Operations Research for Health Care最新文献

We present a new scheduling approach to improve access to care at an inner-city community health centre in Vancouver, Canada, serving marginalised clients with complex biopsychosocial needs. In order to meet the specific care needs of clients, the centre provides a range of services on a booked and walk-in basis, and it is important that clients are seen in a timely manner. To align schedules with client demand, we developed a schedule optimisation model that maximises time nurses spend with clients. This new objective function allows for a simple mixed integer linear programming structure that directly incorporates carryover demand. Client-centred key performance indicators were evaluated using a discrete event simulation model. Optimisation aligns schedules to demand, leading to fewer clients who leave without being seen due to an extended wait. This increases the number of clients receiving care by up to 9 per week, without compromising wait times. Furthermore, our approach addresses service delivery concerns, including baseline nurse coverage for triage and weekly variability in total nurse hours. Strategically aligning nurse shifts to demand is an effective approach to better meet client needs without increasing total nurse staffing levels in a community health centre context.

Background:

The effect of policy and managerial decisions on nurse-workload, and subsequent quality-of-care, are difficult to quantify in advance. A tool is needed that can proactively test these changes — Discrete Event Simulation (DES) may help. While computerized simulation models have existed before, there remains a gap to affirm the validity of these models.

Objective:

Develop an approach to creating a valid computerized simulation model that quantifies the effects of operational decisions on nurse-workload and quality-of-care.

Methods:

The DES model simulates the process of care delivery for nurses on a task-by-task basis. In an effort to validate this approach, the DES model was adapted to a real-world medical-surgical unit. Model inputs include: historical patient-care data; unit-layout; and programming logic, developed via focus-groups. Nurse-workload outcomes were distance-walked, task-in-queue, direct-care time, and nurse-movement. Quality-of-care outcomes included missed-care; and care-task waiting-time. The model is validated via internal validity checks and a field study that consisted of a ‘step-counter study’, a ‘MISSCARE survey’, ‘nurse job shadowing’, and a ‘time and motion study’. An Intraclass-correlation (ICC) and Spearman ranked correlation analysis were used to compare modelling outcomes to field-study outcomes.

Results:

The DES model, when adapted to a real-world medical-surgical unit, has been validated. The ICC coefficients show an “excellent” agreement of 0.99, 0.99, 0.85, 0.85, 0.84 between simulation and real-world outcomes, along with a “good” agreement of 0.86 for Spearman ranked correlation. Specific modelling results include a ‘distance walked’ of 7 to 10.6 km with a ‘direct care time’ of 8.3 to 10.4 h with a total of 77 to 84 trips for an average of 12 to 15 ‘tasks in queue’. Quality-of-care was represented by a ‘care task waiting time’ of 0.9 to 1 h that lead to 25 to 31 ‘missed-care’ tasks, where, 27% were ‘non-patient care’; and ‘missed-care delivery time’ was 2 to 2.9 h.

Conclusion:

This research provides a decision support-system that can help test and inform healthcare system policies that support both care quality and safety. By validating the DES model of a medical-surgical unit, we suggest that the modelling approach will also yield valid result when applied in similar settings. However, the modelling approach needs to be adapted to other healthcare settings and tested before concluding that this approach will consistently yield valid models.

The patient transfer, as a common seen and necessary healthcare procedure, plays an important role in maintaining efficient treatment and improving the quality of healthcare. Among various factors impacting and indicating the safety and efficiency of the patient transfer, the length-of-stay (LOS), which is not often studied in this field, is worth investigating. Phase-type (PH) distributions, as one of the popular methods of modeling LOS, will be integrated in an aggregated Markov chain to construct a model to describe the sequences of LOS in hospital units. In this paper, we model the intra-hospital transfer flow routes by fitting aggregated PH distribution and using Maximum Likelihood Estimation to estimate the parameters. Following the results of distribution fitting, the patients can be divided into different groups according to their LOS in the same unit. By analyzing each group to find out its common characteristics, intra-hospital transfer routes, admission and discharge situations, the associations among significant factors, the LOS and the treatment efficiency are evaluated.

Blood transfusion is one of the most crucial and commonly administered therapeutics worldwide. The need for more accurate and efficient ways to manage blood demand and supply is an increasing concern. Building a technology-based, robust blood demand and supply chain that can achieve the goals of reducing ordering frequency, inventory level, wastage and shortage, while maintaining the safety of blood usage, is essential in modern healthcare systems. In this study, we summarize the key challenges in current demand and supply management for red blood cells (RBCs). We combine ideas from statistical time series modeling, machine learning, and operations research in developing an ordering decision strategy for RBCs, through integrating a hybrid demand forecasting model using clinical predictors and a data-driven multi-period inventory problem considering inventory and reorder constraints. We have applied the integrated ordering strategy to the blood inventory management system in Hamilton, Ontario using a large clinical database from 2008 to 2018. The proposed hybrid demand forecasting model provides robust and accurate predictions, and identifies important clinical predictors for short-term RBC demand forecasting. Compared with the actual historical data, our integrated ordering strategy reduces the inventory level by 40% and decreases the ordering frequency by 60%, with low incidence of shortages and wastage due to expiration. If implemented successfully, our proposed strategy can achieve significant cost savings for healthcare systems and blood suppliers. The proposed ordering strategy is generalizable to other blood products or even other perishable products.

Many western countries undergo substantial demographic changes at present. This is particularly challenging for the health care industry since resources have to be set up and arranged well in advance to be able to cover future patient demand. The objective of this article is to present a method for forecasting regional demand for hospital services. The problem of forecasting regional patient volumes is based on three components. First, population forecasts provided by local authorities serve as a basis for the projections. Second, future per-capita demand is forecasted to account for sociological and medical trends. Forecasting methods in this step include autoregressive integrated moving average models, exponential smoothing models, neural nets, and regression models. Third, patient volumes are anticipated merging the projections of the population and per-capita demand for the respective age and sex groups. The proposed method is applied to publicly available data concerning discharges from German hospitals over 18 years. Results indicate that considering the age structure of the population in the catchment area of the hospital and taking into account trends of significantly changing per-capita demand are crucial for accurate forecasts.

Urged by the outbreak of the COVID-19 in Italy, this study aims at helping to tackle the spread of the disease by resorting to operations research techniques. In particular, we propose a mathematical program to model the problem of establishing how many diagnostic tests the Italian regions must perform in order to maximize the overall disease detection capability. An important feature of our approach is its simplicity: data we resort to are easy to obtain and one can employ standard optimization tools to address the problem. The results we obtain when applying our method to the Italian case seem promising.

The primary goal of emergency medical service (EMS) agencies is to effectively allocate the ambulances and personnel required to provide sufficient geographic coverage of a service area while minimizing response times to high-priority call requests. Given that the demand for ambulances is known to fluctuate spatially and temporally based on the time of day and day of the week, EMS practitioners depend on call volume forecasts to develop staffing and dynamic redeployment plans. In this study, a series of daily, hourly, and spatially distributed hourly call volume predictions are generated using a multi-layer perceptron (MLP) artificial neural network model following feature selection using an ensemble-based decision tree model. For spatially distributed predictions, K-Means clustering is applied to produce heterogeneous spatial clusters based on call location and associated call volume densities. The predictive performance of the MLP model is benchmarked against both a selection of traditional time-series forecasting techniques and a common industry method. Results show that MLP models outperform time-series and industry forecasting methods, specifically at finer levels of spatial granularity where the need for more accurate call volumes forecasts is more essential.

We consider the rostering problem for surgeons in residency at the Clinic of Surgery at St. Olav’s Hospital, Trondheim University Hospital, in Trondheim, Norway. Each surgeon in residency has a rank depending on experience. An exact number of surgeons of each rank must work emergency shifts in a cyclic structure. Each surgeon is affiliated to a section, which has a minimum staffing level. Section shifts can be planned in an acyclic structure, thus establishing a semi-cyclic structure in the full roster. The addition of more typical rostering constraints establishes the novel Semi-Cyclic Ranked Physician Rostering Problem. In manually created rosters, the staffing at sections varies greatly, leading to frequent understaffing. With the addition of absence among staff when rosters are executed, this is problematic for the Clinic of Surgery. We present a two-step matheuristic based on mixed integer linear programming to solve the problem for five real-life instances. Comparing our results with a manually created roster demonstrates superior results in terms of staff availability at sections, greatly improving roster resilience to absence. We also introduce shadow shifts designed to increase the flexibility of rosters to cover for absence at emergency night shifts.

Addressing artificial variability in patient flow is an effective approach to improving accessibility and quality of care and reducing waste and cost in healthcare systems. The most significant artificial variability in patient flow is attributable to dysfunctionally scheduled admissions that can be decreased or eliminated by load smoothing. In this study, we examine the impact of load smoothing of scheduled admissions on patient flow performance metrics of an obstetric unit to provide insights for capacity management. We also investigate the relationship between the impact of load smoothing of scheduled admissions on the patient flow performance metrics of the unit and the volume of unscheduled admissions. In doing so, we develop a detailed discrete-event simulation model of the patient flow of the unit in which patients are categorized into different classes. We compare the results of the simulation model before and after implementing different degrees of load smoothing by considering various ratios of average weekend daily load to average weekday daily load. We determine how different degrees of load smoothing reduce the number of beds required, the average waiting time, and the average probability of delay differently while they have no considerable impact on the average bed occupancy rate. Moreover, considering different volumes of unscheduled admissions, we quantify how the reduction in the average waiting time and the average probability of delay by load smoothing increases when the ratio of unscheduled admissions to scheduled admissions decreases.

Providing timely access to costly surgical services in a manner that balances needs of multiple stakeholders (patients, staff, administrators) is made even more challenging by inherent uncertainty. Decisions about clinician scheduling, shift preferences, operating room planning, and patient assignment also often are decentralized or made separately. We develop a robust optimization model that combines staffing and scheduling decisions to minimize the impact of foreseeable variation in surgery durations, staff availability, and urgent or emergency arrivals. Model performance is tested with data from a major academic medical center, resulting in improved service level (% patients served), overtime, utilization, and shift preferences. Although robustness to staffing, duration, and urgent or emergency uncertainty increases operating costs by 6% on average, overtime is reduced by 68% while utilization decreases by only 6%. The number of necessary schedule adjustments on the day of surgery also is reduced by 13% on average in the robust model compared to the nominal model.