1932
0
  1. Home
  2. A-Z Publications
  3. Annual Review of Statistics and Its Application
  4. Volume 7, 2020
  5. Article

Annual Review of Statistics and Its Application

Volume 7, 2020

Modern Algorithms for Matching in Observational Studies

Abstract

Using a small example as an illustration, this article reviews multivariate matching from the perspective of a working scientist who wishes to make effective use of available methods. The several goals of multivariate matching are discussed. Matching tools are reviewed, including propensity scores, covariate distances, fine balance, and related methods such as near-fine and refined balance, exact and near-exact matching, tactics addressing missing covariate values, the entire number, and checks of covariate balance. Matching structures are described, such as matching with a variable number of controls, full matching, subset matching and risk-set matching. Software packages in R are described. A brief review is given of the theory underlying propensity scores and the associated sensitivity analysis concerning an unobserved covariate omitted from the propensity score.

Keyword(s): assignment algorithm, causal inference, design sensitivity, fine balance, integer programming, Mahalanobis distance, near-fine balance, network optimization, optimal matching, principal unobserved covariate, propensity score, refined balance, sensitivity analysis
Loading

Article metrics loading...

/content/journals/10.1146/annurev-statistics-031219-041058
2020-03-07
2024-04-19
Loading full text...

Full text loading...

/deliver/fulltext/statistics/7/1/annurev-statistics-031219-041058.html?itemId=/content/journals/10.1146/annurev-statistics-031219-041058&mimeType=html&fmt=ahah

Literature Cited

  1. Angrist JD, Krueger AB 1999. Empirical strategies in labor economics. Handbook of Labor Economics, Vol. 3 O Ashenfelter, D Card1237–366 Amsterdam: Elsevier
  2. Apel RJ, Sweeten G. 2010. Propensity score matching in criminology and criminal justice. In Handbook of Quantitative Criminologyed AR Piquero, D Weisburd543–62 New York: Springer
     [Google Scholar]
  3. Austin PC, Stuart EA. 2015. Optimal full matching for survival outcomes: a method that merits more widespread use. Stat. Med. 34:3949–67
     [Google Scholar]
  4. Baiocchi M, Cheng J, Small DS. 2014. Instrumental variable methods for causal inference. Stat. Med. 33:2297–340
     [Google Scholar]
  5. Baiocchi M, Small DS, Lorch S, Rosenbaum PR. 2010. Building a stronger instrument in an observational study of perinatal care for premature infants. J. Am. Stat. Assoc. 105:1285–96
     [Google Scholar]
  6. Baiocchi M, Small DS, Yang L, Polsky D, Groeneveld PW. 2012. Near/far matching: a study design approach to instrumental variables. Health Serv. Outcomes Res. Method. 12:237–53
     [Google Scholar]
  7. Bang-Jensen J, Gutin G 2009. Digraphs: Theory, Algorithms and Applications New York: Springer
  8. Bertsekas DP. 1981. A new algorithm for the assignment problem. Math. Prog. 21:152–71
     [Google Scholar]
  9. Bertsekas DP. 1990. The auction algorithm for assignment and other network flow problems: a tutorial. Interfaces 20:133–49
     [Google Scholar]
  10. Bertsekas DP 1998. Network Optimization Belmont, MA: Athena Sci.
  11. Bertsekas DP. 2001. Auction algorithms. Encyclopedia of Optimizationed CA Floudas, PM Pardalos73–77 New York: Springer
     [Google Scholar]
  12. Bertsekas DP, Tseng P. 1988. The relax codes for linear minimum cost network flow problems. Ann. Oper. Res. 13:125–90
     [Google Scholar]
  13. Bertsimas D, Tsitsiklis JN 1997. Introduction to Linear Optimization Belmont, MA: Athena Sci.
  14. Brooks JM, Ohsfeldt RL. 2013. Squeezing the balloon: propensity scores and unmeasured covariate balance. Health Serv. Res. 48:3078–94
     [Google Scholar]
  15. Burkard R, Dellamico M, Martello S 2009. Assignment Problems Philadelphia: SIAM
  16. Card D, Chetty R, Weber A 2007. The spike at benefit exhaustion: leaving the unemployment system or starting a new job. Am. Econ. Rev. 97113–18
  17. Cochran WG. 1965. The planning of observational studies of human populations. J. R. Stat. Soc. A 128:234–66
     [Google Scholar]
  18. Cochran WG, Rubin DB. 1973. Controlling bias in observational studies: a review. Sankhyā 35:417–46
     [Google Scholar]
  19. Crama Y, Spieksma FCR. 1992. Approximation algorithms for three-dimensional assignment problems with triangle inequalities. Eur. J. Oper. Res. 60:273–79
     [Google Scholar]
  20. Crump RK, Hotz J, Imbens GW, Mitnik OA. 2009. Dealing with limited overlap in estimation of average treatment effects. Biometrika 96:187–99
     [Google Scholar]
  21. Daniel RM, Cousens S, De Stavola B, Kenward M, Sterne J. 2013. Methods for dealing with time-dependent confounding. Stat. Med. 32:1584–618
     [Google Scholar]
  22. Daniel SR, Armstrong K, Silber JH, Rosenbaum PR. 2008. An algorithm for optimal tapered matching, with application to disparities in survival. J. Comput. Graph. Stat. 17:914–24
     [Google Scholar]
  23. Dawid AP. 1979. Conditional independence in statistical theory. J. R. Stat. Soc. B 41:1–31
     [Google Scholar]
  24. Derigs U. 1988. Solving non-bipartite matching problems via shortest path techniques. Ann. Oper. Res. 13:225–61
     [Google Scholar]
  25. Ertefaie A, Small DS, Rosenbaum PR. 2018. Quantitative evaluation of the trade-off of strengthened instruments and sample size in observational studies. J. Am. Stat. Assoc. 113:1122–34
     [Google Scholar]
  26. Feuer AJ, Demmer RT, Thai A, Vogiatzi MG. 2015. Use of selective serotonin reuptake inhibitors and bone mass in adolescents. Bone 78:28–33
     [Google Scholar]
  27. Fisher RA 1935. Design of Experiments Edinburgh: Oliver and Boyd
  28. Fogarty CB, Mikkelsen ME, Gaieski DF, Small DS. 2016. Discrete optimization for interpretable study populations and randomization inference in an observational study of severe sepsis mortality. J. Am. Stat. Assoc. 111:447–58
     [Google Scholar]
  29. Frangakis CE, Rubin DB. 2002. Principal stratification in causal inference. Biometrics 58:21–29
     [Google Scholar]
  30. Gelman A, Imbens G 2019. Why high-order polynomials should not be used in regression discontinuity designs. J. Bus. Econ. Stat. 37:447–56
     [Google Scholar]
  31. Glover F. 1967. Maximum matching in a convex bipartite graph. Naval Res. Logist. 14:313–16
     [Google Scholar]
  32. Greevy R, Lu B, Silber JH, Rosenbaum PR. 2004. Optimal multivariate matching before randomization. Biostatistics 5:263–75
     [Google Scholar]
  33. Gross DB, Souleles NS. 2002. Do liquidity constraints and interest rates matter for consumer behavior? Evidence from credit card data. Q. J. Econ. 117:149–85
     [Google Scholar]
  34. Hansen BB. 2004. Full matching in an observational study of coaching for the SAT. J. Am. Stat. Assoc. 99:609–18
     [Google Scholar]
  35. Hansen BB. 2007. Flexible, optimal matching for observational studies. R News 7:18–24
     [Google Scholar]
  36. Hansen BB, Klopfer SO. 2006. Optimal full matching and related designs via network flows. J. Comput. Graph. Stat. 15:609–27
     [Google Scholar]
  37. Hansen BB, Rosenbaum PR, Small DS. 2014. Clustered treatment assignments and sensitivity to unmeasured biases in observational studies. J. Am. Stat. Assoc. 109:133–44
     [Google Scholar]
  38. Haviland A, Nagin DS, Rosenbaum PR, Tremblay RE. 2008. Combining group-based trajectory modeling and propensity score matching for causal inferences in nonexperimental longitudinal data. Dev. Psychol. 44:422–36
     [Google Scholar]
  39. Heller R, Rosenbaum PR, Small DS. 2010. Using the cross-match test to appraise covariate balance in matched pairs. Am. Stat. 64:2990–309
     [Google Scholar]
  40. Hsu JY, Zubizarreta JR, Small DS, Rosenbaum PR. 2015. Strong control of the familywise error rate in observational studies that discover effect modification by exploratory methods. Biometrika 102:767–82
     [Google Scholar]
  41. Huber P 1981. Robust Statistics New York: Wiley
  42. Karmakar B, Small DS, Rosenbaum PR 2019. Using approximation algorithms to build evidence factors and related designs for observational studies. J. Comput. Graph. Stat 28:698–709
     [Google Scholar]
  43. Keele L, Morgan JW. 2016. Strengthening instruments through matching and weak instrument tests. Ann. Appl. Stat. 10:1086–106
     [Google Scholar]
  44. Keele L, Titiunik R, Zubizarreta JR. 2015. Enhancing a geographic regression discontinuity design through matching to estimate the effect of ballot initiatives on voter turnout. J. R. Stat. Soc. A 178:223–39
     [Google Scholar]
  45. Korte B, Vygen J 2012. Combinatorial Optimization New York: Springer
  46. Lalive R, Van Ours J, Zweimuller J. 2015. How changes in financial incentives affect the duration of unemployment. Rev. Econ. Stud. 73:1009–38
     [Google Scholar]
  47. Lee K, Small DS, Rosenbaum PR. 2018. A powerful approach to the study of moderate effect modification in observational studies. Biometrics 74:1161–70
     [Google Scholar]
  48. Li YP, Propert KJ, Rosenbaum PR. 2001. Balanced risk set matching. J. Am. Stat. Assoc. 96:870–82
     [Google Scholar]
  49. Lu B. 2005. Propensity score matching with time-dependent covariates. Biometrics 61:721–28
     [Google Scholar]
  50. Lu B, Greevy R, Xu X, Beck C. 2011. Optimal nonbipartite matching and its statistical applications. Am. Stat. 65:21–30
     [Google Scholar]
  51. Lu B, Rosenbaum PR. 2004. Optimal pair matching with two control groups. J. Comput. Graph. Stat. 13:422–34
     [Google Scholar]
  52. Maritz JS. 1979. A note on exact robust confidence intervals for location. Biometrika 66:163–70
     [Google Scholar]
  53. Meyer BD. 1995. Natural and quasi-experiments in economics. J. Bus. Econ. Stat. 13:151–61
     [Google Scholar]
  54. Ming K, Rosenbaum PR. 2000. Substantial gains in bias reduction from matching with a variable number of controls. Biometrics 56:118–24
     [Google Scholar]
  55. Neyman J. 1990. On the application of probability theory to agricultural experiments. Stat. Sci. 5:465–72
     [Google Scholar]
  56. Nieuwbeerta P, Nagin DS, Blokland AAJ. 2009. Assessing the impact of first-time imprisonment on offenders subsequent criminal career development: a matched samples comparison. J. Quant. Criminol. 25:227–57
     [Google Scholar]
  57. Pimentel SD. 2016. Large, sparse optimal matching with R package rcbalance. Obs. Stud. 2:4–23
     [Google Scholar]
  58. Pimentel SD, Kelz RR, Silber JH, Rosenbaum PR. 2015a. Large, sparse optimal matching with refined covariate balance in an observational study of the health outcomes produced by new surgeons. J. Am. Stat. Assoc. 110:515–27
     [Google Scholar]
  59. Pimentel SD, Small DS, Rosenbaum PR. 2016. Constructed second control groups and attenuation of unmeasured biases. J. Am. Stat. Assoc. 111:1157–67
     [Google Scholar]
  60. Pimentel SD, Small DS, Rosenbaum PR. 2017. An exact test of fit for the Gaussian linear model using optimal nonbipartite matching. Technometrics 59:330–37
     [Google Scholar]
  61. Pimentel SD, Yoon F, Keele L. 2015b. Variable-ratio matching with fine balance in a study of the peer health exchange. Stat. Med. 34:4070–82
     [Google Scholar]
  62. Rosenbaum PR. 1984. Conditional permutation tests and the propensity score in observational studies. J. Am. Stat. Assoc. 79:565–74
     [Google Scholar]
  63. Rosenbaum PR. 1987a. Model-based direct adjustment. J. Am. Stat. Assoc. 82:387–94
     [Google Scholar]
  64. Rosenbaum PR. 1987b. Sensitivity analysis for certain permutation inferences in matched observational studies. Biometrika 74:13–26
     [Google Scholar]
  65. Rosenbaum PR. 1989. Optimal matching for observational studies. J. Am. Stat. Assoc. 84:1024–32
     [Google Scholar]
  66. Rosenbaum PR. 1991a. A characterization of optimal designs for observational studies. J. R. Stat. Soc. B 53:597–610
     [Google Scholar]
  67. Rosenbaum PR. 1991b. Discussing hidden bias in observational studies. Ann. Intern. Med. 115:901–5
     [Google Scholar]
  68. Rosenbaum PR 2002. Observational Studies New York: Springer
  69. Rosenbaum PR. 2004. Design sensitivity in observational studies. Biometrika 91:153–64
     [Google Scholar]
  70. Rosenbaum PR. 2005. Heterogeneity and causality: unit heterogeneity and design sensitivity in observational studies. Am. Stat. 59:147–52
     [Google Scholar]
  71. Rosenbaum PR. 2007. Sensitivity analysis for M-estimates, tests, and confidence intervals in matched observational studies. Biometrics 63:456–64
     [Google Scholar]
  72. Rosenbaum PR 2010. Design of Observational Studies New York: Springer
  73. Rosenbaum PR. 2012. Optimal matching of an optimally chosen subset in observational studies. J. Comput. Graph. Stat. 21:57–71
     [Google Scholar]
  74. Rosenbaum PR. 2013. Impact of multiple matched controls on design sensitivity in observational studies. Biometrics 69:118–27
     [Google Scholar]
  75. Rosenbaum PR. 2015a. How to see more in observational studies: some new quasi-experimental devices. Annu. Rev. Stat. Appl. 2:21–48
     [Google Scholar]
  76. Rosenbaum PR. 2015b. Two R packages for sensitivity analysis in observational studies. Obs. Stud. 1:1–17
     [Google Scholar]
  77. Rosenbaum PR. 2017a. Imposing minimax and quantile constraints on optimal matching in observational studies. J. Comput. Graph. Stat. 26:66–78
     [Google Scholar]
  78. Rosenbaum PR 2017b. Observation and Experiment Cambridge, MA: Harvard Univ. Press
  79. Rosenbaum PR, Ross RN, Silber JH. 2007. Minimum distance matched sampling with fine balance in an observational study of treatment for ovarian cancer. J. Am. Stat. Assoc. 102:75–83
     [Google Scholar]
  80. Rosenbaum PR, Rubin DB. 1983. The central role of the propensity score in observational studies for causal effects. Biometrika 70:41–55
     [Google Scholar]
  81. Rosenbaum PR, Rubin DB. 1984. Reducing bias in observational studies using subclassification on the propensity score. J. Am. Stat. Assoc. 79:516–24
     [Google Scholar]
  82. Rosenbaum PR, Rubin DB. 1985a. Constructing a control group using multivariate matched sampling methods that incorporate the propensity score. Am. Stat. 39:33–38
     [Google Scholar]
  83. Rosenbaum PR, Rubin DB. 1985b. The bias due to incomplete matching. Biometrics 41:103–16
     [Google Scholar]
  84. Rosenbaum PR, Silber JH. 2009. Amplification of sensitivity analysis in matched observational studies. J. Am. Stat. Assoc. 104:1398–405
     [Google Scholar]
  85. Rosenbaum PR, Silber JH. 2013. Using the exterior match to compare two entwined matched control groups. Am. Stat. 67:67–75
     [Google Scholar]
  86. Rubin DB. 1974. Estimating causal effects of treatments in randomized and nonrandomized studies. J. Ed. Psych. 66:688–701
     [Google Scholar]
  87. Rubin DB. 1979. Using multivariate matched sampling and regression adjustment to control bias in observational studies. J. Am. Stat. Assoc. 74:318–28
     [Google Scholar]
  88. Rubin DB. 2007. The design versus the analysis of observational studies for causal effects: parallels with the design of randomized trials. Stat. Med. 26:20–36
     [Google Scholar]
  89. Sanni AM, Groenwold RHH, Klungel OH. 2014. Propensity score methods and unobserved covariate balance. Health Serv. Res. 49:1074–82
     [Google Scholar]
  90. Sekhon JS. 2009. Opiates for the matches: matching methods for causal inference. Annu. Rev. Political Sci. 12:487–508
     [Google Scholar]
  91. Silber JH, Rosenbaum PR, McHugh MD, Ludwig JM, Smith HL 2016. Comparison of the value of nursing work environments in hospitals across different levels of patient risk. JAMA Surg. 151:527–36
     [Google Scholar]
  92. Silber JH, Rosenbaum PR, Ross RN, Ludwig JM, Wang Wet al. 2014. Template matching for auditing hospital cost and quality. Health Serv. Res. 49:1446–74
     [Google Scholar]
  93. Stuart EA. 2010. Matching methods for causal inference. Stat. Sci. 25:1–21
     [Google Scholar]
  94. Stuart EA, Rubin DB. 2008. Matching with multiple control groups with adjustment for group differences. J. Educ. Behav. Stat. 33:279–306
     [Google Scholar]
  95. Traskin M, Small DS. 2011. Defining the study population for an observational study to ensure sufficient overlap: a tree approach. Stat. Biosci. 3:94–118
     [Google Scholar]
  96. Tukey JW. 1980. We need both exploratory and confirmatory. Am. Stat. 34:23–25
     [Google Scholar]
  97. Vandenbroucke JP 2004. When are observational studies as credible as randomised trials. Lancet 3631728–31
  98. Vazirani VV 2010. Approximation Algorithms New York: Springer
  99. Walker AM. 2013. Matching on provider is risky. J. Clin. Epidemiol. 66:565–68
     [Google Scholar]
  100. Williamson DP, Shmoys DB 2011. Design of Approximation Algorithms Cambridge, UK: Cambridge Univ. Press
  101. Wolsey LA 1998. Integer Programming New York: Wiley
  102. Wu CF, Hamada MS 2011. Experiments: Planning, Analysis, and Optimization New York: Wiley
  103. Yang D, Small DS, Silber JH, Rosenbaum PR. 2012. Optimal matching with minimal deviation from fine balance in a study of obesity and surgical outcomes. Biometrics 68:628–36
     [Google Scholar]
  104. Yu R, Silber JH, Rosenbaum PR 2019. Matching methods for observational studies derived from large administrative databases. Stat. Sci. In press
  105. Zubizarreta JR. 2012. Using mixed integer programming for matching in an observational study of kidney failure after surgery. J. Am. Stat. Assoc. 107:1360–71
     [Google Scholar]
  106. Zubizarreta JR, Cerdá M, Rosenbaum PR. 2013. Effect of the 2010 Chilean earthquake on posttraumatic stress: reducing sensitivity to unmeasured bias through study design. Epidemiology 24:79–87
     [Google Scholar]
  107. Zubizarreta JR, Keele L. 2017. Optimal multilevel matching in clustered observational studies: a case study of the effectiveness of private schools under a large-scale voucher system. J. Am. Stat. Assoc. 112:547–60
     [Google Scholar]
  108. Zubizarreta JR, Neuman M, Silber JH, Rosenbaum PR. 2012. Contrasting evidence within and between institutions that provide treatment in an observational study of alternate forms of anesthesia. J. Am. Stat. Assoc. 107:901–15
     [Google Scholar]
  109. Zubizarreta JR, Paredes RD, Rosenbaum PR. 2014a. Matching for balance, pairing for heterogeneity in an observational study of the effectiveness of for-profit and not-for-profit high schools in Chile. Ann. Appl. Stat. 8:204–31
     [Google Scholar]
  110. Zubizarreta JR, Reinke CE, Kelz RR, Silber JH, Rosenbaum PR. 2011. Matching for several nominal variables in a case-control study of readmission following surgery. Am. Stat. 65:229–38
     [Google Scholar]
  111. Zubizarreta JR, Small DS, Rosenbaum PR. 2014b. Isolation in the construction of natural experiments. Ann. Appl. Stat. 8:2096–121
     [Google Scholar]
/content/journals/10.1146/annurev-statistics-031219-041058
Loading
Modern Algorithms for Matching in Observational Studies
Annual Review of Statistics and Its Application 7, 143 (2020); https://doi.org/10.1146/annurev-statistics-031219-041058
/content/journals/10.1146/annurev-statistics-031219-041058
/content/journals/10.1146/annurev-statistics-031219-041058
Loading

Data & Media loading...

  • Article Type: Review Article

Most Cited Most Cited RSS feed

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error