Psychological e-health interventions for children and adolescents
Posted on December 18, 2018 by Sophia Fedorowicz
In this blog, Sophia discusses a recent Cochrane review which asked the question: Are psychological e-health interventions for children and adolescents with long-term physical conditions effective for reducing depression and anxiety?
The prevalence of long-term physical conditions in childhood is increasing, with conditions lasting longer than 3 months that are judged to impair functioning, affecting 10 -12% of children globally (van der Lee et al., 2007; Eiser, 1997). These conditions include things such as asthma, diabetes, epilepsy, Crohns disease, chronic kidney disease, neurological conditions, and cancers (Burkart, 2002). Children and adolescents with long-term physical conditions are at greater risk of psychological problems (Pao & Bosk, 2011) and currently the access to first line treatment, psychological therapy, for these psychological problems is limited (Bennett et al., 2015).
E-health presents a possible solution to address the need for psychological interventions for children and adolescents with long-term physical conditions. E-health is conceptualized as the provision of health care using digital technology; it is interactive and patient-tailored (McLean et al., 2010). The increasing popularity of digital technology has led to the design and release of app-based interventions and the following review by Thabrew, Stasiak, Hetrick, Wong, Huss, & Merry (2018) synthesizes the available evidence for e-health interventions for anxiety and depression in children and adolescents with long-term physical conditions.
The Thabrew et al., (2018) review assessed the effectiveness of e-health interventions in comparison with waiting list controls, attention and psychological placebos, or non-psychological interventions for treating anxiety and depression in children and adolescents with long-term physical conditions.
The review in more detail
Five trials of three interventions were identified; Breathe Easier Online, Web-MAP, and multimodal cognitive behavioural therapy (CBT). Across the five included studies there were 463 participants with an age range of 10 to 18 years. Conditions included chronic headache (migraine, tension headache and others), chronic pain conditions (musculoskeletal, abdominal and others) and respiratory illnesses (asthmas, cystic fibrosis and others) with symptoms of anxiety or depression.
The authors concluded the quality of evidence was low across trials for primary and secondary outcomes due to the introduction of bias. The nature of the interventions means that blinding of participants and research staff was not possible or it was unclear, and only one trial had published a protocol for the trial. All three interventions tested in the five included trials were conducted by developers of those interventions, and sample sizes were small. The high risk of bias means that it could not be determined whether e-health interventions were clearly better than any comparison in reducing depression or anxiety or improving quality of life and functioning or the physical health symptoms.
There was very low quality evidence that e-health was less acceptable than comparators in children and adolescents with long-term conditions. Most of the trials worked with adolescents so it is impossible to comment on the effectiveness of these interventions for children under the age of 10.
What were the characteristics of the participants?
There are some points to be considered regarding the sample size of this review. First, considerably less than half of the total of participants (n=463) were male: 15% (Law, 2015), 25% (Palermo, 2016a), 30% (Palermo, 2009), 45% (Trautmann, 2010), 50% (Newcombe, 2012) creating an overall gender imbalance in the sample. This may be reflective of population-level sex differences in affective disorders in adolescence, whereby rates are much higher for females (Thapar et al., 2012) but this was not commented on in the review, nor were differences between male and female participants explored in the trials.
Secondly, the majority of participants were white: 92% (Law, 2015), 100% (Newcombe, 2012), 90% (Palermo, 2009), 85% (Palermo, 2016a) and Trautmann (2010) did not specify.
Third, all the trials were conducted in high income countries: USA (Law, 2015; Palermo, 2009; Palermo 2016a), Australia (Newcombe, 2012), Germany (Trautmann, 2010)., and fourth some excluded people without access to a computer. Although e-health is championed as cost-effective and could be widely available to difficult to reach populations or those who wish to be private about their therapy, it is impossible to determine if these interventions would be effective in populations from low- and middle-income countries, and other cultures or populations.
Interventions and outcome measures
Three of the five trials evaluated the same intervention: Web-MAP (Law, 2015; Palermo, 2009; Palermo, 2016a) a web-based intervention for managing chronic pain. Two other trials evaluated Breathe Easier Online (Newcombe, 2012) and an online form of multimodal CBT training for reducing headaches (Trautmann, 2010). This makes the pain management intervention more dominant in the sample. Web-MAP also included modules for the parents of the child, 8 modules for the child and 8 modules for parents, and the others did not. This is a fundamental difference in the way the intervention is delivered because involving the parent in the homework tasks may change the way the child or adolescent engages in the intervention. It might be important, therefore, to compare the effectiveness of Web-MAP to the effectiveness of Breathe Easier Online and the multimodal CBT intervention as they address different long-term conditions and are delivered in a different format?
An oversight by four of the trials which is close to my heart (and research) is measurements of quality of life. Only one trial (Trautmann, 2010) included in this review measured improvements in quality of life. Historically, the outcomes measured in trials are chosen by health professionals and methodologists rather than patients. The resulting data often only represent a snap shot in time and are not relevant to the patient. Patients are interested in long term quality of life improvement as part of effective treatment (Haywood, Wilson, Staniszewska, & Salek, 2016). This is particularly important for those living with long-term conditions when complete recovery is not a realistic part of their future.
Discrepancies between healthcare measures and patient measures in research can be resolved by involving the specific patient population concerned in the design of the trials and introducing PROM (patient reported outcome measures) of outcomes such as functional status, well-being and quality of life (Staniszewska, Haywood, Brett, & Tutton, 2012).
One final point, which is perhaps the most important, is that none of the interventions were designed to treat anxiety or depression as a primary focus, exposing a large gap in the evidence. There is need for more evidence in this area.
A strength of this review is that it included trials with many mixed, long-term conditions with both anxiety and depression. It also included participants with comorbidities such as diabetes, asthma or another mental health condition besides anxiety and depression. This was good to see and is reflective of the real-life contexts within which these e-health interventions would be used. Long-term physical health conditions are often extremely complex, as are psychological problems, and co-morbidities are the norm rather than the exception.
In the case of long-term conditions, the aim is often to improve quality of life rather than pursuing an absence of symptoms. It is therefore discouraging that the trials in this review did not collect data regarding the participants’ quality of life before and after using the interventions. Similarly, other secondary outcomes such as school attendance, and functioning, which may have been important to the patient, were either not recorded or not followed up past the end of the trial.
Due to the high risk of bias and the small sample, it is concluded by the authors that there was very low-quality evidence available and so it could not be determined whether these e-health interventions were better than comparators. I completely agree. Perhaps the inclusion of quasi-randomised and non-randomised trials in future reviews could provide a clearer view.