Psychological therapies delivered remotely for the management of chronic pain (excluding headache) in adults
Chronic pain (pain lasting three months or more) is an unpleasant sensory and emotional experience associated with actual or potential tissue damage. Common types (excluding headache) include back pain, fibromyalgia, and neuropathic pain. Access to traditional face‐to‐face therapies can be restricted by healthcare resources, geography, and cost. Remote technology‐based delivery of psychological therapies has the potential to overcome treatment barriers. However, their therapeutic effectiveness compared to traditional delivery methods requires further investigation.
To determine the benefits and harms of remotely‐delivered psychological therapies compared to active control, waiting list, or treatment as usual for the management of chronic pain in adults.
We searched for randomised controlled trials (RCTs) in CENTRAL, MEDLINE, Embase, and PsycINFO to 29 June 2022. We also searched clinical trials registers and reference lists. We conducted a citation search of included trials to identify any further eligible trials.
We included RCTs in adults (≥ 18 years old) with chronic pain. Interventions included psychological therapies with recognisable psychotherapeutic content or based on psychological theory. Trials had to have delivered therapy remote from the therapist (e.g. Internet, smartphone application) and involve no more than 30% contact time with a clinician. Comparators included treatment as usual (including waiting‐list controls) and active controls (e.g. education).
Data collection and analysis
We used standard Cochrane methodological procedures.
We included 32 trials (4924 participants) in the analyses. Twenty‐five studies delivered cognitive behavioural therapy (CBT) to participants, and seven delivered acceptance and commitment therapy (ACT). Participants had back pain, musculoskeletal pain, opioid‐treated chronic pain, mixed chronic pain, hip or knee osteoarthritis, spinal cord injury, fibromyalgia, provoked vestibulodynia, or rheumatoid arthritis. We assessed 25 studies as having an unclear or high risk of bias for selective reporting. However, across studies overall, risk of bias was generally low. We downgraded evidence certainty for primary outcomes for inconsistency, imprecision, and study limitations. Certainty of evidence ranged from moderate to very low. Adverse events were inadequately reported or recorded across studies. We report results only for studies in CBT here.
Cognitive behavioural therapy (CBT) versus treatment as usual (TAU)
Immediately after treatment, CBT likely demonstrates a small beneficial effect compared to TAU (standardised mean difference (SMD) ‐0.28, 95% confidence interval (CI) ‐0.39 to ‐0.16; 20 studies, 3206 participants; moderate‐certainty evidence). Participants receiving CBT are probably more likely to achieve a 30% improvement in pain intensity compared to TAU (23% versus 11%; risk ratio (RR) 2.15, 95% CI 1.62 to 2.85; 5 studies, 1347 participants; moderate‐certainty evidence). They may also be more likely to achieve a 50% improvement in pain intensity (6% versus 2%; RR 2.31, 95% CI 1.14 to 4.66; 4 studies, 1229 participants), but the evidence is of low certainty.
At follow‐up, there is likely little to no difference in pain intensity between CBT and TAU (SMD ‐0.04, 95% CI ‐0.17 to 0.09; 8 studies, 959 participants; moderate‐certainty evidence). The evidence comparing CBT to TAU on achieving a 30% improvement in pain is very uncertain (40% versus 24%; RR 1.70, 95% CI 0.82 to 3.53; 1 study, 69 participants). No evidence was available regarding a 50% improvement in pain.
Immediately after treatment, CBT may demonstrate a small beneficial improvement compared to TAU (SMD ‐0.38, 95% CI ‐0.53 to ‐0.22; 14 studies, 2672 participants; low‐certainty evidence). At follow‐up, there is likely little to no difference between treatments (SMD ‐0.05, 95% CI ‐0.23 to 0.14; 3 studies, 461 participants; moderate‐certainty evidence).
Quality of life
Immediately after treatment, CBT may not have resulted in a beneficial effect on quality of life compared to TAU, but the evidence is very uncertain (SMD ‐0.16, 95% CI ‐0.43 to 0.11; 7 studies, 1423 participants). There is likely little to no difference between CBT and TAU on quality of life at follow‐up (SMD ‐0.16, 95% CI ‐0.37 to 0.05; 3 studies, 352 participants; moderate‐certainty evidence).
Immediately after treatment, evidence about the number of people experiencing adverse events is very uncertain (34% in TAU versus 6% in CBT; RR 6.00, 95% CI 2.2 to 16.40; 1 study, 140 participants). No evidence was available at follow‐up.
Cognitive behavioural therapy (CBT) versus active control
Immediately after treatment, CBT likely demonstrates a small beneficial effect compared to active control (SMD ‐0.28, 95% CI ‐0.52 to ‐0.04; 3 studies, 261 participants; moderate‐certainty evidence). The evidence at follow‐up is very uncertain (mean difference (MD) 0.50, 95% CI ‐0.30 to 1.30; 1 study, 127 participants). No evidence was available for a 30% or 50% pain intensity improvement.
Immediately after treatment, there may be little to no difference between CBT and active control on functional disability (SMD ‐0.26, 95% CI ‐0.55 to 0.02; 2 studies, 189 participants; low‐certainty evidence). The evidence at follow‐up is very uncertain (MD 3.40, 95% CI ‐1.15 to 7.95; 1 study, 127 participants).
Quality of life
Immediately after treatment, there is likely little to no difference in CBT and active control (SMD ‐0.22, 95% CI ‐1.11 to 0.66; 3 studies, 261 participants; moderate‐certainty evidence). The evidence at follow‐up is very uncertain (MD 0.00, 95% CI ‐0.06 to 0.06; 1 study, 127 participants).
Immediately after treatment, the evidence comparing CBT to active control is very uncertain (2% versus 0%; RR 3.23, 95% CI 0.13 to 77.84; 1 study, 135 participants). No evidence was available at follow‐up.
Currently, evidence about remotely‐delivered psychological therapies is largely limited to Internet‐based delivery of CBT. We found evidence that remotely‐delivered CBT has small benefits for pain intensity (moderate certainty) and functional disability (moderate to low certainty) in adults experiencing chronic pain. Benefits were not maintained at follow‐up. Our appraisal of quality of life and adverse events outcomes post‐treatment were limited by study numbers, evidence certainty, or both. We found limited research (mostly low to very low certainty) exploring other psychological therapies (i.e. ACT). More high‐quality studies are needed to assess the broad translatability of psychological therapies to remote delivery, the different delivery technologies, treatment longevity, comparison with active control, and adverse events.
Benjamin A Rosser, Emma Fisher, Sadia Janjua, Christopher Eccleston, Edmund Keogh, Geoffrey Duggan
Plain language summary
Which remotely‐delivered psychological approaches help people with long‐term chronic pain to improve symptoms?
• Online cognitive behavioural therapy represents the most common remotely‐delivered psychological therapy. It may improve pain and disability in individuals experiencing chronic pain.
• It is largely unclear whether remotely‐delivered psychological therapies improve quality of life or cause harmful effects due to limited evidence, of often limited quality.
• We need more and better studies to investigate remotely‐delivered psychological therapies. Future studies should explore a broader range of technologies and therapies, and focus on possible unwanted effects.
Why consider remotely‐delivered psychological therapies for chronic pain?
Chronic pain is pain that lasts three months or longer. It is a common experience that can significantly impact on a person’s everyday life and well‐being. Psychological therapies have been found to improve mood and pain‐related disability. The most common psychological approach for chronic pain is cognitive behavioural therapy (CBT), which focuses on the interrelationship between thoughts, feelings, and actions, to support symptom management.
Unfortunately, gaining access to psychological therapies may be difficult. There are limited numbers of qualified healthcare professionals providing these services, and some people may find it physically difficult to attend clinics. Technologies (such as mobile phones, computers, and the Internet) may offer new ways of delivering psychological therapies directly to people within their everyday environment and without a healthcare professional being present. This approach (known as remote delivery) has the potential to help more people access therapy.
What did we want to find out?
We wanted to find out if remotely‐delivered psychological therapies:
• improve pain, disability, and quality of life (i.e. well‐being across life as a whole);
• cause any unintended harmful effects.
What did we do?
We searched for studies that compared remotely‐delivered psychological therapies with usual care or non‐psychological treatments (such as education about pain). We looked at study results at the end of treatment and up to one year after.
We compared and summarised the results of the studies and rated our confidence in the evidence, based on factors such as study methods and sizes.
What did we find?
We found 32 studies that included 4924 people with a range of chronic pain conditions, such as back pain, osteoarthritis, fibromyalgia, and rheumatoid arthritis. Average ages ranged from 24 to 67 years. Where those taking part were followed up after treatment ended, this follow‐up was between 3 and 12 months later; we did not include results collected after 12 months. Studies included in the review were carried out across 11 countries, with over half attributable to Sweden (9), the USA (6), and Australia (5). All studies were funded by government grants or charities, bar one study that did not state its funding source.
Studies investigated treatments based on the psychological therapies of CBT (25 studies) and acceptance and commitment therapy (ACT; 7 studies). One of the CBT studies included an additional group who received a positive psychology intervention. All therapies were delivered online, except one study using a smartphone app.
Our results only speak to therapy delivered by the Internet due to the lack of alternative forms of remote delivery in the studies.
• Compared to usual treatment (i.e. the standard support typically available), online CBT probably reduces pain and may reduce disability slightly. It is unclear whether online CBT improves quality of life or has unintended harmful effects.
• Compared to non‐psychological treatments for pain (e.g. education, online discussion boards), online CBT also probably reduces pain slightly. However, it probably makes little to no difference to quality of life, may make little or no difference to disability, and it is unclear whether it has unintended harmful effects.
• The benefits of online CBT compared to usual treatment are probably no longer present at 3 to 12 months after treatment ends. We do not know if this finding is also the case when compared to a non‐psychological treatment because the effects are unclear.
It is unclear whether other psychological therapies (such as ACT) lead to improvements because, overall, we are very uncertain of the available results.
What are the limitations of the evidence?
We have moderate confidence that pain is reduced by online CBT by the end of treatment, but this improvement is not present 3 to 12 months later. In addition, we have moderate confidence in our finding of no benefits of online CBT for disability and quality of life at follow‐up. However, we have little to very little confidence in our findings for ACT.
Three main factors reduced our confidence in the evidence. First, some of the studies were very small or there were not enough studies to be certain about their results. Second, where there were small numbers of studies for an outcome, the evidence did not cover a range of pain conditions, so we cannot assume that those findings would be the same across all types of chronic pain. Finally, the results were sometimes inconsistent across studies.
How up to date is this evidence?
The evidence is up to date to 29 June 2022.
Benjamin A Rosser, Emma Fisher, Sadia Janjua, Christopher Eccleston, Edmund Keogh, Geoffrey Duggan
Implications for practice
For adults with chronic pain (excluding headache and migraine)
We are moderately certain that remotely‐delivered cognitive behavioural therapy (CBT) provides small beneficial effects over treatment as usual (TAU) and active control for pain intensity. We found small beneficial effects of CBT over TAU in terms of functional disability, with low‐certainty evidence. We are moderately certain that remote CBT has little to no difference on quality of life compared to active control. Although findings were similar for CBT versus TAU, the evidence was of very low certainty. Overall, all beneficial effects were small, immediately post‐treatment, and were not maintained at follow‐up.
Few trials have evaluated remote delivery of other therapeutic approaches. These trials are predominantly limited to acceptance and commitment therapy (ACT), and the trials typically provide low‐ to very low‐certainty evidence. Consequently, we remain uncertain about the effectiveness of remote delivery of psychological therapies beyond CBT.
The available evidence pertaining to remote delivery of psychological therapies is nearly exclusively limited to Internet‐based interventions. Furthermore, we cannot reliably draw conclusions about potential harm associated with remote delivery of psychological therapies as we found very limited evidence on adverse events.
Remotely‐delivered CBT for adults experiencing chronic pain (excluding headache and migraine) may provide small, short‐term, beneficial effects for pain intensity, functional disability, anxiety, and depression compared to TAU. However, remote delivery evidence is currently predominantly limited to Internet‐based intervention. Current evidence for ACT is limited and of very low certainty. It is unclear whether other psychological therapies can also be successfully translated to remote delivery, given the evidence available.
Policy‐makers may consider remote delivery options for provision of psychological therapies for adults experiencing chronic pain as they may provide opportunities to improve treatment access. We find that evidence currently speaks primarily to CBT and Internet‐based delivery; other therapeutic approaches and technologies remain potentially useful but insufficiently researched. The beneficial effects of remotely delivered CBT appear small and short‐term, whereas face‐to‐face delivery may extend treatment effects (Williams 2020). Consequently, remotely‐delivered CBT may be considered alongside, rather than in replacement of, other evidentially‐established support, such as part of a stepped care approach.
For funders of interventions
The evidence supports the potential usefulness of remotely‐delivered psychological therapies for adults experiencing chronic pain, which may increase access to interventions providing short‐term, small benefits. The greatest evidence lies with translation of CBT to Internet‐based delivery. Notably, for those commissioning remotely‐delivered psychological therapies, we limited our review to interventions developed with the involvement of professionals with qualified expertise in psychological therapy. Whilst regulation exists for certain professional titles in psychology (e.g. Clinical Psychologist) and accreditation in particular therapeutic approaches (e.g. BABCP 2022), psychological interventions themselves are not regulated. Therefore, it remains possible for anyone to claim creation of CBT‐based interventions, irrespective of their knowledge or capability. Consequently, policy‐makers should look to the involvement of suitably‐trained healthcare professionals to determine the likely fidelity of the intervention to the psychological approach utilised.
Implications for research
Although we have found moderate evidence for beneficial effects of remotely‐delivered CBT, we do not consider the body of evidence to have yet reached the saturation point found for face‐to‐face therapies. Questions remain as to how remotely‐delivered interventions perform compared to active controls, the longevity of intervention effects, and their potential for harm. Beyond CBT, research should explore translation of a wider range of psychological therapeutic approaches for the management of chronic pain in adults. Whilst we identified ongoing trials focusing on approaches such as ACT (e.g. Slattery 2019b; Terhorst 2020), and emotional awareness and expression therapy (NCT04751825), overall CBT is still the dominant underpinning therapeutic approach in ongoing trials.
Whilst a broad range of delivery technologies were eligible for inclusion in our review, we found the evidence‐base to be near‐exclusively Internet‐based. Despite expectations of an increasing volume of research utilising smartphone applications (McGuire 2017), we found only one completed trial (Morcillo‐Muñoz 2022) and one ongoing trial (NCT05090683) using this technology. Review of the ongoing trials suggests that planned research of other novel technologies, such as virtual reality (Birckhead 2021; NCT04042090), is limited currently. Consequently, the individual and relative efficacy of different delivery technologies remains insufficiently explored at present.
Given the proposed importance of therapeutic alliance within psychological interventions (Horvath 2011; Zilcha‐Mano 2017), we recommend more direct research exploring relative levels of human involvement and type of involvement in remote delivery of psychological therapies. We identified few studies that did not involve some form of human contact. Even in interventions where that contact was not considered intentionally therapeutic, it remains possible that any human involvement may alter how intervention content is received and experienced. Currently, empirical comparison of different levels and type of involvement remains uncommon (e.g. Dear 2015; Lin 2017), and therefore, inconclusive. Finally, developments in artificial intelligence provide new potential avenues for enhanced emulation of interpersonal therapy components, alongside ethical issues, that warrant consideration (Fiske 2019).
We recommend further randomised controlled trials within this field. Good‐quality designs should be sufficiently powered in all trial arms and routinely include follow‐up as well as post‐treatment assessment. We encourage inclusion of active control comparisons as well as TAU and waiting‐list controls to facilitate determination of effects attributable to the psychological intervention. Consistent with the recommendations of Fisher 2019, researchers should seek to control for the influence of delivery technology in their selection of a suitable control comparison (e.g. Internet‐delivered psychoeducation). Research should provide clear and detailed overviews of intervention content and proposed mechanisms of actions. We recommend routine incorporation of proof of intervention fidelity to psychological therapeutic approach in trial reports. To support generation of efficient and targeted interventions, we also encourage consideration of designs enabling component analysis. The role and type of human involvement incorporated into intervention delivery may comprise one element of such a design and support better prediction of resources required and scalability of interventions. Correspondingly, fully automated interventions involving zero human contact are presently scarce. Finally, an area of common concern identified in GRADE assessment of the current evidence was retrospective trial registration and deviation of outcomes from those specified in the registered protocol. We emphasise the importance of prospective trial registration and consistency in specified outcomes between protocol and final report.
We make the following recommendations in terms of measurement.
- Trials should assess key outcomes associated with chronic pain using validated, standardised measures, as specified by the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) (Dworkin 2005). To facilitate standardisation and comparison, we recommend selection of measures that are most utilised with the field and have the greatest reliability and validity.
- In terms of choice of outcomes, priority should be given to measures capturing meaningful change in the participant's life over the use of symptoms as proxies for such change. Additionally, we encourage more frequent inclusion of behavioural assessment alongside self‐report measures.
- Trials investigating remotely‐delivered therapies should routinely include assessment of engagement and adverse outcomes. Additionally, we encourage collaboration within the research community to develop more standardised methods for capturing these data and improving cross‐study comparison.
- Trials should routinely include a follow‐up assessment at least three months' post‐intervention to support evaluation of the longevity of intervention effects.