Evaluation of clinical guidelines for contracted (frozen) shoulder 12–18 months after publication

Abstract

Background/aim: To obtain an early indication of the dissemination of evidence-based physiotherapy guidelines for contracted (frozen) shoulder and to explore how the guidelines’ recommendations correspond with practice. Methods: The authors monitored downloads of the guidelines for the 18 months after publication. This study also used a self-administered, anonymous online questionnaire survey, which targeted UK-based physiotherapists who had accessed the guidelines by 12–18 months after publication. The survey's questions required ‘yes’ or ‘no’ responses and ‘yes’ responses activated questions asking whether practice had changed. With ethics clearance, and following piloting, the survey was advertised electronically. Analysis was by frequency counts. Results: There were 6 185 de-replicated UK downloads of the guidelines, most via the Chartered Society of Physiotherapy website. There were 366 survey responses. The median correspondence between the guidelines’ recommendations and practice was 66%, mostly reflecting validation of practice, but some change of practice. Notably among the diagnosis and assessment recommendations, passive external rotation was widely used as a pathognomonic test (75%). The proposed pain-predominant/stiffness-predominant classification was widely adopted (66%), and 88% of those who asked their patients said those patients found this terminology meaningful. Only 34% of responders reported using validated outcome measures. Correspondence of practice with the guidelines’ treatment recommendations was high for out-patient physiotherapy and injection but low for shortwave diathermy. Conclusion: The wide distribution and survey results suggest that the guidelines’ development was worthwhile. There were some areas of poor implementation of the guidelines’ recommendations, explicable by known obstacles; however, the findings provide a timely indication of where supplementary efforts should be targeted, e.g. using validated outcome measures.