A rule-based, dose-finding design for use in stroke rehabilitation research: methodological development

Dose-optimisation studies as precursors to clinical trials are rare in stroke rehabilitation.

To develop a rule-based, dose-finding design for stroke rehabilitation research.

3 + 3 rule-based, dose-finding study. Dose escalation/de-escalation was undertaken according to preset rules and a mathematical sequence (modified Fibonacci sequence). The target starting daily dose was 50 repetitions for the first cohort. Adherence was recorded by an electronic counter. At the end of the 2-week training period, the adherence record indicated dose tolerability (adherence to target dose) and the outcome measure indicated dose benefit (10% increase in motor function). The preset increment/decrease and checking rules were then applied to set the dose for the subsequent cohort. The process was repeated until preset stopping rules were met.

Participants had a mean age of 68 (range 48 to 81) years, and were a mean of 70 (range 9 to 289) months post stroke with moderate upper limb paresis.

A custom-built model of exercise-based training to enhance ability to open the paretic hand.

Repetitions per minute of extension/flexion of paretic digits against resistance.

Usability of the preset rules and whether the maximally tolerated dose was identifiable.

Five cohorts of three participants were involved. Discernibly different doses were set for each subsequent cohort (i.e. 50, 100, 167, 251 and 209 repetitions/day). The maximally tolerated dose for the model training task was 209 repetitions/day.

This dose-finding design is a feasible method for use in stroke rehabilitation research.