Advertisement

Intensive Care Medicine

, Volume 41, Issue 2, pp 348–350 | Cite as

Multiple imputation: a mature approach to dealing with missing data

  • S. ChevretEmail author
  • S. Seaman
  • M. Resche-Rigon
Statistical Editorial

Missing values in clinical studies are almost unavoidable. When analyzing such data, the standard response is to exclude the patients with missing data. This is known as ‘complete case analysis’ (CCA) and has been shown to be the leading strategy in the epidemiology [1] and intensive care unit (ICU) literature [2]. However, if the excluded patients are not a representative subsample from the whole sample, their exclusion can lead to bias and loss of precision in estimation, both of which can, for example, adversely affect the performance of predictive risk models in the ICU (Supplementary 1). To deal with this issue, numerous imputation methods have been developed. The simplest method is “simple imputation.” This involves replacing each missing value with a single value, such as the mean of the observed data [3]. Thereafter, all patients present in the sample can be included in the analysis. The simplicity and ease of implementation of this method make it attractive. However, it tends...

Notes

Conflicts of interest

None.

Supplementary material

134_2014_3624_MOESM1_ESM.docx (542 kb)
Supplementary material 1 (DOCX 541 kb)

References

  1. 1.
    Eekhout I, de Boer RM, Twisk JW, de Vet HC, Heymans MW (2012) Missing data: a systematic review of how they are reported and handled. Epidemiology 23(5):729–732PubMedCrossRefGoogle Scholar
  2. 2.
    Vesin A, Azoulay E, Ruckly S, Vignoud L, Rusinova K, Benoit D, Soares M, Azeivedo-Maia P, Abroug F, Benbenishty J, Timsit JF (2013) Reporting and handling missing values in clinical studies in intensive care units. Intensive Care Med 39:1396–1404PubMedCrossRefGoogle Scholar
  3. 3.
    Lee CH, Arzeno NM, Ho JC, Vikalo H, Ghosh J (2012) A imputation-enhanced algorithm for ICU mortality prediction. Comput Cardiol 39:253–256Google Scholar
  4. 4.
    Rubin DB (1976) Inference and missing data. Biometrika 63:581–592CrossRefGoogle Scholar
  5. 5.
    Rubin DB (1987) Multiple imputation for nonresponse in surveys. John Wiley, New YorkCrossRefGoogle Scholar
  6. 6.
    Pérez A, Dennis RJ, Gil JF, Rondón MA, López A (2002) Use of the mean, hot deck and multiple imputation techniques to predict outcome in intensive care unit patients in Colombia. Stat Med 21(24):3885–3896PubMedCrossRefGoogle Scholar
  7. 7.
    Debaty G, Maignan M, Savary D, Koch FX, Ruckly S, Durand M, Picard J, Escallier C, Chouquer R, Santre C, Minet C, Guergour D, Hammer L, Bouvaist H, Belle L, Adrie C, Payen JF, Carpentier F, Gueugniaud PY, Danel V, Timsit JF (2014) Impact of intra-arrest therapeutic hypothermia in outcomes of prehospital cardiac arrest: a randomized controlled trial. Intensive Care Med 40(12):1832–1842PubMedCrossRefGoogle Scholar
  8. 8.
    van Buuren S. Flexible imputation of missing data. Chapman & Hall/CRC, 2012, Boca Raton. ISBN 9781439868249. CRC PressGoogle Scholar
  9. 9.
    Little RJA, Rubin DB (2002) Statistical analysis with missing data, 2nd edn. Wiley, HobokenCrossRefGoogle Scholar
  10. 10.
    van Buuren S, Boshuizen HC, Knook DL (1999) Multiple imputation of missing blood pressure covariates in survival analysis. Stat Med 18:681–694PubMedCrossRefGoogle Scholar
  11. 11.
    Schaffer JL (2008) Software for multiple imputation. http://www.stat.psu.edu/~jls/misoftwa.html. Accessed 28 Dec 2014
  12. 12.
    White I, Royston P, Wood AM (2011) Multiple imputation using chained equations: issues and guidance for practice. Stat Med 30(4):377–399PubMedCrossRefGoogle Scholar
  13. 13.
    Sterne JA, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, Wood AM, Carpenter JR (2009) Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ 29(338):b2393CrossRefGoogle Scholar
  14. 14.
    Carpenter JR, Kenward MG, White IR (2007) Sensitivity analysis after multiple imputation under missing at random: a weighting approach. Stat Methods Med Res 16(3):259–275PubMedCrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg and ESICM 2015

Authors and Affiliations

  1. 1.Service de Biostatistique et Information MédicaleHôpital Saint-Louis, AP-HP, ECSTRA Team, Inserm UMR-1153, Université Paris DiderotParisFrance
  2. 2.MRC Biostatistics UnitInstitute of Public Health, Robinson WayCambridgeUK

Personalised recommendations