PREDIGT Trial: Piloting an Unsupervised, Home-Based Toolkit to Screen for Parkinson Disease

BackgroundEffective screening for Parkinson disease (PD) is important for both symptomatic treatment and recruitment into intervention trials. We recently developed a toolkit to quantify PD risk. Here, we examined the PREDIGT models diagnostic performance when deployed at home. MethodsWe contacted 613 subjects following outpatient clinic encounters. Between 2022-2024, 305 participants (range, 40-85 years) were recruited: 93 with typical PD; 66 had other neurological diseases (OND); 146 were neurologically healthy. Two versions of the toolkit were completed: First, an original, 69-item-long questionnaire paired with a 40-scent smell test; thereafter, a simplified, 11-item-long questionnaire and a newly developed, 8-scent smell test. PREDIGT summary scores were calculated for each subject to examine diagnostic classifications. Area-under-the-ROC-curve, sensitivity, specificity, and likelihood ratios were used to evaluate performances and to determine clinically relevant thresholds. ResultsIn both versions, PD patients had higher questionnaire scores and lower smell test scores than neurologically healthy controls (p<0.001); scores for OND subjects ranked at intermediate levels. The simplified questionnaire outperformed the original version in diagnostic accuracy. The abbreviated smell test performed as well as the 40-item version in identifying hyposmia. At a value of 22.94 (range 0-100) for the threshold that separates PD subjects from other participants, the simplified PREDIGT summary score showed a sensitivity of 0.98, a specificity of 0.83, and revealed positive and negative likelihood ratios of 5.88 and 0.02, respectively. InterpretationOur study reveals that unsupervised screening for typical PD can be effectively carried out at home using an 11-item questionnaire and 8-scent smell test.