Application of survival analysis techniques to determine the optimal number of acupuncture therapy sessions for stroke patients.

Background: Acupuncture has shown clinical benefits in stroke patients but there is currently no clear recommendation on the number of acupuncture sessions that should be offered. The aim of this study was to determine the optimal number of acupuncture therapy sessions, following which significant clinical improvement could be anticipated.

Methods: We performed a retrospective observational study in stroke patients who had received acupuncture therapy as an adjunct treatment at Trang Hospital-a non-academic provincial hospital in Southern Thailand-between 1 July 2016 and 30 June 2017. Significant clinical improvement in activities of daily living (ADLs) was defined as an increase in the Barthel Index of ⩾10 points from baseline. Survival analysis techniques were applied to determine the optimal number of acupuncture therapy sessions.

Results: Of 89 patients, 78% had a significant clinical improvement. Kaplan-Meier analysis estimated that 50% of the patients exhibited significant improvements by the eighth acupuncture therapy session. By the 16th session, the proportion of patients increased to 81%. Subgroup analyses revealed that the clinical improvement did not depend on baseline Barthel index, whereas early treatment (within 3 months of stroke onset) was associated with a faster response than late treatment. The median numbers of sessions required to achieve a significant improvement in the early and late treatment groups were 8 and 11, respectively.

Conclusion: Stroke patients who were going to improve saw a significant improvement in Barthel index by the 16th acupuncture session and the marginal improvement declined substantially afterward. Our approach required minimal data available from the medical records of a non-academic facility, yet practical implication could be achieved.

Trial Registration: The study was retrospectively registered on 18 August 2017 with the Thai Clinical Trials Registry (ref. TCTR20170822002).