University of Twente Student Theses

Login

Decision support for design, implementation, and feasibility of an admission lounge

Veneklaas, W. (2019) Decision support for design, implementation, and feasibility of an admission lounge.

[img] PDF
2MB
Abstract:Background: Dutch hospitals are transforming the elective patient admissions process with a new type of ward: the Admission Lounge (AL). The AL facilitates elective patient admissions for relatively low-complex high-volume patient populations. With the AL, hospitals reduce the number of patient admissions that take place at the Clinical Ward (CW). Thereby improving the efficacy and efficiency of the perioperative process. In a preliminary field study we found that while many hospitals in the Netherlands have established or are establishing the AL, there is no systematic approach for its implementation. This raises the opportunity to improve the establishment of an AL through systematic strategic decision support. Potentially, this could lead to improved operational performances, better patient satisfaction, cost savings, and better care. As the Dutch market leader in the field of hospital information systems, ChipSoft sees opportunities to implement methods from the field of operations research and management sciences into their hospital information system HiX. Part of their mission is to increase the efficiency and efficacy of hospitals. ChipSoft facilitates this research project to support hospital managers in the decision making process involved with the implementation process of an AL. Goal and method: The goal of this research is to facilitate systematic decision making regarding design, implementation and feasibility of an AL. We aim to quantify those decisions and give insights into the relations between the patient selection criteria, the potential bed reduction for the CW, and the required number of AL beds. We use the taxonomic classification of planning decisions of Hulshof et al. (2012) to propose a five phase stepwise approach for strategic decision making. Using data visualisation we demonstrate the effects of inclusion and exclusion criteria for the AL on the volume and complexity of the AL’s patient population. On the basis of the attributes priority, age, ASA classification, and specialty, patients are assigned to the AL or CW. Patients fall within a grey area when they could be assigned to both. With literature research, we find the Erlang loss model (see e.g., De Bruin et al. (2010)) as the most suitable model to determine the potential bed reduction for the CW. Inputs for the bed reduction are the patient assignments to the AL or CW, and a blocking probability of 5% for the CW. The potential bed reduction for the CW is determined for a risk pooling strategy. We assess the AL bed requirements and performance for a set service level, using our own deterministic model. By enumeration of the possible patient assignment rules for patients assigned to the grey area, we give suggestions for improved efficacy and efficiency of the AL and CW. Our proposed five phase stepwise approach and the described algorithms are integrated into a decision support system (DSS) and we provide a mock-up design of the DSS in the HiX environment. For validation purposes, we perform two case studies: the first to verify and validate our algorithms, and the second to validate interpretability and use of the DSS at the strategic level. Results: The proposed stepwise approach consists of the following phases: 1. Set up inclusion and exclusion criteria for the AL 2. Determine appropriate staff, equipment, and supporting processes for the AL and CW 3. Analysis of potential bed reductions for the CW and required capacity for the AL 4. Analysis of feasibility within the facility layout 5. Optimisation: assignment of the grey area patients to the AL or CW In the first case study we provide the inputs for Phase 1. In the second case study, the case hospital’s representatives provide the input. The visualisation of the inclusion and exclusion criteria for the AL provides insights about the impact of the criteria for each attribute on the distribution of AL, CW, and grey area patients. In the first case study, the patient population contains 19% AL patients, and 12% grey area patients. In the second case study, the population contains 44% AL patients, and 35% grey area patients. During Phase 2, hospital management derives the appropriate staff, equipment, and supporting process in correspondence with the complexity profiles of AL and CW patients. In Phase 3 of the first case study, the CW can potentially reduce its capacity by 4 beds while the AL requires 2 required beds, resulting in an overall reduction of 2 beds. For the second case study, the CW bed reduction amounts 2 while the AL requires 3 beds, resulting in an overall increase of 1 bed. Hospital management assesses the feasibility of the AL requirements in Phase 4. During Phase 5, we enumerate the assignment of patients of 1 specialty, 1 ASA class, and 3 age ranges from the grey area to the AL. In both case studies, the outcomes of the enumeration indicate potential for improved performance of the AL and CW by assigning patients from the grey area to the AL. For the first case study, we find a solution that reduces the CW capacity by 5 beds, while the AL requires 3 beds. For the second case study, we find a solution that reduces the CW’s capacity by 3 beds while the AL requires 4 beds. In both case studies, the AL gets assigned a bigger patient population. This indicates a bigger effect on the hospital’s efficiency than the solution before enumeration. The runtime of the DSS, performing the enumeration for 12 assignment combinations, is shorter than one minute. The results of the first case study are presented in a mock-up design in the ChipSoft’s HiX environment. Conclusions and recommendations The DSS facilitates systematic strategic decision making by following the developed five phase stepwise approach. The visualisation of the results successfully provides insights into the relations between patient selection and capacity requirements for both the AL and CW. The optimisation method in our DSS enumerates AL assignment rules successfully and is able to indicate a solution that is effective and efficient for both the AL and the CW. Insights generated by the DSS are well interpretable and useful, according to hospital management. Moreover, ChipSoft sees potential in the developed tool. In our case studies we consider one specialty for assignment to the AL. Consideration of more (sub)specialties or a variety of other patient selection criteria for the AL can exponentially increase the solution space and runtime for the enumeration method. However, the current short runtime allows enumerating relatively large solution spaces. This indicates potential for application of our algorithms to bigger instances. We recommend ChipSoft to relate the DSS to the tools that are currently developed for forecasting the outflow of the operating room to the clinical wards and other hospital units, as a result of the master surgery schedule. Our model is capable of determining the expected inflow of the AL with a relatively simple method. This method could be refined with more accurate forecasts that are related to the outflow of the OR as a results of the MSS. Another recommendation is to include a field within the preoperative screening form of the anaesthesiologist which automatically indicates whether a patient is suitable for AL admission in compliance with the patient’s characteristics. For further research, we recommend to explain the variability of the load on the AL and CW using discrete event simulation. We also see potential in applying a time-dependent Erlang loss model to determine the potential bed reduction at the CW. For AL patients, arrival at the CW is prolonged, meaning that the peak load is potentially reduced. The time-dependent loss model allows to incorporate peak and off peak arrival rates for the CW and thereby the effect of the AL on the CW can be explained further.
Item Type:Essay (Master)
Clients:
ChipSoft, Amsterdam, Nederland
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:50 technical science in general, 58 process technology, 85 business administration, organizational science
Programme:Industrial Engineering and Management MSc (60029)
Link to this item:https://purl.utwente.nl/essays/79023
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page