Can Surgeons or Patients Predict the Likelihood of Improvement With Nonoperative Treatment of Chronic Tennis Elbow?

Persisting symptoms after an attempt of nonoperative treatment represents one of the most common indications for surgery in many musculoskeletal conditions, such as tennis elbow. The rationale behind the practice of resorting to surgery in individuals with long-standing symptoms is that resolution of symptoms is believed to be unlikely without surgery after a certain period, and surgeons can identify a subgroup to benefit from surgery. For this approach to be sound, surgeons must be able to reliably distinguish between patients unlikely to improve without surgery and those who are likely to benefit from it.

(1) Do patients with persistent tennis elbow symptoms (lasting > 10 months) who are referred to a surgeon improve without surgery over a 24-month follow-up period? (2) Are surgeons' or patients' predictions about improvement associated with actual improvement? (3) What patient characteristics, if any, are associated with predictions of improvement made by either surgeons or patients?

Between 2016 and 2018, we prospectively recruited 97 patients with persistent tennis elbow symptoms (> 10 months) who were dissatisfied with nonsurgical treatment and referred for surgical consultation at five secondary or tertiary public hospitals. Of these, 89% (86 of 97) agreed to continued nonoperative treatment and were included in this observational cohort study. To evaluate the outcomes of continued nonoperative treatment, we measured the Oxford Elbow Score (OES) and global improvement at 6 weeks and at 3, 6, 12, and 24 months. To assess whether either the surgeons or the patients could predict the likelihood of symptom improvement, we asked both parties at baseline to predict whether each patient would be satisfied with their symptom state without surgery within the next 6 months. We then evaluated the prognostic value of these predictions by comparing the OES and global improvement scores between (1) patients who believed that they would improve versus patients who did not and (2) patients whom the surgeons predicted would improve versus those whom the surgeons predicted would not. To explore factors that might explain the predictions, we assessed the correlation between the predictions and baseline characteristics, including age, sex, affected side, smoking status, duration of symptoms, disability (OES score), Pain Catastrophizing Scale score, prior corticosteroid injections, and any planned injection treatments. Data from patients who underwent surgery during the follow-up period were included only up to the time of surgery. The mean ± SD age of the patients was 49 ± 5.4 years, and prior to the initial consultation, they had had symptoms for a mean ± SD of 20 ± 12 months. One-half of the patients were female.

Nine percent (8 of 86) of patients eventually underwent operation during the 2-year follow-up period. The mean total OES of the cohort (range 0 to 100, with higher scores indicating better outcomes) increased from approximately 50 points at baseline to 80 points at the final 24-month follow-up visit. Surgeons' predictions about likelihood of improvement were not associated with the observed improvement, while patients who were more pessimistic about their likelihood of recovery at baseline had slightly inferior outcomes compared with patients who were more optimistic about their likelihood of recovery. As for factors associated with patients' predictions of recovery, both patients who had previously received corticosteroid injections (relative risk [RR] 1.4 [95% confidence interval (CI) 1.1 to 1.7]; p = 0.03) and those scheduled to receive botulinum toxin or platelet-rich plasma injections (RR 3.8 [95% CI 2.0 to 7.3]; p < 0.001) were more likely to predict improvement compared with those who opted to wait and see. Surgeons' predictions about the recovery were not associated with any of the measured patient characteristics, indicating that the predictions were based on heuristics, that is, mental shortcuts or rules of thumb that clinicians commonly use in clinical decision-making.

Our findings suggest that persistent tennis elbow symptoms are a poor indication for surgery, as the majority of patients experience symptom resolution without it, and surgeons are unable to reliably predict who will or will not improve with nonoperative treatment. Therefore, treatment decisions should not be based on the clinician's perception of the disease course. Patients' predictions, especially more pessimistic views, were found to more accurately reflect the likely recovery trajectory. Finally, despite evidence demonstrating the ineffectiveness of injections, they elevated patients' expectations for improvement.

Level II, therapeutic study.