Development and pilot of a multicriteria decision analysis (MCDA) tool for health services administrators.

INTRODUCTION: Health administration is complex and serves many masters. Value, quality, infrastructure and reimbursement are just a sample of the competing interests influencing executive decision-making. This creates a need for decision processes that are rational and holistic. METHODS: We created a multicriteria decision analysis tool to evaluate six fields of healthcare provision: return on investment, capacity, outcomes, safety, training and risk. The tool was designed for prospective use, at the beginning of each funding round for competing projects. Administrators were asked to rank their criteria in order of preference. Each field was assigned a representative weight determined from the rankings. Project data were then entered into the tool for each of the six fields. The score for each field was scaled as a proportion of the highest scoring project, then weighted by preference. We then plotted findings on a cost-effectiveness plane. The project was piloted and developed over successive uses by the hospital's executive board. RESULTS: Twelve projects competing for funding at the Royal Brisbane and Women's Hospital were scored by the tool. It created a priority ranking for each initiative based on the weights assigned to each field by the executive board. Projects were plotted on a cost-effectiveness plane with score as the x-axis and cost of implementation as the y-axis. Projects to the bottom right were considered dominant over projects above and to the left, indicating that they provided greater benefit at a lower cost. Projects below the x-axis were cost-saving and recommended provided they did not harm patients. All remaining projects above the x-axis were then recommended in order of lowest to highest cost-per-point scored. CONCLUSION: This tool provides a transparent, objective method of decision analysis using accessible software. It would serve health services delivery organisations that seek to achieve value in healthcare.

Epidemiologic modeling with FluSurge for pandemic (H1N1) 2009 outbreak, Queensland, Australia.

At the beginning of the pandemic (H1N1) 2009 outbreak, we estimated the potential surge in demand for hospital-based services in 4 Health Service Districts of Queensland, Australia, using the FluSurge model. Modifications to the model were made on the basis of emergent evidence and results provided to local hospitals to inform resource planning for the forthcoming pandemic. To evaluate the fit of the model, a comparison between the model's predictions and actual hospitalizations was made. In early 2010, a Web-based survey was undertaken to evaluate the model's usefulness. Predictions based on modified assumptions arising from the new pandemic gained better fit than results from the default model. The survey identified that the modeling support was helpful and useful to service planning for local hospitals. Our research illustrates an integrated framework involving post hoc comparison and evaluation for implementing epidemiologic modeling in response to a public health emergency.