Skip to main content
Log in

Methods to Account for Attrition in Longitudinal Data: Do They Work? A Simulation Study

  • Methods
  • Published:
European Journal of Epidemiology Aims and scope Submit manuscript

Abstract

Attrition threatens the internal validity of cohort studies. Epidemiologists use various imputation and weighting methods to limit bias due to attrition. However, the ability of these methods to correct for attrition bias has not been tested. We simulated a cohort of 300 subjects using 500 computer replications to determine whether regression imputation, individual weighting, or multiple imputation is useful to reduce attrition bias. We compared these results to a complete subject analysis. Our logistic regression model included a binary exposure and two confounders. We generated 10, 25, and 40% attrition through three missing data mechanisms: missing completely at random (MCAR), missing at random (MAR) and missing not at random (MNAR), and used four covariance matrices to vary attrition. We compared true and estimated mean odds ratios (ORs), standard deviations (SDs), and coverage. With data MCAR and MAR for all attrition rates, the complete subject analysis produced results at least as valid as those from the imputation and weighting methods. With data MNAR, no method provided unbiased estimates of the OR at attrition rates of 25 or 40%. When observations are not MAR or MCAR, imputation and weighting methods may not effectively reduce attrition bias.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

Abbreviations

CI:

confidence interval

MAR:

missing at random

MCAR:

missing completely at random

MNAR:

missing not at random

MTBI:

mild traumatic brain injury

OR:

odds ratio

SD:

standard deviation

References

  1. PM Pennefather W Tin MP Clarke J Dutton S Fritz EN Hey (1999) ArticleTitleBias due to incomplete follow up in a cohort study Br J Ophthalmol 83 643–645 Occurrence Handle10340968

    PubMed  Google Scholar 

  2. G Touloumi SJ Pocock AG Babiker JH Darbyshire (2002) ArticleTitleImpact of missing data due to selective dropouts in cohort studies and clinical trials Epidemiology 13 347–355 Occurrence Handle10.1097/00001648-200205000-00017 Occurrence Handle11964938

    Article  PubMed  Google Scholar 

  3. S Greenland (1977) ArticleTitleResponse and follow-up bias in cohort studies Am J Epidemiol 106 184–187 Occurrence Handle900117

    PubMed  Google Scholar 

  4. JL Schafer JW Graham (2002) ArticleTitleMissing data: Our view of the state of the art Psychol Methods 7 147–177 Occurrence Handle10.1037//1082-989X.7.2.147 Occurrence Handle12090408

    Article  PubMed  Google Scholar 

  5. JM Engels P Diehr (2003) ArticleTitleImputation of missing longitudinal data: A comparison of methods J Clin Epidemiol 56 968–976 Occurrence Handle10.1016/S0895-4356(03)00170-7 Occurrence Handle14568628

    Article  PubMed  Google Scholar 

  6. AC Heath PA Madden NG Martin (1998) ArticleTitleAssessing the effects of cooperation bias and attrition in behavioral genetic research using data-weighting Behav Genet 28 415–427 Occurrence Handle10.1023/A:1021633127604 Occurrence Handle9926611

    Article  PubMed  Google Scholar 

  7. MB Landrum MP Becker (2001) ArticleTitleA multiple imputation strategy for incomplete longitudinal data Stat Med 20 2741–2760 Occurrence Handle10.1002/sim.740 Occurrence Handle11523080

    Article  PubMed  Google Scholar 

  8. CM Musil CB Warner PK Yobas SL Jones (2002) ArticleTitleA comparison of imputation techniques for handling missing data West J Nurs Res 24 815–829 Occurrence Handle10.1177/019394502762477004 Occurrence Handle12428897

    Article  PubMed  Google Scholar 

  9. SL Crawford SL Tennstedt JB McKinlay (1995) ArticleTitleA comparison of analytic methods for non-random missingness of outcome data J Clin Epidemiol 48 209–219 Occurrence Handle10.1016/0895-4356(94)00124-9 Occurrence Handle7869067

    Article  PubMed  Google Scholar 

  10. JL Schafer (1999) ArticleTitleMultiple imputation: A primer Stat Methods Med Res 8 3–15 Occurrence Handle10.1191/096228099671525676 Occurrence Handle10347857

    Article  PubMed  Google Scholar 

  11. F Barzi M Woodward (2004) ArticleTitleImputations of missing values in practice: Results from imputations of serum cholesterol in 28 cohort studies Am J Epidemiol 160 34–45 Occurrence Handle10.1093/aje/kwh175 Occurrence Handle15229115

    Article  PubMed  Google Scholar 

  12. DB Rubin (1996) ArticleTitleMultiple imputation after 18+ years J Am Stat Assoc 91 473–489

    Google Scholar 

  13. RJA Little DB Rubin (1987) Statistical Analysis with Missing Data John Wiley & Sons New York

    Google Scholar 

  14. KM Bisgard AR Folsom CP Hong TA Sellers (1994) ArticleTitleMortality and cancer rates in nonrespondents to a prospective study of older women: 5-year follow-up Am J Epidemiol 139 990–1000 Occurrence Handle8178787

    PubMed  Google Scholar 

  15. RR McLean MT Hannan BE Epstein et al. (2000) ArticleTitleElderly cohort study subjects unable to return for follow-up have lower bone mass than those who can return Am J Epidemiol 151 689–692 Occurrence Handle10752796

    PubMed  Google Scholar 

  16. V Kristman M Manno P Côté (2004) ArticleTitleLoss to follow-up in cohort studies: How much is too much? Eur J Epidemiol 19 751–760 Occurrence Handle10.1023/B:EJEP.0000036568.02655.f8 Occurrence Handle15469032

    Article  PubMed  Google Scholar 

  17. J Twisk W Vente Particlede (2002) ArticleTitleAttrition in longitudinal studies: How to deal with missing data J Clin Epidemiol 55 329–337 Occurrence Handle10.1016/S0895-4356(01)00476-0 Occurrence Handle11927199

    Article  PubMed  Google Scholar 

  18. RS Rao AJ Sigurdson MM Doody BI Graubard (2005) ArticleTitleAn application of a weighting method to adjust for nonresponse in standardized incidence ratio analysis of cohort studies Ann Epidemiol 15 129–136 Occurrence Handle10.1016/j.annepidem.2004.05.007 Occurrence Handle15652718

    Article  PubMed  Google Scholar 

  19. P Makela (2003) ArticleTitleImpact of correcting for nonresponse by weighting on estimates of alcohol consumption J Stud Alcohol 64 589–596 Occurrence Handle12921202

    PubMed  Google Scholar 

  20. SG Gerberich TR Church PM McGovern et al. (2004) ArticleTitleAn epidemiological study of the magnitude and consequences of work related violence: The Minnesota Nurses’ Study Occup Environ Med 61 495–503 Occurrence Handle10.1136/oem.2003.007294 Occurrence Handle15150388

    Article  PubMed  Google Scholar 

  21. WM Hopman C Berger L Joseph et al. (2004) ArticleTitleStability of normative data for the SF-36 Can J Public Health 95 387–391 Occurrence Handle15490932

    PubMed  Google Scholar 

  22. GD Mishra AJ Dobson (2004) ArticleTitleMultiple imputation for body mass index: Lessons from the Australian Longitudinal Study on Women’s Health Statist Med 23 3077–3087 Occurrence Handle10.1002/sim.1911

    Article  Google Scholar 

  23. SAS Publishing. SAS/STAT Software: Changes and Enhancements, Release 8.1. Cary, NC: SAS Institute, 2000

  24. JD Elashoff RM Elashoff (1974) ArticleTitleTwo-sample problems for a dichotomous variable with missing data Appl Stat 23 26–34

    Google Scholar 

  25. SAS Publishing. SAS/STAT User’s Guide, Version 8, Volumes 1, 2 and 3. Cary, NC: SAS Institute, 1999

  26. JPC Kleijnen (1979) The role of statistical methodology in simulation BP Zeigler (Eds) Methodology in Systems Modelling and Simulation North-Holland Amsterdam

    Google Scholar 

  27. AO Youk RA Stone GM Marsh (2004) ArticleTitleA method for imputing missing data in longitudinal studies Ann Epidemiol 14 354–361 Occurrence Handle10.1016/j.annepidem.2003.09.010 Occurrence Handle15177275

    Article  PubMed  Google Scholar 

  28. JM Taylor KL Cooper JT Wei AV Sarma TE Raghunathan SG Heeringa (2002) ArticleTitleUse of multiple imputation to correct for nonresponse bias in a survey of urologic symptoms among African-American men Am J Epidemiol 156 774–782 Occurrence Handle10.1093/aje/kwf110 Occurrence Handle12370166

    Article  PubMed  Google Scholar 

  29. RJA Little (1995) ArticleTitleModeling the drop-out mechanism in repeated-measures studies J Am Stat Assoc 90 1112–1121 Occurrence HandleMR1354029

    MathSciNet  Google Scholar 

  30. G Maldonado S Greenland (1997) ArticleTitleThe importance of critically interpreting simulation studies Epidemiology 8 453–456 Occurrence Handle9209864

    PubMed  Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Rights and permissions

Reprints and permissions

About this article

Cite this article

Kristman, V.L., Manno, M. & Côté, P. Methods to Account for Attrition in Longitudinal Data: Do They Work? A Simulation Study. Eur J Epidemiol 20, 657–662 (2005). https://doi.org/10.1007/s10654-005-7919-7

Download citation

  • Accepted:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10654-005-7919-7

Keywords

Navigation