OBJECTIVES: A key challenge for detecting efficacy in Parkinson’s disease (PD)-modifying therapy trials is a lack of easy-to-implement, sensitive outcome measures to detect changes in disease progression. Digital at-home assessments can be implemented frequently and have a potential to assess progression in shorter times and/or with smaller sample sizes than traditional in-clinic outcome measures. However, the implementation of digital assessments requires empirical data to guide evidence-based clinical trial design.

METHODS: To develop recommendations for digital assessment implementation in PD clinical trials, we use simulations based on a previously developed and parameterized model of PD progression¹ informed by PPMI data. We consider three designs to detect disease progression within one year: (1) Individual measurements at baseline and end of study, (2) A burst of measurements at baseline and end of study to assess differences in change from baseline; (3) Evenly spaced measurements to assess the difference in the rate of change from baseline.

RESULTS: For study design (1), 260 participants are required to detect a treatment effect with 80% power. Including a burst of assessments (study design 2) reduces the necessary sample size to 150. In this scenario, we include 7 daily assessments per cluster; increasing the number of assessments per cluster beyond ~5 does not increase power. For scenario (3), only 70 participants are required to detect a strong treatment effect with weekly assessments; increasing the number of measurements from 6 to 48 shows a continuous increase in power.

CONCLUSIONS: We find that frequent measurements in PD greatly increase the ability to detect a treatment effect with fewer participants. Additionally, a study design with evenly spaced assessments is superior to a burst design. Further work to understand the sensitivity of these results to non-linear disease progression and varying characteristics of at-home assessments is warranted.