This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.

Four strategies for the prevention of diabetes among high-risk individuals were examined.

The "DPP lifestyle program": people were immediately exposed to lifestyle modification (diet and exercise), such as that described for the lifestyle arm of the Diabetes Prevention Program (DPP) study. People who developed diabetes were maintained on the intensive lifestyle modification and followed for disease progression, while those whose glycosylated haemoglobin (HbA1c) level exceeded 7% were entered into an intensive diabetes treatment protocol designed to reduce their HbA1c level to below 7%.

The "baseline strategy": no lifestyle or other intervention was given initially. If people developed diabetes they were given dietary advice, but not entered into an intensive lifestyle programme, and were monitored for progression of the disease. If their HbA1c level exceeded 7%, they were entered into an intensive management programme with a goal of controlling HbA1c level to below 7%.

The "lifestyle when FPG > 125" (i.e. fasting plasma glucose > 125 mg/dL or 6.9375 mmol/L): no lifestyle or other intervention was given initially. If people developed diabetes they were entered into the intensive DPP lifestyle programme and monitored. If their HbA1c level increased to greater than 7%, they began intensive treatment to control their HbA1c levels to a goal of less than 7%.

The "metformin" strategy: the patients were put on metformin as soon as they were determined to be at high risk. If diabetes was diagnosed in these patients, they continued receiving metformin, were given dietary advice, and continued to be monitored. If their HbA1c level exceeded 7%, their drug treatment was intensified to control the HbA1c level to less than 7%. The DPP study was described in another publication (Knowler et al. 2002, see 'Other Publications of Related Interest' below for bibliographic details).

Primary prevention and secondary prevention.

Cost-utility analysis.

The study population comprised a hypothetical cohort of patients at high-risk for developing diabetes. The inclusion criteria were a body mass index greater than 24 kg/m2, an FPG level of 5.2725 to 6.9375 mmol/L (95 to 125 mg/dL), and a 2-hour oral glucose tolerance test result of 7.77 to 11.0445 mmol/L (140 to 199 mg/dL). Diabetes was considered to be present if the FPG level was greater than 6.9375 mmol/L (>125 mg/dL) or the 2-hour oral glucose tolerance test result was greater than 11.0445 mmol/L (>199 mg/dL).

The setting was primary care. The economic study was carried out in the USA.

Some clinical data were taken from studies published between 1998 and 2005, while the dates for resource use were not explicitly reported. The price year was 2000.

The effectiveness data were derived from a synthesis of completed studies.

The framework of the published Archimedes model was used to assess the medium-term cost-effectiveness of the four alternative preventive strategies in a hypothetical cohort of 10,000 people. The simulation model was first built up to model diabetes, congestive heart failure, coronary artery disease, stroke, hypertension and asthma in a single integrated model. The model included the biological variables and outcomes relevant to diabetes and its complications. Within the model, biological variables were changing and interacting continuously, and the natural histories and severity of conditions progressed smoothly. In addition, any clinical event could occur at any time, and the timing of events was as condensed or drawn out as occurs in reality. The model also included a detailed representation of the processes and logistics of clinical care and their related costs. Further details of the model were provided in an appendix. The model was validated using data extracted from clinical trials. Different time horizons (5, 10, 20 and 30 years) were explored.

The outcomes assessed from the literature were:

the incidence of diabetes among Hispanic men;

the progression of pre-diabetes to diabetes with no intervention, metformin and lifestyle modifications;

the rate of myocardial infarction (MI) in people with newly diagnosed diabetes with conventional and intensive treatment;

the rate of MI in people with diabetes and a high risk for coronary disease;

the rate of albuminuria, proteinuria, end-stage renal disease (ESRD), 2-step retinopathy, legal blindness and amputations in people with newly diagnosed diabetes with conventional and intensive treatment; and

the rates of development of ESRD in people with diabetes and microalbuminuria.

The utility values associated with specific health states were also estimated from the literature. A detailed list of inputs was not provided.

The authors stated that each variable was estimated from at least one empirical source. However, it was unclear whether a systematic review of the literature had been undertaken to identify the primary studies. It was noted that the clinical data were derived from basic physiologic studies, surveys, epidemiologic studies and clinical trials. The utility values came from two studies that used the Quality of Well-Being Index.

MEDLINE was searched from 1970 to 28 February 2005.

Not stated.

Not stated.

Five primary studies used to derive clinical data were explicitly reported. However, many other studies are likely to have been used to populate the model.

Not stated.

Not stated.

Only some of the clinical data estimated from the literature were reported.

The incidence of diabetes among Hispanic men was 0.00795.

The progression of pre-diabetes to diabetes was 0.201 with no intervention, 0.079 with metformin, and 0.046 with lifestyle modifications.

The rate of MI in people with newly diagnosed diabetes was 0.018 with conventional treatment and 0.0138 with intensive treatment.

The rate of MI in people with diabetes and a high risk for coronary artery disease was 0.034 with placebo and 0.0235 with ramipril.

The rate of development of albuminuria in people with newly diagnosed diabetes was 0.0338 with conventional treatment and 0.0198 with intensive treatment.

The rate of development of proteinuria in people with newly diagnosed diabetes was 0.0086 with conventional treatment and 0.0065 with intensive treatment.

The rate of development of ESRD in people with diabetes and microalbuminuria was 0.101 with placebo, 0.054 with irbesartan 150 mg, and 0.0297 with irbesartan 300 mg.

The rate of development of ESRD in people with newly diagnosed diabetes was 0.00081 with conventional treatment and 0.0001 with intensive treatment.

The rate of development of 2-step retinopathy in people with newly diagnosed diabetes was 0.056 with conventional treatment and 0.04 with intensive treatment.

The rate of development of legal blindness in people with newly diagnosed diabetes was 0.00486 with conventional treatment and 0.00336 with intensive treatment.

The rate of development of amputations in people with newly diagnosed diabetes was 0.0004 with conventional treatment and 0.000134 with intensive treatment.

The measure of benefits used was the number of quality-adjusted life-years (QALYs). These were estimated by combining utility weights and survival in the decision model. Reductions in diabetes incidence and other events, such as MI, were also reported since they were relevant from the perspective of an individual high-risk person. The short-term benefits observed in the clinical trials persisted as long as patients were receiving the lifestyle intervention. A discount rate of 3% was applied to future benefits.

The cost analysis was performed from the perspectives of the health plan and society. However, the same categories of costs were taken into consideration. The difference was the size of the relevant population. The economic analysis considered only the direct medical costs associated with the interventions examined in the study. A detailed breakdown of the cost items was not provided. For example, it was stated that lifestyle and metformin interventions included the costs of personnel, health education materials, medications and laboratory tests. The generic for metformin was considered. The unit costs were not presented separately from the quantities of resources used. Most of the costs were estimated from Kaiser Permanente using a micro-costing method. These costs represented real costs rather than charges or reimbursement rates. The other costs came from clinical trials. The resource use data were presumably derived from the same source as the costs. No other information on the costs was provided. Discounting was relevant and an annual rate of 3% was applied. The price year appears to have been 2000.

The costs were treated deterministically.

The indirect costs were not explicitly taken into consideration in the economic analysis.

US dollars ($).

Sensitivity analyses were performed to examine the robustness of the cost-utility ratios to variations in baseline model inputs. The ranges of values used were generally derived from the literature. The type of analysis appears to have been univariate. The authors stated that the issue of uncertainty around the model structure was addressed by model calibration.

From an individual high-risk person perspective and over a 30-year time horizon, with lifestyle modification, diabetes would be prevented in about 11% of cases and postponed in about 61%. With metformin, about 4% of diabetes cases would be prevented over a 30-year period, a relative reduction of about 5.5%.

The use of lifestyle interventions reduced the incidence of diabetes and other events (MI, stroke, retinopathy, neuropathy and death) in comparison with the baseline strategy.

From a societal perspective and over a 30-year time horizon, the estimated QALYs were 11.319 with the baseline strategy, 11.444 with the "lifestyle when FPG > 125 mg/dL" strategy, 11.478 with the DPP lifestyle strategy, and 11.432 with metformin.

Using a health plan perspective, the lifestyle programme reduced office visits and procedures (i.e. electrocardiography, coronary artery bypasses, and photocoagulation surgery), but increased hospital admissions, medications and ongoing programmes (i.e. case management) in comparison with the baseline strategy. Thus, the charge per member per month of an insured individual (not only high-risk individuals but also the entire membership) increased.

Using a societal perspective and a 30-year time horizon, the estimated costs were $24,523 with the "lifestyle when FPG > 125 mg/dL" strategy, $62,602 with the DPP lifestyle strategy, and $35,523 with metformin.

The incremental cost-utility ratios (i.e. cost per QALY) were calculated to combine the costs and benefits of the alternative strategies.

From the perspective of a health plan, the incremental cost per QALY for the DPP lifestyle programme compared with no intervention (assuming a 10% member turnover rate per year) was $143,000 after 30 years. The cost per QALY increased with shorter time horizons: $2.7 million for 5 years after the start of the programme, $1.2 million for 10 years, and $180,000 for 20 years.

Using a societal perspective and a 30-year time horizon, the incremental cost per QALY was $24,523 with the "lifestyle when FPG > 125 mg/dL" strategy and $201,818 with the DPP lifestyle strategy. The metformin strategy was dominated by the "lifestyle when FPG > 125 mg/dL" strategy, which was both more effective and less expensive. The cost per QALY increased with shorter time horizons, as in the analysis from the health plan perspective. The average cost per QALY was also calculated. This was $24,523 with the "lifestyle when FPG > 125 mg/dL" strategy, $62,602 with the DPP strategy, and $35,523 with the metformin strategy.

The sensitivity analysis showed that it was quite unlikely for the lifestyle intervention to be cost-effective when considering a threshold of $50,000 per QALY. A probability distribution showed that there was only a 0.1% chance for the incremental cost per QALY of the lifestyle programme versus no intervention being lower than $50,000 per QALY.

Only substantial reductions in the cost of the intervention would make the lifestyle intervention cost-effective. The greatest uncertainty was around the utility values and associated QALYs obtained from the DPP study.

The lifestyle intervention was effective in reducing the risk of diabetes among high-risk individuals. However, neither health plans nor the society could afford it because of the high cost of the intervention itself. Further, the authors stated that cheaper options were available, such as delaying the intervention until people actually got diabetes.

The authors provided a justification for the choice of the comparators, which were based on the results of the DPP study. However, the analysis focused on the lifestyle intervention. You should decide whether they are valid comparators in your own setting.

The effectiveness evidence came from a synthesis of published studies. However, no details on the design and characteristics of the primary studies were reported. Some of the data were taken from clinical trials, but no other details were given. There was limited information on the sources of the utility values. Some data were estimated from the DPP study. Full details of the primary sources were reported in a separate publication (available online). Since only some clinical inputs were reported, it was not possible to assess the validity of the primary estimates. Sensitivity analyses were carried out on a few parameters only.

The use of QALYs as the summary benefit measure was appropriate as they capture the most important dimensions of care (i.e. survival and quality of life). The use of QALYs ensures the comparability of the benefits of the current study with those associated with other interventions. Discounting was applied, as recommended in US guidelines for economic evaluations.

The authors stated that the perspectives of the health plan and the society were adopted. However, only the direct medical costs were taken into consideration. It was pointed out that other costs were implicitly considered in the calculation of the QALYs, although this represents a controversial issue. The unit costs were not presented separately from the quantities of resources used. There was limited information on the costs included in the analysis and a detailed breakdown of the cost items was not given. This limits the possibility of replicating the results of the analysis in other settings. The costs were treated deterministically, but some key estimates were varied in the sensitivity analysis. The source of the cost data was reported, while that for resource consumption was unclear. The price year was reported, although not explicitly, which aids reflation exercises in other time periods.

The authors compared their findings extensively with those from a published study that reported a more favourable cost-effectiveness ratio. The potential explanations for this difference were discussed in an appendix. The issue of the generalisability of the study results to other settings was not explicitly addressed. Some sensitivity analyses were performed, which partially enhance the external validity of the study. The authors provided an extensive description of the model used in the analysis. Strengths and weaknesses of their model, in comparison with a more traditional Markov model, were presented in an appendix. The authors noted that their analysis would apply only to patient populations similar to those enrolled in the DPP study.

The study results suggested that the programme used in the DPP study may be too expensive for health plans. Thus, less expensive methods should be developed.

Funded by Kaiser Permantente, the American Diabetes Association and Bristol-Myers Squibb.

Knowler WC, Barrett-Connor E, Fowler SE, et al. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002;346:393-403.

Herman WH, Hoerger TJ, Brandle M, et al. The cost-effectiveness of lifestyle modification or metformin in preventing type 2 diabetes in adults with impaired glucose tolerance. Ann Intern Med 2005;142:323-32.

UK Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 1998;352:837-53.