A hierarchy of patient-reported outcome measures for meta-analysis of knee osteoarthritis trials: Empirical evidence from a survey of high impact journals

Carsten Bogh Juhl, Hans Lund, GH Guyatt, Weiya Zhang, Ewa M. Roos, Robin Daniel Kjersgaard Christensen

Research output: Contribution to journalConference abstract in journalResearchpeer-review

Abstract

Title
A hierarchy of patient-reported outcome measures for meta-analysis of knee osteoarthritis trials: empirical evidence from a survey of high impact journals

Objective
To develop a prioritized list for extracting patient-reported outcomes (PROs) measuring pain and disability for meta-analyses in knee osteoarthritis (OA).

Methods
A systematic literature search was conducted in high impact factor journals. Eligible were randomized controlled trials, using two or more PROs measuring pain or disability. A prioritized list was developed based on the capacity to discriminate between intervention and control treatment and how often the PROs were used in the included trials.

Results
The literature search identified 402 hits and 38 trials were included. Thirty-five of these trials had sufficient data on at least two PROs on pain and 15 had sufficient data on disability. The most frequently used composite score for both pain and disability were the WOMAC subscales, used in 77% and 100% of the included trials, respectively. These subscales were the most responsive composite score in 83% and 62% times, when used respectively. The following list was developed:
Pain: 1. WOMAC subscale ‘pain’; 2. Pain during activity (VAS); 3. Pain during walking (VAS); 4. General knee pain (VAS); 5. Pain at rest (VAS); 6. Other composite pain scales; 7. Other single item measures.
Disability: 1. WOMAC (subscale ‘function’); 2. SF-36 (subscale ‘physical function’); 3. Physical composite score – (SF-36); 4. Other composite disability scores.

Conclusions
As choosing the most favorable PROs from individual trials can overestimate the effect compared to a systematic approach, using a prioritized list as presented in this study is recommended to reduce reviewer's likelihood of biased selection of PROs in meta-analyses.
Original languageEnglish
JournalOsteoarthritis and Cartilage
Volume18
Issue numberSupplement 2
Pages (from-to)S 130
ISSN1063-4584
Publication statusPublished - 2010
EventOARSI World Congress on Osteoarthritis - Bruxelles, Belgium
Duration: 23. Sep 201026. Sep 2010

Conference

ConferenceOARSI World Congress on Osteoarthritis
CountryBelgium
CityBruxelles
Period23/09/201026/09/2010

Cite this

@article{c4f63f222b2c4d91894c24e0045320c3,
title = "A hierarchy of patient-reported outcome measures for meta-analysis of knee osteoarthritis trials: Empirical evidence from a survey of high impact journals",
abstract = "Title A hierarchy of patient-reported outcome measures for meta-analysis of knee osteoarthritis trials: empirical evidence from a survey of high impact journals Objective To develop a prioritized list for extracting patient-reported outcomes (PROs) measuring pain and disability for meta-analyses in knee osteoarthritis (OA). Methods A systematic literature search was conducted in high impact factor journals. Eligible were randomized controlled trials, using two or more PROs measuring pain or disability. A prioritized list was developed based on the capacity to discriminate between intervention and control treatment and how often the PROs were used in the included trials. Results The literature search identified 402 hits and 38 trials were included. Thirty-five of these trials had sufficient data on at least two PROs on pain and 15 had sufficient data on disability. The most frequently used composite score for both pain and disability were the WOMAC subscales, used in 77{\%} and 100{\%} of the included trials, respectively. These subscales were the most responsive composite score in 83{\%} and 62{\%} times, when used respectively. The following list was developed: Pain: 1. WOMAC subscale ‘pain’; 2. Pain during activity (VAS); 3. Pain during walking (VAS); 4. General knee pain (VAS); 5. Pain at rest (VAS); 6. Other composite pain scales; 7. Other single item measures. Disability: 1. WOMAC (subscale ‘function’); 2. SF-36 (subscale ‘physical function’); 3. Physical composite score – (SF-36); 4. Other composite disability scores. Conclusions As choosing the most favorable PROs from individual trials can overestimate the effect compared to a systematic approach, using a prioritized list as presented in this study is recommended to reduce reviewer's likelihood of biased selection of PROs in meta-analyses.",
keywords = "Knee Osteoarthritis, Patient-Reported Outcome,, Meta-analysis, Prioritized list",
author = "Juhl, {Carsten Bogh} and Hans Lund and GH Guyatt and Weiya Zhang and Roos, {Ewa M.} and Christensen, {Robin Daniel Kjersgaard}",
year = "2010",
language = "English",
volume = "18",
pages = "S 130",
journal = "Osteoarthritis and Cartilage",
issn = "1063-4584",
publisher = "Elsevier",
number = "Supplement 2",

}

A hierarchy of patient-reported outcome measures for meta-analysis of knee osteoarthritis trials : Empirical evidence from a survey of high impact journals. / Juhl, Carsten Bogh; Lund, Hans ; Guyatt, GH; Zhang, Weiya; Roos, Ewa M.; Christensen, Robin Daniel Kjersgaard.

In: Osteoarthritis and Cartilage, Vol. 18, No. Supplement 2, 2010, p. S 130.

Research output: Contribution to journalConference abstract in journalResearchpeer-review

TY - ABST

T1 - A hierarchy of patient-reported outcome measures for meta-analysis of knee osteoarthritis trials

T2 - Empirical evidence from a survey of high impact journals

AU - Juhl, Carsten Bogh

AU - Lund, Hans

AU - Guyatt, GH

AU - Zhang, Weiya

AU - Roos, Ewa M.

AU - Christensen, Robin Daniel Kjersgaard

PY - 2010

Y1 - 2010

N2 - Title A hierarchy of patient-reported outcome measures for meta-analysis of knee osteoarthritis trials: empirical evidence from a survey of high impact journals Objective To develop a prioritized list for extracting patient-reported outcomes (PROs) measuring pain and disability for meta-analyses in knee osteoarthritis (OA). Methods A systematic literature search was conducted in high impact factor journals. Eligible were randomized controlled trials, using two or more PROs measuring pain or disability. A prioritized list was developed based on the capacity to discriminate between intervention and control treatment and how often the PROs were used in the included trials. Results The literature search identified 402 hits and 38 trials were included. Thirty-five of these trials had sufficient data on at least two PROs on pain and 15 had sufficient data on disability. The most frequently used composite score for both pain and disability were the WOMAC subscales, used in 77% and 100% of the included trials, respectively. These subscales were the most responsive composite score in 83% and 62% times, when used respectively. The following list was developed: Pain: 1. WOMAC subscale ‘pain’; 2. Pain during activity (VAS); 3. Pain during walking (VAS); 4. General knee pain (VAS); 5. Pain at rest (VAS); 6. Other composite pain scales; 7. Other single item measures. Disability: 1. WOMAC (subscale ‘function’); 2. SF-36 (subscale ‘physical function’); 3. Physical composite score – (SF-36); 4. Other composite disability scores. Conclusions As choosing the most favorable PROs from individual trials can overestimate the effect compared to a systematic approach, using a prioritized list as presented in this study is recommended to reduce reviewer's likelihood of biased selection of PROs in meta-analyses.

AB - Title A hierarchy of patient-reported outcome measures for meta-analysis of knee osteoarthritis trials: empirical evidence from a survey of high impact journals Objective To develop a prioritized list for extracting patient-reported outcomes (PROs) measuring pain and disability for meta-analyses in knee osteoarthritis (OA). Methods A systematic literature search was conducted in high impact factor journals. Eligible were randomized controlled trials, using two or more PROs measuring pain or disability. A prioritized list was developed based on the capacity to discriminate between intervention and control treatment and how often the PROs were used in the included trials. Results The literature search identified 402 hits and 38 trials were included. Thirty-five of these trials had sufficient data on at least two PROs on pain and 15 had sufficient data on disability. The most frequently used composite score for both pain and disability were the WOMAC subscales, used in 77% and 100% of the included trials, respectively. These subscales were the most responsive composite score in 83% and 62% times, when used respectively. The following list was developed: Pain: 1. WOMAC subscale ‘pain’; 2. Pain during activity (VAS); 3. Pain during walking (VAS); 4. General knee pain (VAS); 5. Pain at rest (VAS); 6. Other composite pain scales; 7. Other single item measures. Disability: 1. WOMAC (subscale ‘function’); 2. SF-36 (subscale ‘physical function’); 3. Physical composite score – (SF-36); 4. Other composite disability scores. Conclusions As choosing the most favorable PROs from individual trials can overestimate the effect compared to a systematic approach, using a prioritized list as presented in this study is recommended to reduce reviewer's likelihood of biased selection of PROs in meta-analyses.

KW - Knee Osteoarthritis

KW - Patient-Reported Outcome,

KW - Meta-analysis

KW - Prioritized list

M3 - Conference abstract in journal

VL - 18

SP - S 130

JO - Osteoarthritis and Cartilage

JF - Osteoarthritis and Cartilage

SN - 1063-4584

IS - Supplement 2

ER -