Using the Ontario Ministry of Health perspective and a 40-year time horizon, and cost and effectiveness data obtained from the multidisciplinary programme, the ODEM calculated cost-effectiveness data. The model was an adaptation of another model, the United Kingdom Prospective Diabetes Study (UKPDS) outcome model, to Ontario. The UKPDS model is a probabilistic, discrete-time model that predicts diabetes-related complications (i.e. myocardial infarction, ischaemic heart disease, stroke, heart failure, amputation, renal failure and blindness) based on individual patient characteristics and time-varying risk factors.
A retrospective chart review (historical case series) was performed on 401 patients undergoing the programme, both before and after its implementation. Data gathered to populate the model included demographic characteristics, diabetes medical history, history of other medical conditions and five key intermediate outcomes (haemoglobin A1C, blood-pressure, total cholesterol, high-density lipoprotein cholesterol and smoking status). Multiple imputation techniques were used for missing data. The main clinical effectiveness estimates were based on the difference for each participant in each of the intermediate outcome, from the beginning- to end-of-study value (average changes were statistically compared using paired t-tests). The base-case analysis assumed a 1-year treatment effect. The equations predicting long-term outcomes were already contained in the model.
Monetary benefit and utility valuations:
Diabetes-related complications and quality of life values were taken from a survey of 3,192 patients in the UKPDS that used the EQ-5D health status questionnaire.
Measure of benefit:
Intermediate outcomes were translated into two primary measures of benefit: life-years (LY) and quality-adjusted life-years (QALYs). A discount rate of 3% was used.
Costs categories included diabetes-related complications and programme implementation costs (including patient and provider education, specialty nurse liaison, diabetes template programming in the electronic record, and medication change during the programme).
The costs of complications were extracted from actual Ontario resource use profiles for a large prospective cohort of individuals with diabetes (n=734,113) over a 10-year time period. The cost of inpatient and outpatient hospital, outpatient physician visits, prescription drugs and home care services were included. The unit costs were collected from various Canadian sources. Costs included those of the first year (immediate) as well as those of long-term follow-up. The costs were reported in 2004 Canadian dollars, using a discount rate of 3%. Various sources were used to collect costing data relating to programme implementation. These ranged from the change in diabetes-related medication costs extracted from the chart review, to wages and salaries in the other components.
Analysis of uncertainty:
One-way, two-way and scenario analyses were evaluated. These included simultaneous variation of both the programme and treatment effect duration, and different durations of treatment and effectiveness.