Stochastic modeling and decision making in two healthcare applications: inpatient flow management and influenza pandemics

摘要: Delivering health care services in an efficient and effective way has become a great challenge for many countries due to the aging population worldwide, rising expenses, increasingly complex healthcare delivery systems. It is widely recognized that models analytical tools can aid decision-making at various levels of process, especially when decisions have be made under uncertainty. This thesis employs stochastic improve uncertainty two specific settings: inpatient flow management infectious disease modeling. In Part I this thesis, we study patient from emergency department (ED) hospital wards. line research aims develop insights into reduce waiting time admission wards ED. Delayed wards, also known as ED boarding, been identified key contributor overcrowding big hospitals. consists three main chapters. In Chapter 2 present extensive empirical our collaborating hospital. Motivated by study, 3 high fidelity processing network model capture with focus on transfer process 4 devise new framework, two-time-scale analysis, predict time-dependent performance measures some simplified versions proposed model. We explore both exact Markov chain analysis diffusion approximations. Part makes contributions dimensions. First, identify several novel features need built With these features, able dynamics hourly resolution reproduce measures, whereas traditional time-varying queueing fail do so. These include unconventional non-i.i.d. (independently identically distributed) service times, overflow mechanism, allocation delays. Second, framework overcomes number challenges faced existing methods analyzing features. arrivals extremely long times. Third, developed generates set useful managerial insights, which allow managers (i) strategies (ii) evaluate trade-off between benefit reducing congestion cost implementing certain policies. particular, early discharge policies eliminate excessively times patients requesting beds morning. In II spread influenza pandemics identifying factors may lead multiple waves outbreak. of…