Background People diagnosed with schizophrenia and other severe mental illness (SMI) have a reduced life expectancy compared to the general population, which can partly be explained by their increased risk of diabetes and associated poor diabetes management. A growing number of interventions are targeting this patient group to reduce physical health inequalities and improve life expectancy, which is around 20 years shorter compared to the general population. Several systematic reviews have investigated the effectiveness of interventions for managing body weight and other anthropometric markers. Less is known about their effectiveness for improving diabetes outcomes, or the quality of evidence focusing on this co-morbidity. This systematic review aimed to evaluate the clinical effectiveness of pharmacological and non-pharmacological interventions for improving diabetes outcomes in adults with SMI.

Methods We searched key databases from inception to October 2015 to identify randomised controlled trials measuring the effects of interventions for improving diabetes outcomes. The primary outcome was blood glucose control measured using HbA1c or fasting blood glucose. Trials which measured one of these outcomes and included adults with SMI, with or without diabetes, were included. Screening and data extraction were carried out independently by two researchers. A meta-analysis and narrative synthesis were performed to investigate the effects and components of different interventions. Risk of bias was assessed using the Cochrane Collaboration’s tool for critical appraisal. For non-pharmacological interventions, which targeted behaviour change, we also coded the behaviour change techniques (e.g. goal setting, rewards) employed.

Results We included 53 studies, comprising 39 pharmacological (intervention n = 1497, control n = 1346), 13 non-pharmacological (intervention n = 661, control n = 641), and one mixed intervention (intervention n = 96, control n = 32). Studies included lifestyle and other behavioural interventions, metformin, anti-psychotic switching, and a range of adjunctive pharmacological therapies to existing anti-psychotic medications. Many studies included small sample sizes and targeted different sub-groups of the population; for example inclusion criteria commonly set limits regarding body weight, medications, and presence of diabetes and other co-morbidities. Compared to usual care, both pharmacological and non-pharmacological interventions showed a small clinical improvement in both HbA1c and fasting glucose levels in people with SMI; however, there was considerably heterogeneity between studies, and few conclusions can be drawn about the effectiveness of interventions to improve diabetes outcomes for people with co-existing SMI and diabetes, due to poor reporting of sample characteristics and study findings.