Patient-Reported Outcome Measure PRO-PD Reliable for Monitoring Symptoms of Parkinson Disease
PRO-PD showed good 6-month test-retest reliability, as well as good convergent validity between other widely used validated scales of PD.
Filip Bergquist, MD, PhD
Findings from a recently published study in Movement Disorders showed that the Patient-Reported Outcomes in Parkinson’s Disease (PRO-PD) is a reliable and valid scale for the detection and monitoring of PD symptoms, and may be useful in clinical practice. Investigators concluded that the information on the size of clinically important changes as well as predicted changes over time can help guide the interpretation of the scale in clinical settings.1
The analysis included 286 patients with PD who visited outpatient clinics in West Sweden who underwent a baseline clinic visit, a 1-year telephone follow-up, and a 3-year telephone follow-up. Four PRO measures were administered repeatedly: Parkinson’s Disease Questionnaire (PDQ)-8, Non-Motor Symptoms Questionnaire (NMSQ), EuroQoL Five-Dimension Five-Level Scale (EQ-5D-5L), and PRO-PD. Patient-reported Clinical Global Impression of Change (PGI-C) was administered at 1-year follow-up.
Led by senior investigator Filip Bergquist, MD, PhD, department of pharmacology, University of Gothenburg, PRO-PD ratings were available from 716 of 747 (96%) study visits. In terms of item evaluation, all items on PRO-PD exhibited positive skewness, ranging from 0.1 in “stooped posture” to 2.7 in “nausea.” Of note, no ceiling effects were found, and the items with highest proportion of values of 100 were “urinary symptoms" (11%) and "sexual dysfunction” (6%).
The internal consistency of the broad scale at baseline was excellent, reflected by 0.93 on Cronbach’s α. At the conclusion of the 3-year period, the highest item-item correlations were seen between items "motivation” and “loss of interest” (r = 0.80), items “walking” and “slowness” (r = 0.71), and items “depression” and “anxiety” (r = 0.71).
PRO-PD showed good test-retest reliability as well, as investigators recorded an ICC coefficient of 0.87 for 82 patients who provided 2 PRO-PD measures within 6 months (median interval, 152.5 days). ICC coefficients for individual items ranged from 0.45 in “control of body temperature” to 0.84 in “nausea.” In addition, the scale was not influenced by sex (Mann-Whitney, P = .34) and did not correlate with age (rs = 0.08; P = .20). The correlation between total PRO-PD and time since diagnosis was weak (rs = 0.26; P <.0001), even after omitting the items with the lowest corrected item-total correlations.
READ MORE: Incorporating Augmented Reality to Treat Parkinson Disease Gait Problems
Convergency validity of PRO-PD was good, with correlation coefficients of 0.70 with PDQ-9, 0.70 with NMSQ, 0.71 with EQ-5D-5L, and 0.69 with Clinical Impression of Severity Index for Parkinson Disease. The sum of the 21 nonmotor items in the PRO-PD also had a strong correlation with NMSQ (rs = 0.71; P <.001). Overall, the median increase in total PRO-PD score was 61 (–107 to 244) in the first year in the group of 191 patients, with complete measures both from baseline and the 1-year follow-up.
In a subgroup of patients who filled out PGI-C at the 1-year follow-up (n = 153), 32 reported clear deterioration, indicated by PGI-C scores above 5, and 30 reported no change. The minimal clinically important change, as determined by the anchor-based distribution method, was 119. Investigators concluded that this value represented the change in total PRO-PD score that most accurately separates patients reporting no change from patients reporting minimal clinically important deterioration (sensitivity, 0.66; specificity, 0.73).
"Future improvements of PRO-PD may include a revision of included items to improve the coverage of clinically important symptoms and avoid redundancy,” Bergquist et al wrote. "Furthermore, intermediate ticks on the VAS line could be evaluated, as well as how anchors are best defined for individual items. If the scale is to be used for research, validation of translations and clinimetric performance is essential."
