TITLE

Comparison of imputation methods for handling missing covariate data when fitting a Cox proportional hazards model: a resampling study

AUTHOR(S)
Marshall, Andrea; Altman, Douglas G.; Holder, Roger L.
PUB. DATE
January 2010
SOURCE
BMC Medical Research Methodology;2010, Vol. 10, p112
SOURCE TYPE
Academic Journal
DOC. TYPE
Article
ABSTRACT
Background: The appropriate handling of missing covariate data in prognostic modelling studies is yet to be conclusively determined. A resampling study was performed to investigate the effects of different missing data methods on the performance of a prognostic model. Methods: Observed data for 1000 cases were sampled with replacement from a large complete dataset of 7507 patients to obtain 500 replications. Five levels of missingness (ranging from 5% to 75%) were imposed on three covariates using a missing at random (MAR) mechanism. Five missing data methods were applied; a) complete case analysis (CC) b) single imputation using regression switching with predictive mean matching (SI), c) multiple imputation using regression switching imputation, d) multiple imputation using regression switching with predictive mean matching (MICE-PMM) and e) multiple imputation using flexible additive imputation models. A Cox proportional hazards model was fitted to each dataset and estimates for the regression coefficients and model performance measures obtained. Results: CC produced biased regression coefficient estimates and inflated standard errors (SEs) with 25% or more missingness. The underestimated SE after SI resulted in poor coverage with 25% or more missingness. Of the MI approaches investigated, MI using MICE-PMM produced the least biased estimates and better model performance measures. However, this MI approach still produced biased regression coefficient estimates with 75% missingness. Conclusions: Very few differences were seen between the results from all missing data approaches with 5% missingness. However, performing MI using MICE-PMM may be the preferred missing data approach for handling between 10% and 50% MAR missingness.
ACCESSION #
57829264

 

Related Articles

  • SilverPlatter Publishes Patient Advice Database.  // Information Today;May1994, Vol. 11 Issue 5, p36 

    The article presents information about the Health and Drug Information Library on SilverPlatter CD-ROM, which is a compendium of four separate products from Clinical Reference Systems Ltd., Greenwood Village, Colorado. This database represents a comprehensive collection of full-text information...

  • Data Sources and Information to be Collected for a Romanian Diabetes Register. Leucuţa, Daniel-Corneliu; Turdeanu, Cristian; Cadariu, Andrei Achimaş // Applied Medical Informatics;Jul2010, Vol. 26 Issue 1/2, p101 

    Diabetes is a serious condition that damages many systems and organs, mainly the blood vessels and the nerves. The importance of the diabetes and its complications, along with the prevalence of the disease makes it a major public health problem worldwide. The spread of diabetes and its...

  • View from the frontline: Does IT make patient care worse? Brown, Harry // Health Information on the Internet;Apr2008, Issue 62, p9 

    The article discusses the benefits of using information technology (IT) to provide patient care. It states that the accumulated patient data and medical histories from several sources also provide high quality knowledge bases containing evidence-based information that could be used to optimally...

  • The way forward. Pardes, Herbert // Modern Healthcare;5/19/2003, Vol. 33 Issue 20, p41 

    Presents an article about the importance of an electronic health information infrastructure to hospitals in the U.S. Benefits from medical information technology; Influence of electronic health information system on the quality of patient care; Proposed bill to govern the use of electronic...

  • Healthcare Informatics Research: From Data to Evidence-Based Management. Wan, Thomas T. H. // Journal of Medical Systems;Feb2006, Vol. 30 Issue 1, p3 

    Healthcare informatics research is a scientific endeavor that applies information science, computer technology, and statistical modeling techniques to develop decision support systems for improving both health service organizations' performance and patient care outcomes. The analytical...

  • A Clinical Information Management Platform for Semantic Exploitation of Clinical Data. Seebode, Christian; Trautwein, Martin; Ort, Matthias; Lehmann, Jan-Marcus // International MultiConference of Engineers & Computer Scientists;2013, Vol. 1, p1 

    We present a platform that combines an approach to semantic extraction of medical information from clinical free-text documents with the processing of structured information from HIS records. The information extraction process uses a fine-grained linguistic analysis, and maps preprocessed terms...

  • Social and Contextual Factors in the Analysis of Mortality in End-Stage Renal Disease Patients: Implications for Health Policy. Plough, Alonzo L.; Salem, Susanne // American Journal of Public Health;Nov82, Vol. 72 Issue 11, p1293 

    A sample of medical records of deceased End-Stage Renal Disease (ESRD) patients was reviewed by a panel of experienced clinicians. The panel's determination of cause of death was compared to that reported for these patients in the Health CaFe Financing Administration Management Information...

  • Cochrane update: identifying health-related research resources relevant to low- and middle-income countries. Stansfield, Claire; Weightman, Alison L; Kavanagh, Josephine; Johansen, Marit // Journal of Public Health;Sep2013, Vol. 35 Issue 3, p477 

    The article reports on the establishment of the low- and middle-income countries (LMIC) resource list by the Cochrane Collaboration's Information Retrieval Methods Group (IRMG). It notes that the group aims to enhance access to public health services. It also mentions the group's effort on...

  • If You Build It, Don't Expect Them to Come. Connely, Robert E. // Health Management Technology;Aug2008, Vol. 29 Issue 8, p32 

    The article offers information on the challenges faced by health information exchanges (HIE) and the reasons for its failure in the U.S. Many HIE efforts start with the goal of sharing patient information across the community anywhere, anytime, which requires a significant investment in...

Share

Read the Article

Courtesy of THE LIBRARY OF VIRGINIA

Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics