Download PDF
Methods Inf Med 2009; 48(05): 438-443
DOI: 10.3414/ME9241
Original Articles
Schattauer GmbH

Efficient Risk Set Sampling when a Time-dependent Exposure Is Present

Matching for Time to Exposure Versus Exposure Density Sampling
M. Wolkewitz
1   Institute of Medical Biometry and Medical Informatics, University Medical Center Freiburg, Freiburg, Germany
,
J. Beyersmann
1   Institute of Medical Biometry and Medical Informatics, University Medical Center Freiburg, Freiburg, Germany
,
P. Gastmeier
2   Institute for Medical Microbiology and Hospital Epidemiology, Medical School Hannover, Hannover, Germany
,
M. Schumacher
1   Institute of Medical Biometry and Medical Informatics, University Medical Center Freiburg, Freiburg, Germany
› Author Affiliations
Further Information

Publication History

05 August 2009

Publication Date:
20 January 2018 (online)

    Permissions and Reprints

Summary

Objectives: The impact of time-dependent exposures on the time until study endpoint may correctly be analyzed with data of a full cohort. Ignoring the time-dependent nature of these exposures leads to time-dependent bias. Matching for time to exposure is often applied to take the time-dependency into account, but prefixed sets of exposed and unexposed may still create bias. This approach is attractive since a subcohort would also save resources, especially when exposure and outcome data are only available in the full cohort but further covariate information is required. The first objective is to show to which extent matching for time to exposure yields biased results. Secondly, exposure density sampling is introduced and explored.

Methods: To evaluate how both sampling methods perform, they are compared to the correct method as well as to the approach in which the time-dependent nature of the exposure is ignored. Real data of the SIR-3 study (Germany, 2000–2001) and a simulation study are used.

Results: Simulations show that matching may reduce the time-dependent bias but still there is a bias. The matching bias decreases if fewer patients are exposed. Exposure density sampling yields unbiased results.

Conclusions: Results from studies in which matching for time to exposure was applied are only tolerable for rare exposures. Whenever subcohorting is the intention in order to save resources, exposure density sampling should be preferred instead.

Keywords

Cohort study - nosocomial infection - multistate model - survival bias - time-dependent bias - case-cohort design
 

  • References

  • 1 van Walraven C, Davis D, Forster A, Wells G. Time-dependent bias was common in survival analyses published in leading clinical journals. J Clin Epidemiol 2004; 57: 672-682.
    CrossrefPubMedGoogle Scholar
  • 2 Zhou Z, Rahme E, Abrahamowicz M, Pilote L. Survival bias associated with time-to-treatment initiation in drug effectiveness evaluation: a comparison of methods. Am J Epidemiol 2005; 162: 1016-1023.
    CrossrefPubMedGoogle Scholar
  • 3 Suissa S. Immortal time bias in pharmacoepidemiology. Am J. Epidemiol 2008; 167: 492-499.
    PubMedGoogle Scholar
  • 4 Clark D, Stinson E, Griepp R, Schroeder J, Shumway N, Harrison D. Cardiac transplantation in man. VI. Prognosis of patients selected for cardiac transplantation. Ann Intern Med 1971; 75: 15-21.
    CrossrefPubMedGoogle Scholar
  • 5 Sylvestre M, Huszti E, Hanley J. Do OSCAR winners live longer than less successful peers? A reanalysis of the evidence. Ann Intern Med 2006; 145: 361-363.
    CrossrefPubMedGoogle Scholar
  • 6 Redelmeier D, Singh S. Survival in Academy Award-winning actors and actresses. Ann Intern Med 2001; 134: 955-962.
    CrossrefPubMedGoogle Scholar
  • 7 Korenman S, Goldman N, Fu H. Misclassification bias in estimates of bereavement effects. Am J Epidemiol 1997; 145: 995-1002.
    CrossrefPubMedGoogle Scholar
  • 8 Beyersmann J, Gastmeier P, Wolkewitz M, Schumacher M. An easy mathematical proof showed that time-dependent bias inevitably leads to biased effect estimation. J Clin Epidemiol 2008; 61: 1216-1221.
    CrossrefPubMedGoogle Scholar
  • 9 Beyersmann J, Wolkewitz M, Schumacher M. The impact of time-dependent bias in proportional hazards modelling. Stat Med 2008; 27: 6439-6454.
    CrossrefPubMedGoogle Scholar
  • 10 Blot S, Vandewoude K, Colardyn F. Nosocomial bacteremia involving Acinetobacter baumannii in critically ill patients: a matched cohort study. Intensive Care Med 2003; 29: 471-475.
    CrossrefPubMedGoogle Scholar
  • 11 Bercault N, Boulain T. Mortality rate attributable to ventilator-associated nosocomial pneumonia in an adult intensive care unit: a prospective case-control study. Crit Care Med 2001; 29 (12) 2303-2309.
    CrossrefPubMedGoogle Scholar
  • 12 Lee NY, Lee HC, Ko NY, Chang CM, Shih HI, Wu CJ, Ko WC. Clinical and economic impact of multidrug resistance in nosocomial Acinetobacter baumannii bacteremia. Infect Control Hosp Epidemiol J 2007; 28 (06) 713-719.
    PubMedGoogle Scholar
  • 13 Ernst EJ, Raley G, Herwaldt LA, Diekema DJ. Importance of control group selection for evaluating antimicrobial use as a risk factor for methicillin-resistant Staphylococcus aureus bacteremia. Infect Control Hosp Epidemiol 2005; 26 (07) 634-637.
    CrossrefPubMedGoogle Scholar
  • 14 Brawley R, Weber D, Samsa G, Rutala W. Multiple nosocomial infections. An incidence study. Am J Epidemiol 1989; 130: 769-780.
    CrossrefPubMedGoogle Scholar
  • 15 Schumacher M, Wangler M, Wolkewitz M, Beyers-mann J. Attributable mortality due to nosocomial infections. A simple and useful application of multistate models. Methods Inf Med 2007; 46: 595-600.
    Article in Thieme ConnectPubMedGoogle Scholar
  • 16 Kalbfleisch J, Prentice R. The statistical analysis of failure time data. Wiley Series in Probability and Mathematical Statistics. John Wiley and Sons; 2002
    Google Scholar
  • 17 Schröder M, Hüsing J, Jöckel K. An implementation of automated individual matching for observational studies. Methods Inf Med 2004; 43 (05) 516-520.
    Article in Thieme ConnectPubMedGoogle Scholar
  • 18 Stürmer T, Gefeller O, Brenner H. A computer program to estimate power and relative efficiency of flexibly matched case-control studies. Methods Inf Med 2005; 44: 693-696.
    Article in Thieme ConnectPubMedGoogle Scholar
  • 19 Girou E, Stephan F, Novara A, Safar M, Fagon J. Risk factors and outcome of nosocomial infections: results of a matched case-control study of ICU patients. Am J Respir Crit Care Med 1998; 157: 1151-1158.
    CrossrefPubMedGoogle Scholar
  • 20 Renaud B, Brun-Buisson C. Outcomes of primary and catheter-related bacteremia. A cohort and case-control study in critically ill patients. Am J Respir Crit Care Med 2001; 163: 1584-1590.
    CrossrefPubMedGoogle Scholar
  • 21 Heyland D, Cook D, Griffith L, Keenan S, Brun-Buisson C. The attributable morbidity and mortality of ventilator-associated pneumonia in the critically ill patient. The Canadian Critical Trials Group. Am J Respir Crit Care Med 1999; 159: 1249-1256.
    CrossrefPubMedGoogle Scholar
  • 22 Blot S, De Bacquer D, Hoste E, Depuydt P, Vandewoude K, De Waele J, Benoit D, De Schuijmer J, Colardyn F, Vogelaers D. Influence of matching for exposure time on estimates of attributable mortality caused by nosocomial bacteremia in critically ill patients. Infect Control Hosp Epidemiol 2005; 26 (04) 352-356.
    CrossrefPubMedGoogle Scholar
  • 23 Schulgen G, Kropec A, Kappstein I, Daschner F, Schumacher M. Estimation of extra hospital stay attributable to nosocomial infections: heterogeneity and timing of events. J Clin Epidemiol 2000; 53: 409-417.
    CrossrefPubMedGoogle Scholar
  • 24 Beyersmann J, Gastmeier P, Grundmann H, Baer-wolff S, Geffers C, Behnke M, Rueden H, Schumacher M. Use of multistate models to assess prolongation of intensive care unit stay due to nosocomial infection. Infect Control Hosp Epidemiol 2006; 27: 493-499.
    CrossrefPubMedGoogle Scholar
  • 25 Richardson D. An incidence density sampling program for nested case-control analyses. Occup Environ Med 2004; 61: e59.
    CrossrefPubMedGoogle Scholar
  • 26 Asgharian M, M’Lan C, Wolfson D. Length-Biased Sampling with Right Censoring: An Unconditional Approach. J Am Stat Assoc 2002; 97 457 201-209.
    CrossrefPubMedGoogle Scholar
  • 27 Barlow W. Robust variance estimation for the case-cohort design. Biometrics 1994; 50: 1064-1072.
    CrossrefPubMedGoogle Scholar
  • 28 Andersen P, Borgan O, Gill RD, Keiding N. Statistical models based on counting processes. Springer Series in Statistics. New York: Springer-Verlag Inc; 1993
    Google Scholar
  • 29 Hjort NL. On inference in parametric survival data models. International Statistical Review 1992; 60: 355-387.
    CrossrefPubMedGoogle Scholar
  • 30 Claeskens G, Hjort N. Model Selection and Model Averaging. Cambridge University Press; 2008
    Google Scholar
  • 31 Faddy M, McClean S. Analysing data on lengths of stay of hospital patients using phasetype distributions. Appl Stochastic Models Bus Ind 1999; 15: 311-317.
    CrossrefPubMedGoogle Scholar
  • 32 Faddy M, McClean S. Markov chain modelling for geriatric patient care. Methods Inf Med 2005; 44: 369-373.
    Article in Thieme ConnectPubMedGoogle Scholar
  • 33 Aalen O, Borgan Ø, Gjessing H. Event History Analysis: A Process Point of View. New York: Springer-Verlag; 2007
    Google Scholar
  • 34 Olsson M. Estimation of phase-type distributions from censored data. Scandinavian Journal of Statistics 1996; 23 (04) 443-460.
    PubMedGoogle Scholar