A Kinematic Sensor and Algorithm to Detect Motor Fluctuations in Parkinson Disease: Validation Study Under Real Conditions of Use.

The study included 23 patients selected by convenience sampling, so the results may not apply to other Parkinson disease populations. Patients had to recognize their On and Off states and walk without help, so people who could not do these tasks were not represented. The sensor was compared with patient diaries, not direct observation, and the diary can miss episodes, so sensitivity and specificity were not calculated. The sensor cannot classify motor state during periods without gait or dyskinesia, so unknown time slots and missed Off periods are likely. The study did not validate whether the sensor can separate On-state with dyskinesia from On-state without dyskinesia. Six patients stopped diary recording before the third day because of inconvenience, but those results were not shown. No prospective study showed that using the sensor improves therapy adjustment or patient outcomes.

ARM, AS, CPL, and AC are shareholders of Sense4Care, a spin-off company that may commercialize the device. They state that possible commercialization is a research outcome and did not affect the study design, analysis, interpretation of results, or conclusions.

Do not read the 92.20% result as proof that the sensor catches every Off period. It only scored when it made a decision, and it cannot classify states during periods without gait or dyskinesia. The study also did not show that the sensor improves therapy adjustment or patient outcomes.