The economics of Parkinson monitoring are a warning for wearable-health startups

Published October 28, 2026 · Early health technology assessment and simulation · Negative economic base case
Primary source: Parkinson ML remote-monitoring assessment, PMID 42947537

Wearables and machine learning could identify progression to advanced Parkinson disease sooner. But detecting progression is not itself a patient outcome. Investigators modeled a 16-year Finnish care pathway comparing standard management with machine-learning-enabled remote monitoring, assuming 88% sensitivity and 62% specificity in the base case.

The modeled system generated 0.0044 additional quality-adjusted life-years at €761 additional cost per patient, an incremental cost-effectiveness ratio of €172,023 per QALY. At a €35,000/QALY willingness-to-pay threshold, net monetary benefit was negative. Some scenarios became favorable if monitoring produced direct quality-of-life gains or delayed entry into institutional care.

Strength and limitations. This is a transparent model-based assessment, not an outcome trial. Many inputs are assumed; results depend on the health system, intervention costs, prevalence, and how clinicians respond to alerts. It should be treated as a design challenge rather than a verdict against all Parkinson wearables.

Clinical practice: Monitoring needs a defined action pathway and evidence that action changes outcomes.