|Home | About | Journals | Submit | Contact Us | Français|
Chronic pain affects many children, who report severe pain, distressed mood, and disability. Psychological therapies are emerging as effective interventions to treat children with chronic or recurrent pain. This update adds recently published randomised controlled trials (RCTs) to the review published in 2009.
To assess the effectiveness of psychological therapies, principally cognitive behavioural therapy and behavioural therapy, for reducing pain, disability, and improving mood in children and adolescents with recurrent, episodic, or persistent pain. We also assessed the risk of bias and methodological quality of the included studies.
Searches were undertaken of MEDLINE, EMBASE, and PsycLIT. We searched for RCTs in references of all identified studies, meta-analyses and reviews. Date of most recent search: March 2012.
RCTs with at least 10 participants in each arm post-treatment comparing psychological therapies with active treatment were eligible for inclusion (waiting list or standard medical care) for children or adolescents with episodic, recurrent or persistent pain.
All included studies were analysed and the quality of the studies recorded. All treatments were combined into one class: psychological treatments; headache and non-headache outcomes were separately analysed on three outcomes: pain, disability, and mood. Data were extracted at two time points; post-treatment (immediately or the earliest data available following end of treatment) and at follow-up (at least three months after the post-treatment assessment point, but not more than 12 months).
Eight studies were added in this update of the review, giving a total of 37 studies. The total number of participants completing treatments was 1938. Twenty-one studies addressed treatments for headache (including migraine); seven for abdominal pain; four included mixed pain conditions including headache pain, two for fibromyalgia, two for pain associated with sickle cell disease, and one for juvenile idiopathic arthritis. Analyses revealed five significant effects. Pain was found to improve for headache and non-headache groups at post-treatment, and for the headache group at follow-up. Mood significantly improved for the headache group at follow-up, although, this should be interpreted with caution as there were only two small studies entered into the analysis. Finally, disability significantly improved in the non-headache group at post-treatment. There were no other significant effects.
Psychological treatments are effective in reducing pain intensity for children and adolescents (<18 years) with headache and benefits from therapy appear to be maintained. Psychological treatments also improve pain and disability for children with non-headache pain. There is limited evidence available to estimate the effects of psychological therapies on mood for children and adolescents with headache and non-headache pain. There is also limited evidence to estimate the effects on disability in children with headache. These conclusions replicate and add to those of the previous review which found psychological therapies were effective in reducing pain intensity for children with headache and non-headache pain conditions, and these effects were maintained at follow-up.
Psychological therapies (relaxation, hypnosis, coping skills training, biofeedback, cognitive behavioural therapy) are treatments that may help people manage pain and its disabling consequences. For children and adolescents there is good evidence that both relaxation and cognitive behavioural therapy (treatment that helps people test and revise their thoughts and actions) are effective in reducing the severity and frequency of pain in chronic headache, recurrent abdominal pain, fibromyalgia, sickle cell disease, and juvenile idiopathic arthritis immediately after treatment is delivered. Psychological therapies also have a lasting effect for improving mood and reducing pain for chronic headache. Forty-nine per cent of children who received psychological therapies reported less pain compared with 17% of children who did not receive a psychological therapy. Disability is improved immediately after treatment for many pain conditions (not chronic headache) which helps young people to participate in important daily activities. More studies are needed to understand whether psychological therapies can improve mood and have more lasting effects on pain and disability in other groups of young people who have chronic pain.
This review is an update of a previously published review in The Cochrane Database of Systematic Reviews (Eccleston 2009) on ’Psychological therapies for the management of chronic and recurrent pain in children and adolescents’. Chronic and recurrent pain (pain lasting more than three months) is a common problem in young people. Recent epidemiology gives the prevalence at 15% to 30%, with 8% of children described as having severe and frequent pain (Perquin 2000; Perquin 2001; Stanford 2008). The most common location for pain is in the head, abdomen, and limbs (Perquin 2000). All types of chronic and recurrent pain are more commonly reported by girls, and peak in incidence at ages 14 to 15 years (Stanford 2008). Young people report pain to be distressing and interfering, and in some cases this can be severely debilitating, affecting all aspects of a child’s life (Bursch 1998; Palermo 2000) and the lives of their parents and family members (Walker 1989; Palermo 2005). The deleterious effects of untreated pain in childhood can also extend to adulthood (Fearon 2001).
There is a broad family of treatments included in the general term “psychological”. In essence, treatments have been developed that are specifically designed to alter psychological processes thought to underlie or significantly contribute to pain, distress, and/or disability. The design of psychological treatments is normally informed by specific theories of the etiology of human behaviour, or have developed pragmatically through observation and study of response to intervention. Behavioural and cognitive treatments designed to ameliorate pain, distress and disability were first introduced with adults over 40 years ago and have become well established (Fordyce 1968; Keefe 2004). A companion review of psychological treatments for the management of chronic pain in adults is also published (Eccleston 2009a).
In paediatric practice, however, the treatments have a shorter history and different therapeutic aims and components than those used with adults. In general, psychological treatments have aimed to control pain and modify situational, emotional, familial, and behavioural factors that play a role in pain or related consequences (e.g., McGrath 1990). A variety of intervention strategies have been designed to reduce pain sensations, increase comfort, and/ or reduce associated disability and dysfunction in children with pain conditions. Behavioural strategies include relaxation training, biofeedback, and behavioural management programs (e.g., teaching parents operant strategies to reinforce adaptive behaviours such as school attendance). Cognitive strategies include hypnosis, stress management, guided imagery, and cognitive coping skills (Palermo 2012).
Cognitive behavioural programs incorporate elements of both behavioural and cognitive strategies. Given that headache and abdominal pain are the most common types of recurrent pain in children, most of the treatment literature has focused on these two populations. By far the most commonly described treatment is relaxation training and/or biofeedback for headache, and recommendations have been made to offer psychological treatment as a matter of routine care for children with headaches (Masek 1999). In an effort to enhance the efficiency of psychological treatments for children with headache, more recent treatment developments have compared different elements of relaxation training and biofeedback with a variation in treatment formats (individual and group), treatment dose, and treatment setting (clinic, school, and home).
Psychological therapies have also been developed to treat children with non-headache chronic and recurrent pain including children with abdominal, musculoskeletal, and disease-related pain. Multidisciplinary pain treatment programs for children have recently become a standard of care (McGrath 1999a), and now many specialised pain clinics are available for children with chronic or recurrent pain, which may involve outpatient care or intensive inpatient rehabilitation. Such programs offer physical rehabilitation, psychological treatment, and medical strategies, and aim to restore function rather than provide pain relief. There is evidence from case series and uncontrolled studies for the effectiveness of multidisciplinary treatment with psychological therapy for paediatric chronic and recurrent pain (Eccleston 2003b).
Several reviews have documented the effectiveness of psychological therapies for children with headache, abdominal, and disease-related pain (Janicke 1999; Kibby 1998; Holden 1999; Huertas-Ceballos 2008; Walco 1999; Weydert 2003). Three reviews have used data pooling techniques for studies of children with headache (Eccleston 2009; Hermann 1995; Trautmann 2006). In their review of paediatric migraine Hermann 1995 found that biofeedback and muscle relaxation are more effective than placebo treatments and prophylactic drug treatments in controlling headache. In the previous published Cochrane review (Eccleston 2009), we found that psychological treatments were effective in reducing pain intensity in youths with headache. Trautmann 2006 conducted a meta-analysis of psychological treatment for recurrent headache in children finding small effect sizes across three headache variables: frequency, duration, and intensity, although reduction in pain intensity at post-treatment was a statistically significant effect. A large binomial effect size of 50% or greater reduction in headache symptoms was reported.
The previous Cochrane review (Eccleston 2009) is now out of date (Shojania 2007). Developments in paediatric psychology have led to new populations of children being treated. The aim of this review is to further update the published evidence on the efficacy of psychological treatments for chronic and recurrent pain in children and adolescents in outcomes with chronic, non-headache, pain.
Randomised controlled trials (RCTs) comparing a credible psychological treatment, or a compound treatment with credible primary psychological content, active treatment, treatment as usual, or waiting list control, in paediatric chronic or recurrent pain. Content was judged credible if it referred to a scientific theory of psychological action. Studies were excluded if the pain was associated with a malignant life-threatening disease.
Studies were included if they:
Children and adolescents (<18 years) reporting persistent, recurrent or episodic pain in any body site, not associated with cancer or similar life threatening malignant disease.
Studies were included if at least one trial arm consisted of a psychological intervention, and a comparator arm used, active treatment, treatment as usual, or waiting list control.
Randomised controlled trials (RCTs) of any psychological therapy for paediatric chronic or recurrent pain were identified by searching MEDLINE, EMBASE and PsycINFO from their inception to March 2012. The Cochrane Central Register of Controlled Trials (CENTRAL) was searched in the previous versions of this review (Eccleston 2003a; Eccleston 2009). However, for this update of the review it was agreed by all authors that the sensitivity of CENTRAL for pain trials was low and that all updated trials could be identified through the alternative databases and by contacting study authors for further trials. Three separate searches have been undertaken. The first search was undertaken from inception of the abstracting services to the end of 1999 (Eccleston 2003a), the second searched databases from 1999 to 2008 (Eccleston 2009), and the third searched databases from 2008 to March 2012.
The selection of included studies was made using the following criteria; the study had to be RCT in design and published in a peer-reviewed journal, include children (<18 years of age) who have chronic pain (non-malignant), include a psychological intervention as an active treatment, and have > 10 participants in each arm at post-treatment. Psychological interventions were considered for inclusion if they had credible, recognisable psychological/psychotherapeutic content and were specifically designed to change the child’s behaviour, cognition, and/or outcomes. There were no additional criteria (regarding sample size or other) added for the purposes of this update, therefore, the trials used in the previous systematic review and meta-analysis (Eccleston 2009) were considered automatically eligible and included.
Data were extracted on details relating to the design of the study, the participants, primary diagnosis, method of treatment, outcome measurement tools used, and outcome data for computation of effect sizes. When data were missing on primary outcomes of interest, we contacted trial authors via email to obtain data necessary for effect size calculations. Data suitable for pooling were entered into RevMan 5.1.
The risk of bias was measured using the recommended Cochrane ’Risk of bias’ tool (Higgins 2011). We selected five categories from this tool; random sequence generation (selection bias), allocation concealment (selection bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), and selective reporting (reporting bias). Two categories were excluded for the purposes of this review as they were deemed redundant because of the nature of delivering or receiving a psychological intervention. The categories were ’blinding (performance bias and detection bias)’ and ’blinding participants and personnel’. Judgements were made on the categories using the following rules. Random sequence generation judgements were based on whether authors gave a convincing method of randomisation. Allocation concealment bias judgements were based on whether there were convincing methods used for random allocation to take place. Participants being stratified by age or gender were not deemed as biased. Blinding of outcome assessment was judged on whether the measures were taken by a third party who was blind to the treatment condition. Incomplete outcome data bias judgements were based on whether attrition was fully reported. Authors had to report attrition at each measurement time point (post-treatment and follow-up), and state whether there were any significant differences between completers and non-completers. Finally, selective reporting bias was judged on whether data could be fully extracted for analyses in this review. If authors provided data when requested, this category would be marked as ’low bias’.
A quality rating scale designed specifically for studies of psychological treatments in chronic pain was applied (Yates 2005). All papers were scored by two of the review authors and consensus reached after initial comparison of ratings. Interrater reliability was calculated. The quality rating scale provides an overall total quality score (zero to 35) consisting of two subscales: a treatment quality scale (zero to nine) covering stated rationale for treatment, manualisation, therapist training, and patient engagement; and a design and methods scale (zero to 26) covering inclusion/exclusion criteria, attrition, sample description, minimisation of bias (randomisation method, allocation bias, blinding of assessment, and equality of treatment expectations), selection of outcomes, length of follow-up, analyses, and choice of control. This comprehensive quality assessment method was used in place of the single item of allocation bias.
All treatments labelled as psychological were combined in the following meta-analyses, and designated “Treatment”. Similarly, all control conditions were combined and designated “Control”. Where more than one intervention or control group was reported the intervention or control arms were combined to create a single pair-wise comparison in accordance with the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). The studies were divided into two groups. The first group was labelled “headache” and the second group was labelled “non-headache”. Two assessment points were also selected: post-treatment and follow-up. Post-treatment is the assessment point occurring soonest following treatment (often after a delay of several weeks to allow for recording of episodic pain), and follow-up is the assessment point at least three months after the post-treatment assessment point, but not more than 12 months, and the longer time point was selected if there were two follow-up assessments within this time frame. Therefore, four separate comparisons were designed comprising two forms of comparator (Treatment, Control) and two assessment time points (post-treatment and follow-up). They were labelled as follows.
Multiple measurement tools were typically used in each study. For each comparison, three outcomes were identified and labelled “Pain”, “Disability”, and “Mood”. From each trial we selected the measure considered most appropriate for each of the three outcomes. To guide the choice of outcome measure, we applied two rules. First, if an outcome measure was established and occurred frequently among studies it was selected over more novel instruments. Second, given a choice between single item and multi-item self-report tools, multi-item tools were chosen on the basis of inferred increased reliability. Studies did not necessarily report data in all three outcomes. For headache treatments, the data for pain outcomes were dichotomous so relative ratios or risk ratios (RR) were used, and we calculated numbers-needed-to-treat-to-benefit (NNTBs). For disability and mood outcomes, continuous data were used. Continuous data were used for pain, disability, and mood for non-headache studies. Effect sizes can be interpreted as follows; small = 0.2, medium = 0.5, large = 0.8 (Cohen 1992).
For dichotomous outcomes, such as achieved (or failed to achieve) 50% reduction in pain, we calculated the odds ratio (OR) using 95% confidence intervals (CI) and a random-effects model. For ease of interpretation, the risk ratio (RR), and NNTB are also reported. For continuous outcomes (such as rating scales) we calculated the standardized mean differences using 95% CI and a random-effects model. The heterogeneity of the findings is also reported.
Three separate searches have been undertaken. The first search was undertaken from inception of the abstracting services to the end of 1999 (Eccleston 2003a). This yielded 3715 abstracts, of which 123 were read in full, identifying 18 RCTs. The second search, which updated the original review, was undertaken focusing on the 10 years since the previous search, overlapping by one year; from 1999 to 2008 and was later published (Eccleston 2009). This yielded 1319 abstracts, of which 45 papers were read in full, identifying a further 16 RCTs, giving a total set of 34. However, five studies were later excluded because they did not meet the minimum criteria of 10 participants in each arm, therefore, leaving 29 studies. The third, which searched databases from 2008 to March 2012 yielded 851 abstracts, of which 25 papers were read in full, and eight further RCTs were included in the review, giving a total of 37 RCTs (Abram 2007; Alfven 2007; Barakat 2010; Barry 1997; Bussone 1998; Connelly 2006; Duarte 2006; Fichtel 2001; Gil 1997; Griffiths 1996; Hicks 2006; Humphreys 2000; Kashikar-Zuck 2005; Kashikar-Zuck 2012; Kroener-Herwig 2002; Labbe 1984; Labbe 1995; Larsson 1987a; Larsson 1987b; Larsson 1990; Larsson 1996; Levy 2010; Palermo 2009; McGrath 1988; McGrath 1992; Osterhaus 1997; Passchier 1990; Richter 1986;Robins 2005; Sanders 1994; Sartory 1998; Scharff 2002; Stinson 2010; Trautmann 2010; van Tilburg 2009; Vlieger 2007; Wicksell 2009). Nine further studies did not meet the inclusion criteria and were excluded (Fentress 1986; Kroener-Herwig 1998; Larsson 1986; Olness 1987;Sanders 1989; Trautmann 2008; Vlieger 2012; Weydert 2006; Youssef 2009).
Of the 37 included studies, eight are new to this update (Barakat 2010; Kashikar-Zuck 2012; Levy 2010; Palermo 2009; Stinson 2010; Trautmann 2010; van Tilburg 2009; Wicksell 2009). The total number of participants completing treatments was 1938, 506 more than the previous published analyses. Of the 37 studies, one had four treatment arms, 11 had three arms, and 25 had two arms. The mean number of participants per study at the end of treatment was 52 (sd = 34). Girls outnumbered boys in 26 studies, and boys outnumbered girls in nine, (65% girls, range 22 to 100%). Age was reported in 35 studies (mean = 11.6 years, sd = 4.1 years). Only 16 studies reported the duration of pain, with a mean of 3.8 years.
Participants were recruited from a range of healthcare settings and other sources. Nineteen studies recruited from hospital or clinic settings, four from schools, six from direct advertisement, one study recruited from hospital/clinics and schools, two studies recruited from hospitals and direct advertisements, and five did not report their source. There were 21 studies of treatments for children with headache (including migraine). Of the remainder, seven were for abdominal pain (Alfven 2007; Duarte 2006; Humphreys 2000; Levy 2010; Robins 2005; Sanders 1994; van Tilburg 2009), and one study treated participants with either a primary diagnosis of abdominal pain or a primary diagnosis of irritable bowel syndrome (Vlieger 2007). Two studies treated children with fibromyalgia (Kashikar-Zuck 2005; Kashikar-Zuck 2012), two were for the treatment of pain associated with sickle cell disease (Barakat 2010; Gil 1997) and a further three studies included mixed pain conditions including headache and non-headache pain (Hicks 2006; Palermo 2009; Wicksell 2009) and so were included in both analyses where appropriate. Five study authors provided additional data upon request.
Treatment arms were classified on the basis of their content and of the label given by the study authors. The interventions were categorised into two broad groups. The first is best described as behavioural, typically relaxation based, with or without biofeedback, and including autogenic or hypnotherapeutic content (Bussone 1998; Fichtel 2001; Labbe 1984; Labbe 1995; Larsson 1987a; Larsson 1987b; Larsson 1990; Larsson 1996; McGrath 1988; McGrath 1992; Passchier 1990; Vlieger 2007). The second is best described as cognitive behavioural therapy, including cognitive coping, coping skills training, and parent operant strategies (Abram 2007; Alfven 2007; Barakat 2010; Barry 1997; Connelly 2006; Duarte 2006; Gil 1997; Griffiths 1996; Hicks 2006; Humphreys 2000; Kashikar-Zuck 2005; Kashikar-Zuck 2012; Kroener-Herwig 2002; Levy 2010; McGrath 1992; Osterhaus 1997; Palermo 2009; Richter 1986; Robins 2005; Sanders 1994; Sartory 1998; Scharff 2002; Stinson 2010; Trautmann 2010; van Tilburg 2009; Wicksell 2009). Different control conditions were employed that were categorised into either active control (e.g. education, n = 19) or wait-list/treatment as usual control (n = 18). Ten studies reported post-treatment data only; five studies reported follow-up of three months or less and therefore were not included in analyses. Fifteen studies reported follow-up data between three months and a year. Thirty studies reported the treatment length which was typically short in duration (M = 5.3 hours for headache studies, M = 3.5 hours for non-headache studies, Table 1). Seven studies did not report on the duration of treatment (Alfven 2007; Connelly 2006; Hicks 2006; Humphreys 2000; Kashikar-Zuck 2005; Sartory 1998; Trautmann 2010).
Treatment delivery varied between studies (Table 1). Eighteen studies delivered treatment in a clinic, eight studies delivered treatment at home, of which five used an internet- or computer-based delivery, three were based either in a clinic or at home, so exposure to treatment was uncontrolled. A further three were based in schools, and five were unknown. Home maintenance or practice of treatment was a common and important feature of many studies, but overall treatment exposure including home practice was not reported. Supervised treatment or the introduction of treatment in self-delivery conditions was, however, commonly reported.
Nine studies were excluded, of which four are new to this update (Trautmann 2008; Vlieger 2012; Weydert 2006; Youssef 2009). Seven had fewer than 10 participants in a treatment arm at the end of treatment (Fentress 1986; Kroener-Herwig 1998; Larsson 1986; Sanders 1989; Trautmann 2008; Weydert 2006; Youssef 2009), one study was judged to have insufficient psychological content in the treatment (Olness 1987), and one study reported only follow-up data of more than one year (Vlieger 2012).
All included studies were rated for risk of bias on five categories; random sequence generation (selection bias), allocation concealment (selection bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), and selective reporting (reporting bias) (Figure 1; Figure 2). Thirteen studies were scored as low risk of bias and gave a convincing method of randomisation, a further 24 studies were judged unclear on random sequence generation as they did not provide an adequate method of randomisation. There were none that were scored as having high risk of bias. There were four studies that were judged to have a low risk of bias and gave a convincing method for the allocation, 22 studies that were unclear and 11 studies that had a high risk of bias. Six studies used a third person who was blinded to the group allocation when taking measurements, 31 studies did not report this and so were unclear. Fifteen studies reported attrition fully, reporting that there was no significant difference between completers and non-completers. Seventeen studies only partially reported attrition and so were judged to be unclear and five studies did not report attrition so were judged to have a high risk of bias. Twenty-two studies reported data fully that could be extracted and used in analyses; 15 studies did not provide full extractable data and were judged to have high risk of bias for selective reporting.
All 37 studies were rated for their quality. For the previously included 29 studies, the quality was rated by two review authors blind to each other’s ratings. Intraclass correlation coefficients (absolute agreement) for the two raters on each of the three quality scores (trial quality, study design, and total) ranged from 0.87 for trial quality, 0.89 for study design, and 0.90 for total quality. The quality for the updated eight studies was also rated by two review authors (EF, EL) and disagreements were arbitrated by a third review author (CE) meaning that agreement was met on all items. For the 37 included studies, the mean overall quality of the studies was 20.3 (sd = 5.6, range = 11 to 32). The mean design quality score was 14.4 (sd = 4.1, range = 8 to 23) and the mean treatment quality score was 5.9 (sd = 2.0, range = 2 to 9). The ’Risk of bias’ tables’ graphics show the overall total quality rating, and summary scores for the quality of the treatment and the quality of the design (Figure 1; Figure 2). A Spearman’s correlation was conducted to investigate the association between year of study and treatment quality score, design quality score, and total quality score (sum of treatment quality and design quality). The total quality score was associated with the year of study (rho = 0.598, P < 0.001). Treatment quality and design quality were also associated with the year of study (rho = 0.596, P < 0.001; rho = 0.556, P < 0.001) respectively. Similarily, design quality was associated with treatment quality (rho = 0.662, P < 0.001). There were no differences in overall quality, treatment or design quality between headache and non-headache studies. The N at the end of treatment was not associated with total quality score, or design and treatment quality scores when considered independently. This finding remains after accounting for trial complexity (i.e., the number of arms of a trial). Comparing the results of the quality of the studies reported in the previous published review (n = 29) and for this update (n = 8), those published since Eccleston 2009 are of better overall quality (Mann Whitney U = 18.5, P < 0.001), treatment quality (Mann Whitney U = 21.0, P < 0.001) and design quality (Mann Whitney U = 28.5, P < 0.001). Finally, comparing those published before and since the previous version of this review (Eccleston 2009) there was no significant change in the size of the trial, that is, there was not an increase in the end of treatment N.
Twelve analyses were attempted (pain, disability, and mood outcomes were analysed for headache and non-headache conditions at post-treatment and follow-up). One comparison had only one study contributing and so was excluded. Of the remaining 11 comparisons, five showed low heterogeneity (I2 < 25%), four showed modest heterogeneity (I2 > = 25% − < 50%) and two showed large heterogeneity (I2 > = 50%).
See: Summary of findings for the main comparison Psychological therapies for the management of chronic pain (headache) in children and adolescents; Summary of findings 2 Psychological therapies for the management of chronic (non-headache) in children and adolescents
Eighteen studies of 748 participants were entered into an analysis of the effects of psychological treatment on pain immediately post-treatment (Barry 1997; Connelly 2006; Fichtel 2001; Griffiths 1996; Hicks 2006; Kroener-Herwig 2002; Labbe 1984; Labbe 1995; Larsson 1987a; Larsson 1987b; Larsson 1990; Larsson 1996; McGrath 1992; Osterhaus 1997; Palermo 2009; Sartory 1998; Scharff 2002; Trautmann 2010). This analysis gave an risk ratio (RR) of 2.90 (95% confidence interval (CI) 2.25 to 3.73; z = 8.25, P < 0.05) for a clinically significant change in pain (NNTB = 2.72; CI 2.32 to 3.29) (Figure 3; Figure 4). Three studies of 108 participants were entered into analysis of the effects of treatment on disability (Connelly 2006; Palermo 2009; Wicksell 2009). A small effect was found standardised mean difference (SMD) -0.30 (95% CI −0.85 to 0.24); and this was not significant (z = 1.09, P > 0.05). Five studies of 204 participants were entered into an analysis of the effects of treatment on mood (Bussone 1998; Griffiths 1996; Palermo 2009; Trautmann 2010; Wicksell 2009). A small effect was found SMD −0.16 (95% CI −0.45 to 0.13); and this was not significant (z = 1.07, P > 0.05).
Six studies of 196 participants, were entered into analysis of the effects of treatment on pain at follow-up (Connelly 2006; Hicks 2006; Labbe 1984; Larsson 1987a; Larsson 1987b;Larsson 1996). This analysis gave an RR of 3.34, (95% CI 2.01 to 5.53; z = 4.68, P < 0.05), for a clinically significant change in pain (NNTB = 2.01; CI 1.62 to 2.64). There was only one study that could be analysed for disability at follow-up (Wicksell 2009), therefore no conclusion could be drawn. Two studies with 59 participants was entered into an analysis of the effects of treatment on mood at follow-up (Bussone 1998; Wicksell 2009) finding a large effect of SMD −0.60 (95% CI −1.13 to −0.07; z = 2.24, P < 0.05).
Twelve studies of 709 participants were entered into an analysis of the effects of psychological treatment on continuous pain outcomes immediately post-treatment (Barakat 2010; Hicks 2006; Humphreys 2000; Kashikar-Zuck 2005; Kashikar-Zuck 2012; Levy 2010; Palermo 2009; Robins 2005; Stinson 2010; van Tilburg 2009; Vlieger 2007; Wicksell 2009). A large effect size SMD -0.55 (95% CI −0.84 to −0.26; z = 3.70, P < 0.05) was found (Figure 5). Nine studies involving 588 participants were entered into analysis of the effects of treatment on disability (Humphreys 2000; Kashikar-Zuck 2005; Kashikar-Zuck 2012; Levy 2010; Palermo 2009; Robins 2005; Stinson 2010; van Tilburg 2009; Wicksell 2009). A small but significant effect was found SMD −0.29 (95% CI −0.49 to −0.10; z = 2.99, P < 0.05). Six studies of 435 participants were entered into analysis of the effects of treatment on mood (Kashikar-Zuck 2005; Kashikar-Zuck 2012; Levy 2010; Palermo 2009; Stinson 2010; Wicksell 2009). A small effect size was found SMD −0.14 (95% CI −0.42 to 0.15); and this was not significant (z = 0.92, P > 0.05).
Five studies of 357 participants, had data available for analysis of the effects of treatment on pain at follow-up (Barakat 2010; Hicks 2006; Kashikar-Zuck 2012; Levy 2010; Wicksell 2009). A small effect size was found SMD −0.17 (95% CI −0.60 to 0.26) and this was not significant (z = 0.77, P > 0.05). Three studies of 292 participants were entered into an analysis of the effects of treatment on disability (Kashikar-Zuck 2012; Levy 2010; Wicksell 2009). A small effect size was found SMD −0.19 (95% CI −0.51 to 0.13) and this was not significant (z = 1.14; P > 0.05). Three studies of 292 participants were entered into an analysis of the effects of treatment on mood (Kashikar-Zuck 2012; Levy 2010; Wicksell 2009). A small effect size was found SMD −0.09 (95% CI −0.32 to 0.14) and this was not significant (z = 0.74, P > 0.05).
Eight additional studies (end of treatment n = 506) were added in this update providing a total of 37 studies and an overall sample of 1938 participants. The outcomes measured and the categories of measurement remain the same. In multi-arm trials involving more than one treatment or control group, similar treatments or control groups were combined for the purposes of the analyses. The majority of studies used one or two treatment conditions in comparison to a waiting list or standard care control group. Similar to the previous review, the treatments could still be categorised as behavioural or cognitive behavioural although the data were not analysed in this way. The average length of treatment for headache studies was 5.3 hours, and for non-headache studies it was considerably shorter (M = 3.5 hours). Follow-up data are increasingly being reported in more recent studies and were included when relevant. The overall quality of the studies has improved since the previous review (Eccleston 2009), although the size of the samples did not increase over time.
The inclusion of further studies has extended the evidence base. Of the 12 possible analyses, five were significant, and three were different from the previous review. Pain intensity was found to improve for headache and non-headache groups at post-treatment, and for the headache group at follow-up. Forty-nine per cent of children with headaches significantly improved their pain scores at post-treatment after receiving psychological therapy compared with only 17% in the control group. However, the addition of more studies to the non-headache pain intensity analysis at follow-up meant this was no longer significant. Disability was significantly improved in non-headache analyses at post-treatment, although the effect size was small. This was also true for mood in the headache group at follow-up. However, this should be interpreted with caution as there were only two studies available for analysis.
Pain intensity was the most common treatment outcome assessed, with 18 studies of children with headache and 12 of children with non-headache pain providing data on pain intensity that could be analysed. An NNTB of 2.72 for psychological therapies to produce more than 50% relief in pain in children with headaches was found. An NNTB of 2.01 was found for the smaller number of trials reporting follow-up. Large effect sizes were also found for reduction in pain intensity in non-headache chronic and recurrent pain at post-treatment. However, the confidence intervals around the effects are large.
Further to the previous update, there have been developments in the field which are reflected in the updated trials. All of the new trials used cognitive behavioural therapy compared to purely behavioural therapy. Five of the eight updated studies were non-headache trials, and a further two included mixed samples and so were included in both headache and non-headache analyses. All the updated studies provide data on mood and disability, and four studies provided data that could be entered into the follow-up analyses thereby allowing a better estimate of the maintenance of interventions. Reflecting advances in treatment delivery, four updated studies (six in total) used computer applications (internet and/or CD-ROM) to provide treatment to the children.
In regard to condition, this review included 21 headache studies, seven abdominal pain studies, two sickle cell disease studies, two fibromyalgia studies, four mixed pain studies and one juvenile idiopathic arthritis study. Psychological treatment still appears to have a minimal impact on disability and mood despite reporting two significant effects from this update. Previously, we reported that mood and disability outcomes in trials with children with chronic pain were a focus for trials (McGrath 2008) and this seems to be the case as seven of the eight updated studies had extractable data for disability analyses and five out of eight studies provided extractable data for mood analyses. As these outcomes are being measured and reported more frequently, standardised measurements should be used consistently across the field of paediatric chronic pain and treatments should specifically target changes in these areas.
One limitation of this review is that we are unable to fully discuss the effectiveness of psychological interventions as they were compared with a control group that combined active (e.g. education) and wait-list controls. Just over half the studies used active controls, yet we did not feel that it was an appropriate sample to separate for analysis as has been done in a companion review of treatments for adults with pain (Eccleston 2009a). This limitation may contribute to an overestimation of the treatment effects since it is not possible to separate differences specific to treatment versus active treatment or waiting list control.
Psychological treatments, principally relaxation and cognitive behavioural therapies, are effective treatments of childhood headache pain at post-treatment and follow-up, and mood at follow-up. Behavioural and cognitive behavioural treatments are also effective in reducing non-headache pain, and disability at post-treatment. There is insufficient evidence to explain the effectiveness of psychological interventions for individual non-headache pain conditions due to the limited number of studies for each condition.
Since the original version of this review there has been an improvement in the evidence base by the addition of new studies, and the extension into non-headache pain conditions, and treatments that rely on more complex methods. This update adds studies to the previous review so readers are able to understand the effects that psychological interventions have for headache and non-headache pain. However, this structure limits our understanding of whether psychological therapies are unique in their improvement of symptoms in comparison to active or wait-list control groups, yet we felt this was important to present before introducing further analyses. The author team are considering the following possible changes for the next version of the review.
Measurement of non-pain outcomes is increasing and should continue to be included and reported in trials with children and adolescents. Recommendations are now available with regards to evidence based assessment of relevant domains for trials of chronic and recurrent pain in children (McGrath 2008). Measures have been developed to assess the majority of these domains. However, additional validation work is needed particularly on child disability outcome measures. Currently available measures have several limitations including a lack of data on sensitivity to change and factor structure, and do not have established cut-off scores to aid in interpretation of scores as situated in the clinical versus normative range (Palermo 2008). It is currently unclear whether treatments might be effective at changing particular aspects of disability because the measures lump together various concepts such as vigorous physical activity, role functioning (e.g., school attendance), and regular activity participation. Future validation work may help disentangle this complex and multidimensional domain.
Studies in this review are dominated by behavioural or cognitive behavioural treatments. However, more recent studies show a shift towards cognitive behavioural therapy. There is now a call for more varied psychological therapies that have been used in other pae-diatric chronic illnesses such as diabetes (multisystemic therapy), cancer (problem-solving therapy) and asthma (family therapy). Such therapies in painful conditions would give a broader understanding of which therapies can work for a particular population. There is also a need for more trials in non-headache conditions as we are as yet unable to understand the effectiveness of psychological interventions in specific conditions such as sickle cell disease. There have been developments in the delivery of treatments with technology advances. Five trials delivered their therapy using a computer- or internet-based programme (Connelly 2006; Hicks 2006; Palermo 2009; Trautmann 2010; Wicksell 2009). This delivery mode has the ability to reach more patients quicker, and in this digital age, these treatments are fast becoming the most accessible. As reported in the previous review, important information is still lacking in trials such as treatment expectations, treatment history, baseline pain and disability which is important to understand how and why some treatments are more effective than others. Similarly, treatment response is precluded by individual differences of the trials and samples, such as small sample sizes, mixed illness groups and wide age ranges included in the current trials.
Trials have improved in quality and updated trials include follow-up periods so we are better able to understand the maintenance of treatment effects. However, the number of participants entering treatments is still small and should be increased in future trials, and multi-site collaboration is needed to achieve this. More trials are needed in non-headache pain conditions so as to determine the effectiveness of psychological therapies for individual conditions. Therapies should also be tailored to the child and families’ needs, and sociodemographic differences should be explored and reported. Treatments delivered are to date dominated by behavioural or cognitive behavioural therapy. Future trials should expand beyond these therapy groups to identify if other therapies are effective for chronic pain conditions and all treatments should include specific content to improve disability and mood. Finally, new trials should use active comparators over wait-list control groups.
Last assessed as up-to-date: 12 November 2012.
|24 October 2012||New search has been performed||New authors have been added to this review. A new|
search was run in March 2012. Eight new studies were
added (Barakat 2010; Kashikar-Zuck 2012; Levy 2010;
Palermo 2009; Stinson 2010; Trautmann 2010; van Tilburg 2009; Wicksell 2009), and four new studies
were excluded (Trautmann 2008; Vlieger 2012; Weydert 2006; Youssef 2009).
|24 October 2012||New citation required and conclusions have changed||The previous review reported psychological treatments|
were effective for headache and non-headache groups at
post-treatment and effects were maintained at follow-up.
Updated studies have altered previous results. The cur-
rent update found pain improved at post-treatment for
headache and non-headache groups, and for headache
groups at follow-up. An additional significant finding for
disability at post-treatment for the non-headache group
was found. Conclusions have been updated accordingly
We would like to thank Kiki Mastroyannopoulou and Louise Yorke for their contributions to the original version of this review. Thank you also to Hannah Somhegyi for help with coding and data management and to Jane Hayes for running the updated search. Finally, thanks also go to the PaPaS review group team and to the peer referees for their helpful comments.
SOURCES OF SUPPORT
Cochrane Highly Sensitive Search Strategy for identifying randomized trials in MEDLINE: sensitivity maximising version (2008 revision) as referenced in Chapter 6.4.11 and detailed in boxes 6.4.c of The Cochrane Handbook for Systematic Reviews of Interventions Version 5.0.0 [updated February 2008].
CONTRIBUTIONS OF AUTHORSChristopher Eccleston oversaw the project and contributed to the design, analysis and authoring of the text and is responsible for any future update of this review.
Amy Lewandowski, Emma Fisher, Emily Law, Stephen Morley, Tonya Palermo, and Amanda Williams, all contributed to the design, analysis and authoring of the text.
DECLARATIONS OF INTEREST
DIFFERENCES BETWEEN PROTOCOL AND REVIEW
* Indicates the major publication for the study