Unifying instrumental variable and inverse probability weighting approaches for inference of causal treatment effect and unmeasured confounding in observational studies
Restricted accessResearch articleFirst published online March, 2021
Unifying instrumental variable and inverse probability weighting approaches for inference of causal treatment effect and unmeasured confounding in observational studies
Confounding is a major concern when using data from observational studies to infer the causal effect of a treatment. Instrumental variables, when available, have been used to construct bound estimates on population average treatment effects when outcomes are binary and unmeasured confounding exists. With continuous outcomes, meaningful bounds are more challenging to obtain because the domain of the outcome is unrestricted. In this paper, we propose to unify the instrumental variable and inverse probability weighting methods, together with suitable assumptions in the context of an observational study, to construct meaningful bounds on causal treatment effects. The contextual assumptions are imposed in terms of the potential outcomes that are partially identified by data. The inverse probability weighting component incorporates a sensitivity parameter to encode the effect of unmeasured confounding. The instrumental variable and inverse probability weighting methods are unified using the principal stratification. By solving the resulting system of estimating equations, we are able to quantify both the causal treatment effect and the sensitivity parameter (i.e. the degree of the unmeasured confounding). We demonstrate our method by analyzing data from the HIV Epidemiology Research Study.
RosenbaumPR.Observational studies.
New York, NY:
Springer, 2002.
2.
VanderWeeleTJShpitserI.On the definition of a confounder.Ann Stat2013;
41: 196–220.
3.
RosenbaumPRRubinDB.Reducing bias in observational studies using subclassification on the propensity score. J Am Stat Assoc1984;
79: 516–524.
4.
RobinsJMRotnitzkyAZhaoLP.Estimation of regression coefficients when some regressors are not always observed. J Am Stat Assoc1994;
89: 846–866.
5.
D’AgostinoRB.Tutorial in biostatistics: propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med1998;
17: 2265–2281.
6.
RobinsJMHernánMABrumbackB.Marginal structural models and causal inference in epidemiology. Epidemiology2000;
11: 550–560.
7.
HoganJWLancasterT.Instrumental variables and inverse probability weighting for causal inference from longitudinal observational studies.Stat Meth Med Res2004;
13: 17–48.
8.
BangHRobinsJM.Doubly robust estimation in missing data and causal inference models.Biometrics2005;
61: 962–972.
9.
ColeSRHernánMAMargolickJB, et al.Marginal structural models for estimating the effect of highly active antiretroviral therapy initiation on CD4 cell count.Am J Epidemiol2005;
162: 471–478.
10.
KangJDYSchaferJL.Demystifying double robustness: a comparison of alternative strategies for estimating a population mean from incomplete data. Stat Sci2007;
22: 523–539.
11.
StuartEA.Matching methods for causal inference: a review and a look forward.Stat Sci2010;
25: 1–21.
12.
WrightS.Appendix to the tariff on animal and vegetable oils. vol.
26.
New York, NY:
Macmillan; 1928.
13.
StockJHTrebbiF.Retrospectives: who invented instrumental variable regression?J Econ Perspect2003;
17: 177–194.
14.
AngristJDImbensGW.Two-stage least squares estimation of average causal effects in models with variable treatment intensity. J Am Stat Assoc1995;
90: 431–442.
15.
ImbensGWAngristJD.Identification and estimation of local average treatment effects. Econometrica1994;
62: 467–475.
16.
AngristJDImbensGWRubinDB.Identification of causal effects using instrumental variables. J Am Stat Assoc1996;
91: 444–455.
17.
RobinsJMGreenlandS.Identification of causal effects using instrumental variables: comment. J Am Stat Assoc1996;
91: 456–458.
18.
Robins JM. The Analysis of Randomized and Non-Randomized AIDS Treatment Trials Using a New Approach to Causal Inference in Longitudinal Studies. In Sechrest L, Freeman H and Bailey A (eds) Health Service Research Methodology: A Focus on AIDS. Washington, DC: U. S. Public Health Service, 1898, pp. 113–159.
19.
ManskiCF.Nonparametric bounds on treatment effects. Am Econ Rev1990;
80: 319–323.
20.
BalkeAPearlJ.Bounds on treatment effects from studies with imperfect compliance. J Am Stat Assoc1997;
92: 1171–1176.
21.
JoffeMM.Using information on realized effects to determine prospective causal effects. J R Stat Soc Ser B2001;
63: 759–774.
22.
ChengJSmallDS.Bounds on causal effects in three-arm trials with non-compliance. J R Stat Soc Ser B2006;
68: 815–836.
23.
ZhangJLRubinDB.Estimation of causal effects via principal stratification when some outcomes are truncated by death. J Edu Behav Stat2003;
28: 353–368.
24.
SmithDKWarrenDVlahovP, et al.
Design and baseline participant characteristics of the Human Immunodeficiency Virus Epidemiology Research (HER) Study: a prospective cohort study of human immunodeficiency virus infection in US women.Am J Epidemiol1997;
146: 459–469.
25.
KoHHoganJWMayerKH.Estimating causal treatment effects from longitudinal {HIV} natural history studies using marginal structural models.Biometrics2003;
59: 152–162.
26.
WalkerAM.Confounding by indication.Epidemiology1996;
7: 335–336.
27.
BrookhartMASchneeweissS.Preference-based instrumental variable methods for the estimation of treatment effects: assessing validity and interpreting results. Int J Biostat2007;
3: 14.
28.
RobinsJMRotnitzkyAScharfsteinDO.Sensitivity analysis for selection bias and unmeasured confounding in missing data and causal inference models.Stat Model Epidemiol: Environ Clin Trials1999;
116: 1–92.
29.
BrumbackBAHernánMAHaneuseSJPA, et al.Sensitivity analyses for unmeasured confounding assuming a marginal structural model for repeated measures.Stat Med2004;
23: 749–767.
30.
FrangakisCERubinDB.Principal stratification in causal inference.Biometrics2002;
58: 21–29.
31.
CarpenterCCJCooperDAFischlJM, et al.
Antiretroviral therapy in adults:updated recommendations of the International AIDS Society-USA Panel.J Am Med Assoc2000;
283: 381–391.
32.
Ten HaveTRNormandSLTMarcusSM, et al.Intent-to-treat vs. non-intent-to-treat analyses under treatment non-adherence in mental health randomized trials.Psych Ann2008;
38: 772.
33.
JohnstonSC.Combining ecological and individual variables to reduce confounding by indication. J Clin Epidemiol2000;
53: 1236–1241.
34.
BrookhartMAWangPSSolomonDH, et al.Evaluating short-term drug effects using a physician-specific prescribing preference as an instrumental variable. Epidemiology2006;
17: 268–275.
35.
NeymanJ.On the application of probability theory to agricultural experiments: essay on principles. Stat Sci1923;
1990: 465–472.
36.
RubinDB.Estimating causal effects of treatments in randomized and nonrandomized studies. J Educ Psychol1974;
66: 668–701.
37.
GangeSJKitahataMMSaagMS, et al.
Cohort profile: The North American AIDS Cohort Collaboration on Research and Design (NA-ACCORD).Int J Epidemiol2007;
36: 294–301.
38.
RobinsJM.Association, causation, and marginal structural models. Synthese1999;
121: 151–179.
39.
RosenbaumPRRubinDB.The central role of the propensity score in observational studies for causal effects. Biometrika1983;
70: 41–55.
40.
MiguelHABabetteBRobinsJM.Marginal structural models to estimate the joint causal effect of nonrandomized treatments. J Am Stat Assoc2001;
96: 440–448.
41.
WooldrigeJM.Econometric analysis of cross section and panel data.
Cambridge, MA:
MIT Press, 2002.
42.
HernanMARobinsJM.Instruments for causal inference: an epidemiologist’s dream?Epidemiology2006;
17: 360–372.
43.
DavidsonRMacKinnonJ.Estimation and inference in econometrics.
Oxford, UK:
Oxford University Press, 1993.
44.
StockJH.Instrumental variables in statistics and econometrics.
Amsterdam:
Elsevier, 2001.
45.
FreedmanD.Statistical models: theory and practice.
Cambridge, UK:
Cambridge University Press, 2009.
46.
BhattacharyaJShaikhAMVytlacilE.Treatment effect bounds under monotonicity assumptions: an application to Swan-Ganz Catheterization. Am Econ Review2008;
98: 351–356.
47.
VansteelandtSBowdenJBabanezhadM, et al.On instrumental variables estimation of causal odds ratios. Stat Sci2011;
26: 403–422.
48.
SiddiqueZ. Partially identified treatment effects under imperfect compliance: the case of domestic violence. J Am Stat Assoc2014; 108: 504–513.
49.
VansteelandtSGoetghebeurandEKenwardMG, et al.Ignorance and uncertainty regions as inferential tools in a sensitivity analysis. Stat Sinica2006;
16: 953–979.
50.
HorowitzJLManskiCF.Nonparametric analysis of randomized experiments with missing covariate and outcome data. J Am Stat Assoc2000;
95: 77–84.
51.
BickelPJFreemanDA.Some asymptotic theory for the bootstrap. Ann Stat1981;
9: 1196–1217.
52.
DiNardoJLeeDS.Program evaluation and research designs. In: Ashenfelter O and Card D (eds)Handbook of labor economics. vol.
4. Amsterdam, the Netherlands:
Elsevier, 2011, pp.463–536.
53.
de ChaisemartinC.Tolerating defiance? Local average treatment effects without monotonicity. Quantitative Economics2017;
8: 367–396.
54.
RoyJHoganJWMarcusBH.Principal stratification with predictors of compliance for randomized trials with 2 active treatments.Biostatistics2008;
9: 277–289.
55.
SmallDSTanZRamsahaiRR, et al.
Instrumental variable estimation with a stochastic monotonicity assumption. Stat Sci2017;
32: 561–579.
56.
HiranoKImbensGWRubinDB, et al.
Assessing the effect of an influenza vaccine in an encouragement design.Biostatistics (Oxford, England)2000;
1: 69–88.
