The authors concluded that lifestyle interventions for patients at high risk of diabetes, delivered by a variety of healthcare providers in routine clinical settings, were feasible but appeared to be of limited clinical benefit one year after intervention. Given the limitations of the included studies and with the review, the authors' conclusions should be interpreted with caution.

To evaluate the effectiveness of physical activity and nutrition interventions for patients at risk of diabetes, and comment on the transferability of reference trials to routine clinical practice.

MEDLINE, the Cochrane Library, CINAHL, and EMBASE were searched for studies from January 1990 to August 2009 that were published in English. Search terms were reported. Reference lists from identified articles were handsearched. Google Scholar was also searched. Reports were identified from government, non-government, and health service sources by searching the Internet and grey literature.

Eligible for inclusion in the review were interventions that were either replications or modifications of all or some components of the US Diabetes Prevention Program (DPP), Finnish Diabetes Prevention Study (DPS) or any other reference trial, or those including the reduction of diabetes risk or diabetes incidence explicitly. Eligible study designs were RCTs, before-and-after studies, cohort studies with or without a control group, and interrupted-time-series analyses. Eligible studies had to last least three months, with three months follow-up of single (either nutrition or physical activity programmes with or without medication) or combined (nutrition and physical activity) programmes, with or without medication. Eligible participants were adults with impaired glucose regulation at risk of diabetes. Comparators included no intervention or alternative intervention (single or combined).

Eligible studies had to be performed in routine clinical settings, be delivered on-site or in associated facilities, with outcome measures assessed on-site by general medical practitioners, specialist physicians, practice nurses, allied health professionals, community health staff or research staff attached to the service. Study authors were contacted to verify the study setting. Studies that included patients diagnosed with diabetes were only included in the review if outcomes were reported separately from patients without diabetes. Studies of diabetes education materials only, or medication only were excluded. Further details were given in the paper.

Studies had to report at least one of the following outcomes: improvement in objectively measured risk factors, such as weight loss or waist circumference reduction; metabolic outcomes indicative of diabetes reduction; self-reported or objectively measured behavioural outcomes. The secondary outcome was prevention of diabetes (incidence percentage or delay in onset or reduction in incidence over a given follow-up time).

All the included studies were performed in developed countries. Three studies were conducted in the UK. The age of the target population varied between studies, but half of the studies targeted middle-aged people only. Most studies included normoglycaemic partcipants. All interventions combined physical activity and dietary advice, with two studies additionally including medication. Half the physical activity interventions were structured. For group based interventions, the median number of sessions was six; for individual based interventions, the median number of sessions was 13.5. Most of the studies were delivered in specialist service or hospital outpatient clinics, or in general practitioners consulting rooms or community health centres. Nurses or allied health-care professionals usually delivered the interventions. Approximately 60% of the studies attempted replication of the US Diabetes Prevention Program or the Finnish Diabetes Prevention Study (with some modifications). Despite a minimum three-month intervention duration being specified in the inclusion critiera, it was noted that the intervention duration in one study was only one month.

The authors did not state how many reviewers performed study selection.

One author assessed internal validity using the following criteria: evidence of assessment of risk of diabetes at enrolment; explicit eligibility and exclusion criteria; reported participation rate of at least 50% of eligible people; follow-up assessment rates of at least 65%; evidence of measurable or explicit outcome assessment; appropriate statistical methods; adequate control for confounders (in non-RCTs); explicit intervention components; and conclusions supported by findings.

One author assessed the potential for bias using an amalgamation of established tools, such as STROBE (Strengthening the Reporting of Observational studies in Epidemiology), PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) and MOOSE (Meta-analysis Of Observational Studies in Epidemiology). Criteria assessed included: participant recruitment source; selection criteria; treatment allocation; blinding of outcome assessors; simultaneous data collection for intervention and control group; measurement error; subgroup analysis; and discussion of study limitations.

Three authors independently conducted a second check on a sample of studies.

Data were extracted to calculate mean differences and 95% confidence intervals (CIs). Standard deviations were calculated where not reported for individual studies.

Two authors independently performed the data extraction.

Mean differences and 95% confidence intervals at one year follow-up for RCTs that reported outcomes in the same units were combined in a random-effects meta -analysis. The authors did not report whether heterogeneity was assessed. Data from the remaining studies were presented in a narrative synthesis.

Twelve studies were included in the review (sample size ranged from 37 to 3,304 participants) comprising seven RCTs (including one cluster RCT), three before-and-after studies without a control group, and two before-and-after studies with a control group. Study quality scores ranged from 3 to 8; most studies scored 7. Follow-up ranged from six to 60 months.

Four RCTs were included in the meta-analysis (n=1,097 participants, sample size ranged from 37 to 458). All four RCTs were of high quality (three scored 7 points and one scored 8). Loss to follow-up ranged from 11 to 19%.

At 12 month follow-up, there was a significantly greater mean weight reduction (mean difference -1.82kg, 95% CI -2.7 to -0.99; four RCTs) and mean waist reduction (-4.6cm, 95% CI -5.8 to -3.4; three RCTs) in the treatment group compared with the control group. There was no significant difference in fasting plasma glucose (-0.19mmol/L, 95% CI -0.44 to 0.06; three RCTs) or two-hour oral glucose tolerance test (0.04mmol/L, 95% CI -0.49 to 0.42; two RCTs)

Nine studies examined whether translation of the reference trials into clinical care was feasible. Of these, eight studies found that translation was feasible with some modifications.

Lifestyle interventions for participants at high risk of diabetes, delivered by a variety of healthcare providers in routine clinical settings, were feasible but appeared to be of limited clinical benefit one year after intervention. Translation into routine practice had somewhat less of an impact on diabetes risk reduction.

The review addressed a clear research question, which was supported by detailed inclusion criteria. The search strategy included a number of sources, including unpublished material. However, the search was limited to studies published in English, which meant that relevant studies may have been missed. It was unclear how many reviewers performed study selection; the remaining review process appeared to be conducted using measures to reduce reviewer error and bias.

The authors presented overall study quality scores, but did not report the full results for each component; so it was of limited value for assessing the reliability of the included studies. There was a lack of assessment of heterogeneity, so it was unclear whether the chosen method of synthesis was appropriate.

Given the limitations of the included studies (small number of studies, small sample sizes, short follow-up), which were acknowledged by the authors, and limitations of the review (language bias, lack of assessment of heterogeneity, limited reporting of study quality), the authors' conclusions should be interpreted with caution.

Practice: The authors stated that it was still worth promoting the translation of modified, group based lifestyle interventions.

Research: The authors stated that there should be more rigourous evaluations of the interventions in routine clinical settings. There should be an establishment of a register of translation projects using consistent outcomes. Measurable outcomes would add more certainty to the effectiveness of routine practice interventions.

Not stated

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.