Forecasting Hospital Daily Occupancy Using Patient Journey Data - A Heuristic Approach

  • Shaowen QinEmail author
  • Dale Ward
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11323)


Hospitals are dynamic environments that involve many stochastic processes. Each day, some patients are discharged from hospital, emergency patients arrive and require admission, and a variable number of elective admissions are planned for the day. The ability to forecast hospital occupancy will assist hospital managers to balance the supply and demand on inpatient beds on a daily basis, which in turn will reduce the risk of hospital congestion. This study employed a heuristic approach to derive a forecasting model based on hospital patient journey data. Instead of using estimated overall length of stay (LOS) for each patient, the forecasting model relies on daily evaluation of the probabilities of staying or being discharged based on a patient’s current LOS. Patients’ characteristics are introduced as additional model parameters in an incremental manner to balance model complexity and prediction accuracy. It was found that a model without enough details can provide indications of overall trends in terms of the mean occupancy. However, more parameters, such as day of the week, must be considered in order to capture the extremes present in the data. Of course, as more parameters are introduced, less data become available for meaningful analysis. This proof-of-concept study provides a demonstration of a heuristic approach to determine how complex a model needs to be and what factors are important when forecasting hospital occupancy.


Forecast model Hospital occupancy Probability of discharge Patient journey 


  1. 1.
    Van Walraven, C., Forster, A.: The TEND (Tomorrow’s Expected Number of Discharges) model accurately predicted the number of patients who were discharged from the hospital the next day. J. Hosp. Med. 13(3), 158–163 (2017)Google Scholar
  2. 2.
    Fuhs, P., Martin, J., Hancock, W.: The use of length of stay distributions to predict hospital discharges. Med. Care 17(4), 355–368 (1979)CrossRefGoogle Scholar
  3. 3.
    Azari, A., Janeja, V.P., Mohseni, A.: Healthcare data mining: predicting hospital length of stay (PHLOS). Int. J. Knowl. Discov. Bioinform. 3(3), 44–66 (2012)CrossRefGoogle Scholar
  4. 4.
    Xu, H., Wu, W., Nemati, S., Zha, H.: Patient flow prediction via discriminative learning of mutually-correcting processes. IEEE Trans. Knowl. Data Eng. 29(1), 157–171 (2017)CrossRefGoogle Scholar
  5. 5.
    Barado, J., Guergué, J., Esparza, L., Azcárate, C., Mallor, F., Ochoa, S.: A mathematical model for simulating daily bed occupancy in an intensive care unit. Crit. Care Med. 40(4), 1098–1104 (2012)CrossRefGoogle Scholar
  6. 6.
    Ben-Tovim, D., Filar, J., Hakendorf, P., Qin, S., Thompson, C., Ward, D.: Hospital event simulation model: arrivals to discharge - design, development and application. Simul. Model. Pract. Theory 68, 80–94 (2016)CrossRefGoogle Scholar
  7. 7.
    Mackay, M., Lee, M.: Choice of models for the analysis and forecasting of hospital beds. Health Care Manag. Sci. 8(3), 221–230 (2005)CrossRefGoogle Scholar
  8. 8.
    Hastie, T., Tibshirani, R., Friedman, J.: The Elements of Statistical Learning - Data Mining, Inference and Prediction, 2nd edn. Springer, New York (2009). Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  1. 1.College of Science and EngineeringFlinders UniversityTonsleyAustralia

Personalised recommendations