In our latest multicentre study published in Bone & Joint Research, we set out to answer a question that has challenged orthopaedic surgeons for decades: can we predict which older patients with a proximal humerus fracture truly benefit from surgery?
We analysed data from two randomized controlled trials that compared nonoperative treatment with either locking plate fixation or hemiarthroplasty in patients aged 60 years and older. Altogether, we included 248 patients and modelled their two-year outcomes using three widely used measures — the DASH, Oxford Shoulder Score, and Constant-Murley Score. Our goal was to see whether baseline factors such as age, sex, fracture type, dominance, or comorbidities could meaningfully predict recovery — and whether any subgroup of patients might derive a greater benefit from surgery.
The results were sobering but important. The predictability of patient-reported outcomes based on baseline characteristics was only moderate, with R² values between 0.36 and 0.40. Dizziness at baseline emerged as the single strongest predictor for DASH and Oxford Shoulder Score outcomes, while dominance of the injured arm mattered most for Constant-Murley. Yet these associations were weak and clinically uncertain.
Crucially, when we examined heterogeneous treatment effects — in other words, whether any patient profile consistently benefitted more from surgery — we found no credible evidence of such interactions. The apparent differences were small and highly uncertain. Simply put, our usual “front-door” variables did not help identify who should undergo surgery and who should not.
These findings have clear implications for everyday fracture care. Despite the increasing interest in “precision orthopaedics” and personalized treatment, our study shows that, with current baseline data, reliable prediction of recovery after proximal humerus fracture is not yet possible. For most older patients, nonoperative management remains the safest and most evidence-based approach.
We believe this is an important message for both clinicians and patients. Surgical treatment should not be justified on the assumption that certain visible characteristics at presentation guarantee better outcomes.
Study can be read at: https://boneandjoint.org.uk/article/10.1302/2046-3758.1410.BJR-2024-0218.R2
