Criterion related validity of self-reported knee symptoms among athletes

The objective of this study was to investigate the validity of self-reported knee symptoms among athletes. During 12 months, 252 athletes answered a weekly questionnaire about concurrent sports activity, knee pain, knee instability, knee swelling and knee locking. Finally, the athletes answered a retrospective questionnaire summing up symptoms during the preceding 12 months. Accordingly, two 12 month period prevalences of the same period were calculated for each knee symptom. The proportion of agreement between the retrospective 12 month questionnaire and the concurrent weekly questionnaires was for knee pain, knee instability, knee swelling and knee locking, 86%, 89%, 89% and 90%, respectively. Using weekly reports as reference, the corresponding sensitivity of the 12 month period prevalence question was 0.86, 0.71, 0.66 and 0.56, respectively. The sensitivities were all significantly associated with the number of weeks since the last episode of the corresponding knee symptom. As concerns estimation of 12 month period prevalences, a retrospective summary questionnaire at the end of the period is a valid substitute for concurrent weekly questionnaire information on knee pain, knee instability and knee swelling, but only a moderate substitute for information on knee locking. A reduction in the length of the period the athlete should recall knee symptoms might improve the questionnaire validity.