Background Children with chronic physical illness are significantly more likely to develop common psychiatric symptoms than otherwise healthy children. These children therefore warrant effective integrated healthcare yet it is not established whether the known, effective, psychological treatments for symptoms of common childhood mental health disorders work in children with chronic physical illness.
Methods EMBASE, MEDLINE, PsycINFO and CINAHL databases were searched with predefined terms relating to evidence-based psychological interventions for psychiatric symptoms in children with chronic physical illness. We included all studies (randomised and non-randomised designs) investigating interventions aimed primarily at treating common psychiatric symptoms in children with a chronic physical illness in the review. Two reviewers independently assessed the relevance of abstracts identified, extracted data and undertook quality analysis.
Results Ten studies (209 children, including 70 in control groups) met the criteria for inclusion in the review. All studies demonstrated some positive outcomes of cognitive behavioural therapy for the treatment of psychiatric symptoms in children with chronic physical illness. Only two randomised controlled trials, both investigating interventions for symptoms of depression, were found.
Conclusions There is preliminary evidence that cognitive behavioural therapy has positive effects in the treatment of symptoms of depression and anxiety in children with chronic physical illness. However, the current evidence base is weak and fully powered randomised controlled trials are needed to establish the efficacy of psychological treatments in this vulnerable population.
- Adolescent Health
- Child Psychiatry
- Child Psychology
- General Paediatrics
- Paediatric Practice
This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Rates of psychiatric disorder are up to four times greater in children with chronic physical illness than in children who are physically well.1–3 Psychiatric symptoms have considerable consequences for a child's quality of life, their behavioural, emotional, educational and social functioning,4 ,5 and mental ill health has, in turn, been shown to impact upon management and medical consequences of the physical illness.6–10 Delivery of effective psychological treatment to this population is therefore a priority. In the UK, government bodies such as the National Health Service (NHS) Confederation have highlighted the social, health and economic benefits that arise from integration of physical and mental health treatments.11 The US National Center for Chronic Disease Prevention and Health Promotion's ‘Public Health Action Plan to Integrate Mental Health Promotion and Mental Illness Prevention with Chronic Disease Prevention’,12 similarly includes an objective to develop strategies for integrating mental health and mental illness and public health systems.
There are highly effective evidence-based psychological treatments for some of the common psychiatric disorders in children and young people.13 However, guidelines regarding evidence-based interventions for common mental health disorders in children with physical illness are scarce, and in many cases there remains a large unmet need. For example one study14 found that of 114 children with epilepsy, 61% had psychiatric diagnoses, but, of these, only 33% had received treatment, despite regularly attending clinics for their epilepsy. Clinicians do not have adequate guidance to support them in making decisions regarding effective interventions in this population and thus children are not able to access appropriate and timely interventions for their mental health disorder.14 ,15
It appears that children with physical and mental health conditions are viewed as complex; the care of their physical health may be prioritised, inadvertently leading to neglect of their mental health needs. If clinicians who work with children in mental health and paediatric services are aware of the effectiveness of mental health treatments in this population, and the best ways for families to access them, then services can be organised to meet the need. It is inequitable that at the present time children who are already disadvantaged by their physical illnesses are not able to access appropriate services.
This systematic review therefore aimed to investigate the evidence for the effectiveness of psychological therapies for symptoms of common mental health disorders in children and young people with chronic physical illnesses. In addition, we aimed to conduct a meta-analysis of the findings if the data were appropriate. Finally, the review aimed to understand any key factors associated with the success of an intervention and the ability of children/young people to access it.
Systematic review methods were used in accordance with Cochrane guidelines.15a
Electronic Searches, citation searches, reference list searches and grey literature searches were independently undertaken by AC and SB.
EMBASE, MEDLINE, PsycINFO and CINAHL databases were searched from inception to February 2014. Grey/unpublished literature was also included, through searches of Google and Google Scholar. Broadly, the search terms were categorised into three primary areas; (1) Chronic illness, (2) Impairing psychiatric symptoms, (3) Psychotherapeutic intervention. See online supplementary appendix 1 for full list of search terms.
Other search resources
Citation lists and reference lists of identified papers were also searched for relevant papers. Additional literature was found through personal contact with researchers in the area.
Study eligibility criteria were:
(1) Randomised controlled trials (RCTs), controlled trials, cohort studies, case control studies and multiple-baseline studies; (2) Studied participants aged 0–18 years with a chronic physical illness and symptoms of mental health disorder (anxiety, depression or disruptive behaviour symptoms; defined by Diagnostic and Statistical Manual of Mental Disorders IV16 and Diagnostic and Statistical Manual of Mental Disorders 5 17); (3) Reported a child-related mental health measure as the primary outcome (The measure had to relate to the mental health of the child and not the parent/carer, although parent-reports of child health/behaviour were acceptable); (4) Studied a psychotherapeutic intervention (defined as an intervention in which a therapist purposively and systematically attempts to influence a patient by psychological means so that the patients’ symptoms decrease or there is a positive change in behaviour’; as used in Yorke et al, 200718). At present, there is no consensus regarding the definition of chronic physical illness. Van der Lee et al19 conducted a systematic review of the definitions and measurement of chronic illness, and found three commonly used definitions for ‘chronic illness’ or ‘chronic health conditions’ (those of Pless and Douglas20; Perrin et al21 and Stein et al22). All define chronic physical illnesses as lasting for at least 3 months (some define longer periods) and causing functional impairment. As definitions vary, so too do the lists of possible conditions that fall under these definitions. We derived our list of illnesses (and thus search terms) from those used in previous reviews of chronic physical illnesses in children.2 ,23 Conditions included: AIDS and HIV, asthma, cancer, chronic fatigue syndrome, cleft palate, cystic fibrosis, deafness/hearing impairment, diabetes, epilepsy, heart disease, inflammatory bowel disease (IBD), kidney disease, liver disease, migraine, sickle cell anaemia, spina bifida and visual impairment.
We excluded those interventions that had a primary aim of increasing self-efficacy or treatment adherence related to the physical illness. Additionally, we excluded papers where the psychiatric symptoms were directly related to the physical illness, such as interventions for anxious breathing in asthma. We excluded children who were ‘survivors of cancer’, as under definitions of chronic physical illness, it is not clear that this is a current illness, causing functional impairment within the last 3 months. We also excluded chronic pain (including headache), in line with previous reviews23 and as this has been the topic of a recent distinct Cochrane review.24
Data collection and analysis
Study selection was performed independently by two reviewers (AC and SB). Where disagreements arose about whether a study fitted with the inclusion criteria, this was resolved through discussion with RS as appropriate.
A data extraction form was developed, covering study characteristics and main results. Data was independently extracted by two reviewers (SB and SW). Data were inputted into EndNote X5 software.
Methodological quality assessment
Study quality was independently assessed by two reviewers (AC and SB) with the Effective Public Health Practice Project Quality Assessment Tool.25 ,26 This tool was chosen for its suitability in assessing a range of study designs within the area of public health research. Studies are rated as strong, moderate or weak, using predefined criteria, on a range of areas: selection bias, study design, confounders, blinding, data collection methods, withdrawals and dropouts. Total sample size is not considered. An overall total for study quality is also calculated by assessing the number of areas rated weakly (strong studies have no weak ratings, moderate ones have one weak rating and weak studies have two or more weak ratings).
The initial search identified 1966 independent papers. A total of 10 studies, and 2 follow-up studies, were found to fit with the criteria of the review.27–38 A total of 209 participants (173 participants with a chronic physical illness and impairing psychiatric symptoms), took part in the studies. See figure 1 for flow chart of study selection, tables 1 and 2 for summaries of included studies and table 3 for comprehensive recruitment figures.
All studies investigated interventions for anxiety or depression. Study participants included children with epilepsy (n=2), IBD (n=3), diabetes (n=3), asthma (n=1) and cystic fibrosis (n=1). All interventions were based on a cognitive-behavioural framework; most had been previously used and evaluated in cohorts of children without a chronic physical illness. Two RCTs were found to fit with the criteria of the review.32 ,37 The remaining studies were non-randomised designs (see tables 1 and 2). Due to the range of study designs, it was not possible to conduct a meta-analysis.
Within the bounds of the research design, studies were, for the most part, well executed, and 9 out of 10 studies were rated strongly or moderately with respect to quality (see online supplementary appendix 2). One rated weakly28, due to a lack of blinding, high rate of withdrawals/dropouts and the likely presence of selection bias of participants. No studies were rated ‘strongly’ with respect to blinding. As all studies demonstrated positive effects on mental health outcomes, it is not possible to analyse whether there is an association between methodological quality and study outcomes. However, we note that as this tool was designed to assess a range of study designs (randomised and non-randomised), it is possible for a study to rate strongly or moderately with respect to overall quality despite only moderate ratings across the categories (including study design). In addition, sample size is not accounted for and thus some strongly rated studies have small sample sizes and non-randomised designs. Positive study quality assessments therefore need to be interpreted with caution.
All five depression interventions were 12-session cognitive behavioural therapy (CBT) protocols. Three studies32 ,35 ,37 used versions of protocols which have been well validated in children without physical illness (Treatment for Adolescents with Depression Study39 and Primary and Secondary Control Enhancement Training40). Standard CBT strategies were delivered, such as mood monitoring, problem solving and behavioural activation. One study reports34 on CBT delivered in a group format. Other programmes worked primarily one to one with the child. Szigethy et al35 ,36 also offered three family sessions of 60 min; 40 min with the parents alone and 20 min with the family. The purpose of the family sessions was to review the perspectives of parents, review the skills learnt in the young person's session, review family coping skills and review homework tasks.
Four33–35 ,37 of the five depression interventions (for children with diabetes or IBD) included protocol modifications related to physical health. For example, all four included psychoeducation about the relationship between specific physical illnesses and mood. One34 also covered setting personal goals for diabetes self-care, diabetic barriers to behavioural activation and what to tell others about having diabetes. Martinović et al32 did not report any specific modifications for the physical health comorbidity (epilepsy).
Anxiety protocols varied in format, although all were based on basic principles of CBT for anxiety (eg, cognitive restructuring and exposure exercises) and many used adapted versions of previously validated protocols. Four were delivered in a one-to-one format; one31 also added three parent sessions and one30 combined the results of an individual and group intervention.
Again, four of the studies adapted the intervention to account for the physical illness. Three28 ,29 ,31 related the material to illness-specific stressors (for IBD, diabetes and cystic fibrosis). In one, the intervention was particularly revised to account for the increased rates of learning problems found in children with epilepsy.27 Such alterations included longer sessions (to allow for a slower pace), additional written materials, more concrete language and a focus on behavioural rather than cognitive elements.
It is difficult to interpret the results of these studies as a whole, due to the large variety of methodologies, generally small sample sizes and variety of outcome measures. As the two RCTs32 ,37 have the highest quality rating and least bias, we consider them to have the most valid and reliable results regarding efficacy. We note that there are no RCTs focused on anxiety.
Both trials demonstrated statistically significant results, with large effect sizes. In their depression treatment study, Szigethy et al37 state that they did not correct for Type 1 error, despite multiple comparisons, because ‘the decision was made to err on the side of detecting versus not detecting a difference in treatment effect in this exploratory study’. In addition, a greater number of participants in the control group had their IBD rated as moderate/severe, compared with the intervention group. This means results may be confounded by illness severity.
All other studies demonstrated positive results for the interventions in terms of reductions in anxiety/depression, despite using different definitions of improvement (eg, presence of psychiatric disorder, change in clinical category or change on a symptom measure). Where analysis was undertaken, studies reported a statistically significant benefit for at least one outcome. In interpreting the outcomes of these studies, we note that Szigethy et al36 offered additional sessions and/or psychotropic medications as necessary between end of treatment and follow-up. Gains at follow-up may not be due to the initial intervention alone. Tables 4 and 5 provide details on the main outcomes relating to mental health, physical health and other secondary outcomes.
Varied outcomes were demonstrated in relation to physical health measures. In general, where outcomes related to physical health showed significant improvement, these were related to subjective measures (eg, pain scales, self-reported self-management); no consistent significant difference was found for objective measures of physical health, such as glycaemic control.28 ,29 ,31 ,33–35
Practical adaptations for delivery within a physical healthcare setting
Most studies made accommodations for young people who had a physical illness, and thus a number of medical appointments. For example, studies conducted sessions in participants’ homes,28 outpatient settings that were either attached to a hospital, or were in the hospital,29 ,30 ,32 ,33 or in other medical care settings.27
Studies of young people with IBD were particularly accommodating of medical appointments, through offering convenient time slots, telephone appointments and intervention locations. For example, appointments were coordinated with physical health appointments where necessary and some appointments were offered at the same time as a medical procedure (an infusion).
It was concluded that a meta-analysis would not be informative as there were only two RCTs, which reported different outcomes at different time points. The observational studies did not report appropriate data to undertake a meta-analysis. It was similarly not possible to fully investigate factors associated with the success of an intervention.
This review shows that children may benefit from cognitive behavioural interventions for depression and anxiety in the context of a comorbid chronic physical health problem. However, it also emerged that there is a significant lack of studies evaluating treatment of psychiatric symptoms in children and young people with chronic physical illnesses, despite 435 studies demonstrating their efficacy in otherwise healthy children.41 Methodologies, measures and methodological quality were variable, sample sizes were small and inclusion criteria differed, with studies investigating a variety of combinations of physical illness and psychiatric symptoms. This variability meant that a meta-analysis was not statistically appropriate and that the results are difficult to generalise.
While the significant results of all studies included in the search may represent an element of publication bias, full searches of trial databases were carried out prior to the review being undertaken. No currently running trials of interventions for common impairing psychiatric symptoms in children with long term conditions were found. Thus, it would appear that there is a true deficit in the literature, and that the available studies are representative of the little data available. It is possible that our search terms biased the findings towards cognitive behavioural interventions, however.
Specific adaptations to young people with a physical illness were generally included but were relatively minor and typically did not require significant specialist knowledge about the illness. Many child and adolescent mental health professionals are trained in the delivery of evidence-based cognitive behaviour therapies for anxiety and depression and therefore should be able to deliver these without significant additional training in paediatrics. Where specific information is needed to provide appropriate psychoeducation, Child and Adolescent Mental Health Service clinicians can liaise with the child's paediatrician.
Many studies made allowances for physical illness through the treatment location. Some studies allowed for the use of telephone sessions, sessions at home, or sessions at the same time/venue as medical appointments, to reduce the burden on families. This more flexible approach was particularly seen in the IBD studies,31 ,35 which also showed good patient satisfaction. Clinically, a more flexible approach would be a step towards creating services that are more accessible for this population.
Directions for future research
Larger well controlled trials in the wider area of mental health interventions for children with physical illness are needed. Experimental studies are also needed since it is possible that some elevated level of anxiety regarding the physical illness may be beneficial and may contribute to good illness management. Existing studies have generally focused on adolescent populations and it would be useful to investigate the effects of interventions in younger age groups including disruptive behaviour. There were no studies of, for example, the efficacy of parenting programmes, a strongly evidence-based intervention for children with oppositional defiant disorder. Additional research to understand the effects of these interventions on physical health outcomes is also needed.
Together, these results suggest that it is possible to use evidence-based cognitive behavioural interventions to effectively treat anxiety and depressive symptoms in young people with chronic physical illnesses. Standard protocols developed for children and young people without physical illness can be used, with the same outcome measurement strategies. However, larger RCTs are needed. The results of this review suggest that this should ideally be a trial of a cognitive behavioural intervention, compared with treatment as usual. The cognitive behavioural intervention may need slight adaptation for use in children with physical illnesses—in particular flexibility around times and locations of appointments may be useful.
The authors thank Professor Peter Bower and Dr Craig Whittington for their guidance in drafting the manuscript. The authors also thank Mr John Clarke for his advice regarding the search strategy.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Contributors SB, RS, AC and IH developed the search strategy. AC and SB ran the database searches and contacted researchers to identify relevant papers. AC and SB identified studies matching inclusion criteria; RS was consulted where there was disagreement or ambiguity regarding whether studies met inclusion criteria. SB and SW extracted data from the studies. AC and SB undertook quality analysis. RS, SB and IH drafted the manuscript. All authors read and approved the final manuscript.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.