Pace 2013.

Methods	Randomised trial of cognitive‐based compassion training vs wait list
Participants	Included (n = 71) Adolescents in the Georgia foster care system with a documented history of trauma and neglect Mean age: 14.7 years. Female: 31. Ethnicity: African American 74%, Caucasian 13%, multi‐racial 7%, Asian 4%. Mean number of psychiatric diagnoses: 1.74. Trauma was neglect 40.20%, physical abuse 25.07%, sexual abuse 11.66% Excluded Adolescents taking medications known to influence immune and endocrine functioning, including corticosteroids and non‐steroidal anti‐inflammatory compounds; medical illness including cancer, cardiovascular disease, diabetes and autoimmune disorders, schizophrenia, bipolar I disorder, eating disorders and major depression severe enough to require hospitalisation Setting Georgia State foster care system, USA, 2010
Interventions	Cognitive‐based compassion training (n = 37) Participants in the cognitive‐based compassion training (CBCT) group attended classes of 1 hour twice a week for 6 weeks. CBCT is a secular, analytical meditation‐based programme derived from Tibetan Buddhist mind‐training. The goal of CBCT is to challenge unexamined assumptions regarding feelings and actions toward others, with a focus on generating spontaneous empathy and compassion for the self as well as others Wait list (n = 34) Control group was on a 6‐week wait list Therapists Not described
Outcomes	Depression Scale: Quick Inventory of Depressive Symptomatology Rater: child/adolescent Anxiety Scale: State‐Trait Anxiety Inventory Rater: child/adolescent When Post therapy
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	'Participants were evenly randomized by a list of random numbers, generated by computer'
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding of participants (performance bias	High risk	Participants probably were aware of whether they were in the active or control group
Blinding of outcome assessment (detection bias) All outcomes	High risk	Not stated, but self report measures were used
Incomplete outcome data (attrition bias) All outcomes	High risk	Loss to follow‐up: 22.5%
Selective reporting (reporting bias)	Low risk	All outcomes appear to have been reported
Other bias	High risk	Percentage of participants who had undergone sexual abuse was much higher in the intervention group than in the control group (18.52% vs 4.00%), and neglect was much higher in the control group (52.00% vs 29.63%)