Validation of a Patient-Reported Outcome Measure in Parkinson's Disease.

BACKGROUND

The Patient-Reported Outcomes in Parkinson's Disease (PRO-PD) scale is a 35-item visual analog measure designed to quantify symptom severity across motor and non-motor domains. Developed as a continuous, patient-centered outcome, PRO-PD captures patient-perceived change over time and is suitable for remote longitudinal assessment. This study evaluated the psychometric properties of PRO-PD across two independent datasets, including reliability, validity, factor structure, and minimal clinically important difference (MCID), and assessed its relevance to nutrition- and lifestyle-focused research.

METHODS

Convergent validity was evaluated in a cross-sectional clinical dataset (n = 46) using established clinician-rated and patient-reported instruments, including Hoehn and Yahr, Unified Parkinson's Disease Rating Scale (UPDRS), Parkinson's Disease Questionnaire-39 (PDQ-39), Montreal Cognitive Assessment (MoCA), and PROMIS measures. Internal consistency, temporal stability, factor structure, and known-groups validity were assessed in a large remote-monitoring cohort (n = 2612). MCID thresholds were estimated in a longitudinal subsample (n = 390) using anchor-based methods, multinomial regression, and receiver operating characteristic analyses.

RESULTS

PRO-PD demonstrated strong convergent validity with established clinical measures, excellent internal consistency (Cronbach's α = 0.93 (95% CI: 0.90-0.96) for Small Data and 0.95 (95% CI: 0.947-0.951) for Big Data"), and good test-retest reliability (ICC = 0.78 overall; 0.89 at 6 months). Confirmatory testing of a previously proposed eight-factor structure showed suboptimal fit, leading to a parsimonious four-factor solution (Neurobehavioral, Autonomic, Motor, Mood/Motivation) explaining 47.6% of variance. PRO-PD scores increased significantly with advancing disease duration and stage. MCID thresholds were +53.5 points for worsening and -78.5 points for improvement (AUC = 0.64 for worsened vs. not worsened; AUC = 0.71 for improved vs. worsened), with greater sensitivity for detecting deterioration than improvement. PRO-PD scores demonstrated sensitivity to patient-perceived symptom change over time, supporting its utility for longitudinal monitoring and potential application in lifestyle-focused intervention research.

CONCLUSIONS

These findings support PRO-PD as a psychometrically robust outcome measure that can be completed remotely without trained administrators.