Problematic formulations of SAS PROC.MIXED models for repeated measurements

J. E. Overall, C. Ahn, C. Shivakumar, Y. Kalburgi

Research output: Contribution to journalArticle

26 Citations (Scopus)

Abstract

The work reported in this article was undertaken to evaluate the utility of SAS PROC.MIXED for testing hypotheses concerning GROUP and TIME x GROUP effects in repeated measurements designs with dropouts. If dropouts are not completely at random, covariate control over informative individual differences on which dropout data patterns depend is widely recognized to be important. However, the inclusion of baseline scores and time-in-study as between-subject covariates in an otherwise well formulated SAS PROC.MIXED model resulted in inadequate control over type I error in simulated data with or without dropouts present. The inadequate model formulations and resulting deviant test sizes are presented here as a warning for others who might be guided by the same information sources to employ similar model specifications when analyzing data from actual clinical trials. It is important that the complete model specification be provided in detail when reporting applications of the general linear mixed-model procedure. A single random- coefficients model produced appropriate test sizes, hut it provided inferior power when informative covariates were added in the attempt to adjust for dropouts. As an alternative, the incorporation of covariate controls in simpler two-stage endpoint or random regression analyses is documented to be effective in dealing with dropouts under specifiable conditions.

Original languageEnglish (US)
Pages (from-to)189-216
Number of pages28
JournalJournal of Biopharmaceutical Statistics
Volume9
Issue number1
DOIs
StatePublished - 1999

Fingerprint

Repeated Measurements
Drop out
Mixed Model
Individuality
Linear Models
Regression Analysis
Clinical Trials
Formulation
Covariates
Model Specification
Repeated Measurement Designs
Random Coefficient Models
Linear Mixed Model
Testing Hypotheses
Individual Differences
Type I error
Power (Psychology)
Baseline
Inclusion
Regression

Keywords

  • Covariates
  • Missing data
  • Mixed models
  • Regression on time
  • Repeated measures
  • Test size

ASJC Scopus subject areas

  • Pharmacology (medical)
  • Pharmacology, Toxicology and Pharmaceutics(all)

Cite this

Problematic formulations of SAS PROC.MIXED models for repeated measurements. / Overall, J. E.; Ahn, C.; Shivakumar, C.; Kalburgi, Y.

In: Journal of Biopharmaceutical Statistics, Vol. 9, No. 1, 1999, p. 189-216.

Research output: Contribution to journalArticle

Overall, J. E. ; Ahn, C. ; Shivakumar, C. ; Kalburgi, Y. / Problematic formulations of SAS PROC.MIXED models for repeated measurements. In: Journal of Biopharmaceutical Statistics. 1999 ; Vol. 9, No. 1. pp. 189-216.
@article{b11c483cbca44e0fbaf98a9e795208a3,
title = "Problematic formulations of SAS PROC.MIXED models for repeated measurements",
abstract = "The work reported in this article was undertaken to evaluate the utility of SAS PROC.MIXED for testing hypotheses concerning GROUP and TIME x GROUP effects in repeated measurements designs with dropouts. If dropouts are not completely at random, covariate control over informative individual differences on which dropout data patterns depend is widely recognized to be important. However, the inclusion of baseline scores and time-in-study as between-subject covariates in an otherwise well formulated SAS PROC.MIXED model resulted in inadequate control over type I error in simulated data with or without dropouts present. The inadequate model formulations and resulting deviant test sizes are presented here as a warning for others who might be guided by the same information sources to employ similar model specifications when analyzing data from actual clinical trials. It is important that the complete model specification be provided in detail when reporting applications of the general linear mixed-model procedure. A single random- coefficients model produced appropriate test sizes, hut it provided inferior power when informative covariates were added in the attempt to adjust for dropouts. As an alternative, the incorporation of covariate controls in simpler two-stage endpoint or random regression analyses is documented to be effective in dealing with dropouts under specifiable conditions.",
keywords = "Covariates, Missing data, Mixed models, Regression on time, Repeated measures, Test size",
author = "Overall, {J. E.} and C. Ahn and C. Shivakumar and Y. Kalburgi",
year = "1999",
doi = "10.1081/BIP-100101008",
language = "English (US)",
volume = "9",
pages = "189--216",
journal = "Journal of Biopharmaceutical Statistics",
issn = "1054-3406",
publisher = "Taylor and Francis Ltd.",
number = "1",

}

TY - JOUR

T1 - Problematic formulations of SAS PROC.MIXED models for repeated measurements

AU - Overall, J. E.

AU - Ahn, C.

AU - Shivakumar, C.

AU - Kalburgi, Y.

PY - 1999

Y1 - 1999

N2 - The work reported in this article was undertaken to evaluate the utility of SAS PROC.MIXED for testing hypotheses concerning GROUP and TIME x GROUP effects in repeated measurements designs with dropouts. If dropouts are not completely at random, covariate control over informative individual differences on which dropout data patterns depend is widely recognized to be important. However, the inclusion of baseline scores and time-in-study as between-subject covariates in an otherwise well formulated SAS PROC.MIXED model resulted in inadequate control over type I error in simulated data with or without dropouts present. The inadequate model formulations and resulting deviant test sizes are presented here as a warning for others who might be guided by the same information sources to employ similar model specifications when analyzing data from actual clinical trials. It is important that the complete model specification be provided in detail when reporting applications of the general linear mixed-model procedure. A single random- coefficients model produced appropriate test sizes, hut it provided inferior power when informative covariates were added in the attempt to adjust for dropouts. As an alternative, the incorporation of covariate controls in simpler two-stage endpoint or random regression analyses is documented to be effective in dealing with dropouts under specifiable conditions.

AB - The work reported in this article was undertaken to evaluate the utility of SAS PROC.MIXED for testing hypotheses concerning GROUP and TIME x GROUP effects in repeated measurements designs with dropouts. If dropouts are not completely at random, covariate control over informative individual differences on which dropout data patterns depend is widely recognized to be important. However, the inclusion of baseline scores and time-in-study as between-subject covariates in an otherwise well formulated SAS PROC.MIXED model resulted in inadequate control over type I error in simulated data with or without dropouts present. The inadequate model formulations and resulting deviant test sizes are presented here as a warning for others who might be guided by the same information sources to employ similar model specifications when analyzing data from actual clinical trials. It is important that the complete model specification be provided in detail when reporting applications of the general linear mixed-model procedure. A single random- coefficients model produced appropriate test sizes, hut it provided inferior power when informative covariates were added in the attempt to adjust for dropouts. As an alternative, the incorporation of covariate controls in simpler two-stage endpoint or random regression analyses is documented to be effective in dealing with dropouts under specifiable conditions.

KW - Covariates

KW - Missing data

KW - Mixed models

KW - Regression on time

KW - Repeated measures

KW - Test size

UR - http://www.scopus.com/inward/record.url?scp=0033015444&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033015444&partnerID=8YFLogxK

U2 - 10.1081/BIP-100101008

DO - 10.1081/BIP-100101008

M3 - Article

C2 - 10091918

AN - SCOPUS:0033015444

VL - 9

SP - 189

EP - 216

JO - Journal of Biopharmaceutical Statistics

JF - Journal of Biopharmaceutical Statistics

SN - 1054-3406

IS - 1

ER -