Inconsistent Movement Disorders Society–Unified Parkinson’s Disease Rating Scale Part III Ratings in the Parkinson’s Progression Marker Initiative

The Movement Disorders Society version of the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) was developed to improve the measurement of Parkinson’s disease (PD) symptoms.^1,2 In the Parkinson’s Progression Markers Initiative (PPMI)³ study, the motor examination of Part III is performed by a site investigator who has completed the MDS-UPDRS online training. Ratings are recorded on paper case report form and subsequently entered into an electronic data capture portal. However, transcription errors can occur during examination or transfer. Some transcription errors could be avoided if ratings were entered directly into an electronic data capture portal that would flag inconsistent ratings. Specifically, this would include the following: (1) resting tremor amplitude without constancy (and vice versa); (2) Hoehn and Yahr stage 2 rating with unilateral symptoms; and (3) opposite lateralization of resting tremor or rigidity between upper and lower extremities (eg, left lateralized resting tremor in upper extremities, right-lateralized resting tremor in lower extremities).

In this study, we explored the frequency of inconsistencies listed previously from the MDS-UPDRS Part III in PPMI. The current sample included 414 recently diagnosed and untreated PD participants with positive DaTscans who completed the MDS-UPDRS at baseline and a follow-up visit at 3 or 6 months (https://www.ppmi-info.org). The SPSS (version 25; IBM Corp., Armonk, NY) syntax used for identifying inconsistent ratings is included in the supplementary materials.

At baseline, 1 or more inconsistencies were found in 31 participants (7.5%). As shown in Table 1, most inconsistencies at baseline were reconciled at the first follow-up visit. However, at follow-up, new inconsistencies emerged for 19 participants who had no inconsistent ratings at baseline. Across baseline and follow-up, 1 or more inconsistencies were identified in 49 participants (11.8%).

TABLE 1.

Frequencies of inconsistent Movement Disorders Society–Unified Parkinson’s Disease Rating Scale Part III ratings at baseline and first follow-up visit for 414 Parkinson’s disease patients enrolled in Parkinson’s Progression Markers Initiative

Status	Baseline, n (%)	Persisting Baseline Inconsistency at Follow-Up, n	New at Follow-Up, n (%)	Combined Baseline and Follow-Up, n (%)
No inconsistencies	383 (92.5)		390 (94.4)	365 (88.2)
One or more inconsistency	31 (7.5)	6	19 (5.6)	49 (11.8)
Hoehn and Yahr inconsistency	13 (3.1)	2	9 (2.1)	21 (5.1)
Resting tremor inconsistencies	12 (2.9)	2	4 (1.0)	16 (4.0)
Positive tremor, no constancy	5	1	2	7
No tremor, positive constancy	7	1	2	7
Opposite lateralization inconsistencies	7 (1.7)	0	6 (1.4)	13 (3.1)
Resting tremor	3	0	4	7
Rigidity	4	0	2	6

One participant had inconsistent ratings at baseline for both opposite lateralization of rigidity and tremor.

This investigation revealed that inconsistent ratings on Part III of the MDS-UPDRS are not uncommon in PPMI. Correction of most Hoehn and Yahr inconsistencies at follow-up suggests that the majority represent error rather than alternative interpretation of Hoehn and Yahr criteria or considerations of features not recorded by the MDS-UPDRS. Similarly, correction at follow-up of all inconsistencies involving opposite lateralization between the upper and lower extremities for resting tremor and rigidity indicates error rather than a highly unusual presentation. These inconsistent ratings likely reflect transcription errors made during examination and/or electronic database entry. Within PPMI, inconsistencies increase estimates of PD motor progression variability and diminish apparent relationships with biomarkers and other measures.⁴ Given that the centers involved in PPMI represent institutions renowned for excellence in PD research, these estimates may well underestimate occurrence in research with less rigorous protocols and inexperienced examiners. Of particular concern are clinical trials, where such measurement error could obscure potentially meaningful treatment effects.

Errors attributed to inconsistent ratings could be reduced through electronic data capture on a tablet or similar device during examination. For instance, algorithms could alert examiners when inconsistent ratings were entered and prompt correction or confirmation. Highly unusual symptom combinations (eg, severe postural instability without gait disturbance) or improbable changes over time (eg, reversed lateralization) could also be flagged. Tablet-based, MDS-UPDRS data-capture tools are currently available. Further study could determine if real-time auditing decreases inconsistencies in study data.