A diagnostic clinical prediction rule for predicting hip subluxation/dislocation in patients with cerebral palsy

Objective - Hip subluxation/dislocation, a common problem in patients with cerebral palsy (CP), needs to be diagnosed with hip radiography. However, patients with cerebral palsy in a rural or country border areas may not have access to a radiographic screening program due to transportation difficulties and cost. This study aims to develop a clinical prediction rule (CPR) for diagnostic prediction of hip subluxation/dislocation in patients with CP for use as a risk-screening tool. Methods - This is a cross-sectional diagnostic CPR development study. Data were obtained from medical and radiologic records of patients with CP who had undergone outpatient follow-up at a 750-bed general hospital between January 2017 and December 2023. Clinical predictive factors were medical records plus hip subluxation/dislocation diagnoses using the migration percentage (MP), with ≥ 33% indicating hip subluxation and ≥ 90% indicating hip dislocation. Multivariable logistic regression analysis was used for choosing predictive variables and rating their coefficient. Both discriminative and calibration aspects of the performance of the CPR were evaluated using both a development and an internal validity model. Results - Among the 69 patients with CP in the study, the mean (SD) age was 113 (242) months. Of the 69 patients, 30 were diagnosed with hip subluxation/dislocation, a prevalence of 43%. Using multivariable logistic regression analysis, a simple CPR performance calibration system was developed which included three factors: age ≥ three years (1 point), female sex (1 point), non-ambulatory status (Gross Motor Function Classification System (GMFCS) levels IV and V) (2 points). The discriminative ability of the CPR, evaluated using the area under the receiver operating characteristic curve (AuROC), was 0.776 (95%CI: 0.668-0.884) and the calibration curve showed acceptable performance in both the development and the internal validation models. Conclusions - Our diagnostic CPR for predicting hip subluxation/dislocation in patients with CP provides acceptable discriminative and calibration performance. This CPR may be used to evaluate the risk of hip subluxation/dislocation in settings where hip radiography is not available. Further external validation studies are needed to confirm the robustness of the performance before applying this CPR in other clinical settings.