Validation of a French patient-reported outcome measure for patello-femoral disorders: The Lille Patello-Femoral Score

Abstract Background The diagnosis of patello-femoral instability (PFI) relies chiefly on the patient's clinical findings. Nevertheless, few clinical scores specifically designed to evaluate the patello-femoral joint are available. The Lille scoring system is a 12-item self-questionnaire yielding a score from 0 to 100 that is used in France but has not been validated. We therefore conducted a validation study in a population of younger patients with PFI. Hypothesis The Lille scoring system meets validation criteria for patient-reported outcome measures (PROMs). Material and method A retrospective study done in two centres identified 136 patients with objective ( n = 109) or potential ( n = 27) PFI. Before and after surgery, the Lille score was determined by all patients and the Kujala score in 61 patients. The Lille score was also determined by 30 controls free of patello-femoral disorders to allow an evaluation of discrimination between PFI and other knee disorders in individuals of similar age. Results The response rate was 100%, indicating that the Lille questionnaire was easy to complete. Consistency was established: (a) the global score showed no floor or ceiling effect (in no questionnaires were over 85% of items given the highest or lowest possible score), and saturation occurred neither for the global score nor for the item sub-scores (fewer than 85% of patients had the lowest or highest possible score); (b) a single redundancy was found, between the items ‘pain’ and ‘locking’, for which the correlation coefficient was ≥ 0.7 ( P < 0.0001). Discriminating performance was assessed by comparing the mean Lille score values in the controls (67.8 ± 9.2) and patients (38.1 ± 10.4); the difference was significant ( P < 0.05) and the estimated effect size was > 0.8, indicating strong discrimination by the Lille scoring system. Item uniformity, with all items measuring the same phenomenon, was established by the Cronbach alpha coefficient value > 0.7. External consistency between the Lille and Kujala scoring systems was confirmed in the 61 patients for whom both scores were available (Pearson correlation coefficient, 0.5). Sensitivity to change was established by the > 0.8 effect size of surgical treatment. Discussion The Lille scoring system deserves to be used routinely in clinical practice as a patient-reported outcome measure. A prospective study will assess intra-observer reproducibility and sensitivity to change in patients treated non-operatively. Alth