Table 6.

Effectiveness of user involvement in adolescent mental healthcare

Reference	Study design	Participant characteristics	Intervention, study setting	Trial/study armsa	Results	Internal and external validity assessment b
Walker 2017 [38], USA	Randomized controlled trial	Age: 12–18 years (x̄ = 14.2, SD1.3), female 42%, serious mental health problems	Wraparound: team working with adolescents, their family members and the family’s social support network, determining the primary needs, service and support strategies to be included in the care plan AMP: Achieve My Plan, enhances Wraparound through multi-system involvement with caregivers and service providers Outpatient CAMHS	I: Wraparound with AMP: n = 35 C: Wraparound without AMP: n = 20	Primary outcomes: Youth Participation in Planning Scale (YPP): Youth participation in preparation and planning in favour of the intervention at 3–4 weeks and 10–12 weeks (p < 0.01). Accountability in favour of the intervention at 3–4 weeks (p < 0.03), but not at 10–12 weeks (p = 0.10) Youth Empowerment Scale (YES): No significant effects Secondary outcome: Intervention group participants were 2.35 times more likely to rate care planning meetings as much better than control group participants (p < 0.001)	Internal validity: Overall risk of bias: high Performance bias: high Selection, detection, reporting and other forms of bias: unclear External validity: More pragmatic than explanatory
Jager 2017 [34], c Netherlands	Longitudinal prospective cohort study	Age: 12–18 years (x̄ = 15.2, SD1.7) female 61%, adolescents who signed up for psychosocial care (77% in mental health care)	Psychosocial care, mostly delivered by a mental healthcare organization (77%). Care with patient-centred communication, including shared decision-making. Intervention duration: 6 months Specialist healthcare services	T1 (baseline): N = 416 T2 (3mo.) + T3 (1 year): n = 315(76%)	Strengths and Difficulties Questionnaire (SDQ) with changes in Total Difficulties Score (TDS) from T1 to T3: Experience of shared decision-making associated with larger improvement in TDS scores, irrespective of adolescents’ expectations. Unmet shared decision-making communication needs associated with lower improvement in self-confidence (p < 0.001)	Internal validity: Overall risk of bias: high Selection, performance, detection and attrition bias: high. Reporting and other forms of bias: low External validity: Equally pragmatic and explanatory
Jager 2014 [33], Netherlands	Longitudinal prospective cohort study	Age: 12–18 years, female 65%, adolescents who signed up for psychosocial care (76% in mental health care)	Psychosocial care, mostly delivered by a mental healthcare organization (76%). Care with patient-centred communication, including shared decision-making Duration 3 months Specialist healthcare services	T1 (baseline): N = 416 T2 (3mo.): n = 211 (51%) (min. 2 appointments)	Shared decision-making on the Consumer Quality Index (CQI) at 3 months: Adolescents who considered shared decision-making to be important (expectations), but experienced it to less extent, had lower degree of improved understanding of mental health problems and how to handle them, compared to those who had agreement between expectations and experiences (OR 4.2, 95% CI 1.7–10.8, p < 0.01)	Internal validity: Overall risk of bias: high Selection, performance, detection and attrition bias: high. Reporting and other forms of bias: low External validity: Equally pragmatic and explanatory
Simmons 2017 [36], Australia	Non-randomized comparative study	Age: 16–25 years (x̄ = 17.8, SD 2.9), female 63%, adolescents who attended a youth mental health service clinic	Peer workers engaged with adolescents during intake assessment and online shared decision-making tool, prior to individual counseling session with a clinician Historical comparison group without peer workers and online shared decision-making tool E-health in primary & secondary care	I: n = 149 Response to SDMQ-9: n = 78 (52%) C: n = 80 Response to SDMQ-9: n = 61(76%)	Shared Decision-Making Questionnaire (SDMQ-9) (clinician rated) on day 1: In favour of the intervention group (p = 0.015), but limited clinical effect (mean difference 2.4 on a 54 point scale)	Internal validity: Overall risk of bias: high Selection, performance, detection and attrition bias: high. Reporting and other forms of bias: unclear External validity: More pragmatic than explanatory
Simmons 2017 [37], Australia	Cohort study with pre- to post-assessment	Age: 12–25 years (x̄ = 18.5, SD3.4), female 82%, depression (PHQ-9): mild (min.5 points)(18%), mild–moderate (26%), moderate–severe (56%)	Online decision aid to help adolescents make decisions in line with evidence and their personal preferences and values Primary care	T1 (before decision aid): N = 66 T2 (after decision aid): n = 57 (86%) T3 (8 weeks): n = 48 (73%)	Patient Health Questionnaire (PHQ-9) from T1 to T3: mean reduction of 2.7 points (95% CI, 1.3;4.0) Decisional Conflict Scale (DCS) from T1 to T2: mean reduction 17.8 points (95% CI 13.3;22.9, p < 0.001)	Internal validity: Overall risk of bias: high Selection, performance, detection, attrition and reporting bias: high. Other forms of bias: low External validity: More pragmatic than explanatory
Walker 2010 [39], USA	Cross-sectional and repeated measures survey	Age 14–21 years (x̄ = 16.2,SD1.7), female 43%, mental health difficulties: ADHD, depression, bipolar disorder, PTSD, ODD, conduct disorder	Testing of a Youth Empowerment Scale–Mental Health (YES–MH), adapted from the Family Empowerment Scale (FES), services provided by multiple child- and family-serving agencies, primary & secondary care	T1 (baseline): N = 185 T2 (6 weeks): n = 60	Results based on exploratory factor analysis of YES–MH suggest three levels of empowerment: a) system: confidence & capacity to help providers improve services and help other youth with emotional/mental health difficulties b) services: confidence & capacity to work with service providers to select and optimize services c) self: confidence & capacity to cope with or manage one’s own condition Positive correlation between YES-MH and a 6-item Participation in Planning Scale (PPS)(p < 0.01)	Internal reliability: very good for both YES-MH (Cronbach’s alpha 0.85 – 0.91) and PPS (0.90) Test–retest reliability good for all three sub-scales of YES-MH (p < 0.01). No other forms of psychometric tests were applied
Nolkemper 2019 [35], Germany	Cross-sectional survey	Age: 12–18 years (x̄ = 14.8, SD1.5), female 42%, adolescents who have been hospitalized for mental health conditions	Psychiatric treatment Child and adolescent psychiatry medical college & child and adolescent psychiatry university hospital	Experience of participation in psychiatric treatment: N = 114	Self-developed questionnaire focusing on feeling of being able to participate in decision-making (6 items, Likert scale): Yes, very much: 12% Yes: 40% Partially: 25% Not really: 13% Not at all: 10% No significant age, gender or clinic differences	Internal validity: Overall risk of bias: high Selection, performance, detection, attrition and other forms of bias: high. Reporting bias: low External validity: Equally pragmatic and explanatory

^aI = Intervention, C = Control

^bInternal validity: Cochrane Collaboration’s guidelines for risk of bias assessment [23]. External validity: The PRECIS tool for assessing studies on a pragmatic-explanatory continuum was used [26]. Validity assessment for Walker 2010 [39] focuses solely on criteria of relevance to psychometric tests

^cJager 2017 [34] builds on the same data as Jager 2014 [33], but assesses different outcomes and includes long-term follow-up