Compliance and Adherence to Occlusal Splint Therapy and Physiotherapy in Temporomandibular Disorders: A Retrospective Cohort Study Using Narrative‐Derived Global Impression of Change
Nadav Grinberg, Ofek Shpigler, Oren Peleg, Clariel Ianculovichi, Shlomi Kleinmann, Reema Mahmoud, Ayelet Zlotogorski-Hurvitz, Gilad WassermanBackground
Occlusal splints and physiotherapy are the mainstays of conservative management for temporomandibular disorders (TMDs), but how often patients follow these recommendations, and what follows when they do not, is poorly quantified.
Methods
From 3317 patients with a TMD diagnosis and at least two clinic visits, 800 records were randomly selected and screened; 273 met inclusion criteria and contributed 443 recommendation events. Adherence was classified by initiation and persistence. A proxy Patient Global Impression of Change (PGIC) score was derived from the follow‐up narrative under a predefined rule. The primary outcome was PGIC ≤2.
Results
Persistence after initiation was 70% for splints, 88% for self‐administered and 90% for professional physiotherapy. Splint losses were characteristically after acquisition (40 of 133 acquirers discontinued), whereas professional physiotherapy failed mostly before initiation (66 of 174 never attended). Among 66 patients recommended both physiotherapy types, the home programme was performed far more often (20 versus 1 discordant pairs, p < 0.001). Of 133 obtaining a splint, 122 (92%) received the prescribed appliance; deviation was concentrated in community‐fabricated splints (odds ratio [OR] 6.37, 95% confidence interval [CI] 1.60–25.39). Adherence was associated with improvement in every modality (splint OR 10.35, 95% CI 4.31–24.82; professional physiotherapy OR 5.05, 1.97–12.92; self‐administered physiotherapy 37% versus 0%, p < 0.001; pooled OR 8.24, 4.05–16.79); response rose from 4% to 45% across adherence strata. Apart from a joint‐only diagnosis, no baseline characteristic predicted response after adjustment (adherence adjusted OR [aOR] 12.00, 5.40–26.67). Restricting the contrast to never‐initiated versus adherent events, which excludes response‐dependent discontinuation, gave OR 9.11 (4.26–19.50).
Conclusions
About 40% of conservative TMD recommendations were not followed, and non‐adherence was associated with a substantially lower probability of meaningful improvement, although residual confounding, including a healthy‐adherer effect, cannot be excluded. Non‐initiation and discontinuation are distinct problems with distinct causes.