Page 28 - The efficacy and effectiveness of psychological treatments for eating disorders - Elske van den Berg

28 Chapter 2
advice and psychoeducational interventions, and placebo conditions. The non-spe- cific treatment modality Specialist Supportive Clinical Management (SSCM) is regard- ed as a control condition, although recent guidelines consider SSCM as a psycholog- ical treatment (NICE, 2017). Treatment and control conditions could be individually or group based, inpatient as well as outpatient settings were included. During the screening phases, the references were independently rated by three researchers (LH, EB, and GK). Discrepancies were resolved by consensus and if needed, a senior reviewer was consulted (JD).
Risk of bias
The quality of the studies was assessed using The Cochrane Collaboration’s tool for assessing risk of bias in randomized trials (Higgins et al., 2011). The following domains on study and/or outcome level were independently assessed for each study by two researchers (EB and ID): sequence generation, allocation concealment, incomplete data and selective outcome reporting. Assessing incomplete data includ- ed screening using intention-to-treat (ITT) data. For each domain of potential bias, a value was given; ‘0’ indicating low risk of bias, ‘2’ indicating high risk and ‘1’ indicat- ing lack of sufficient information for a value to be given. For each study, a global risk of bias score was computed by adding up the values of all domains. Studies with a very low risk of bias (i.e. no potential bias at all domains) were considered high-qual- ity studies. Discrepancies were resolved by consensus within the review team and if needed, a senior reviewer was consulted (EK).
Outcomes
Meta-analyses were performed at post-treatment on intention-to-treat data. Primary outcome in this meta-analysis was weight gain. Weight gain could be meas- ured in terms of kilogrammes (kg), body mass index (BMI: body weight in kg divid- ed by height in m squared; kg/m2), ideal body weight, adjusted body weight, mean matched population weight (expressed as the percentage of the average weight for age, height and sex) or as a percentage of body fat. Eating disorder pathology was measured by a structured interview or a patient-reported measure, providing either a global score or separate scores for different domains. A measure of QoL was defined as any patient-reported measure assessing perceived quality of life or social impair- ment due to eating disorder pathology. With regard to variables related to patient characteristics at baseline (BMI, onset age, duration of illness), an average of the treatment and control condition for each arm was used.