Abstract

Purpose Surgical options for end-stage knee osteoarthritis (OA) include total and medial unicompartmental knee replacement (TKR and UKR). Deciding which surgery to perform is complex and ill-defined, yet it has important implications for patients and the health service. The study aimed to identify clinical and surgeon factors predicting surgeons' preferences. Methods Based on a preliminary survey of 162 UK surgeons, we identified clinical features frequently considered when deciding between TKR and UKR. By systematically varying patient age, obesity, site of pain, anaesthetic risk and anterior cruciate ligament (ACL) integrity, we constructed 32 clinical vignettes. We used these in a new survey, where surgeons indicated which surgery they would recommend on an 11-point rating scale with end points anchored at ‘definitely TKR’ and ‘definitely medial UKR’. Data were analysed with mixed-effects linear regressions. Results Eighty-three UK arthroplasty surgeons completed the vignettes. Preference for UKR over TKR was significantly lower for patients over 50 years (b = −0.57 [−0.82 to −0.33], p < 0.001) with abnormal ACL (b = −1.93 [−2.17 to −1.68], p < 0.001) and severe systemic disease (b = −0.46 [−0.70 to −0.21], p < 0.001). Obesity was a weak and unreliable predictor, and we did not detect any influence of site of pain. The surgeons' habitual practice (proportion of UKRs over all knee replacements performed in a typical year) was the second strongest predictor after ACL (b = 1.26 [0.54–1.99], p = 0.001). Conclusions ACL integrity was the most important determinant of surgeons' preferences between TKR and UKR. Their habitual practice was also a strong predictor, outweighing most clinical factors in the vignettes. Level of Evidence Level II, prospective cohort study.