Many people experience pain that lasts more than three months that is neither cancer‐related nor a headache. The search for a diagnosis and pain relief is often long and can be discouraging. In this interview, lead author of the recently updated Cochrane Review Psychological therapies for the management of chronic pain (excluding headache) in adults Amanda C de C Williams discusses the findings of this review.
Can you tell us about this Cochrane Review and what is new about this update?
This review is the third in Cochrane, and the fourth overall, attempting to answer the question of efficacy of psychological interventions to help people with chronic (persistent) pain. These interventions use a variety of methods targeting behaviour, thoughts, and emotions, to try to enable people with chronic pain to lead more satisfying lives, less limited by pain. These interventions have been used now for about 50 years, and there have been many trials. This review included 75 studies, with 9401 patients completing treatment. This was much larger than our previous Cochrane review in 2012, so we hoped to get a clearer answer as to whether these interventions reduced disability, distress, and pain.
What does the evidence tell us?
We divided the interventions into three groups. The largest was cognitive behavioural therapy (CBT), which aims to help people identify and change unhelpful habits of thinking, and over-pessimistic beliefs about their pain and what it means, alongside ways of encouraging them to increase their range and extent of activities. Then there were two small groups: behaviour therapy (BT), which aims to change people’s habits directly, without cognitive methods, and acceptance and commitment therapy (ACT), which takes a similar behavioural approach to CBT but a different cognitive approach, teaching people to disengage from negative thinking patterns (rather than changing them) and to focus on pursuing valued activities.
For CBT, where we had 59 studies to combine, there was a negligible positive effect of the intervention on pain levels, and small but robust improvements on disability and distress, all assessed by the individuals’ own reports. Evidence quality was mostly moderate, some low. The improvements in all outcomes were more marked, lasting to follow-up at 6 to 12 months, when treatment was compared with waiting list or treatment as usual than when it was compared with another treatment, such as an exercise programme, educational intervention, or support group.
We found only 8 studies of BT, with little change in pain, disability, or distress. The quality of evidence was variable and the treatment methods used were quite diverse, so it is hard to draw any helpful conclusions. There were 5 studies of ACT, and all evidence was of very low quality; the results showed a mix of benefits and no effects, but are so uncertain that we have no confidence interpreting them.
None of the interventions provided even moderate quality evidence on adverse effects, which for psychological interventions are likely to be worsening of symptoms, or dropout because of worsening or disappointment with extent or rate of progress.
What implications are there for further research?
After our 2012 Cochrane review, we wrote an editorial in the journal PAIN saying that we did not need more studies of CBT, testing minor variations in elements of treatment. However, that is largely what we found in our 2020 update. We now know that there are robust average effects, in that it would take a large number of trials finding no benefit to overturn them, but we know little about the reasons for variation around these averages. We are planning another editorial …
For BT, although the interventions are in principle distinct from one another, and have a strong basis in acute pain studies, the trials for chronic pain often combine several elements, making effects even harder to interpret. We need to start with well-powered trials that test each technique rather than combining them. Our results for ACT are in marked contrast to the several systematic reviews and meta-analyses that included many small trials (we excluded any with less than 20 people in any trial arm at the end of treatment): these report it to be very effective, albeit mainly for self-reported acceptance of pain rather than for disability or distress. ACT methods have been widely adopted, often replacing CBT, but the evidence we found just does not support this. One of the problems in psychology trials, particularly with newer treatments such as ACT, is that trials are run by enthusiasts for the new treatment. For the next generation of trials, we need more even-handedness.
What message can policy-makers take from this study?
The diversity of psychological interventions can be bewildering for policy-makers, but some distinctions are important: not all ‘talking therapies’ are the same, or equally effective. Our evidence supports the use of CBT to reduce disability and distress associated with chronic pain, when delivered by trained psychologists, one of the criteria for including trials in our review. Psychology is not ‘common sense’, nor can it be applied without some understanding of psychological processes and without skills to deal with patients’ distress. For the other two types of intervention, BT and ACT, outcomes are uncertain, so effects, both beneficial and adverse, should be monitored and reported.