Effect of yoga on chronic non-specific neck pain: An unconditional growth model
- Publication Type:
- Journal Article
- Complementary Therapies in Medicine, 2018, 40 pp. 237 - 242
- Issue Date:
Copyright Clearance Process
- Recently Added
- In Progress
- Open Access
This item is currently unavailable due to the publisher's embargo.
The embargo period expires on 31 Dec 2018
© 2017 Elsevier Ltd Objective: Chronic neck pain is a common problem that affects approximately half of the population. Conventional treatments such as medication and exercise have shown limited analgesic effects. This analysis is based on an original study that was conducted to investigate the physical and behavioral effects of a 9-week Iyengar yoga course on chronic non-specific neck pain. This secondary analysis uses linear mixed models to investigate the individual trajectories of pain intensity in participants before, during and after the Iyengar yoga course. Method: Participants with chronic non-specific neck pain were selected for the study. The participants suffered from neck pain for at least 5 days per week for at least the preceding 3 months, with a mean neck pain intensity (NPI) of 40 mm or more on a Visual Analog Scale of 100 mm. The participants were randomized to either a yoga group (23) or to a self-directed exercise group (24). The mean age of the participants in the yoga group was 46, and ranged from 19 to 59. The participants in the yoga group participated in an Iyengar yoga program designed to treat chronic non-specific neck pain. Our current analysis only includes participants who were initially randomized into the yoga group. The average weekly neck pain intensity at baseline, during and post intervention, comprising 11 total time points, was used to construct the growth models. We performed a step-up linear mixed model analysis to investigate change in NPI during the yoga intervention. We fit nested models using restricted maximum-likelihood estimation (REML), tested fixed effects with Wald test p-values and random effects with the likelihood ratio test. We constructed 10 REML models. Results: The model that fit the data best was an unconditional random quadratic growth model, with a first-order auto-regressive structure specified for the residual R matrix. Participants in the yoga group showed significant variation in NPI. They demonstrated variation in their intercepts, in their linear rates of change, and most tellingly, in their quadratic rates of change. Conclusions: While all participants benefitted from the yoga intervention, the degree to which they benefitted varied. Additionally, they did not experience a consistent rate of reduction in NPI − their NPI fluctuated, either increasing and then decreasing, or vice-versa. We comment on the clinical and research implications of our findings.
Please use this identifier to cite or link to this item: