Combining directed acyclic graphs and the change-in-estimate procedure as a novel approach to adjustment-variable selection in epidemiology

BMC Medical Research Methodology

Volumn 12, Issue , 2012, Pages

  Evans, David ^a,b,c,d   Chaix, Basile ^a,c   Lobbedez, Thierry ^d,e   Verger, Christian ^d   Flahault, Antoine ^a,b  

^a INSERM   (France)

^b EHESP School of Public Health   (France)

^c SORBONNE UNIVERSITÉ   (France)

^d Registre de Dialyse Peritoneale de Langue Française   (France)

^e CHU Clemenceau   (France)

Author keywords

Adjustment variable selection; Change in estimate; Directed acyclic graph; Peritoneal dialysis

Indexed keywords

ARTICLE; COMPUTER GRAPHICS; EPIDEMIOLOGY; HUMAN; MULTIVARIATE ANALYSIS; REGRESSION ANALYSIS; REPRODUCIBILITY; STATISTICAL ANALYSIS; STATISTICAL MODEL;

BIAS (EPIDEMIOLOGY); CAUSALITY; COMPUTER GRAPHICS; CONFOUNDING FACTORS (EPIDEMIOLOGY); DATA INTERPRETATION, STATISTICAL; EPIDEMIOLOGIC RESEARCH DESIGN; HUMANS; MODELS, STATISTICAL; MULTIVARIATE ANALYSIS; REGRESSION ANALYSIS; REPRODUCIBILITY OF RESULTS; SELECTION BIAS;

MLCS; MLOWN;

EID: 84867236463     PISSN: None     EISSN: 14712288     Source Type: Journal    
DOI: 10.1186/1471-2288-12-156     Document Type: Review

References (59)

1. Causal diagrams for epidemiologic research. Greenland S, Pearl J, Robins JM, Epidemiology 1999 10 37 48 10.1097/00001648-199901000-00008 9888278 (Pubitemid 29014427)
2. Causal diagrams. Glymour M, Greenland S, Modern epidemiology Philadelphia, PA: Lippincott Williams &Wilkins 3rd 2008 183 209
3. Pearl J, Causality: models, reasoning, and inference Cambridge: Cambridge University Press 2nd 2009
4. Modeling and variable selection in epidemiologic analysis. Greenland S, Am J Public Health 1989 79 340 349 10.2105/AJPH.79.3.340 2916724 (Pubitemid 19077878)
5. On model selection and model misspecification in causal inference. Vansteelandt S, Bekaert M, Claeskens G, Stat Methods Med Res 2012 21 7 30 10.1177/0962280210387717 21075803
6. Methods of covariate selection: directed acyclic graphs and the change-in-estimate procedure. Weng HY, Hsueh YH, Messam LLM, Hertz-Picciotto I, Am J Epidemiol 2009 169 1182 1190 10.1093/aje/kwp035 19363102
7. Spirtes P, Glymour C, Scheines R, Causation, prediction, and search, second edition Cambridge: The MIT Press 2nd 2001
8. Rejoinder to glymour and spirtes. Computation, causation, and discovery Cambridge MA: AAAI Press/The MIT Press, Glymour C, Cooper G, 1999 333 342
9. Management practices and risk of occupational blood exposure in U.S. Paramedics: Non-intact skin exposure. Leiss JK, Ann Epidemiol 2009 19 884 890 10.1016/j.annepidem.2009.08.006 19944350
10. Prevalence of and risk factors for anal human papillomavirus infection in Men Who have Sex with women: a cross national study. Nyitray AG, Smith D, Villa L, Lazcano Ponce E, Abrahamsen M, Papenfuss M, Giuliano AR, J Infect Dis 2010 201 1498 1508 10.1086/652187 20367457
11. Sleep disturbances and cause-specific mortality: results from the GAZEL cohort study. Rod NH, Vahtera J, Westerlund H, Kivimaki M, Zins M, Goldberg M, Lange T, Am J Epidemiol 2010 173 300 309 21193534
12. The effect of highly active antiretroviral therapy on the survival of HIV-infected children in a resource-deprived setting: a cohort study. Edmonds A, Yotebieng M, Lusiama J, Matumona Y, Kitetele F, Napravnik S, Cole SR, Van Rie A, Behets F, PLoS Med 2011 8 1001044 10.1371/journal.pmed.1001044 21695087
13. Assessing perceived risk and STI prevention behavior: a national population-based study with special reference to HPV. Leval A, Sundström K, Ploner A, Arnheim Dahlström L, Widmark C, Sparén P, PLoS One 2011 6 20624 10.1371/journal.pone.0020624 21674050
14. Whole grains Are associated with serum concentrations of high sensitivity C-reactive protein among premenopausal women. Gaskins AJ, Mumford SL, Rovner AJ, Zhang C, Chen L, Wactawski-Wende J, Perkins NJ, Schisterman EF, for the BioCycle Study Group, J Nutr 2010 140 1669 1676 10.3945/jn.110.124164 20668255
15. Effect of daily fiber intake on reproductive function: the BioCycle study. Gaskins AJ, Mumford SL, Zhang CL, Wactawski-Wende J, Hovey KM, Whitcomb BW, Howards PP, Perkins NJ, Yeung E, Schisterman EF, Am J Clin Nutr 2009 90 1061 1069 10.3945/ajcn.2009.27990 19692496
16. Confounding equivalence in observational studies (or, when are two measurements equally valuable for effect estimation?). Pearl J, Paz A, Proceedings of the twenty-sixth conference on uncertainty in artificial intelligence Corvallis: AUAI 2010 433 441
17. Adjustments and their consequences-Collapsibility analysis using graphical models. Greenland S, Pearl J, Int Stat Rev 2011 79 401 426 10.1111/j.1751-5823.2011.00158.x
18. Reducing bias through directed acyclic graphs. Shrier I, Platt RW, BMC Med Res Methodol 2008 8 70 10.1186/1471-2288-8-70 18973665
19. Using directed acyclic graphs to guide analyses of neighbourhood health effects: an introduction. Fleischer NL, Diez Roux AV, J Epidemiol Community Health 2008 62 842 846 10.1136/jech.2007.067371 18701738
20. Using directed acyclic graphs to consider adjustment for socioeconomic status in occupational cancer studies. Richiardi L, Barone-Adesi F, Merletti F, Pearce N, J Epidemiol Community Health 2008 62 14 10.1136/jech.2007.065581 18572430
21. A New criterion for confounder selection. Vanderweele TJ, Shpitser I, Biometrics 2011 67 1406 1413 10.1111/j.1541-0420.2011.01619.x 21627630
22. Methodological considerations, such as directed acyclic graphs, for studying "acute on chronic" disease epidemiology: chronic obstructive pulmonary disease example. Tsai C-L, Camargo CA Jr, J Clin Epidemiol 2009 62 982 990 10.1016/j.jclinepi.2008.10.005 19211222
23. Beware of the DAG! Dawid AP, Journal of Machine Learning Research Workshop and Conference Proceedings 2010 6 59 86
24. Influence diagrams for causal modelling and inference. Dawid AP, Int Stat Rev 2002 70 161 189
25. Estimation of direct causal effects. Petersen ML, Sinisi SE, van der Laan MJ, Epidemiology 2006 17 276 284 10.1097/01.ede.0000208475.99429.2d 16617276
26. Identifiability and exchangeability for direct and indirect effects. Robins JM, Greenland S, Epidemiology 1992 3 143 155 10.1097/00001648-199203000- 00013 1576220
27. A complete graphical criterion for the adjustment formula in mediation analysis. Shpitser I, Vanderweele TJ, Int J Biostat 2011 7 16 10.2202/1557-4679.1297 21556286
28. Shpitser I, VanderWeele TJ, Robins JM, On the validity of covariate adjustment for estimating causal effects. Proceedings of the Twenty-Sixth Conference Annual Conference on Uncertainty in Artificial Intelligence (UAI-10) Corvallis: AUAI 2010 527 536
29. Confounding and collapsibility in causal inference. Greenland S, Robins JM, Pearl J, Stat Sci 1999 14 29 46 10.1214/ss/1009211805
30. A suite of R functions for directed acyclic graphs. Breitling L, Epidemiology 2010 21 586 587 10.1097/EDE.0b013e3181e09112 20539116
31. DAG program: identifying minimal sufficient adjustment sets. Knueppel S, Stang A, Epidemiology 2010 21 159 20010223
32. Rothman K, Greenland S, Lash T, Modern epidemiology Philadelphia: Lipincott Williams &Wilkins 3rd 2008
33. Marginalia: comparing adjusted effect measures. Kaufman JS, Epidemiology 2010 21 490 493 10.1097/EDE.0b013e3181e00730 20539110
34. Absence of confounding does not correspond to collapsibility of the rate ratio or rate difference. Greenland S, Epidemiology 1996 7 498 501 10.1097/00001648-199609000-00007 8862980 (Pubitemid 26277694)
35. Confounding-essence and detection. Miettinen OS, Cook EF, Am J Epidemiol 1981 114 593 603 7304589
36. A tutorial on methods to estimating clinically and policy-meaningful measures of treatment effects in prospective observational studies: a review. Austin PC, Laupacis A, Int J Biostat 2011 7 1 6 22848188
37. Absolute risk reductions, relative risks, relative risk reductions, and numbers needed to treat can be obtained from a logistic regression model. Austin PC, J Clin Epidemiol 2010 63 2 6 10.1016/j.jclinepi.2008.11.004 19230611
38. Logistic regression was preferred to estimate risk differences and numbers needed to be exposed adjusted for covariates. Gehrmann U, Kuss O, Wellmann J, Bender R, J Clin Epidemiol 2010 63 1223 1231 10.1016/j.jclinepi. 2010.01.011 20430578
39. Estimating the relative risk in cohort studies and clinical trials of common outcomes. McNutt L-A, Wu C, Xue X, Hafner JP, Am J Epidemiol 2003 157 940 943 10.1093/aje/kwg074 12746247 (Pubitemid 36555054)
40. Simulation study of confounder-selection strategies. Maldonado G, Greenland S, Am J Epidemiol 1993 138 923 936 8256780 (Pubitemid 24002359)
41. Invited Commentary: causal diagrams and measurement bias. Hernán MA, Cole SR, Am J Epidemiol 2009 170 959 962 10.1093/aje/kwp293 19755635
42. Bias due to non-differential misclassification of polytomous confounders. Brenner H, J Clin Epidemiol 1993 46 57 63 10.1016/0895-4356(93)90009-P 8433115 (Pubitemid 23084432)
43. On the nondifferential misclassification of a binary confounder. Ogburn EL, VanderWeele TJ, Epidemiology 2012 23 433 439 10.1097/EDE.0b013e31824d1f63 22450692
44. Intuitions, simulations, theorems: the role and limits of methodology. Greenland S, Epidemiology 2012 23 440 442 10.1097/EDE.0b013e31824e278d 22475828
45. Theorems, proofs, examples, and rules in the practice of epidemiology. Vanderweele TJ, Ogburnb EL, Epidemiology 2012 23 443 445 10.1097/EDE. 0b013e31824e2d4e 22475829
46. On a class of bias-amplifying covariates that endanger effect estimates. Pearl J, Proceedings of the twenty-sixth conference on uncertainty in artificial intelligence, 417 - 424. AUAI, Corvallis, OR, 2010 2010 417 424 Technical report (R-356)
47. Should instrumental variables be used as matching variables? Wooldridge J, Tech. Rep. Michigan state university 2006
48. Effects of adjusting for instrumental variables on bias and precision of effect estimates. Myers JA, Rassen JA, Gagne JJ, Huybrechts KF, Schneeweiss S, Rothman KJ, Joffe MM, Glynn RJ, Am J Epidemiol 2011 174 1213 1222 10.1093/aje/kwr364 22025356
49. Invited commentary: understanding bias amplification. Pearl J, Am J Epidemiol 2011 174 1223 1227 10.1093/aje/kwr352 22034488
50. Instrumental variables I: instrumental variables exploit natural variation in nonexperimental data to estimate causal relationships. Rassen JA, Brookhart MA, Glynn RJ, Mittleman MA, Schneeweiss S, J Clin Epidemiol 2009 62 1226 1232 10.1016/j.jclinepi.2008.12.005 19356901
51. Peritoneal dialysis in polycystic kidney disease patients. Report from the French peritoneal dialysis registry (RDPLF). Lobbedez T, Touam M, Evans D, Ryckelynck J-P, Knebelman B, Verger C, Nephrol Dial Transplant 2011 26 2332 2339 10.1093/ndt/gfq712 21115669
52. A modified least-squares regression approach to the estimation of risk difference. Cheung YB, Am J Epidemiol 2007 166 1337 1344 10.1093/aje/kwm223 18000021 (Pubitemid 350129422)
53. Quantitative assessment of unobserved confounding is mandatory in nonrandomized intervention studies. Groenwold RHH, Hak E, Hoes AW, J Clin Epidemiol 2009 62 22 28 10.1016/j.jclinepi.2008.02.011 18619797
54. Data, design, and background knowledge in etiologic inference. Robins JM, Epidemiology 2001 12 313 320 10.1097/00001648-200105000-00011 11338312 (Pubitemid 32336395)
55. Understanding human functioning using graphical models. Kalisch M, Fellinghauer BAG, Grill E, Maathuis MH, Mansmann U, Bühlmann P, Stucki G, BMC Med Res Methodol 2010 10 14 10.1186/1471-2288-10-14 20149230
56. Some surprising results about covariate adjustment in logistic-regression models. Robinson LD, Jewell NP, Int Stat Rev 1991 59 227 240 10.2307/1403444
57. Transportability across studies: a formal approach. Pearl J, Bareinboim E, Technical report R-372 2011
58. Signed directed acyclic graphs for causal inference. VanderWeele TJ, Robins JM, Journal of the Royal Statistical Society Series B-Statistical Methodology 2009 72 111 127
59. Directed acyclic graphs, sufficient causes, and the properties of conditioning on a common effect. VanderWeele TJ, Robins JM, Am J Epidemiol 2007 166 1096 1104 10.1093/aje/kwm179 17702973 (Pubitemid 47574683)