, Mathematial Statistics: Basic Ideas and Selected Topics (1977)
[4] Breslow, Statistical Methods in Cancer Research, Volume 1: The Analysis of Case-Control Studies (1980)
[5] Cox, The Plannng of Experiments (1958)
[6] Graybill, Theory and Application of the Linear Model (1976) · Zbl 0371.62093
[7] Greenland, Causal analysis in the health sciences, Journal of the American Statistical Association 95 pp 286– (2000) · doi:10.2307/2669548
[8] Greenland, Control of confounding in the assessment of medical technology, International Journal of Epidemiology 9 pp 361– (1980) · doi:10.1093/ije/9.4.361
[9] Greenland, The interpretation of multiplicative model parameters as standardized parameters, Statistics in Medicine 13 pp 989– (1994) · doi:10.1002/sim.4780131002
[10] Hernán, Marginal structural models to estimate the joint causal effect of non-randomized treatments, Journal of the American Statistical Association 96 pp 440– (2001) · Zbl 1018.62098 · doi:10.1198/016214501753168154
[11] Hernán, Estimating the causal effect of zidovudine on CD4 count with a marginal structural model for repeated measures, Statistics in Medicine 21 pp 1689– (2002) · doi:10.1002/sim.1144
[12] Lampinen, Association of antiretroviral therapy with detection of HIV-1 RNA and DNA in the anorectal mucosa of homosexual men, AIDS 14 pp F69– (2000) · doi:10.1097/00002030-200003310-00001
[13] Liang, Longitudinal data analysis using generalized linear models, Biometrika 73 pp 13– (1986) · Zbl 0595.62110 · doi:10.1093/biomet/73.1.13
[14] Newey, Semiparametric Efficiency Bounds, Journal of Applied Econometrics 5 pp 99– (1990) · Zbl 0705.62033 · doi:10.1002/jae.3950050202
[15] Newey, Large Sample Estimation and Hypothesis Testing. Handbook of Econometrics, (1994)
[16] Pearl, Causality: Models, Reasoning, and Inference (2000) · Zbl 0959.68116
[17] Robins, The control of confounding by intermediate variables, Statistics in Medicine 8 pp 679– (1989) · doi:10.1002/sim.4780080608
[18] Robins, Estimation of the time-dependent accelerated failure time model in the presence of confounding factors, Biometrika 79 pp 321– (1992) · Zbl 0753.62076 · doi:10.1093/biomet/79.2.321
[19] Robins, Correcting for non-compliance in randomized trials using structural nested mean models, Communications in Statistics, Theory and Methods 23 pp 2379– (1994) · Zbl 0825.62203 · doi:10.1080/03610929408831393
[20] Robins, Latent Modelling with Applications to Causality pp 69– (1997) · doi:10.1007/978-1-4612-1842-5_4
[21] Robins, The Encyclopedia of Biostatistics pp 4372– (1998)
[22] Robins, Robust estimation in sequentially ignorable missing data and causal inference models, Proceedings of the American Statistial Association Section on Bayesian Statistical Science pp 6– (2000)
[23] Robins, Comment on the Bickel and Kwon article, ”Inference for semiparametric models: Some questions and an answer.”, Statistica Sinica 11 pp 920– (2001)
[24] Robins, Estimating exposure effects by modelling the expectation of exposure conditional on confounders, Biometrics 48 pp 479– (1992) · Zbl 0768.62099 · doi:10.2307/2532304
[25] Robins, Marginal structural models and causal inference in epidemiology, Epidemiology 11 pp 550– (2000) · doi:10.1097/00001648-200009000-00011
[26] Rosenbaum, The central role of the propensity score in observational studies of causal effects, Biometrika 70 pp 41– (1983) · Zbl 0522.62091 · doi:10.1093/biomet/70.1.41
[27] Rubin, Comment: Neyman (1923) and causal inference in experiments and observational studies, Statistical Science 5 pp 472– (1990) · Zbl 0955.01559
[28] Seber, Linear Regression Analysis (1977)
[29] Stone, The assumptions on which causal inferences rest, Journal of the Royal Statistical Society, Series B 55 pp 455– (1993) · Zbl 0794.62003
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.