Objective: To pilot the 2-step PREDIGT Score model for feasibility and improve it by simplifying both steps.

Background: We previously demonstrated that the original PREDIGT model [1], which combined a self-reported questionnaire with a comprehensive smell test, could accurately identify incident Parkinson’s disease (PD). Using data from two case-control cohorts (PPMI; DeNoPa), PREDIGT 1.0 distinguished subjects with newly diagnosed PD from controls with AUC values between 0.89 and 0.9 [2].

Method: The Ottawa Trial cohort represents a cross-sectional, observational study to pilot PREDIGT Score models for at-home screening of PD. Between 2023 and 2024, study participants were recruited at two outpatient neurology clinics. PREDIGT 2.0 included three changes to potentially improve diagnostic performances: the questionnaire was updated by removing less-informative variables for environmental exposure history; three questions related to fatigue, handwriting, and slowness were added; and an abbreviated smell-test card of 8 scents [3] replaced the 40-scent UPSIT kit. Performances of PREDIGT versions 1.0 and 2.0 were compared within our cohort.

Results: Among 305 participants, 163 have completed all components (healthy controls (HC): n = 93; PD/DLB: n = 49; other neurological diseases (OND): n = 21) [Table 1]. When judged by group classifications, PREDIGT 2.0 showed better performances [Figure 1]: for PD/DLB vs. HC, the AUC value with its 95% confidence interval (CI) increased from 0.93 (0.89-0.97) to 0.97 (0.95-0.99); for PD/DLB vs. OND, the AUC value increased from 0.77 (0.65-0.9) to 0.81 (0.68-0.94). Notably, the 11-item questionnaire of PREDIGT 2.0 showed marked improvement from its original version with the AUC value increasing from 0.76 (0.68-0.84) to 0.96 (0.93-0.98) for PD/DLB vs. HC. The abbreviated 8-scent smell test (AUC: 0.94 (0.9-0.98)) showed a similar performance to the 40-scent UPSIT kit (AUC: 0.93 (0.89-0.97)) for PD/DLB vs. HC.

Conclusion: The results of our pilot study revealed robust performances by PREDIGT 2.0 in separating patients with typical PD from those with other neurological conditions and healthy controls. We posit that the home-based, simple-to-complete PREDIGT 2.0 model could evolve into a screening tool to identify subjects with PD before their first assessment by a neurologist.

Table 1

Figure 1

References: [1] Schlossmacher MG, Tomlinson JJ, Santos G, et al. Modelling idiopathic Parkinson disease as a complex illness can inform incidence rate in healthy adults: the PR EDIGT score. Eur J Neurosci. 2017 Jan;45(1):175-191.
[2] Li J, Mestre TA, Mollenhauer B, et al. Evaluation of the PREDIGT score’s performance in identifying newly diagnosed Parkinson’s patients without motor examination. NPJ Parkinsons Dis. 2022 Jul 29;8(1):94.
[3] Li J, Grimes K, Saade J, et al. (2024). Development of a Simplified Smell Test to Identify Parkinson’s Disease Using Multiple Cohorts, Machine Learning and Item Response Theory. npj Parkinsons Dis. In revision. Submission ID: 7173cf5c-f6c1-450e-aeeb-69ea8eddd21e. Preprint on medRxiv https://doi.org/10.1101/2024.08.09.24311696

« Back to 2025 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/predigt-2-0-a-self-administered-two-step-screening-tool-to-identify-individuals-with-parkinsons-disease/