140 likes | 243 Views
Evaluation of Healthcare Coverage Efficiency Using Agent-Based Simulation . Joe Dinius ECE508 15 Oct 2009. Agenda. Introduction Review of the Literature Hypotheses Model Description Parameters Results Sensitivity Analysis Conclusions. Introduction.
E N D
Evaluation of Healthcare Coverage Efficiency Using Agent-Based Simulation Joe Dinius ECE508 15 Oct 2009
Agenda • Introduction • Review of the Literature • Hypotheses • Model Description • Parameters • Results • Sensitivity Analysis • Conclusions
Introduction • Healthcare is a pressing issue in American society • Economic Recovery • Public Welfare • Current public health data is insufficient to evaluate impact to public health impact due to uninsured • Agent-based simulation provides flexible framework for evaluating impact of uninsured to overall public health
Introduction • Parameters of interest will be means and standard deviations of • Cost • Resource Utilization (Hospitalization) • Time for epidemic to end • Agents are simple • Interaction rule is nearest-neighbor disease propagation • Hypothetical disease presented requiring hospitalization (resource utilization)
Review of the Literature • HUS08 publishes average data • Percentage of Americans who are uninsured • Cost of private vs. public insurance • Average results are misleading as not every condition requires treatment • Need data from catastrophic life events requiring hospitalization • Provides comparison between privately insured and uninsured • Other studies completed comparing FFS vs. managed care case studies for pregnant women in California
Hypotheses • Lower percentage of uninsured agents should lead to lower epidemic time • Should be difference in cost structure as number of uninsured agents increases • Non-profit insurance should ensure better care at less net cost
Model Description • One agent is infected at initialization • Random draw for susceptibility of nearest-neighbors is performed and agents are infected accordingly • Agents are hospitalized one day after being infected and social network is broken • Agents are hospitalized until treatment time ends • If released before fully treated, time to be cured of the disease increases by a scale factor
Model Description • After infection occurs, agents’ susceptibility goes to 0 • Simulation runs until number of infected agents is 0
Parameters • Insurance status • Cost • Treatment time • Susceptibility • Cure time • Scale factor for cure time
Results • Refer to paper for statistical tables of output
Conclusions • Less statistically significant differences than expected • Cost • Resource Utilization • Epidemic Duration • Metrics focused on were averaging • Less impact on results from transients • Hypothetical disease model suggests macroscopic view of contemporary healthcare problem is incomplete