Recruitment bias resulted in poorer overall health status in a community-based control group

K. S. M. Taylor, J. C. Gordon, C. E. Harris, C. E. Counsell

Research output: Contribution to journalArticle

11 Citations (Scopus)

Abstract

Objective: During a prospective community-based incidence study of parkinsonism, a control group was recruited for comparison with the incident patients. This study compared the demographic and health status of recruited vs. nonrecruited controls.

Study Design and Setting: For each incident patient. attempts were made to recruit an age-gender matched control from the same general practice or, failing that, from a previously identified community cohort of people aged over 64 years who had expressed an interest in taking part in future research. Recruited controls were compared with those who were approached but not recruited in terms of age, socioeconomic status, gender. several measures of health status, and survival.

Results: A total of 74 controls (40%) were recruited Out Of 186 potential controls who were approached. Recruited controls scored slightly worse than nonrecruited controls on every measure of health status. which reached statistical significance for numbers of acute prescriptions and major surgical procedures. There were no significant differences in age, gender, socioeconomic status. or survival.

Conclusion: The control cohort was affected by recruitment bias, which suggested that recruited controls had slightly poorer health compared to nonrecruited controls. This bias may reduce differences in health when comparisons are made between the controls and the parkinsonian patients. (c) 2008 Elsevier Inc. All rights reserved.

Original languageEnglish
Pages (from-to)890-895
Number of pages6
JournalJournal of Clinical Epidemiology
Volume61
Issue number9
DOIs
Publication statusPublished - Sep 2008

Keywords

  • control group
  • patient recruitment
  • selection bias
  • parkinsonian disorders
  • health status indicators
  • cohort studies
  • consent

Cite this

Recruitment bias resulted in poorer overall health status in a community-based control group. / Taylor, K. S. M.; Gordon, J. C.; Harris, C. E.; Counsell, C. E.

In: Journal of Clinical Epidemiology, Vol. 61, No. 9, 09.2008, p. 890-895.

Research output: Contribution to journalArticle

@article{9860d1ddccb54b1c84718be20999ae2e,
title = "Recruitment bias resulted in poorer overall health status in a community-based control group",
abstract = "Objective: During a prospective community-based incidence study of parkinsonism, a control group was recruited for comparison with the incident patients. This study compared the demographic and health status of recruited vs. nonrecruited controls.Study Design and Setting: For each incident patient. attempts were made to recruit an age-gender matched control from the same general practice or, failing that, from a previously identified community cohort of people aged over 64 years who had expressed an interest in taking part in future research. Recruited controls were compared with those who were approached but not recruited in terms of age, socioeconomic status, gender. several measures of health status, and survival.Results: A total of 74 controls (40{\%}) were recruited Out Of 186 potential controls who were approached. Recruited controls scored slightly worse than nonrecruited controls on every measure of health status. which reached statistical significance for numbers of acute prescriptions and major surgical procedures. There were no significant differences in age, gender, socioeconomic status. or survival.Conclusion: The control cohort was affected by recruitment bias, which suggested that recruited controls had slightly poorer health compared to nonrecruited controls. This bias may reduce differences in health when comparisons are made between the controls and the parkinsonian patients. (c) 2008 Elsevier Inc. All rights reserved.",
keywords = "control group, patient recruitment, selection bias, parkinsonian disorders, health status indicators, cohort studies, consent",
author = "Taylor, {K. S. M.} and Gordon, {J. C.} and Harris, {C. E.} and Counsell, {C. E.}",
year = "2008",
month = "9",
doi = "10.1016/j.jclinepi.2007.10.020",
language = "English",
volume = "61",
pages = "890--895",
journal = "Journal of Clinical Epidemiology",
issn = "0895-4356",
publisher = "Elsevier USA",
number = "9",

}

TY - JOUR

T1 - Recruitment bias resulted in poorer overall health status in a community-based control group

AU - Taylor, K. S. M.

AU - Gordon, J. C.

AU - Harris, C. E.

AU - Counsell, C. E.

PY - 2008/9

Y1 - 2008/9

N2 - Objective: During a prospective community-based incidence study of parkinsonism, a control group was recruited for comparison with the incident patients. This study compared the demographic and health status of recruited vs. nonrecruited controls.Study Design and Setting: For each incident patient. attempts were made to recruit an age-gender matched control from the same general practice or, failing that, from a previously identified community cohort of people aged over 64 years who had expressed an interest in taking part in future research. Recruited controls were compared with those who were approached but not recruited in terms of age, socioeconomic status, gender. several measures of health status, and survival.Results: A total of 74 controls (40%) were recruited Out Of 186 potential controls who were approached. Recruited controls scored slightly worse than nonrecruited controls on every measure of health status. which reached statistical significance for numbers of acute prescriptions and major surgical procedures. There were no significant differences in age, gender, socioeconomic status. or survival.Conclusion: The control cohort was affected by recruitment bias, which suggested that recruited controls had slightly poorer health compared to nonrecruited controls. This bias may reduce differences in health when comparisons are made between the controls and the parkinsonian patients. (c) 2008 Elsevier Inc. All rights reserved.

AB - Objective: During a prospective community-based incidence study of parkinsonism, a control group was recruited for comparison with the incident patients. This study compared the demographic and health status of recruited vs. nonrecruited controls.Study Design and Setting: For each incident patient. attempts were made to recruit an age-gender matched control from the same general practice or, failing that, from a previously identified community cohort of people aged over 64 years who had expressed an interest in taking part in future research. Recruited controls were compared with those who were approached but not recruited in terms of age, socioeconomic status, gender. several measures of health status, and survival.Results: A total of 74 controls (40%) were recruited Out Of 186 potential controls who were approached. Recruited controls scored slightly worse than nonrecruited controls on every measure of health status. which reached statistical significance for numbers of acute prescriptions and major surgical procedures. There were no significant differences in age, gender, socioeconomic status. or survival.Conclusion: The control cohort was affected by recruitment bias, which suggested that recruited controls had slightly poorer health compared to nonrecruited controls. This bias may reduce differences in health when comparisons are made between the controls and the parkinsonian patients. (c) 2008 Elsevier Inc. All rights reserved.

KW - control group

KW - patient recruitment

KW - selection bias

KW - parkinsonian disorders

KW - health status indicators

KW - cohort studies

KW - consent

U2 - 10.1016/j.jclinepi.2007.10.020

DO - 10.1016/j.jclinepi.2007.10.020

M3 - Article

VL - 61

SP - 890

EP - 895

JO - Journal of Clinical Epidemiology

JF - Journal of Clinical Epidemiology

SN - 0895-4356

IS - 9

ER -