Instrumental variable estimation with a stochastic monotonicity assumption. (English) Zbl 1383.62290

Small, Dylan S.; Tan, Zhiqiang; Ramsahai, Roland R.; Lorch, Scott A.; Brookhart, M. Alan

Stat. Sci. 32, No. 4, 561-579 (2017).

Summary: The instrumental variables (IV) method provides a way to estimate the causal effect of a treatment when there are unmeasured confounding variables. The method requires a valid IV, a variable that is independent of the unmeasured confounding variables and is associated with the treatment but which has no effect on the outcome beyond its effect on the treatment. An additional assumption often made is deterministic monotonicity, which says that for each subject, the level of the treatment that a subject would take is a monotonic increasing function of the level of the IV. However, deterministic monotonicity is sometimes not realistic. We introduce a stochastic monotonicity assumption, a relaxation that only requires a monotonic increasing relationship to hold across subjects between the IV and the treatments conditionally on a set of (possibly unmeasured) covariates. We show that under stochastic monotonicity, the IV method identifies a weighted average of treatment effects with greater weight on subgroups of subjects on whom the IV has a stronger effect. We provide bounds on the global average treatment effect under stochastic monotonicity and a sensitivity analysis for violations of stochastic monotonicity. We apply the methods to a study of the effect of premature babies being delivered in a high technology neonatal intensive care unit (NICU) vs. a low technology unit.

MSC:

62P10	Applications of statistics to biology and medical sciences; meta analysis
62J15	Paired and multiple comparisons; multiple testing

Keywords:

causal inference; observational study; instrumental variable; two stage least squares

Software:

DOS

References:

[1]	Abadie, A. (2003). Semiparametric instrumental variable estimation of treatment response models.J. Econometrics113231-263. · Zbl 1038.62113
[2]	Angrist, J. D., Imbens, G. W. and Rubin, D. B. (1996). Identification of causal effects using instrumental variables.J. Amer. Statist. Assoc.91444-455. · Zbl 0897.62130
[3]	Angrist, J. D. and Pischke, J.-S. (2009).Mostly Harmless Econometrics:An Empiricist’s Companion. Princeton Univ. Press, Princeton, NJ. · Zbl 1159.62090
[4]	Bagnoli, M. and Bergstrom, T. (2005). Log-concave probability and its applications.Econom. Theory26445-469. · Zbl 1077.60012
[5]	Baiocchi, M., Small, D. S., Lorch, S. and Rosenbaum, P. R. (2010). Building a stronger instrument in an observational study of perinatal care for premature infants.J. Amer. Statist. Assoc.1051285-1296. · Zbl 1388.62311
[6]	Balke, A. and Pearl, J. (1997). Bounds on treatment effects for studies with imperfect compliance.J. Amer. Statist. Assoc.921171-1176. · Zbl 0888.62049
[7]	Boef, A. G. C., Gussekloo, J., Dekkers, O. M., Frey, P., Kearney, P. M., Kerse, N., Mallen, C. D., McCarthy, V. J. C., Mooijaart, S. P., Muth, C., Rodondi, N., Rosemann, T., Russell, A., Schers, H., Virgini, V., de Waal, M. W. M., Warner, A., le Cessie, S. and den Elzen, W. P. J. (2016). Physician’s prescribing preference as an instrumental variable: Exploring assumptions using survey data.Epidemiology27276-283.
[8]	Brookhart, M. A. and Schneeweiss, S. (2007). Preference-based instrumental variable methods for the estimation of treatment effects: Assessing validity and interpreting results.Int. J. Biostat.3Art. 14, 25.
[9]	Brookhart, M. A., Wang, P., Solomon, D. H. and Schneeweiss, S. (2006). Evaluating short-term drug effects using a physician-specific prescribing preference as an instrumental variable.Epidemiology17268-275.
[10]	Cavolina, M. J. F., Kelly, M. A. T., Stone, J. A. J. and Davis, R. G. M. (2000).Growing up Catholic:The Millennium Edition:An Infinitely Funny Guide for the Faithful,the Fallen and Everyone in-Between. Image.
[11]	Chalak, K. (2017). Instrumental variables methods with heterogeneity and mismeasured instruments.Econometric Theory3369-104. · Zbl 1441.62631
[12]	Cheng, J. and Small, D. S. (2006). Bounds on causal effects in three-arm trials with non-compliance.J. R. Stat. Soc. Ser. B Stat. Methodol.68815-836. · Zbl 1110.62152
[13]	Cox, D. R. (1958).Planning of Experiments. Wiley, New York. · Zbl 0084.15802
[14]	de Chaisemartin, C. (2017). Tolerating defiance: Local average treatment effects without monotonicity.Quant. Econ.8367-396. · Zbl 1396.62269
[15]	Deaton, A. (2010). Instruments, randomization and learning about development.J. Econ. Lit.48424-455.
[16]	Dijkstra, E. W. (1959). A note on two problems in connexion with graphs.Numer. Math.1269-271. · Zbl 0092.16002
[17]	DiNardo, J. and Lee, D. S. (2011). Program evaluation and research designs. InHandbook of Labor Economics,Vol.4 463-536. Elsevier, Amsterdam.
[18]	Finkelstein, A., Taubman, S., Wright, B., Bernstein, M., Gruber, J., Newhouse, J. P., Allen, H., Baicker, K. and Group Oregon Health Study (2012). The Oregon health insurance experiment: Evidence from the first year.Q. J. Econ.1271057-1106.
[19]	Guo, Z., Cheng, J., Lorch, S. A. and Small, D. S. (2014). Using an instrumental variable to test for unmeasured confounding.Stat. Med.333528-3546.
[20]	Hernán, M. A. and Robins, J. M. (2006). Instruments for causal inference: An epidemiologist’s dream?Epidemiology17360-372.
[21]	Horowitz, J. L. and Manski, C. F. (2000). Nonparametric analysis of randomized experiments with missing covariate and outcome data.J. Amer. Statist. Assoc.9577-88. · Zbl 0996.62054
[22]	Huber, M. and Mellace, G. (2012). Relaxing monotonicity in the identification of local average treatment effects. Working paper.
[23]	Imbens, G. W. (2014). Instrumental variables: An econometrician’s perspective.Statist. Sci.29323-358. · Zbl 1331.62471
[24]	Imbens, G. W. and Angrist, J. D. (1994). Identification and estimation of local average treatment effects.Econometrica61467-476. · Zbl 0800.90648
[25]	Klein, T. J. (2010). Heterogeneous treatment effects: Instrumental variables without monotonicity?J. Econometrics15599-116. · Zbl 1431.62641
[26]	Korn, E. L. and Baumrind, S. (1998). Clinician preferences and the estimation of causal treatment differences.Statist. Sci.13209-235. · Zbl 1099.62542
[27]	Lehmann, E. L. (1966). Some concepts of dependence.Ann. Math. Statist.371137-1153. · Zbl 0146.40601
[28]	Lorch, S. A., Baiocchi, M., Ahlberg, C. E. and Small, D. S. (2012a). The differential impact of delivery hospital on the outcomes of premature infants.Pediatrics130270-278.
[29]	Lorch, S. A., Kroelinger, C. D., Ahlberg, C. and Barfield, W. D. (2012b). Factors that mediate racial/ethnic disparities in US fetal death rates.Am. J. Publ. Health1021902-1910.
[30]	Manski, C. F. (1990). Non-parametric bounds on treatment effects.Am. Econ. Rev.80351-374.
[31]	McClellan, M., McNeil, B. J. and Newhouse, J. P. (1994). Does more intensive treatment of acute myocardial infarction in the elderly reduce mortality? Analysis using instrumental variables.J. Am. Med. Dir. Assoc.272859-866.
[32]	Neal, D. (1997). The effects of Catholic secondary schooling on educational achievement.J. Labor Econ.1498-123.
[33]	Ogburn, E. L., Rotnitzky, A. and Robins, J. M. (2015). Doubly robust estimation of the local average treatment effect curve.J. R. Stat. Soc. Ser. B. Stat. Methodol.77373-396. · Zbl 1414.62114
[34]	Pearl, J. (2011). Principal stratification—a goal or a tool?Int. J. Biostat.7Art. 20, 15.
[35]	Phibbs, C. S., Mark, D. H., Luft, H. S., Peltzman-Rennie, D. J., Garnick, D. W., Lichtenberg, E. and McPhee, S. J. (1993). Choice of hospital for delivery: A comparison of high-risk and low-risk women.Health Serv. Res.28201.
[36]	Ramsahai, R. R. (2012). Causal bounds and observable constraints for non-deterministic models.J. Mach. Learn. Res.13829-848. · Zbl 1283.68299
[37]	Robins, J. M. (1989). The analysis of randomized and nonrandomized AIDS treatment trials using a new approach to causal inference in longitudinal studies. InHealth Service Research Methodology:A Focus on AIDS(L. Sechrest, H. Freeman and A. Mulley, eds.) 113-159.
[38]	Robins, J. M. (1994). Correcting for non-compliance in randomized trials using structural nested mean models.Comm. Statist. Theory Methods232379-2412. · Zbl 0825.62203
[39]	Rosenbaum, P. R. (2010).Design of Observational Studies. Springer, New York. · Zbl 1308.62005
[40]	Rubin, D. B. (1986). Statistics and causal inference: Comment: Which ifs have causal answers.J. Amer. Statist. Assoc.81961-962.
[41]	Small, D. S., Tan, Z., Ramsahai, R. R., Lorch, S. A. and Brookhart, M. A. (2017). Supplement to “Instrumental variable estimation with a stochastic monotonicity assumption.”DOI:10.1214/17-STS623SUPP. · Zbl 1383.62290
[42]	Stock, J. H. (2001). Instrumental variables in economics and statistics. InInternational Encyclopedia of the Social & Behavioral Sciences(N. J. Smelser and P. B. Baltes, eds.) 7577-7582. Elsevier, Amsterdam.
[43]	Swanson, S. A. and Hernán, M. A. (2014). Think globally, act globally: An epidemiologist’s perspective on instrumental variable estimation [discussion of MR3264545].Statist. Sci.29371-374. · Zbl 1331.62487
[44]	Swanson, S. A., Miller, M., Robins, J. M. and Hernán, M. A. (2015). Definition and evaluation of the monotonicity condition for preference-based instruments.Epidemiology26414-420.
[45]	Tan, Z. (2006). Regression and weighting methods for causal inference using instrumental variables.J. Amer. Statist. Assoc.1011607-1618. · Zbl 1171.62364
[46]	Tan, Z. (2010). Marginal and nested structural models using instrumental variables.J. Amer. Statist. Assoc.105157-169. · Zbl 1397.62137
[47]	VanderWeele, T. J. and Shpitser, I. (2013). On the definition of a confounder.Ann. Statist.41196-220. · Zbl 1347.62017
[48]	Yang, F., Lorch, S. A. and Small, D. S. (2014). Estimation of causal effects using instrumental variables with nonignorable missing covariates: Application to effect of type of delivery NICU on premature infants.Ann. Appl. Stat.848-73. · Zbl 1454.62422