Tailored meta-analysis: an investigation of the correlation between the test positive rate and prevalence

Research output: Contribution to journalArticlepeer-review

Standard

Harvard

APA

Vancouver

Author

Bibtex

@article{3882c524da6349c6b987eb2c9c6a8ac4,
title = "Tailored meta-analysis: an investigation of the correlation between the test positive rate and prevalence",
abstract = "Background and Objective: Meta-analysis may produce estimates that are unrepresentative of a test{\textquoteright}s performance in practice. Tailored meta-analysis (TMA) circumvents this by deriving an applicable region for the practice and selecting the studies compatible with the region. It requires the test positive rate, r and prevalence, p being estimated for the setting but previous studies have assumed their independence. The aim is to investigate the effects a correlation between r and p has on estimating the applicable region and how this affects TMA. Methods: Six methods for estimating 99% confidence intervals (CI) for r and p were investigated: Wilson{\textquoteright}s 6 Bonferroni correction, Clopper-Pearson{\textquoteright}s 6 Bonferroni correction, and Hotelling{\textquoteright}s T2 statistic 6 continuity correction. These were analyzed in terms of the coverage probability using simulation trials over different correlations, sample sizes, and values for r and p. The methods were then applied to two published meta-analyses with associated practice data, and the effects on the applicable region, studies selected, and summary estimates were evaluated. Results: Hotelling{\textquoteright}s T2 statistic with a continuity correction had the highest median coverage (0.9971). This and the Clopper-Pearson method with a Bonferroni correction both had coverage consistently above 0.99. The coverage of Hotelling{\textquoteright}s CI{\textquoteright}s varied the least across different correlations. For both meta-analyses, the number of studies selected was largest when Hotelling{\textquoteright}s T2 statistic was used to derive the applicable region. In one instance, this increased the sensitivity by over 4% compared with TMA estimates using other methods. Conclusion: TMA returns estimates that are tailored to practice providing the applicable region is accurately defined. This is most likely when the CI for r and p are estimated using Hotelling{\textquoteright}s T2 statistic with a continuity correction. Potentially, the applicable region may be obtained using routine electronic health data. ",
keywords = "Data interpretation, Statistical, Decision making, Diagnosis tests, Routine, Mass screening, Meta-analysis, Models, Statistical 1. Introduction It is of interest to policy makers and clinicians to ensure that the",
author = "Willis, {Brian H.} and Dyuti Coomar and Mohammed Baragilly",
note = "Copyright {\textcopyright} 2018 The Authors. Published by Elsevier Inc. All rights reserved.",
year = "2019",
month = feb,
doi = "10.1016/j.jclinepi.2018.09.013",
language = "English",
volume = "106",
pages = "1--9",
journal = "Journal of Clinical Epidemiology",
issn = "0895-4356",
publisher = "Elsevier",

}

RIS

TY - JOUR

T1 - Tailored meta-analysis

T2 - an investigation of the correlation between the test positive rate and prevalence

AU - Willis, Brian H.

AU - Coomar, Dyuti

AU - Baragilly, Mohammed

N1 - Copyright © 2018 The Authors. Published by Elsevier Inc. All rights reserved.

PY - 2019/2

Y1 - 2019/2

N2 - Background and Objective: Meta-analysis may produce estimates that are unrepresentative of a test’s performance in practice. Tailored meta-analysis (TMA) circumvents this by deriving an applicable region for the practice and selecting the studies compatible with the region. It requires the test positive rate, r and prevalence, p being estimated for the setting but previous studies have assumed their independence. The aim is to investigate the effects a correlation between r and p has on estimating the applicable region and how this affects TMA. Methods: Six methods for estimating 99% confidence intervals (CI) for r and p were investigated: Wilson’s 6 Bonferroni correction, Clopper-Pearson’s 6 Bonferroni correction, and Hotelling’s T2 statistic 6 continuity correction. These were analyzed in terms of the coverage probability using simulation trials over different correlations, sample sizes, and values for r and p. The methods were then applied to two published meta-analyses with associated practice data, and the effects on the applicable region, studies selected, and summary estimates were evaluated. Results: Hotelling’s T2 statistic with a continuity correction had the highest median coverage (0.9971). This and the Clopper-Pearson method with a Bonferroni correction both had coverage consistently above 0.99. The coverage of Hotelling’s CI’s varied the least across different correlations. For both meta-analyses, the number of studies selected was largest when Hotelling’s T2 statistic was used to derive the applicable region. In one instance, this increased the sensitivity by over 4% compared with TMA estimates using other methods. Conclusion: TMA returns estimates that are tailored to practice providing the applicable region is accurately defined. This is most likely when the CI for r and p are estimated using Hotelling’s T2 statistic with a continuity correction. Potentially, the applicable region may be obtained using routine electronic health data.

AB - Background and Objective: Meta-analysis may produce estimates that are unrepresentative of a test’s performance in practice. Tailored meta-analysis (TMA) circumvents this by deriving an applicable region for the practice and selecting the studies compatible with the region. It requires the test positive rate, r and prevalence, p being estimated for the setting but previous studies have assumed their independence. The aim is to investigate the effects a correlation between r and p has on estimating the applicable region and how this affects TMA. Methods: Six methods for estimating 99% confidence intervals (CI) for r and p were investigated: Wilson’s 6 Bonferroni correction, Clopper-Pearson’s 6 Bonferroni correction, and Hotelling’s T2 statistic 6 continuity correction. These were analyzed in terms of the coverage probability using simulation trials over different correlations, sample sizes, and values for r and p. The methods were then applied to two published meta-analyses with associated practice data, and the effects on the applicable region, studies selected, and summary estimates were evaluated. Results: Hotelling’s T2 statistic with a continuity correction had the highest median coverage (0.9971). This and the Clopper-Pearson method with a Bonferroni correction both had coverage consistently above 0.99. The coverage of Hotelling’s CI’s varied the least across different correlations. For both meta-analyses, the number of studies selected was largest when Hotelling’s T2 statistic was used to derive the applicable region. In one instance, this increased the sensitivity by over 4% compared with TMA estimates using other methods. Conclusion: TMA returns estimates that are tailored to practice providing the applicable region is accurately defined. This is most likely when the CI for r and p are estimated using Hotelling’s T2 statistic with a continuity correction. Potentially, the applicable region may be obtained using routine electronic health data.

KW - Data interpretation

KW - Statistical

KW - Decision making

KW - Diagnosis tests

KW - Routine

KW - Mass screening

KW - Meta-analysis

KW - Models

KW - Statistical 1. Introduction It is of interest to policy makers and clinicians to ensure that the

UR - http://www.scopus.com/inward/record.url?scp=85055522419&partnerID=8YFLogxK

U2 - 10.1016/j.jclinepi.2018.09.013

DO - 10.1016/j.jclinepi.2018.09.013

M3 - Article

C2 - 30278213

VL - 106

SP - 1

EP - 9

JO - Journal of Clinical Epidemiology

JF - Journal of Clinical Epidemiology

SN - 0895-4356

ER -