Predicting patient arrivals to an accident and emergency department

Au-Yeung, Susanna W. M.; Harder, Uli; McCoy, E J; Knottenbelt, William J.

doi:10.1136/emj.2007.051656

Cited by 32 publications

(24 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A number of authors have therefore fitted and tested nonhomogeneous Poisson models to patient arrival processes for example, (1) Barnett et al [2002] for intensive care admissions and daily variations in the Poisson rate; (2) Shanmugam et al [2007] have used Poisson regression to fit a nonhomogeneous Poisson process to stroke incidence data; (3) Harrison et al [2005] for intensive care admissions and daily variations in the Poisson rate; and (4) Alexopoulos et al [2008] for patient arrivals at community clinics. Also, a number of authors distinguish between emergency (unscheduled) and scheduled arrivals for example, Alexopoulos et al [2008] and Au-Yeung et al [2009]. In the latter case, it is unlikely that arrivals are random while unscheduled arrivals may be subject to effects of grouping of arrivals, fluctuations by time of day, and dependence between arrivals in different periods of the day [Alexopoulos et al 2008].…”

Section: Methodsmentioning

confidence: 99%

A modeling framework that combines markov models and discrete-event simulation for stroke patient care

McClean

Barton

Garg

et al. 2011

ACM Trans. Model. Comput. Simul.

View full text Add to dashboard Cite

Stroke disease places a heavy burden on society, incurring long periods of hospital and community care. Also stroke is a highly complex disease with heterogeneous outcomes and multiple strategies for therapy and care. In this article we develop a modeling framework that clusters patients with respect to their length of stay (LOS); phase-type models are then used to describe patient flows for each cluster. In most cases, there are multiple outcomes, such as discharge to normal residence, nursing home, or death. We therefore derive a novel analytical model for the distribution of LOS in such situations. A model of the whole care system is developed, based on Poisson admissions to hospital, and results obtained for expected numbers in different states of the system at any time. We can thus describe the whole integrated system of stroke patient care, which will facilitate planning of services. We also use the basic model to build a discrete-event simulation, which incorporates back-up queues to model delayed discharge. Based on stroke patients' data from the Belfast City Hospital, various scenarios are explored with a focus on the potential efficiency gains if LOS, prior to discharge to a private nursing home, can be reduced. Predictions for bed occupancy are also provided. The overall modeling framework characterizes the behavior of stroke patient populations, with a focus on integrated system-wide planning, encompassing hospital and community services. Within this general framework we can develop either analytic or simulation models that take account of patient heterogeneity and multiple care options. ACM Reference Format:McClean, S., Barton, M., Garg, L., Fullerton, K. 2011. A modeling framework that combines Markov models and discrete-event simulation for stroke patient care. ACM Trans. Model. Comput.

show abstract

Section: Methodsmentioning

confidence: 99%

A modeling framework that combines markov models and discrete-event simulation for stroke patient care

McClean

Barton

Garg

et al. 2011

ACM Trans. Model. Comput. Simul.

View full text Add to dashboard Cite

show abstract

“…A key concern is the issue of validation of these models. Some known systems, with many users, have been modeled, strategies have been adopted that are not supported by smart environments (see, for example [1]) and their results have been compared with real observations. A future issue is to develop these empirical analyses in relation to the proposed smart environments.…”

Section: Discussion and Future Workmentioning

confidence: 99%

Engineering crowd interaction within smart environments

Harrison

Massink²,

Latella³

2009

Proceedings of the 1st ACM SIGCHI Symposium on Engineering Interactive Computing Systems

View full text Add to dashboard Cite

Smart environments (e.g., airports, hospitals, stadiums, and other physical spaces using ubiquitous computing to empower many mobile people) provide novel challenges for usability engineers. Firstly, interaction can be implicit and therefore unintentional on the part of its users. Secondly, the impact of a smart environment can influence the collective or crowd behavior of those immersed within it. These challenges lead to requirements for complementary analyses which must be combined with the more typical focus on the interaction between user and device. The paper explores a family of stochastic models aimed at analyzing these features with a particular focus on crowd interaction.

show abstract

“…Variable selection was consistent with previous research. 18,21,22,[26][27][28]30,31 Some authors found evidence of the effects of weatherrelated factors for specific pathologic findings [32][33][34][35][36] ; however, these cannot be obtained with sufficient accuracy in advance and most often have little effect on total visits. 18,21 The dependent variable was the total number of ED visits, which was then transformed into staffing requirements.…”

Section: Variablesmentioning

confidence: 99%

“…Also, most often, monthly 25 or daily visits were predicted. 21,22,[26][27][28][29] Because nursing personnel is usually organized in 8-hour shifts, daily predictions are of little use for realistic staff planning. Patient arrival volume often varies dramatically between these shifts, as fewer patients seek ED care at night or during morning working hours than during the afternoon shift.…”

mentioning

confidence: 99%

Emergency Department Visit Forecasting and Dynamic Nursing Staff Allocation Using Machine Learning Techniques With Readily Available Open-Source Software

Zlotnik

Gallardo-Antolín

Alfaro

et al. 2015

CIN: Computers, Informatics, Nursing

View full text Add to dashboard Cite

Although emergency department visit forecasting can be of use for nurse staff planning, previous research has focused on models that lacked sufficient resolution and realistic error metrics for these predictions to be applied in practice. Using data from a 1100-bed specialized care hospital with 553,000 patients assigned to its healthcare area, forecasts with different prediction horizons, from 2 to 24 weeks ahead, with an 8-hour granularity, using support vector regression, M5P, and stratified average time-series models were generated with an open-source software package. As overstaffing and understaffing errors have different implications, error metrics and potential personnel monetary savings were calculated with a custom validation scheme, which simulated subsequent generation of predictions during a 4-year period. Results were then compared with a generalized estimating equation regression. Support vector regression and M5P models were found to be superior to the stratified average model with a 95% confidence interval. Our findings suggest that medium and severe understaffing situations could be reduced in more than an order of magnitude and average yearly savings of up to €683,500 could be achieved if dynamic nursing staff allocation was performed with support vector regression instead of the static staffing levels currently in use.

show abstract

Predicting patient arrivals to an accident and emergency department

Cited by 32 publications

References 15 publications

A modeling framework that combines markov models and discrete-event simulation for stroke patient care

A modeling framework that combines markov models and discrete-event simulation for stroke patient care

Engineering crowd interaction within smart environments

Emergency Department Visit Forecasting and Dynamic Nursing Staff Allocation Using Machine Learning Techniques With Readily Available Open-Source Software

Contact Info

Product

Resources

About