Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

One of the biggest challenges in estimating the cost-effectiveness of health care interventions is the dichotomy between controlled trials and real life health care systems. Most published economic evaluations are based on data taken directly from trial settings, but as a number of prominent commentators have suggested, the trial setting is geared up to removing context from a situation so that the 'efficacy' of an intervention can be measured in isolation. Health care systems, though do not work in isolation, and so the resulting measures of cost-effectiveness are limited in their ability to be generalized beyond their setting. Alternatively the problem with the evaluation on working program or health care systems lies in the lack of comparable output measures. Some public health programs collect process outcome data such as attendances, contacts, products delivered or people treated, but this data is not of limited use across different scales, settings and health areas.

This study looks at the value of stochastic evidence-based modeling as a tool to bridge the gap between proven efficacy and scale and scope effected effectiveness of various public health interventions addressing infectious disease and maternal conditions across multiple countries. The models themselves utilize efficacy data drawn from published systematic reviews, along with country and region specific epidemiological and demographic data where available. This data is then combined with a series of process outcomes generated from public health interventions and complex multi-intervention programs on the ground, incorporating coverage and costs related to scale and relative burden within the specific location. The models combine deterministic and stochastic components to allow for parameter variability. The variability is incorporated via probability density functions or by using appropriate intervals based on mean or median parameter estimates. Both analytical and monte-carlo methods are used to combine the random elements in order to obtain confidence intervals on the model outcomes. Sensitivity analysis can also be performed to identify how the precision of parameter estimates impacts on the accuracy of the outcome predictions.

The results are a matrix of health outcomes and incremental cost-effectiveness ratios that allow for a more realistic estimate of true cost-effectiveness and a better understanding of economies of scale and scope across a number of delivery mechanisms of public health interventions. Its is hoped that the model will allow a better understanding of the relationships between the cost-effectiveness of public health interventions in developing countries, and allow for better resource utilization by both donor agencies and practicing public health and development organizations.