Table 3.

Summary characteristics of the three reviews, review challenges and approaches and strengths and limitations of chosen methods

Report and no. RCTs	Populations, interventions and comparators	Review challenges and approaches	Strengths and limitations of chosen method
Kaltenthaler et al. 2014 [12] Sexual health of people with severe mental illness 13 RCTs	Review aims: summarise effectiveness evidence, determine applicability in UK NHS setting and identify key areas for primary research. Population: people with severe mental illness Interventions: strategies to increase knowledge, assess and reduce sexual health risk, change behaviour and develop condom skills Comparators: educational sessions on HIV, money management or substance abuse, health promotion, wait list or no treatment	Challenges due to evidence base:  • Wide variation in populations and settings (patients in psychiatric clinics, residential centres and homeless shelters)  • Wide range of outcomes including: biological (sexually transmitted infections, pregnancy), behavioural (number of partners, uptake of services, use of contraception/condoms) and proxy (knowledge, attitudes, behaviours, facilitators and barriers etc.) Approaches:  • Focussed definition of severe mental illness  • Brief summary of results presented, narrative synthesis, grouping of results from included studies by outcome (biological, behavioural and proxy)	Strengths:  • Enabled rapid synthesis of a disparate evidence base to ensure policy makers were aware of areas where evidence was available. This informed the design of relevant RCTs Limitations:  • Quantitative data synthesis not generated for use by policy makers (only effect size by intervention and outcome)  • In-depth narrative synthesis not possible  • Non-RCT evidence excluded
Cooper et al. (2015) [13] Premature ejaculation 101 RCTs and 1 CT (65 RCTs from existing reviews and 36 new RCTs and 1 new CT reports)	Review aims: synthesise effectiveness evidence for behavioural, topical and systemic treatments. Population: men with premature ejaculation Interventions: topical anaesthetics, antidepressants, phosphodiesterase-5 inhibitors, opioid analgesics, behavioural therapies, acupuncture, Chinese medicine Comparators: placebo, wait list, other therapies	Challenges due to evidence base:  • Very large number of RCTs (over 100) and existing systematic reviews covering wide range of interventions (several drug classes plus behavioural approaches)  • Several existing systematic reviews Approaches:  • Meta-analysis of primary outcome using data extracted from existing systematic reviews, with new primary study data added  • Narrative synthesis of secondary outcomes	Strengths:  • Meta-analysis able to be used for primary outcome (consistent primary outcome)  • Use of data from existing reviews enabled meta-analysis of large dataset in shorter time Limitations:  • Potential for data errors or synthesis errors in original reviews to be repeated in new report  • Methodological quality of studies extracted from existing reviews not assessed separately  • Although use of data from existing reviews saved some time, triangulation of data from multiple reviews was still time-consuming  • Original RCT publications not revisited for data extraction and quality assessment.
Cooper et al. (2015) [14] Cannabis cessation 33 RCTs	Review aims: summarise effectiveness evidence for psychological and psychosocial interventions and identify key areas for primary research. Population: adults who use cannabis regularly Interventions: cognitive behavioural therapy, motivational interviewing, motivational enhancement therapy, supportive-expressive dynamic psychotherapy, social support groups, case management, contingency management (vouchers as incentive/reward) Comparators: waitlist, treatment as usual, other interventions, assessment only, education controls, written cannabis information, cannabis education	Challenges due to evidence base:  • Wide variation in study populations (extent of cannabis dependence), interventions (type, duration) and comparators  • Very little consistency in outcome measures, time points, and statistics reported  • Large number of RCTs for a short report Approaches:  • For each pair of interventions compared, narrative summary of outcomes reported and how many showed a statistically significant effect	Strengths:  • Inclusive approach, covering a wide range of populations, interventions and outcomes  • Included all psychosocial or psychological interventions undertaken in the adult, community dwelling population of cannabis users Limitations:  • Detailed numerical outcome data not presented, since outcome measures and statistics reported were so disparate  • Outcome measures in RCTs not converted to consistent measures to compare across studies as not feasible in timeframe

CT controlled trial