Evaluation of a Community-wide Diabetes Prevention Program

Abstract

This thesis is an evaluation of the effectiveness of a community-wide diabetes prevention program conducted in three Divisions of General Practice in Sydney, Australia. The aims were to assess whether translation of diabetes prevention programs was feasible in real-life settings and whether results achieved were comparable with those of randomised trials on which this intervention was based. Its primary goals were to assess whether the lifestyle intervention could increase participation in moderate-to-vigorous physical activity to 210 minutes per week, reduce total fat and saturated fat consumption to 30% and 10% of total daily energy intake, increase fibre consumption to 15 g/1,000 kcal/day, and lead to 5% weight loss over one year. The background section covers the physiopathology of type 2 diabetes, its risk factors, and the available population screening tools to identify people at risk. The growing morbidity and mortality burden, the economic implications of this public health problem, and the importance and feasibility of preventing or delaying the onset by intervening in the precursor stages are then summarised. Evidence for preventability is examined through a literature review of lifestyle interventions in research settings comprising highly structured and closely monitored physical activity and dietary programs under controlled conditions. Examples of the effectiveness of translation of randomised controlled trials (RCTs) into less stringent programs in community settings such as workplaces, churches, indigenous communities and whole-of-country initiatives are presented. A systematic review and meta-analysis of effectiveness of the lifestyle approaches in routine clinical practice supplements the evidence for application of prevention principles in real-life settings. The main chapters of the thesis centre on process and impact evaluation of the semi-structured Sydney-based intervention, which recruited 1,250 participants from the mainstream Australian 29 public using general practitioner services in the study area, who were followed for 12 months. The intervention’s goals aligned with those of the Finnish Diabetes Prevention Program but with less stringent entry criteria and less intensive intervention components delivered by purpose-trained lifestyle officers. The Program included an initial individual assessment and coaching session, three subsequent group sessions in the following three months, then three follow-up coaching calls at three, six and nine months. A final assessment at one year, using the same objective and self-reported measures as in the initial assessment, captured changes in body weight, physical activity and dietary habits. The process evaluation showed that it is feasible and effective to use targeted screening to identify and recruit high-risk individuals into a free-of-charge program in the general practice setting, however a quarter of participants were lost to follow-up by one year. While minor variations in aspects of the Program were required to meet local need, Program fidelity in delivering components, and self-reported adherence to diet and physical activity was high. Using a before-after study design, the impact evaluation measured 1-year changes in key Program parameters in relation to baseline. These comprised: measured weight, waist circumference, BMI, and glycaemia measurements; and self-reported dietary intake and structured physical activity, using a 3-day food record and the Physical Activity Scale for the Elderly (PASE) questionnaire, respectively. The main findings at 12 months for the 586 completers as at December 2010 were: a mean weight loss of 2.1 kg; waist circumference reduction of 2.5 cm; no significant change in glycaemia; 3% reduction of fat and saturated fat intake; 16% increase in fibre intake; and mean increase in moderate-to-vigorous physical activity of 13.7 minutes/week. All these changes were smaller than those achieved by the RCTs in research settings, most likely due to the lower intensity and monitoring of the Sydney intervention. Weight loss and waist circumference reductions were similar for participants in 30 group session and those who received telephone-only coaching. Diabetes incidence was 1% at the end of the first year. An economic appraisal of the Program implementation completes the evaluation. A cost of A$400 per kg lost among people achieving the weight goal was estimated on Program completion, but the cost was double for the overall group that included non weight losers. The cost of achieving the physical activity goal and the dietary goals was not feasible or sustainable with resources available in routine clinical settings. The costs per outcome were similar for participants not attending group sessions, who received only telephone coaching. Hence it is worth exploring this less labour-intensive modality if a general practice based Program were to be delivered as routine preventive care. In sum, the evaluation of this community-wide diabetes prevention program showed that translation of diabetes prevention programs into routine practice, while feasible at less intensive levels than in RCTs, has a somewhat lower effect on diabetes risk reduction and it can still be a financial burden in clinical settings. However, given the potential for population-wide benefit, the effectiveness of alternative delivery modes, number and duration of program components and more targeted patient sub-groups should be investigated.