Acute rehabilitation following traumatic anterior shoulder dislocation (ARTISAN): pragmatic, multicentre, randomised controlled trial.
Kearney RS, Ellard DR, Parsons H, Haque A, Mason J, Nwankwo H, Bradley H, Drew S, Modi C, Bush H, Torgerson D, Underwood M, The ARTISAN collaborators, ARTISAN collaborators
- DOI
- 10.1136/bmj-2023-076925
- Record issued
- 2026-08-16
- Engine
- 7.39.0
- Exported
- 2026-09-21
Prepared by Alpha1. This document is confidential: it is intended for the recipient it was shared with and must not be redistributed. The live record at alpha1science.com/verify/1c9218b5-a8d7-4d44-bcd2-02989ad96cbe is authoritative.
How this rating was calculated
- IntegrityIntegrity concern ×2−1★
- ReportingData & code availability partially met−0.25★
- The numeric-impossibility checks (GRIM/GRIMMER/DEBIT/SPRITE) did not run: 16 reported means were read, and their group size is not stated where the values are printed (this source has no machine-readable table structure). These checks need the count the mean was averaged over, so none was performed.
- No data or code availability links were detected to verify.
This Kaimen Rigor review uses Kaimen Rigor reviewers trained on a curated corpus of high-fidelity and retracted papers, with expert supervision and curation. It can still make mistakes; verify each finding against the source before relying on it.
This is a well-conducted and well-reported pragmatic RCT with rigorous randomisation, allocation concealment, pre-specified analysis, and appropriate handling of missing data. The main weaknesses are a vague data availability statement and minor reporting gaps (no explicit CONSORT statement, minor copyedit issues).
Both reviewers agreed on study type (interventional) and on all dimension statuses. The only divergence was on the limitations addressed sub-criterion in scientific premise (adequate vs. reported but inadequate) and reporting guideline (reported but inadequate vs. not reported), both resolved by weighing the evidence. Statistics verification covered only 5 tests (all consistent); the rest are unverified. Citation check found no retracted or unresolved references.
Numerical inconsistencies
1 finding · worst lowValues that contradict each other or are impossible for the stated sample: recomputed p-values and test statistics, GRIM/GRIMMER checks on summary numbers, percentages against their own counts, totals against their parts, and estimates against their own confidence intervals.
- Internal contradictions in the reported numbersAssessed
Recomputed 5 tests: 5 consistent, 0 inconsistent; 5 via agent-written checks.
- CONSISTENTreported p = .110 · recomputed p = .122Reviewer 1Primary outcome between-group difference at 6 months (adjusted) p-value from CI
“6 months | 180 | 36.2 (10.7) | 174 | 38.4 (9.2) | 1.5 (−0.3 to 3.5) | 0.11”
Taken as given: The estimate is 1.5 and the 95% CI is -0.3 to 3.5.; The CI is two-sided at 95%.; The p-value is for the adjusted analysis.Method: Recomputed p-value from the estimate and 95% CI using the normal approximation.How we recomputed it: pCI(1.5, -0.3, 3.5, 0) - CONSISTENTreported p = .870 · recomputed p = .874Reviewer 1Secondary outcome complication: torn rotator cuff comparison using Fisher's exact test
“Torn rotator cuff | 22 (9) | 21 (9) | 0.87”
Taken as given: The numbers 22 and 21 are the event counts in each group.; The group totals are 240 and 242, so non-events are 218 and 221.; The test used is Fisher's exact test (two-tailed).Method: Recomputed two-tailed Fisher's exact test from the 2x2 table.How we recomputed it: pFisher2x2(22, 218, 21, 221, 0) - CONSISTENTreported p = .220 · recomputed p = .220Reviewer 1Secondary outcome complication: shoulder re-dislocation comparison using Fisher's exact test
“Shoulder re-dislocation | 7 (3) | 3 (1) | 0.22”
Taken as given: The numbers 7 and 3 are the event counts in each group.; The group totals are 240 and 242, so non-events are 233 and 239.; The test used is Fisher's exact test (two-tailed).Method: Recomputed two-tailed Fisher's exact test from the 2x2 table.How we recomputed it: pFisher2x2(7, 233, 3, 239, 0) - CONSISTENTreported p = .110 · recomputed p = .122Reviewer 2Check p-value for 6-month primary outcome between-group difference using CI-based method.
“6 months | 180 | 36.2 (10.7) | 174 | 38.4 (9.2) | 1.5 (−0.3 to 3.5) | 0.11”
Taken as given: The estimate is 1.5 and the 95% CI is -0.3 to 3.5.; The CI is two-sided at 95%.; The p-value is for the between-group difference from the adjusted mixed model.Method: Recomputed p-value from the estimate and 95% CI using the normal approximation for a continuous outcome.How we recomputed it: pCI(1.5, -0.3, 3.5, 0) - CONSISTENTreported p = .310 · recomputed p = .301Reviewer 2Check p-value for 6-week QuickDASH between-group difference using CI-based method.
“6 weeks | 154 | 32.8 (23.2) | 168 | 27.6 (21.4) | −1.9 (−5.5 to 1.7) | 0.31”
Taken as given: The estimate is -1.9 and the 95% CI is -5.5 to 1.7.; The CI is two-sided at 95%.; The p-value is for the between-group difference from the adjusted mixed model.Method: Recomputed p-value from the estimate and 95% CI using the normal approximation.How we recomputed it: pCI(-1.9, -5.5, 1.7, 0)
- lowinternal contradictionIn Table 3, the 12-month EQ-5D-5L confidence interval is reported as -0.26 to 0.047, which is inconsistent with the point estimate 0.009 and the other CIs in the table (which have two decimal places). This appears to be a typographical error.
“0.009 (−0.26 to 0.047)”
Table 3Find in source - lowinternal contradictionThe abstract states 482 participants were recruited from 40 sites, but the methods say 41 UK NHS Trust sites were involved. The results clarify that screening occurred at 41 trusts but recruitment from 40, which is consistent.
482 participants were recruited from 40 sites in the UK. (Abstract) ... Trauma research teams at 41 UK NHS Trust sites screened adults ... (Methods)
Abstractreviewer’s wording
Overstated conclusions
None foundConclusions that reach past what the paper's own results support — including a significance claim that no longer holds when the statistic is recomputed, and efficacy resting on an unvalidated surrogate endpoint.
Checked — nothing surfaced.
7 major claims checked against the paper's own evidence: all adequately supported.
- supportedReviewer 1An additional programme of physiotherapy is not superior to advice alone for first-time shoulder dislocation.The primary outcome analysis shows no significant difference and the confidence interval excludes the predefined worthwhile difference, supporting the claim.Evidence: Primary outcome between-group difference 1.5 (95% CI -0.3 to 3.5) at 6 months, with CI excluding the 4-point target.
“An additional programme of current physiotherapy is not superior to advice, supporting materials, and the option to self-refer to physiotherapy.”
ConclusionFind in source - supportedReviewers 1, 2ARTISAN is the largest randomised controlled trial investigating two different rehabilitation approaches in adults with a first time traumatic shoulder dislocation.The paper states this and no contradicting evidence is presented; the trial recruited 482 participants, which is larger than the previously identified trial of 56.Evidence: Recruitment of 482 participants from 40 sites.
“ARTISAN is the largest randomised controlled trial investigating two different rehabilitation approaches in adults with a first time traumatic shoulder dislocation.”
DiscussionFind in source - supportedReviewer 1No significant differences in complications were found between the two groups.All complication p-values are >0.05, supporting the claim.Evidence: Table 4 shows p-values for each complication all >0.05.
“Complication profiles were similar across the two groups and no significant differences were noted.”
ResultsFind in source - supportedReviewer 1The additional physiotherapy programme was not a worthwhile benefit.The 95% confidence interval for the primary outcome excludes the prespecified 4-point difference, supporting the claim of no worthwhile benefit.Evidence: Primary outcome 95% CI -0.3 to 3.5, excluding 4.
“Furthermore, because the 95% confidence intervals of the estimate of effectiveness excludes the prespecified worthwhile difference of 4 points, the additional physiotherapy programme was not a worthwhile benefit.”
DiscussionFind in source - supportedReviewer 2An additional programme of current physiotherapy is not superior to advice, supporting materials, and the option to self-refer to physiotherapy.The primary outcome analysis shows no significant difference and the confidence interval excludes the prespecified worthwhile difference, supporting the claim.Evidence: Primary outcome between-group difference 1.5 (95% CI -0.3 to 3.5) at 6 months, with CI excluding the 4-point target.
“An additional programme of current physiotherapy is not superior to advice, supporting materials, and the option to self-refer to physiotherapy.”
ConclusionFind in source - supportedReviewer 2No significant differences were reported in the QuickDASH or consistent differences shown from the EQ-5D-5L.Secondary outcome tables show all p-values >0.05 and confidence intervals crossing zero, supporting the claim.Evidence: Table 3 shows all QuickDASH and EQ-5D-5L comparisons with p-values >0.05.
For secondary outcomes, no significant differences were reported in the QuickDASH or consistent differences shown from the EQ-5D-5L.
Resultsreviewer’s wording - supportedReviewer 2Complication profiles were similar across the two groups.Fisher's exact tests for each complication show no significant differences, supporting the claim.Evidence: Table 4 shows p-values >0.05 for all complications.
“Complication profiles were similar across the two groups and no significant differences were noted.”
ResultsFind in source
Data authenticity concerns
None foundAn adversarial read for patterns associated with data that may not be genuine: results that look too clean, implausibly large effects, duplicated data or images, and methods that do not match the results reported.
Checked — nothing surfaced.
Reporting gaps
1 finding · worst mediumRequired detail the manuscript never states — study design, biological variables, ethics approval and consent, key resources, statistical reporting, data and code availability, and overall transparency.
- Data/code availability incompleteAssessed
The introduction cites a 2019 population-based cohort and Cochrane reviews to establish the incidence and lack of evidence for rehabilitation methods. The rationale for comparing advice-only versus additional physiotherapy is clearly linked to the resource implications and the gap in evidence. The paper acknowledges the limitations of prior research, such as the 2014 Cochrane review finding no RCTs, and the 2019 update identifying only one ongoing study.
“A 2019 population based cohort of 16 763 people reported the incidence to be 40.4 per 100 000 person years in men and 15.5 in women.”
“Despite the important role of rehabilitation for this common condition, evidence comparing different rehabilitation methods after the initial period in a sling is scarce.”
“A 2014 Cochrane review on methods of non-operative management showed no randomised controlled trials comparing different rehabilitation methods after the initial two weeks of supporting the arm in a sling.”
“A 2019 population based cohort of 16 763 people reported the incidence to be 40.4 per 100 000 person years in men and 15.5 in women.”
“Considering the large personal and societal cost associated with this injury, the primary objective of the ARTISAN trial was to compare the clinical effectiveness of two rehabilitation interventions in adults with a first time traumatic shoulder dislocation.”
“Despite the important role of rehabilitation for this common condition, evidence comparing different rehabilitation methods after the initial period in a sling is scarce.”
Randomisation used a minimisation algorithm with a random element, stratified by centre, age, and arm dominance, with allocation concealment maintained by an independent team. Blinding of participants and clinicians was not possible, but this is acknowledged and the central team was masked. A power analysis was conducted, requiring 478 participants for 90% power. Inclusion/exclusion criteria were pre-specified, and the analysis population (ITT and per protocol) was defined. Outlier handling is addressed through multiple imputation for missing data.
“Allocation concealment was maintained by an independent randomisation team at Warwick Clinical Trials Unit who were responsible for generation of the sequence and had no role in participant recruitment.”
“As such, 478 participants were required to show a 4 point target difference (ie, a small, standardised mean difference of 0.3) at the 5% significance level, with 90% power, allowing a margin of 20% loss during follow-up.”
“Following randomisation, masking participants or treating clinicians to treatment allocation was not possible. However, the treating clinician and the participant were masked to treatment allocation during the initial advice session.”
“Allocation concealment was maintained by an independent randomisation team at Warwick Clinical Trials Unit who were responsible for generation of the sequence and had no role in participant recruitment.”
“Following randomisation, masking participants or treating clinicians to treatment allocation was not possible.”
“As such, 478 participants were required to show a 4 point target difference (ie, a small, standardised mean difference of 0.3) at the 5% significance level, with 90% power, allowing a margin of 20% loss during follow-up.”
Sex is reported (66% male), age is reported (mean 45 years), and demographics such as race/ethnicity, employment, and comorbidities are detailed in Table 1. Since both sexes are enrolled, a justification for single-sex is not applicable. Species/strain and housing conditions are not applicable for a human trial.
“Participants were mostly male (66%, n=317), with a mean age of 45 years.”
“Race and ethnicity/ancestry†: | | Asian | 18 (8) | 19 (8) | 37 (8) |”
“Participants were mostly male (66%, n=317), with a mean age of 45 years.”
“Race and ethnicity/ancestry†: | | Asian | 18 (8) | 19 (8) | 37 (8) |”
The study reports approval from the National Research Ethics Committee with a protocol number (18/WA/0236) and individual NHS Trust approvals. Written informed consent is described. Regulatory compliance is implied through adherence to NHS research governance.
“The national research ethics committee approved this study on 26 July 2018 (18/WA/0236), with each trial site granting individual NHS Trust approval before recruitment at each site.”
“After assessment, potential participants were provided with verbal and written information before they provided written informed consent.”
“The national research ethics committee approved this study on 26 July 2018 (18/WA/0236)”
“After assessment, potential participants were provided with verbal and written information before they provided written informed consent.”
The interventions (advice session and additional physiotherapy) are described with details on duration, content, and delivery. The statistical software R (version 4.0.3) is identified. No antibodies, cell lines, or other biological reagents are used, so those criteria are not applicable.
“Analyses were conducted in R (version 4.0.3)”
“Each additional session lasted for up to 30 min, over a maximum duration of four months from the date of randomisation.”
“All participants had an initial period in which the injured arm was supported in a sling and then received an appointment for a physiotherapy advice session within six weeks of their injury.”
“Analyses were conducted in R (version 4.0.3)”
The primary analysis used mixed effects linear regression, with adjustment for stratification variables. Exact p-values are reported in Tables 2 and 3. Effect sizes are reported with 95% confidence intervals. The software (R) is identified. Data presentation includes per-group n, means, and SDs. Mathematical plausibility checks were not performed due to continuous outcomes and large N, but no obvious inconsistencies were noted.
“Unadjusted and adjusted mixed effects linear regression models were used to estimate the between group difference.”
“6 months | 180 | 36.2 (10.7) | 174 | 38.4 (9.2) | 1.5 (−0.3 to 3.5) | 0.11”
“between group difference favouring physiotherapy 1.5 (95% confidence interval −0.3 to 3.5)”
“Unadjusted and adjusted mixed effects linear regression models were used to estimate the between group difference.”
“0.7 (−1.0 to 2.4) | 0.44”
“between group difference favouring physiotherapy 1.5 (95% confidence interval −0.3 to 3.5)”
The data availability statement says requests should be submitted to an email address for consideration, but does not specify conditions or a timeframe. This is a managed-access route but lacks detail, making it 'reported_but_inadequate'. No repository deposit or accession numbers are provided, which is acceptable for patient-level data, but the statement could be more concrete.
“All data requests should be submitted to WCTUDataAccess@warwick.ac.uk for consideration. Access to anonymised data may be granted following review.”
“All data requests should be submitted to WCTUDataAccess@warwick.ac.uk for consideration. Access to anonymised data may be granted following review.”
The trial is registered with ISRCTN63184243. Methods are detailed enough for replication. Limitations are discussed, including loss to follow-up. Conclusions are proportional to the results, noting no superiority of additional physiotherapy. Funding and competing interests are declared.
“Trial registration Current Controlled Trials ISRCTN63184243.”
“The main limitation was the 27% loss to follow up, however, the observed standard deviation was much smaller than anticipated.”
“Funding: This trial was funded by the National Institute for Health and Care Research (16/167/56), 01/06/18.”
“Trial registration Current Controlled Trials ISRCTN63184243.”
“The main limitation was the 27% loss to follow up”
“This trial was funded by the National Institute for Health and Care Research (16/167/56), 01/06/18.”
Registered (1 ID: ISRCTN). No reporting guideline cited.
Broken references and links
None found · partly checkedReferences checked against Crossref, OpenAlex and Retraction Watch for retractions and resolvability, plus declared data and code links probed for whether they resolve to content matching the paper.
Nothing surfaced — but not everything feeding this category ran (missing: data/code link verification), so read this as a partial clean bill.
Checked 15 references by DOI: 15 verified.
Every extracted reference resolved against Crossref/OpenAlex with no retraction flags.
Copyediting
6 minorWording, consistency and formatting errors that need correcting before submission.
No major wording or formatting errors. 6 minor suggestions below.
6 copyedit issues flagged: mostly typo, consistency, grammar.
- MINORtypoAcknowledgments“Fracnesca Gowing”→ Francesca GowingMisspelling of a collaborator's name.
- MINORconsistencyTable 3“0.009 (−0.26 to 0.047)”→ 0.009 (−0.026 to 0.047)Inconsistent decimal places in confidence interval.
- MINORconsistencyResults, Participants and adherence“81% (n=194) of 240 participants receiving advice only”→ 81% (n=194) of 240 participants received advice onlyVerb tense inconsistency.
- MINORgrammarDiscussion, Principal findings“We did not report a difference in the mean primary outcome”→ We did not find a difference in the mean primary outcomeAwkward phrasing.
- MINORtypoAcknowledgments“Lynn Osbome”→ Lynn OsborneMisspelling of a collaborator's name.
- MINORgrammarDiscussion, Comparison with other studies“The authors reported that adherence was low in the supervised physiotherapy group, with only 43% of participants compliant (n=12), but a significant improvement was noted in the primary outcome (western Ontario shoulder instability index) from supervised exercise when compared with home exercises.”→ Consider splitting the long sentence for clarity.Long sentence may reduce readability.
The published work is robust and trustworthy; an informed reader should weigh the minor data-access vagueness and the absence of an explicit CONSORT statement as minor transparency limitations, not validity threats. No erratum is warranted for the copyedit issues, though the Table 3 CI typo could be corrected in a future erratum.
- 1.HIGHdata codeIn the Data availability statement, specify the conditions and timeframe for data access, e.g., 'Requests will be reviewed within 30 days and data will be shared after approval and a data sharing agreement.'The current statement is vague and does not meet common transparency expectations for a data-driven trial.
- 2.HIGHreportingIn the Methods or Acknowledgments, explicitly state adherence to the CONSORT reporting guideline.The paper follows CONSORT-like structure but does not explicitly confirm adherence, which is a standard expectation for RCTs.
- 3.MEDIUMreportingIn the Methods, add a link to the statistical analysis plan (Supplement 1) for transparency.Providing a direct link improves reproducibility and reviewer confidence.
- 4.MEDIUMreportingIn the Methods, clarify the role of the data monitoring committee and trial steering committee.Enhances governance transparency for a multicentre trial.
- 5.MEDIUMcopyeditIn Table 3, correct the EQ-5D-5L confidence interval from '0.009 (−0.26 to 0.047)' to '0.009 (−0.026 to 0.047)'.The current CI has inconsistent decimal places and appears to be a typographical error.
- 6.MEDIUMcopyeditIn the Acknowledgments, correct the misspelled names 'Fracnesca Gowing' to 'Francesca Gowing' and 'Lynn Osbome' to 'Lynn Osborne'.Misspelling collaborator names is unprofessional and should be fixed.
- 7.LOWcopyeditIn Results, Participants and adherence, change '81% (n=194) of 240 participants receiving advice only' to '81% (n=194) of 240 participants received advice only'.Verb tense inconsistency reduces readability.
- 8.LOWcopyeditIn Discussion, Principal findings, change 'We did not report a difference' to 'We did not find a difference'.Awkward phrasing; 'find' is more natural.
- 9.LOWcopyeditIn Discussion, Comparison with other studies, consider splitting the long sentence about the supervised physiotherapy group for clarity.Long sentences reduce readability.
- 10.LOWdata codeConsider depositing the statistical analysis plan and any non-identifiable aggregate data in a public repository (e.g., Zenodo) with a DOI.Would further improve transparency and reproducibility.
- 11.LOWdata codeInclude a statement on the availability of the statistical analysis code, even if on request.Clarifies code availability and supports reproducibility.
The star rating is the report’s one-glance summary. Every paper starts at 5★ and loses stars for the concrete problems the review finds — so a rating is never a vague average, it’s a running total you can read line by line under “How this rating was calculated.”
- Reporting — 8 dimensionseach dimension that fully fails−½★
- each dimension partially met−¼★
- Statistics · Integrity · Claimseach serious problem−1★
- each medium problem−½★
- Citationseach retracted or unverifiable reference−¼★
- Copyeditonly when the manuscript needs a full edit−½★
The rating never drops below 1★, and a demonstrable critical failure (an impossible statistic, a proven ethics violation) caps it at 1★ on its own — so the stars can never look healthy when the verdict is CRITICAL.
The rating draws on a panel of agents. Three independent Kaimen Rigor reviewers grade the eight dimensions below across several independent passes (the shown verdict is their majority vote — steadier than any single run), isolate the paper’s major claims and check its own evidence backs them, and flag integrity concerns. Alongside them, a citation agent resolves every reference against Crossref, OpenAlex, and Retraction Watch; a statistics agent recomputes reported tests; and rule-based checks verify that declared data/code links actually resolve. Full text is required — an abstract-only submission is not analyzed.
Graded against NIH, MDAR, ARRIVE 2.0, CONSORT, EQUATOR, and RRID guidelines. A dimension that doesn’t apply to the study type is skipped, never penalized.
This Kaimen Rigor review is model-assisted and is not a substitute for formal expert review. It complements human evaluation by surfacing potential methodological concerns — verify each finding against the source.