×

Effect modification and design sensitivity in observational studies. (English) Zbl 06158331

Summary: In an observational study of treatment effects, subjects are not randomly assigned to treatment or control, so differing outcomes in treated and control groups may reflect a bias from nonrandom assignment rather than a treatment effect. After adjusting for measured pretreatment covariates, perhaps by matching, a sensitivity analysis determines the magnitude of bias from an unmeasured covariate that would need to be present to alter the conclusions of the naive analysis that presumes adjustments eliminated all bias. Other things being equal, larger effects tend to be less sensitive to bias than smaller effects. Effect modification is an interaction between a treatment and a pretreatment covariate controlled by matching, so that the treatment effect is larger at some values of the covariate than at others. In the presence of effect modification, it is possible that results are less sensitive to bias in subgroups experiencing larger effects. Two cases are considered: (i) an a priori grouping into a few categories based on covariates controlled by matching and (ii) a grouping discovered empirically in the data at hand. In case (i), subgroup specific bounds on \(p\)-values are combined using the truncated product of \(p\)-values. In case (ii), information that is fixed under the null hypothesis of no treatment effect is used to partition matched pairs in the hope of identifying pairs with larger effects. The methods are evaluated using an asymptotic device, the design sensitivity, and using simulation. Sensitivity analysis for a test of the global null hypothesis of no effect is converted to sensitivity analyses for subgroup analyses using closed testing. A study of an intervention to control malaria in Africa is used to illustrate.

MSC:

62-XX Statistics

Software:

optmatch; DOS
PDFBibTeX XMLCite
Full Text: DOI Link

References:

[1] Altonji J. G., Journal of Political Economy 113 pp 151– (2005) · doi:10.1086/426036
[2] Bauer P., Biometrika 83 pp 934– (1996) · Zbl 0883.62033 · doi:10.1093/biomet/83.4.934
[3] Berger R. L., Statistical Science 11 pp 283– (1996) · Zbl 0955.62555 · doi:10.1214/ss/1032280304
[4] DOI: 10.1198/016214502753479374 · Zbl 1073.62550 · doi:10.1198/016214502753479374
[5] Breiman L., Classification and Regression Trees (1983)
[6] Brown B. M., Biometrika 68 pp 235– (1981) · doi:10.1093/biomet/68.1.235
[7] DOI: 10.1080/01621459.1994.10476778 · doi:10.1080/01621459.1994.10476778
[8] Conover W. J., Biometrics 44 pp 189– (1988) · Zbl 0707.62265 · doi:10.2307/2531906
[9] Cornfield J., Journal of the National Cancer Institute 22 pp 173– (1959)
[10] Fisher R. A., Statistical Methods for Research Workers (1932) · JFM 58.1161.04
[11] Fisher R. A., The Design of Experiments (1935)
[12] Gabriel K. R., Annals of Mathematical Statistics 40 pp 224– (1969) · Zbl 0198.23602 · doi:10.1214/aoms/1177697819
[13] Gastwirth J. L., Jurimetrics 33 pp 19– (1992)
[14] Hammond E. C., Journal of the National Cancer Institute 32 pp 1161– (1964)
[15] DOI: 10.1198/106186006X137047 · doi:10.1198/106186006X137047
[16] Hansen B. B., R News 7 pp 18– (2007)
[17] DOI: 10.1198/jasa.2009.tm08338 · Zbl 1388.62231 · doi:10.1198/jasa.2009.tm08338
[18] Hodges J. L., Annals of Mathematical Statistics 34 pp 598– (1963) · Zbl 0203.21105 · doi:10.1214/aoms/1177704172
[19] Hosman C. A., Annals of Applied Statistics 4 pp 849– (2010) · Zbl 1194.62089 · doi:10.1214/09-AOAS315
[20] Ichino A., Journal of Applied Econometrics 23 pp 305– (2008) · doi:10.1002/jae.998
[21] Imbens G. W., American Economic Review 93 pp 126– (2003) · doi:10.1257/000282803321946921
[22] Jogdeo K., The Annals of Statistics 5 pp 495– (1977) · Zbl 0401.62028 · doi:10.1214/aos/1176343846
[23] DOI: 10.1080/01621459.1979.10481037 · doi:10.1080/01621459.1979.10481037
[24] Lehmann E. L., Testing Statistical Hypotheses (3rd ed.) (2005) · Zbl 1076.62018
[25] Manski C., American Economic Review 80 pp 319– (1990)
[26] Marcus R., Biometrika 63 pp 655– (1976) · Zbl 0353.62037 · doi:10.1093/biomet/63.3.655
[27] Marcus S. M., Journal of Educational Statistics 22 pp 193– (1997)
[28] Maritz J., Biometrika 66 pp 163– (1979) · doi:10.3102/10769986022002193
[29] Millimet D. L., Journal of Applied Econometrics (2012)
[30] Molineaux L., The GARKI Project: Research on the Epidemiology and Control of Malaria in the Sudan Savanna of West Africa (1980)
[31] Neyman J., Statistical Science 5 pp 463– (1923)
[32] Noether G., Journal of the American Statistical Association 68 pp 716– (1973) · doi:10.1080/01621459.1973.10481411
[33] Pepper J., Journal of Econometrics 166 pp 79– (2012) · Zbl 1441.62711 · doi:10.1016/j.jeconom.2011.06.007
[34] Rosenbaum P. R., Biometrika 74 pp 13– (1987) · Zbl 0605.62130 · doi:10.1093/biomet/74.1.13
[35] Rosenbaum P. R., Journal of the Royal Statistical Society, Series B 53 pp 597– (1991)
[36] Rosenbaum P. R., Journal of the American Statistical Association 88 pp 1250– (1993) · doi:10.1080/01621459.1993.10476405
[37] Rosenbaum P. R., Observational Studies (2nd ed.) (2002)
[38] Rosenbaum P. R., Biometrika 91 pp 153– (2004) · Zbl 1132.62363 · doi:10.1093/biomet/91.1.153
[39] Rosenbaum P. R., Biometrics 63 pp 1164– (2007) · doi:10.1111/j.1541-0420.2007.00783.x
[40] Rosenbaum P. R., Biometrics 63 pp 456– (2007) · doi:10.1007/978-1-4757-3692-2
[41] DOI: 10.1198/jasa.2010.tm09570 · Zbl 1392.62326 · doi:10.1198/jasa.2010.tm09570
[42] Rosenbaum P. R., Design of Observational Studies (2010) · Zbl 1308.62005 · doi:10.1007/978-1-4419-1213-8
[43] Rosenbaum P. R., Looking Back: Proceedings of a Conference in Honor of Paul W. Holland pp 87– (2011) · doi:10.1007/978-1-4419-9389-2_6
[44] Rosenbaum P. R., Biometrics 67 pp 1017– (2011) · Zbl 1226.62125 · doi:10.1111/j.1541-0420.2010.01535.x
[45] DOI: 10.1198/jasa.2011.tm10422 · Zbl 1396.62109 · doi:10.1198/jasa.2011.tm10422
[46] Rosenbaum P. R., Biometrika 99 pp 763– (2012) · Zbl 1323.62038 · doi:10.1093/biomet/ass032
[47] Rosenbaum P. R., Journal of the Royal Statistical Society, Series B 45 pp 212– (1983)
[48] Rosenbaum P. R., Biometrics 41 pp 103– (1985) · Zbl 0607.62137 · doi:10.2307/2530647
[49] DOI: 10.1198/jasa.2009.0016 · Zbl 1388.62333 · doi:10.1198/jasa.2009.0016
[50] Rubin D. B., Journal of Educational Psychology 66 pp 688– (1974) · doi:10.1037/h0037350
[51] Schwartz S., Statistics in Medicine 31 pp 949– (2012) · doi:10.1002/sim.4472
[52] Simes R. J., Biometrika 73 pp 751– (1986) · Zbl 0613.62067 · doi:10.1093/biomet/73.3.751
[53] DOI: 10.1198/016214507000000608 · Zbl 1333.62295 · doi:10.1198/016214507000000608
[54] DOI: 10.1080/01621459.1981.10477749 · doi:10.1080/01621459.1981.10477749
[55] Welch B. L., Biometrika 29 pp 21– (1937)
[56] Yanagawa T., Biometrika 71 pp 191– (1984) · Zbl 0532.62087 · doi:10.1093/biomet/71.1.191
[57] Yu B. B., Biostatistics 6 pp 201– (2005) · Zbl 1071.62112 · doi:10.1093/biostatistics/kxi003
[58] Zaykin D. V., Genetic Epidemiology 22 pp 170– (2002) · doi:10.1002/gepi.0042
[59] DOI: 10.1198/jasa.2011.ap10604 · Zbl 1232.62156 · doi:10.1198/jasa.2011.ap10604
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. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.