Article Text
Statistics from Altmetric.com
Commentary on: OpenUrlCrossRefPubMedWeb of Science
Context
Gustavsson et al aim to compare the long-term effects of a multicomponent pain and stress self-management group intervention (PASS) and individually administered physical therapy (IAPT) on patients with persistent tension-type neck pain in a primary healthcare setting.
Methods
Patients with neck pain seeking physical therapy treatment at nine physiotherapy health centres in Sweden were consecutively recruited from 2004 to 2006. They were examined by a physical therapist and included if they were 18–65 years of age and had persistent tension-type neck pain of 3 months or more. Exclusions were having received a similar programme previously, pregnancy, language and psychosis (one each), depression (n=16) or receipt of other treatment (n=5).
Patients were randomly allocated into two groups and were stratified by treatment centre. The PASS group addressed pain control, self-efficacy and catastrophising in a group educational format consisting of seven 1.5 h lectures with a booster session at 20 weeks whereas the IAPT group was provided with individual pragmatic multimodal physiotherapy interventions including: specific exercises, spinal mobilisation, traction and acupuncture. The participants were not blinded to group allocation.
The groups were self-assessed using several standardised questionnaires including: pain (10 point visual analogue scale), analgesic use, pain Coping Strategies Questionnaire (CSQ), the Self Efficacy Scale (for performing daily activities), the Neck Disability Index and the Hospital Anxiety and Depression Scale. The assessors were not blinded in this study.
Findings
This study presents the data of 1 and 2 years of follow-up from the original randomised controlled trial (RCT) that reported 20-week outcomes. The level of significance was set at p=0.01 in the attempt to present results that were clinically meaningful.
The key finding of the study is that gains noted in the original RCT were largely maintained after 2 years. Both groups improved over time. Linear mixed models for repeated measures analyses showed significant time-by-group interaction effects in favour of PASS regarding the outcomes: ability to control pain (p<0.001), self-efficacy for performing activities in spite of pain (p=0.002), and catastrophic thinking according to CSQ catastrophising subscale (p<0.001). There was no difference in disability in analyses adjusted for baseline differences at the start of the study (less disability in the PASS group).
Pair-wise comparisons of pain coping strategies (measured by the CSQ) by group showed significant differences at 10-week, 20-week and 1-year follow-ups, but not at the 2-year follow-up.
Between-group comparisons for categorical data demonstrated no differences between groups in consumption of analgesics at 1 or 2 years. At 1 year (p=0.001) and 2 years (p=0.001), the PASS group reported a higher satisfaction with treatment received during the intervention, than the IAPT-group.
At the 1-year follow-up the PASS group reported that during treatment they had learnt useful skills, which they could apply in everyday life to cope with pain, to a significantly higher degree than the IAPT group (p<0.002).
Commentary
The cohort used in this study was a pragmatic representation of normal physiotherapy practice. The treatments used were also representative. Some of the outcome effect sizes were small (10, 20 and 52 week follow-up for the PASS-IAST comparisons) suggesting that significance may not be clinically meaningful. In addition, the statistical power was reduced due to drop-outs from the study (35%). To address drop-outs, investigators imputed outcome values for all subjects with missing data by using the last known value. They analysed data according to intention-to-treat (regardless of treatment received) as well as per protocol. Results were similar and they reported the intention to treat results.
There were more participants in the control group that were ‘totally off work’ (32% vs 13%) and used medication ‘everyday’ (22% vs 13%) and had higher disability (mean scores 35.4 vs 30.8) highlighting the likelihood that the control group consisted of more disabled individuals.
The use of a different behavioural treatment has been investigated before with similar results.1 Another study, which examined the role of spinal manipulation with and without exercise, found improvements in outcomes when exercise was added to the manipulation.2 Perhaps a multimodal approach incorporating all forms of intervention would yield better results? Further study is required to investigate these interesting possibilities.
Future research should attempt to determine predictors of outcome using longitudinal designs. The measurement of cervicogenic headache or the inclusion of radicular symptoms would better approximate the scope of cervicogenic conditions seen by physiotherapists.
This study provides good evidence for the inclusion of behavioural control and self-efficacy outcomes by self-management into management by physiotherapists. Finally, if replicated, this study has implications for cost containment through the replacement of more expensive individualised care with more cost effective group interactions.
Footnotes
-
Competing interests None.