Choice of Outcomes Used in Randomized Controlled Trials Measuring the Efficacy and Safety of Massage Therapy for Cervical Spondylotic Radiculopathy

Background Massage is an important treatment for cervical spondylotic radiculopathy (CSR) . As massage-related research advances, numerous randomized controlled trials (RCT) concerning massage therapy in CSR have been published, but high-quality evidence is still limited due to some problems in the choice of outcomes. Objective To evaluate the outcomes used in RCTs in recent 10 years regarding CSR treated using massage, providing a basis for the choice of core outcomes used in studies about massage in CSR. Methods RCTs regarding CSR treated using massage were searched in databases of CNKI, Wanfang Data, CQVIP, SinoMed, PubMed, EMBase and Cochrane Library from January 2011 to May 2021. Literature screening, data extraction, and risk of bias assessment were performed by two researchers separately. A qualitative analysis was conducted to analyze the outcomes used in the RCTs. Results In all, 66 RCTs were included, in which the outcomes were categorized into 7 categories based on functional attributes: quality of life (95 times, 41.48%) , symptoms and signs (64 times, 27.95%) , physical and chemical examinations (39 times, 17.03%) , safety events (12 times, 5.24%) , economic assessment (12 times, 5.24%) , long-term prognosis (5 times, 2.18%) , and TCM symptoms/syndromes (2 times, 0.87%) . The most frequently used five outcomes were overall response rate, Visual Analogue Scale (VAS) score, Neck Disability Index score, adverse reactions and adverse events, symptom and sign score, and the measurement time points for which were 17 in total in a period from one day after treatment to the sixth month of follow-up. The above-mentioned outcome indexes are mainly measured at 14 days (27.08%) , 14 days (26.67%) , 14 days (29.17%) , 28 days (28.57%) and 14 days (33.33%) after treatment. The number of RCTs using one, two, three, and at least four outcomes was 5, 20, 14, and 27, respectively. Forty-eight RCTs reported the overall response rate. Conclusion We found many problems existing in the choice of outcomes used in RCTs assessing massage for CSR. The problems are as follows: primary and secondary outcomes were not defined; alternative outcomes received more attention while endpoint outcomes were neglected; the blinding method was ignored in RCTs using subjective outcomes; there was tremendous heterogeneity between RCTs in the number or combination of outcomes used, composite outcomes were mostly used; there was non-uniformity in measurement times of outcomes. We suggest that further relevant studies should pay attention to sufficient top-level design and preliminary research, and actively take actions to establish a set of core outcomes.