| With the acceleration of population aging,demand on medical service is steadily increasing.Worse still,the uneven distribution of medical resources guide patients flowing into big hospitals in large cities for better services.As societal development advances,all of these factors contribute to difficulties in accessibility and costliness in affordability of medical services.A series of medical reforms have been launched to develop the current health care system,aiming to alleviate patients' financial pressure and at the meantime,to improve the working conditions of medical staff and the accessibility of medical services.Even though more and more government budgets are invested in this field,it is still hard to cover the huge demand.Meanwhile many common diseases are hard to be cured with only advanced facilities and technologies,some other rare resources(e.g.,organs)are needed.Thus,how to optimize the distribution of the input of budgets and rare resources is raised into agenda,the whole society has devoted to improve the utilization of various medical resources to ensure timely,accessible and fair public health service to all citizens.Given the overall context,we aim to address on health care system from both macro level and micro level.This work is devoted to optimize the flow system and related alongside bottleneck resources,aiming to improve the utilization of bottleneck resources and to provide medical services of high quality.Due to lack of accurate information,uncertainties at decision point result in poor allocation of medical resources.Sources of uncertainties can be concluded in these four aspects: 1)resources are not well-prepared for complex diseases;2)patients are affected by infectious diseases and can not accept the treatment as planned;3)physicians are pulled away by other overcrowded departments and clinics and can not timely show up for scheduled tasks;4)team work and physicians' experiences would affect service running time.Our work chooses to analyse the bottleneck resources along the surgical path,which appears to be the information center and gathers various kinds of key resources in hospital.Firstly we will concentrate on a special surgery: living kidney transplantation,we will analyse the proper allocation rule to improve the QALYs of patients in the kidney exchange pool,then we will optimize the scheduling of the general surgeries which share the same operating room:1)We will first analyse the bottleneck of living donor transplantation.Because it has a high priority compared to other general surgeries,operating room is always ready for prepared patient and kidney.Thus we focus on the allocation of kidneys among different patient-donor pairs to improve the QALYs of all patients in the exchange pool.Compatible donors are selected based on the health status of both patient and donor,their blood type,HLA and PRA.The donor of living renal transplantation generally comes from patients' relatives and friends.If the donor and his intended patient are matched in blood type,HLA and PRA,a transplantation can be implemented directly,otherwise patient-donor can be paired to join in a kidney exchange program.In a kidney exchange poor,incompatible pairs might get compatible kidneys by exchanging with other incompatible pairs.The first research topic is to study various dynamic allocation policies with regard to kidney exchange program.(2)Uncertainties during operations are among the most important factors to influence the utilization of surgical resources.These uncertainties are from:the complexity of surgery itself,cooperative work of the surgical team,physicians' experiences,patients' health status changes and other unexpected incidents.These uncertainties may lead to great fluctuations on surgical time,and further result in longer waiting time for patients.Currently,stochastic programming is used quite often to fight against these uncertainties and provide advanced static decisions.On the basis of their work,we propose to keep track of the real-time status and adapt the surgical scheduling accordingly.Specifically,we propose an adaptive scheduling solved by dynamic programming and make decisions at the accomplishment of each surgery.(3)Resource scarcity is another important factor to influence the utilization of surgical resources and service quality.Last-minute cancellations due to huge overrun of OR(operating room)are quite popular in many countries,especially developing countries.To avoid waste of bottleneck resources,physicians are always overloaded,they are scheduled to work at clinics and in OR as well.Any delay in one of these procedures would interrupt the original plan for OR,making resources idle and overwork.In our research,stochastic properties of both surgical times and unavailabilities of surgical resources are taken into account.We propose a two shift scheme: morning shift and afternoon shift.After finishing morning shift,if the OR is overrun too much,hospital is allowed to cancel surgeries in the form of rescheduling and transferring to other departments,aiming to keep the cancellation cost and overtime cost under control.The main innovative thoughts and contributions are reflected in three aspects:(1)Our work is the first to establish a Markov chain to calculate the health related values for patient in each state.Quality of adjusted life years(QALYs)is used to measure the health related values.We propose a new allocation policy to maximize patients' QALYs and compare it to existing policies.After collecting data from the kidney exchange pool of the Netherlands,Markov chain model and monte carlo simulation are used together to calculate the QALYs for each policy.Numerical results show that a prominent improvement in QALYs of 6 % over current practice is possible by our newly proposed policy,meanwhile equality among different categories of patients are not deteriorated.The average QALYs of unmatched patients has improved around 31%.The percentage of gap between Max Qaly policy and a theoretical upper bound is narrowed by around 32%.(2)We propose an adaptive scheduling to keep track of the real-time status of OR,the main purpose is to overcome uncertainties during operation and reduce patients' waiting time.To be more realistic,committings are introduced to reserve longer time for preoperative preparations.Theoretically,pseudo polynomial time algorithm is proposed to solve static knapsack problem,furthermore,heuristic adaptive policy is designed on this basis.Numerical results guarantee the accuracy and efficiency of the heuristic policy.An case study shows that adaptive policy improve the revenue by6% compared to practice.This research also offers a simple and executable suggestion: surgery can be sequenced according to the inverse order of standard deviation.(3)To avoid the last-minute cancellations caused by uncertainties of surgical procedures and shortages of surgical resources,meanwhile to keep the consistence of same-day schedule for single OR,we assume that two shifts are adopted in the hospital and cancellations are allowed after the first shift.This problem is build into a three stage stochastic program and solved by sample average approximation(SAA),L-shaped method and submodular properties.The accuracy of the modified L-shaped method is around 1%.Numerical results show that the policy of Exogenous and Scheduled Cancellations(ESC)can balance the overtime cost and cancellation cost,it reduces the overall cost by 11% at most.When the time resource unavailability goes longer,the probability of exogenous cancellation goes higher,ESC policy performs much better than ECO policy.In summary,dynamic decision-making process would keep track of the hospital system in real-time and better handle all kinds of uncertainties from surgical procedures.It would improve the resource utilization,service quality and physicians' working conditions.Highly developed information technology makes realtime monitoring and scheduling possible,our work is quite meaningful for the hospital operation. |
PDF Full Text Request