Analytical approach:
An epidemiological model was structured using data from an administrative database. This database collected information from the collaborative hypertension intervention conducted by the Utah Department of Health, Heart Disease and Stroke Prevention Programme (HDSPP) and SelectHealth commercial health maintenance organisation. Given the lack of long-term follow-up data, both one-year and 10-year time horizons were used. The authors did not report a study perspective.
Effectiveness data:
The main clinical effect estimates were hypertension control rates and adverse events. Effectiveness data were derived from the 2007 data collected from the administrative database. The database provided summary details of the programme structure, participants and blood pressure control at follow-up. For this cost-effectiveness analysis a subset of 534 high-risk hypertensive plan members who received blood pressure monitors were included from a total of 17,318 members with diagnosis of hypertension identified by claims. Adverse events risks estimates were obtained from published sources which used the Framingham risk calculator.
Monetary benefit and utility valuations:
Not relevant.
Measure of benefit:
Hypertension control, number of adverse events and life-years gained. Life expectancy for the 10-year time horizon scenario was discounted using a discount rate of 3%.
Cost data:
The cost categories included in the study were labour, materials and supplies, and contract services. A site visit was undertaken to collect the cost information required to estimate these cost categories. Average medical costs for each event were taken from the published literature. Some author assumptions were used within the control group for prescription drug claims and physician visits and control rates in the absence of intervention. Long-term costs were discounted using an annual rate of 3%. Costs were presented in 2007 US Dollars ($).
Analysis of uncertainty:
One-way sensitivity analysis was undertaken. Results were presented in a tornado diagram.
Three scenarios were presented in order to show the sensitivity of the results to modelling assumptions, a base case and a pessimistic and an optimistic model. The base model assigned average behaviour to outcomes that were not observed and did not attribute the unanticipated reduction in provider visits to the intervention. The pessimistic model included assumptions that minimised benefits attributed to the intervention. The optimistic model included assumptions that maximised benefits attributed to the intervention.
An appendix explained that 95% sensitivity ranges around the incremental cost-effectiveness ratios (ICER) using probabilistic sensitivity analysis. 10,000 simulations were created and the ICER distribution generated was used to estimate the reported 95% sensitivity range.