Introduction Patellar tendinopathy (PT) has poor outcomes, and we have no robust prognostic guides to understand who gets better, when and why. We aimed to identify which combination of self-reported factors best predicts PT recovery.

Materials and Methods A previously validated, reliable online questionnaire battery yielded data from an international sample of elite and non-elite jumping athletes with a one-year follow-up. Recovery was defined by availability status for training and competition. More than 100 bio-psycho-social, demographic and sports specific factors were collected and a multivariable cox proportional-hazards model constructed.

Results 128 athletes with PT (30.9±8.9 years; 77 males; VISA-P=61.5±16.2) provided 25,284 days at risk (198±141 days) for analysis. Recovery rate was 45% occurring around 6-month. The final multivariable model partially predicted PT recovery with acceptable model performance and internal validation (optimism-corrected Harrell’s C discrimination=0.77 and Calibration Slope=0.86). The model showed that PT recovery is associated with a higher KOOS-PF score (lower severity; HR=1.03, 95%CI=1.02–1.05), a shorter time-off from sport (HR=0.93, 95%CI=0.87–0.99), feeling more rested after sleep (HR=1.93, 95%CI=1.13–3.28), not having concurrent tendon problems (HR=0.23, 95%CI=0.07–0.69), higher training duration (HR=1.05, 95%CI=1.01–1.10) and symptoms that can be modified by movement (HR=2.71, 95%CI=1.21–6.09).

Conclusion This is the first study investigating outcome predictors for PT recovery. The developmental statistical causal model showed that sports specific and biomedical variables partially predicted PT recovery. Demographic or psychosocial variables did not contribute to the model. These findings could enhance professionals’ understanding of PT prognosis and clinical decision making.