NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Depression in children and adolescents: Does psychotherapy lead to better results when compared with other therapies? IQWiG Reports – Commission No. HT19-04 [Internet]. Cologne (Germany): Institute for Quality and Efficiency in Health Care (IQWiG); 2022 Aug 12.
Depression in children and adolescents: Does psychotherapy lead to better results when compared with other therapies? IQWiG Reports – Commission No. HT19-04 [Internet].
Show details4.1. Results of the comprehensive information retrieval
The information retrieval identified 13 systematic reviews as relevant for the research question of the benefit assessment [36-48]. No planned or ongoing systematic reviews were found via PROSPERO. Further, 2 completed systematic reviews were found [49,50] for which no results had been reported by November 2020.
The search strategies for bibliographic databases and trial registries are found in the appendix. The last search was conducted on 8 April 2020. A supplementary focused or scoping search for RCTs in accordance with the HTA report protocol was foregone because all psychotherapy types relevant for this HTA report have been discussed in systematic reviews which are currently sufficient and of adequate quality.
Table 1
Exclusions due to a quality assessment by Oxman-Guyatt score
Table 2
Study pool for the health economic assessment
4.2. Characteristics of the systematic reviews included in the assessment
The 13 systematic reviews included in the HTA report covered a total of 150 nonoverlapping RCTs. For the systematic reviews, the last search for primary studies was carried out between March 2009 and January 2019. Six systematic reviews investigated CBT (monotherapy or, in some cases, in combination with antidepressants) [36,37,40,41,45,46], 2 systematic reviews focused on IPT [43,44], 3 systematic reviews investigated CBT and IPT [38,39,42], and 2 systematic reviews focused on all 3 psychotherapy types relevant for this HTA report: CBT, IPT, and DYN [47,48]. The therapy format and duration and frequency of interventions differed widely between the systematic reviews. Two systematic reviews investigated only one-on-one therapy [37,43], 2 systematic reviews analysed only group therapy [36,46], while the remaining systematic reviews included studies on one-on-one as well as on group therapy or provided no specific information on the form of therapy. The psychotherapy durations in the systematic reviews varied between 4 and 54 weeks, and the number of sessions ranged from 4 to 36. Only in rare cases did the included systematic reviews report on whether the psychotherapy took place in an outpatient or inpatient setting. Where information on this topic was provided, the treatment was performed on an inpatient basis. The systematic reviews investigated the psychotherapy types for effectiveness in comparison with inactive controls (psychological placebo, waiting list, no intervention, usual care) or with antidepressants. “Psychological placebo” and “usual care” were not defined in detail in the systematic reviews.
Forti-Buratti 2016 [41] and Yang 2017 [46] focused on children (7–13 years), while Chi 2018 [36] and Dubicka 2010 [40] investigated only adolescents (13–18years). All other systematic reviews included studies on both children and adolescents and analysed them jointly. None of the systematic reviews included studies on children under 7 years of age. The results sections of the included systematic reviews which focused on the treatment of unipolar depression did not distinguish by severity of depression.
Where studies irrelevant for this HTA report – e.g. studies on depression in adults – were included in a systematic review, the HTA report disregarded these studies. For this HTA report, we examined only the results of the relevant primary studies. The only systematic reviews which were included in their entirety were the ones by Pu 2017 [44] and Yang 2017 [46]. In addition, 3 NMAs by Liang 2020 [42], Zhou 2020 [47], and Zhou 2015 [48] were included: In these cases, the extracted data are based on the entire NMA; while not all included studies met the inclusion criteria defined a priori, they were nevertheless used in the network. The NMA results were extracted because the comparisons relevant for this HTA report were carried out at least indirectly. In the derivation of evidence, these results were used only if no direct comparisons were available from individual studies or pairwise metaanalyses. Studies which were part of the network but failed to meet the inclusion criteria of this HTA report deviated primarily with regard to the intervention to be investigated since the network also comprised studies on other types of psychotherapy or comparisons between types of psychotherapy or antidepressants and placebo [42,47,48]. This approach was necessary to identify indirect comparisons. In Zhou 2020 [47], 15% of the NMA studies (11 of 71) deviated from the inclusion criteria of this HTA report with regard to their population since they included not only children and adolescents with depression, but also those with dysthymia, but the authors did not report the percentage of these participants in the individual studies. From a clinical perspective, this deviation is deemed minor since it concerns only 15% of the studies, of which only some actually enrolled persons with dysthymia.
All 3 NMAs reported that the similarity assumptions (similar study and patient characteristics) were adequate for combining the studies in an NMA [42,47,48].
4.3. Overview of patient-relevant outcomes
From 13 systematic reviews, data were extracted on the patient-relevant outcomes of suicide risk (suicidal ideation and behaviour), treatment response, remission, change in depressive symptoms, and functioning. The systematic reviews did not report any data on mortality (all-cause mortality, suicide mortality), health-related quality of life, or adverse events. Regarding the outcome of patient satisfaction, the included systematic reviews likewise provided no results. Table 3 presents an overview of the available data on patient-relevant outcomes from the included reviews.
Table 3
Matrix of patient-relevant outcomes
4.4. Assessment of the quality of systematic reviews and the risk of bias of the results
Quality was rated very high for 9 systematic reviews (7 of 7 points) [37,38,42-48], while 3 systematic reviews received 6 of 7 points [39,40] and 1 systematic review, 5 of 7 points [41]. All systematic reviews which technically met the inclusion criteria but scored fewer than 5 points in the quality assessment were disregarded in this HTA report (see A3.2.2, Table 8 of the full HTA report).
The risk of bias of the primary studies had been assessed by the authors of the systematic reviews. If they assessed only the bias domains but provided no overall assessment of the primary studies, the authors of this HTA report determined an overall rating of the risk of bias. One person performed the assessment and a 2nd person reviewed it. In 18 out of 152 primary studies reported in the systematic reviews, the risk of bias across outcomes was rated as low by at least 1 systematic review, but as unclear or high by the remaining ones. Some of the primary studies which were included in multiple systematic reviews received different ratings of their risk of bias. In case of different ratings, the authors of this HTA report adopted the stricter rating (see A3.2.2).
The risk of bias on the outcome level was not analysed separately because the systematic reviews did not provide this information in the required detail.
4.5. Results on patient-relevant outcomes
Below, results are briefly presented by patient-relevant outcome. First CBT, followed by IPT and then DYN are discussed. It comes down to first comparing psychotherapy with inactive control interventions (placebo, waiting list, no treatment, typical treatment) and then with antidepressant treatment. Afterwards, the evidence on the use of psychotherapy as an add-on to antidepressant treatment is presented. Where for 1 or more of these comparisons, no evidence is available on an outcome, this is stated.
No evidence was found regarding any outcome for comparisons with other active therapies such as exercise or relaxation exercises. Likewise, no evidence was found on the additional outcome of patient satisfaction (acceptance). Detailed results are presented in the tables in Section A3.3.
4.5.1. Results on mortality
The relevant systematic reviews did not report any results on this outcome.
4.5.2. Results on suicide risk
4.5.2.1. Cognitive behavioural therapy
CBT versus inactive controls
The reported evidence does not allow drawing any conclusions as to whether CBT in comparison with an inactive control intervention (placebo or usual care) increases or reduces the risk of suicide. While 1 direct comparison and 1 associated NMA showed no statistically significant difference with regard to suicide risk between CBT and inactive control, the confidence interval was too wide in each case to allow drawing valid conclusions (CBT versus psychological placebo [PP]: NMA: odds ratio [OR] 11.31; 95% CI: [0.01; 46.11]; RCT: OR 1.02; 95% CI: [0.06; 16.45]; CBT versus usual care: OR 276.9; 95% CI: [0.02; 1163]) [47]. Furthermore, no information was available as to whether these results referred to short-term, medium-term, or long-term suicide risk.
Hence, there is no hint of benefit of CBT.
CBT versus antidepressants
While a direct comparison from a systematic review published in 2020 with data on 220 children and adolescents showed a numerically lower suicide risk in the CBT group, the difference was not statistically significant (OR 0.52; 95% CI: [0.17; 1.62]) [47]. One systematic review from 2014 [37] showed in a direct comparison that children and adolescents treated with CBT exhibited a statistically significantly reduced frequency of suicidal thoughts and behaviour than those on antidepressant therapy (CBT versus AD after 12 weeks: OR 0.26; 95% CI: [0.09; 0.72]; 1 RCT with data on 188 children and adolescents; after 18 weeks: OR 0.26; 95% CI: [0.07; 0.98]; 1 RCT with data on 149 children and adolescents). In absolute terms, after 12 weeks, 6% of patients receiving CBT and 19% of patients on antidepressants had an elevated suicide risk [37].
Based on the most current systematic review [47], no hint of greater benefit of CBT versus antidepressants can be derived.
CBT + antidepressants versus antidepressants
Where CBT was used as an add-on to antidepressant treatment in comparison with antidepressant monotherapy, no statistically significant difference regarding the frequency of suicidal thoughts or suicidal behaviour in children and adolescents was found in either the short term or the long term [37,40,47]. One metaanalysis investigating this topic based on 2 individual studies with 388 children and adolescents arrived at the following results after 12 weeks: CBT + AD: OR 0.75; 95% CI: [0.26; 2.16] [37]. Another metaanalysis of 2 studies with data on 424 children arrived at a similar result: OR: 0.75, 95% CI: [0.35; 1.59] [47]. One metaanalysis of 2 studies with data from 344 children and adolescents showed no statistically significant difference even after 40–50 weeks of follow-up: OR 0.53; 95% CI: [0,06; 4,58] [37]. Only 1 older systematic review, which was disregarded in the derivation of evidence based on the methods described in Section 3.1, reported on a study with 157 enrolled adolescents in which after 36 weeks, suicidal behaviour was more commonly observed in those on antidepressant monotherapy than in those receiving a combination of antidepressants and CBT [40].
Based on 2 systematic reviews (each with an Oxman-Guyatt index of 7 of 7) [37,47], no hint of greater benefit of antidepressant therapy with add-on CBT can be derived in comparison with antidepressant monotherapy.
4.5.2.2. Interpersonal psychotherapy
IPT versus inactive controls
One systematic review including studies where children and adolescents received 10 to 16 sessions of IPT over a period of 6 to 16 weeks showed no statistically significant difference in suicide risk between children and adolescents who received IPT versus those who received psychological placebo or usual care. The confidence interval was too wide to allow deriving valid conclusions: OR 0.70; 95% CI: [0.17; 2.93]; 2 RCTs with data on 112 children and adolescents [44]. These findings were confirmed by 1 NMA and 2 individual studies with direct comparisons from the NMA [47].
This results in no hint of benefit of IPT.
IPT versus antidepressants
No evidence was found on this comparison.
IPT + antidepressants versus antidepressants
No evidence was found on this comparison.
4.5.2.3. Psychodynamic psychotherapy
DYN versus inactive controls
In direct comparison, no statistically significant difference in suicide risk was found between children and adolescents after DYN therapy versus psychological placebo: OR 1.01; 95% CI: [0.06; 16.23]; 1 RCT with data on 315 children and adolescents. Information is missing on the time the outcome was surveyed [47]. One of the NMAs calculated in this systematic review arrived at the same result, likewise with no information being available on the time the outcome was surveyed. In addition, the confidence interval was very wide, making it impossible to draw valid conclusions: OR 8.64; 95% CI: [0.01; 40.05] [47].
There is no hint of benefit of DYN.
DYN versus antidepressants
No evidence was found on this comparison.
DYN + antidepressants versus antidepressants
No evidence was found on this comparison.
4.5.3. Results on treatment response
Treatment response was reported by 2 systematic reviews. Dubicka 2010 defined treatment response as “much improved” or “very much improved” according to the Clinical Global Impression of Improvement Scale (CGI-i scale) [40]. Pu 2017 defined response as a 50% improvement on a depression scale [44].
4.5.3.1. Cognitive behavioural therapy
CBT versus inactive controls
No evidence was found on this comparison.
CBT versus antidepressants
No evidence was found on this comparison.
CBT + antidepressants versus antidepressants
In adolescents, after 12 weeks and after 28 to 36 weeks, the combination of CBT and antidepressants resulted in no statistically significant difference in comparison with antidepressant monotherapy (12 weeks: Treatment for Adolescents with Depression Study [TADS] based on data from 193 adolescents: OR 0.61; 95% CI: [0,33; 1,14]; Adolescent Depression Antidepressant and Psychotherapy Trial [ADAPT] based on data from 202 adolescents: OR 1.09; 95% CI: [0.62; 1.89]; 28–36 weeks (TADS, ADAPT): OR 1.09; 95% CI: [0.68; 1.78]; 2 RCTs with data on 349 adolescents) [40].
Hence, there is no hint of greater benefit of the combination of CBT with antidepressants in comparison with antidepressant monotherapy.
4.5.3.2. Interpersonal psychotherapy
IPT versus inactive controls
After 12 to 16 weeks, nearly twice as many children and adolescents who received IPT achieved response compared with those without psychotherapy (placebo, usual care, waiting list) (OR 1.87; 95% CI: [1.40; 2.51]; 4 RCTs with data on 366 children and adolescents ) [44].
Since all studies used for the response analysis were also used for the outcome of “change in depressive symptoms”, employing the same survey instruments, the option of taking this evidence into account once more for the outcome of response was foregone.
IPT versus antidepressants
No evidence was found on this comparison.
IPT + antidepressants versus antidepressants
No evidence was found on this comparison.
4.5.3.3. Psychodynamic psychotherapy
None of the included systematic reviews reported on response in comparison with DYN.
4.5.4. Results on remission
Remission was reported in 1 systematic review, where it was defined as 8 weeks of freedom from symptoms [37].
4.5.4.1. Cognitive behavioural therapy
CBT versus inactive controls
No evidence was found on this comparison.
CBT versus antidepressants
The comparison of CBT versus antidepressants showed no statistically significant difference in remission among children and adolescents after 12 weeks nor after 6 months; the confidence intervals were very wide (CBT versus AD: 12 weeks: OR 0.62; 95% CI: [0.28; 1.35]; 2 RCTs with data on 186 children and adolescents; CBT versus AD: 6 months: OR 0.83; 95% CI: [0.27; 2.60]; 1 RCT with data on 48 children and adolescents) [37].
There is no hint of greater or lesser benefit of CBT.
CBT + antidepressants versus antidepressants
After 12 weeks, the combination of CBT and antidepressants led to remission more frequently than antidepressant monotherapy, but the difference was not statistically significant (OR 1.50; 95% CI: [0.99; 2.27]; 3 RCTs with data on 317 children and adolescents). Even after 6 months and after about 1 year, no statistically significant difference was found, and confidence intervals were wide [37].
Hence, there is no hint of greater benefit of CBT.
4.5.4.2. Interpersonal psychotherapy
No evidence for a comparison with IPT was found on this outcome.
4.5.4.3. Psychodynamic psychotherapy
No evidence for a comparison with DYN was found on this outcome.
4.5.5. Results on the change in depressive symptoms
4.5.5.1. Cognitive behavioural therapy
CBT versus inactive controls
In comparison with inactive control interventions (psychological placebo, waiting list, usual care), most systematic reviews found CBT to lead to a statistically significant improvement in depressive symptoms in children and adolescents [36,38,39,42,46-48]. The same was found in the metaanalysis, which took into account 28 RCTs (but did not report the number of included children and adolescents): Hedges’ g: 0.44; 95% CI: [0.23; 0.65] [38]. In 2 systematic reviews, the relevant RCTs also revealed advantages of CBT, but the difference to the control group was not statistically significant. However, these systematic reviews contained few studies, which were also included in the other systematic reviews [41,45]. The use of computer-based CBT was more effective than psychological placebo or waiting list at reducing depressive symptoms [42]. Mindfulness-based therapy (a subtype of CBT) likewise showed a statistically significant improvement in depressive symptoms in direct comparison with usual care [36]. In NMAs, mindfulness-based therapy, like problem-solving therapy, was not statistically significantly more effective than psychological placebo or usual care at reducing depressive symptoms [42,47,48].
One systematic review analysed children and adolescents separately. It showed that CBT was more effective than the control both in children and in adolescents [38]. Subgroup analyses showed that CBT reduced depressive symptoms in children without comorbidities, but not in children with comorbidities [46]. A subgroup analysis also found CBT to be more effective without parental involvement than with parental involvement [46]. For subgroup analyses which showed no statistically significant result, however, it must be noted that some subgroups contained a very small number of studies, and consequently, the absence of a statistically significant effect might also be due to lack of statistical power [46].
Based on the results of the most current systematic review containing the most RCTs (Oxman-Guyatt index of 7 of 7) [38], there is an indication of benefit for CBT.
CBT versus antidepressants
The most recent systematic review showed, based on 1 study with 220 participants, a statistically significant advantage of antidepressants versus CBT in direct comparison (SMD 0.67; 95% CI: [0.40; 0.97]). No information was available on the survey time [47]. One older systemic review based on 2 RCTs showed no statistically significant difference in reduction of depressive symptoms after 12 weeks nor after about 6 months when patients self-rated their symptoms [37]. When assessed by a physician, an advantage was found for antidepressant therapy after 12 weeks, based on 1 RCT [37]. In the indirect network comparison, the difference was not statistically significant [47].
Based on the systematic review containing the most RCTs [37], no hint of greater or lesser benefit of CBT versus antidepressant treatment was found.
CBT + antidepressants versus antidepressants
Compared with antidepressant monotherapy, the combination of CBT plus antidepressants showed no statistically significant difference in the short term or medium term [37,40,47]. The metaanalysis which took into account the most RCTs showed the following result after 12 weeks: SMD -0.14; 95% CI: [-0.36; 0.09]; 5 RCTs with data on 383 children and adolescents [37]. Only for the 1-year follow-up were the depressive symptoms in the group receiving additional CBT statistically significantly lower (SMD -0.26, 95% CI: [-0.46; -0.05]; 2 RCTs with data on 268 children and adolescents) [37]. However, the CI includes the irrelevance threshold of 0.2 according to IQWiG methods; therefore, it is not certain whether the statistically significant difference is in fact patient relevant.
No hint of greater or lesser benefit of CBT as an add-on to antidepressant therapy in comparison with antidepressant monotherapy can be derived from this.
4.5.5.2. Interpersonal psychotherapy
IPT versus inactive controls
IPT is statistically significantly more effective than inactive controls (psychological placebo, waiting list, usual care) at reducing depressive symptoms in both the short term and the long term [42,44,47,48]. The systematic review which, in the metaanalysis, took into account most RCTs showed an SMD of -0.74 (95% CI: [-0.91; -0.56], 7 RCTs with data on 527 children and adolescents) after 6 to 16 weeks, and after 6 to 18 months, an SMD of -0.75; 95% CI: [-1.21; -0,29], 2 RCTs with data on 89 children and adolescents [44]. A subgroup analysis showed that IPT, both in one-on-one and in group therapy, was more effective than no active therapy [44].
Based on the systematic review containing the most RCTs (Oxman-Guyatt index of 7 of 7) [44], there is an indication of benefit for IPT.
IPT versus antidepressants
In 1 NMA, IPT was compared indirectly with the antidepressant fluoxetine; this comparison showed no statistically significant difference in the reduction of depressive symptoms (SMD 0.13; 95% CI: [-0.74; 1]) [47].
There is no hint of greater or lesser benefit of IPT in comparison with antidepressant treatment.
IPT + antidepressants versus antidepressants
No evidence was found on this comparison.
4.5.5.3. Psychodynamic psychotherapy
DYN versus inactive controls
Evidence on DYN is available from a direct comparison in a pairwise metaanalysis, which showed no statistically significant difference in comparison with psychological placebo [47]. Indirect comparisons from 2 NMAs likewise showed no statistically significant advantage of DYN versus psychological placebo, waiting list, usual care, or no treatment [47,48].
There is no hint of benefit of DYN.
DYN versus antidepressants
In an indirect comparison from 1 NMA, the comparison of DYN versus the antidepressant fluoxetine revealed no statistically significant difference [47].
There is no hint of greater or lesser benefit of DYN in comparison with antidepressant treatment.
DYN + antidepressants versus antidepressants
No evidence was found on this comparison.
4.5.5.4. Psychotherapy in general
One systematic review with 43 RCTs combined various types of psychotherapy and analysed their effectiveness in comparison with usual care, waiting list, or other treatment. In the subgroup of children as well as in the subgroup of adolescents, psychotherapy was statistically significantly more effective than the control intervention at reducing depressive symptoms. However, 6 to 24 months after the intervention, these differences were no longer statistically significant [38].
This results in an indication of short-term benefit for psychotherapy in general in comparison with inactive controls, but not a medium-term or long-term benefit.
4.5.6. Results on functioning
This outcome comprises the mental, social, and school functioning of children and adolescents.
4.5.6.1. Cognitive behavioural therapy
CBT versus inactive controls
No evidence was found on this comparison.
CBT versus antidepressants
The comparison of CBT versus antidepressants showed no statistically significant difference in experienced functioning of children and adolescents after 12 weeks or after 6 months (based on 1 study with 42 participants) [37].
This results in no hint of greater or lesser short-term or medium-term benefit of CBT in comparison with antidepressant treatment.
CBT + antidepressants versus antidepressants
The comparison of CBT + antidepressants versus antidepressant monotherapy likewise showed no statistically significant difference in experienced functioning in children and adolescents after 12 weeks or after 6 months [37,40]. One-year follow-up results from 1 RCT with 152 participants show statistically significantly better functioning in children and adolescents with add-on CBT than in those with antidepressant monotherapy (MD 3.00, 95% CI: [0.40; 5.60] [37].
This results in no hint of greater or lesser short-term or medium-term benefit of add-on CBT in comparison with antidepressant monotherapy [37,40]. In the long term, however, a hint of greater benefit was found for add-on CBT compared with antidepressant monotherapy [37].
4.5.6.2. Interpersonal psychotherapy
IPT versus inactive controls
In children and adolescents, IPT led to a statistically significant improvement in functioning in comparison with no active intervention (waiting list, psychological placebo, or usual care) 10 to 16 weeks after the intervention (SMD 0.53, 95% CI: [0.21; 0.85]; 5 RCTs with data on 407 children and adolescents) [44].
There is a hint of greater short-term benefit of IPT.
IPT versus antidepressants
No evidence was found on this comparison.
IPT + antidepressants versus antidepressants
No evidence was found on this comparison.
4.5.6.3. Psychodynamic psychotherapy
No evidence for a comparison with DYN was found on this outcome.
4.5.7. Results on health-related quality of life
The relevant systematic reviews did not report any results on this outcome.
4.5.8. Results on adverse events
The relevant systematic reviews did not report any results on this outcome.
4.6. Evidence map
Table 4 below shows the evidence map regarding patient-relevant outcomes.
Table 4
Evidence map regarding patient-relevant outcomes
4.7. Outcomes from the perspectives of patient representatives
The interviewed experts confirmed the outcomes used in this HTA report to be relevant. They explained that adolescents predominantly suffer from symptoms such as difficulty concentrating, avolition, low self-esteem, and low regulation of negative emotions such as sadness, anxiety, and (auto)aggressiveness as well as from loneliness. Parents reportedly experience the children’s and adolescents’ social withdrawal, avolition, and listlessness as particularly dramatic. These described changes in experiences (thoughts and feelings) as well as in behaviour are reflected by the scales used for recording depressive symptoms.
The experts deemed improvement in depressive symptoms, quality of life, and functioning to be the most relevant outcomes. With regard to the children’s and adolescents’ quality of life, they also highlighted the quality of relationships with relevant attachment figures and with peers. One expert added cognitive performance, inpatient stays as well as daily media consumption and drug use as further relevant outcomes. In qualitative interviews with adolescents, the alleviation of depressive symptoms and improved functioning were likewise found to be important outcomes of psychotherapeutic treatment of depressive disorders. From the perspective of this study’s adolescent participants, however, personal growth, the improvement of relationships, coping behaviours, and self-management as well as of their own well-being played important roles as well [94]. While adolescents aimed to achieve, in particular, changes in coping behaviours such as behavioural activation and resilience, parents as well as therapists placed greater value on school and work-related functioning [94]. An international expert committee has developed a consensus-based standard set of outcomes for the measurement of treatment results in depressive and other mental disorders. It defined the measurement of symptoms of disease, suicidal thoughts and behaviours as well as functioning [95].
- Results of the comprehensive information retrieval
- Characteristics of the systematic reviews included in the assessment
- Overview of patient-relevant outcomes
- Assessment of the quality of systematic reviews and the risk of bias of the results
- Results on patient-relevant outcomes
- Evidence map
- Outcomes from the perspectives of patient representatives
- Results: Benefit assessment - Depression in children and adolescents: Does psych...Results: Benefit assessment - Depression in children and adolescents: Does psychotherapy lead to better results when compared with other therapies?
Your browsing activity is empty.
Activity recording is turned off.
See more...