Quantitative Modeling Software with Applications to Medical Decision Making

Project Summary Abstract In recent yearshealth care systems and health care providers have made concerted efforts to practice evidence based medicine to provide patients with the best available information when making choices about their medical decisionsHoweverthese decisions are often complex with many uncertainties and potential outcomessome beneficial others direA popular tool used to help identify best treatment strategies is a decision treewhich outlines a patientandapos s potential outcomes given an initial medical choiceComplex decision trees are evaluated via Monte Carlo microsimulation to trace a patientandapos s path through the treeThis movement is inherently stochastic because outcomes are probabilistichoweverwhen the microsimulation is repeated many timesit provides the probability of each associated outcome resulting from the initial medical decisionFrom this probability distributionquantitative measures associated with each medical decision can be calculated including beneficial as well as adverse eventslife yearsquality adjusted life yearsa generic measure of disease burdenand othersWhen outcome costs are known and incorporated into the modelcost effectiveness analysisCEAcan be used to readily compute the relative costseffectivenessand incremental cost effectiveness for each health outcomeWe propose to add functionality to the mathematical modeling software Berkeley Madonna to allow users to build decision trees and carry out Monte Carlo microsimulations on these treesAimBerkeley Madonnaandapos s interface was designed to gently introduce students from non technical fields into mathematical modeling by using a simple syntax and graphical images to construct sophisticated equationsWe will leverage this easy to use interface to introduce medical researchers to microsimulationThe software will be adapted to build decision trees with built in functions and customized graphics specific to this fieldincluding measures from CEAPatients moving through a decision tree using microsimulation must be simulated hundreds of thousands to millions of times to arrive at statistically significant outcome probabilitiesand these simulations are computationally intense often requiring months of computer timeWe will harness the power of graphics processing unitsGPUsto parallelize these simulations to achieve tremendous speedups compared to commercially available softwarewhich have not taken advantage of these hardware capabilitiesWe show that a simple Monte Carlo microsimulation can be simulatedx faster on a GPU compared to a CPUand we will optimize the code to fully realize these speedups when simulating complex decision treesAimSuccessful completion of our goals will provide a powerful research tool to the medical decision making fieldwhich will positively impact health outcomes researchWe expect that our software will be particularly beneficial to the cardiovascular research communitywhich has a history of practicing evidence based medicine that includes simulation modeling Project Narrative This proposal will expand the functionality of the mathematical software Berkeley Madonna to perform medical decision analysis and cost effectiveness analysis using microsimulationWe will implement GPU accelerated algorithms to dramatically increase calculations speedswhich will far outpace current industry standardsSuccessful completion of our goals will provide a powerful and user friendly piece of software that will allow cardiovascular and other medical researchers to more easily construct and simulate decision modelswhich will positively impact health outcomes