• 1

    Brown DC. Control of selection bias in parallel-group controlled clinical trials in dogs and cats: 97 trials (2000–2005). J Am Vet Med Assoc 2006;229:990993.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 2

    Lachin JM. Statistical considerations in the intent-to-treat principle. Control Clin Trials 2000;21:167189.

  • 3

    Nich C, Carroll KM. Intention-to-treat meets missing data: implications of alternate strategies for analyzing clinical trials data. Drug Alcohol Depend 2002;68:121130.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 4

    Tierney JF, Stewart LA. Investigating patient exclusion bias in meta-analysis. Int J Epidemiol 2005;34:7987.

  • 5

    Grasdal A. The performance of sample selection estimators to control for attrition bias. Health Econ 2001;10:385398.

  • 6

    Streiner D, Geddes J. Intention to treat analysis in clinical trials when there are missing data. Evid Based Ment Health 2001;4:7071.

  • 7

    Norman G, Streiner D. Biostatistics: the bare essentials. 2nd ed. Toronto: BC Decker, 2000;260267.

  • 8

    Unnebrink K, Windeler J. Intention-to-treat: methods for dealing with missing values in clinical trials of progressively deteriorating diseases. Stat Med 2001;20:39313946.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 9

    STATA. User's guide. 8th ed. College Station, Tex: STATA Press, 2003.

  • 10

    Tang L, Song J, Belin TR, et al. A comparison of imputation methods in a longitudinal randomized clinical trial. Stat Med 2005;24:21112128.

  • 11

    Lee YJ, Ellenberg JH, Hirtz DG, et al. Analysis of clinical trials by treatment actually received: is it really an option? Stat Med 1991;10:15951605.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 12

    Moher D, Schulz KF, Altman D, et al. The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomized trials. JAMA 2001;285:19871991.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 13

    Montori VM, Guyatt GH. Intention-to-treat principle. Can Med Assoc J 2001;165:13391341.

  • 14

    DeMets DL. Statistical issues in interpreting clinical trials. J Intern Med 2004;255:529537.

  • 15

    Ware JH. Interpreting incomplete data in studies of diet and weight loss. N Engl J Med 2003;348:21362137.

  • 16

    Gadbury GL, Coffey CS, Allison DB. Modern statistical methods for handling missing repeated measurements in obesity trial data: beyond LOCF. Obes Rev 2003;4:175184.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 17

    Mallinckrodt CH, Clark WS, David SR. Accounting for dropout bias using mixed-effects models. J Biopharm Stat 2001;11:921.

Advertisement

Sources and handling of losses to follow-up in parallel-group randomized clinical trials in dogs and cats: 63 trials (2000–2005)

Dorothy Cimino BrownDepartment of Clinical Studies, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, PA 19104-6010

Search for other papers by Dorothy Cimino Brown in
Current site
Google Scholar
PubMed
Close
 DVM, MSCE

Abstract

Objective—To determine the sources and handlingof losses to follow-up (LTF) in parallel-group randomized clinical trials (RCTs).

Sample Population—63 parallel-group RCTs of > 24 hours' duration published from January 2000 through December 2005.

Procedures—Journals were hand searched for eligible reports. Details concerning the presence, cause, and amount of LTF; statistical handlingof data missingbecause of LTF; type of analyses performed; number of animals randomly allocated and analyzed; and the acknowledgement of the potential impact of LTF were recorded.

Results—In 81% (51/63) of trials, LTF were reported. In 80% (41/51) of those studies, losses in the analysis were ignored, and in only 18% (9/51) was the potential impact of LTF on study results acknowledged. Of the 47 studies in which sources of LTF were reported, 72% had loss of subjects because of investigator withdrawals, 30% because of deaths, and 26% because of owner withdrawals. Median loss of subjects for those studies was 12% because of investigator withdrawal (range, 2% to 52%), 8% because of death (1% to 28%), and 4% because of owner withdrawal (2% to 33%).

Conclusions and Clinical Relevance—Most RCTs had LTF, most of which were attributable to investigators removing randomly allocated animals from the study. In most studies, data from animal LTF were ignored and, therefore, only a subgroup of randomly allocated subjects was included in the data analysis. Most reports did not address the potential for a postrandomization selection bias associated with ignoring LTF and did not acknowledge the potential impact of the missingdata on their results.

Abstract

Objective—To determine the sources and handlingof losses to follow-up (LTF) in parallel-group randomized clinical trials (RCTs).

Sample Population—63 parallel-group RCTs of > 24 hours' duration published from January 2000 through December 2005.

Procedures—Journals were hand searched for eligible reports. Details concerning the presence, cause, and amount of LTF; statistical handlingof data missingbecause of LTF; type of analyses performed; number of animals randomly allocated and analyzed; and the acknowledgement of the potential impact of LTF were recorded.

Results—In 81% (51/63) of trials, LTF were reported. In 80% (41/51) of those studies, losses in the analysis were ignored, and in only 18% (9/51) was the potential impact of LTF on study results acknowledged. Of the 47 studies in which sources of LTF were reported, 72% had loss of subjects because of investigator withdrawals, 30% because of deaths, and 26% because of owner withdrawals. Median loss of subjects for those studies was 12% because of investigator withdrawal (range, 2% to 52%), 8% because of death (1% to 28%), and 4% because of owner withdrawal (2% to 33%).

Conclusions and Clinical Relevance—Most RCTs had LTF, most of which were attributable to investigators removing randomly allocated animals from the study. In most studies, data from animal LTF were ignored and, therefore, only a subgroup of randomly allocated subjects was included in the data analysis. Most reports did not address the potential for a postrandomization selection bias associated with ignoring LTF and did not acknowledge the potential impact of the missingdata on their results.