Methodological quality of systematic reviews on neck pain treatments: a cross-sectional appraisal using AMSTAR 2
Leonard Ho1, Ka Wai Yuen1, Cyrus Chung Ching Choi1
1Jockey Club School of Public Health and Primary Care, Faculty of Medicine, The Chinese University of Hong Kong, Shatin, New Territories, Hong Kong.
Objective:
This critical appraisal study aimed to evaluate the methodological quality of recent systematic reviews on neck pain treatments.
Introduction:
Systematic reviews offer the best evidence on neck pain treatments. However, the quality of reviews can vary, and methodological flaws can limit their reliability in clinical and policy decision-making.
Eligibility Criteria:
We included systematic reviews with at least 1 meta-analysis on randomized controlled trials on treatments for neck pain. Reviews evaluating pharmacological, non-pharmacological, or combined pharmacological and non-pharmacological treatments were eligible. Systematic reviews on the treatment of neck pain complications were excluded, as were editorials, protocols, systematic reviews without randomized controlled trials, and network meta-analyses.
Methods:
We searched MEDLINE (Ovid), Embase (Ovid), PsycINFO (Ovid), and the Cochrane Database of Systematic Reviews for eligible systematic reviews on neck pain treatments published between 2019 and 2024 in English or Chinese. We extracted bibliographical data and critically appraised the methodological quality using the AMSTAR 2 (A MeaSurement Tool to Assess systematic Reviews) tool. Logistic regression analyses were conducted to evaluate the predictability of bibliographical characteristics for overall methodological quality.
Results:
Of the 63 systematic reviews, 2 (3.2%) were high quality, 2 (3.2%) were moderate, 5 (7.9%) were low, and 54 (85.7%) were critically low. The reviews performed poorly across 4 critical domains. The majority of the reviews (n=53; 84.1%) did not provide lists of excluded studies; 29 (46.0%) did not account for the risk of bias in primary studies when interpreting the results; 38 (60.3%) did not investigate and discuss publication bias, and 26 (41.3%) did not publish review protocols. While 6 bibliographical characteristics showed significant associations with overall methodological quality in univariate logistic regression, no associations were observed after adjusting for confounders in the subsequent multivariate analysis.
Conclusions:
The methodological quality of systematic reviews on neck pain treatments is unsatisfactory. Future reviewers should pay attention to the aforementioned critical domains. Resources should also be allocated for upscaling the infrastructure of evidence synthesis and improving the critical appraisal skills of all evidence users.
