PLOS One published and subsequently retracted the article titled “Effect of sensory art therapies on root canal treatment anxiety and high dental anxiety in adults: A systematic review with meta-analysis.” The retraction notice was issued by the PLOS One Editors and published with a DOI. The notice lists a sequence of methodological concerns identified after publication that together prompted independent expert and statistical advice and led the Editors to conclude that the study’s validity and reliability were in question.
A central concern was the review’s use of the umbrella term sensory art therapies to group a wide range of highly variable interventions. The review combined interventions including virtual reality, aromatherapy, music therapy, chromotherapy (color therapy), as well as yoga and aerobic exercise. The Editors and external advisors noted that this heterogeneity makes it unclear which intervention type, if any, is responsible for observed effects. Pooling such diverse modalities undermines the interpretability of pooled effect estimates and weakens the validity of the meta-analysis.
The review’s Methods specified participant eligibility within a PICOS framework as adult patients with symptoms related to dental root canal treatment, high dental anxiety, or dental phobia. The retraction notice identifies at least one included study that did not meet these predefined participant criteria: that study enrolled patients undergoing routine dental care and included a range of dental anxiety levels from none or low to high, rather than exclusively adults undergoing root canal treatment or adults with high dental anxiety or phobia. This discrepancy between eligibility criteria and included populations raises concerns about selection and the applicability of pooled results.
The Methods also defined eligible interventions as art therapy, including sensory-based art therapies targeting the five senses. Two studies included in the review evaluated interventions outside that scope—specifically, aerobic exercise and yoga. The retraction notice indicates those two studies did not meet the predefined intervention (I) criterion. The notice clarifies that these two studies were not included in the meta-analysis, but their presence in the systematic review itself exposed inconsistencies in how inclusion criteria were applied.
The authors conducted eight outcome-specific meta-analyses. Each pooled analysis relied on a small number of studies (reported n = 2, 4, 2, 4, 3, 6, 6, and 3 for the eight outcomes). The Editors noted that meta-analyses based on few studies are less robust, and the concern was compounded by inclusion of multiple subgroups from the same studies across six outcomes. Repeated subgroup entries from the same trial can disproportionately influence pooled estimates and reduce the reliability of conclusions drawn from the meta-analyses.
The retraction notice documents a discrepancy in the GRADE rating for heart rate (HR) between Fig 5 (moderate) and Table 5 (high). Editorial follow-up with the first author revealed that the authors had upgraded the certainty of evidence for some outcomes (visual analogue scale [VAS], HR, and salivary cortisol [SC]) on the basis of effect size magnitude. According to the Cochrane GRADE guidance cited by the editors, upgrading certainty on that basis is not permitted when included studies are randomized controlled trials. Correcting for inappropriate upgrading alters the reported certainty range: the review’s claim of “very low to high” certainty would instead read “very low to low.”
One included study was entered as four separate entries in the meta-analyses for both systolic and diastolic blood pressure, each compared against the same 30-person control group. The first author stated these entries correspond to two intervention groups (Chinese and Western music), each measured at two appointments within a single root canal treatment course. Treating repeated measurements on the same individuals—compared repeatedly against the same control group—as independent data points violates the independence assumption of meta-analysis. This practice inflates the effective sample size and the weight attributed to that study in pooled estimates, biasing the meta-analytic results.
Following identification of these methodological issues, the PLOS One Editors consulted an independent expert and a Statistical Advisor with expertise in systematic reviews and meta-analyses. Both advisors agreed with the Editors that the concerns compromised the validity and reliability of the published results. Based on this assessment, the PLOS One Editors retracted the article.
Three authors (XL, SG, and GH) responded to the retraction notice but did not state whether they agreed or disagreed with the editorial decision. One author (JW) either did not respond directly or could not be reached. The retraction notice cites the original (retracted) article and the specific studies implicated in the methodological concerns. The notice and citation are publicly available in the PLOS One record and are distributed under an open license.
The retraction indicates that conclusions from the withdrawn systematic review and meta-analysis should not be used to guide clinical practice or policy. The primary methodological problems identified—heterogeneous intervention grouping, breaches of predefined PICOS criteria, inappropriate handling of repeated measures, small numbers of studies in pooled analyses, and incorrect application of GRADE—are core elements of systematic review validity. Future syntheses on nonpharmacologic interventions for dental anxiety should apply clear, consistent eligibility criteria, avoid pooling highly dissimilar interventions without appropriate subgrouping, correctly account for repeated measures and nonindependence, and adhere to accepted GRADE methods for rating certainty of evidence.