The self-regulating nature of occupancy in ICUs: stochastic homoeostasis


As pressure on the health system grows, intensive care units (ICUs) are increasingly operating close to their capacity. This has led a number of authors to describe a link between admission and discharge behaviours, labelled variously as: ‘bumping’, ‘demand-driven discharge’, ‘premature discharge’ etc. These labels all describe the situation that arises when a patient is discharged to make room for the more acute arriving patient. This link between the admission and discharge behaviours, and other potential occupancy-management behaviours, can create a correlation between the arrival process and LOS distribution. In this paper, we demonstrate the considerable problems that this correlation structure can cause capacity models built on queueing theory, including discrete event simulation (DES) models; and provide a simple and robust solution to this modelling problem. This paper provides an indication of the scope of this problem, by showing that this correlation structure is present in most of the 37 ICUs in Australia. An indication of the size of the problem is provided using one ICU in Australia. By incorrectly assuming that the arrival process and LOS distribution are independent (i.e. that the correlation structure does not exist) for an occupancy DES model, we show that the crucial turn-away rates are markedly inaccurate, whilst the mean occupancy remains unaffected. For the scenarios tested, the turn-away rates were over-estimated by up to 46 days per year. Finally, we present simple and robust methods to: test for this correlation, and account for this correlation structure when simulating the occupancy of an ICU.

This is a preview of subscription content, log in to check access.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7


  1. 1.

    Kendall’s notation is used to characterise queueing systems. It is given by A/S/c/K/N/D, where A describes the inter-arrival process, S indicates the distribution of service time, c indicates the number of servers, K indicates the system capacity, N indicates the size of the customer population, and D indicates the queue’s discipline.

  2. 2.

    Also called critical care units in the literature.

  3. 3.

    APAHCE stands for Acute Physiology and Chronic Health Evaluation. It is a severity-of-disease classification system, and this is the third iteration of this system

  4. 4.

    A&E: accident and emergency

  5. 5.

    OT: operating theatre

  6. 6.

    Note the shift pattern in the study ICU is: Day shift - 7:00a.m. to 3:00p.m., Afternoon shift - 3:00p.m. to 11:00p.m., Night shift - 11:00p.m. to 7:00a.m..

  7. 7.

    Bai et al. [3] report patients grouped into 2, 3, 4, 8 or m classes, where no value for m is given. This is because a number of papers do not provide the exact number of patient classes.

  8. 8.

    We used both the graphical method and the confidence interval method described by Robinson [45] to determine that n = 100 was sufficient for our simulation. The confidence interval for the mean at n = 100 is 1%, sufficiently less than the 5% interval we were aiming for.

  9. 9.

    This was arbitrarily chosen as sufficiently large to say a correlation existed, and was not spurious. Other thresholds could have been considered.

  10. 10.

    The ‘standard’ model used is the DES model described in Section 5. That is, the method described in Section 5 without using the SD-ratio to ‘shrink’ the ICU-occupancy standard deviation.

  11. 11.

    Night time discharges were discharges from 10pm to 6:59am.

  12. 12.

    Day light hour discharges are all those which are not night time discharges.


  1. 1.

    Asaduzzaman M, Chaussalet TJ, Robertson NJ (2010) A loss network model with overflow for capacity planning of a neonatal unit. Ann Oper Res 178(1):67–76.

    Article  Google Scholar 

  2. 2.

    Bahou N, Fenwick C, Anderson G, van der Meer R, Vassalos T (2017) Modeling the critical care pathway for cardiothoracic surgery. Health Care Manag Sci:1–120.

    Article  Google Scholar 

  3. 3.

    Bai J, Fügener A, Schoenfelder J, Brunner JO (2016) Operations research in intensive care unit management: a literature review.

    Article  Google Scholar 

  4. 4.

    Baker DR, Pronovost PJ, Morlock LL, Geocadin RG, Holzmueller CG (2009) Patient flow variability and unplanned readmissions to an intensive care unit, vol 37.

    Article  Google Scholar 

  5. 5.

    Barado J, Guergué JM, Esparza L, Azcárate C, Mallor F, Ochoa S (2012) A mathematical model for simulating daily bed occupancy in an intensive care unit. Crit Care Med 40(4):1098–1104.

    Article  Google Scholar 

  6. 6.

    Barer ML, Evass RG, Hertzman C, Lomas J (1987) Aging and health care utilization: new evidence on old fallacies. Soc Sci Med 24(10):851–862.

  7. 7.

    Capuzzo M, Moreno RP, Alvisi R (2010) Admission and discharge of critically ill patients. Curr Opin Crit Care 16(5):499–504.

    Article  Google Scholar 

  8. 8.

    Chalfin DB, Trzeciak S, Likourezos A, Baumann BM, Dellinger RP (2007) Impact of delayed transfer of critically ill patients from the emergency department to the intensive care unit. Crit Care Med 35 (6):1477–1483.

    Article  Google Scholar 

  9. 9.

    Chan CW, Farias VF, Bambos N, Escobar GJ (2012) Optimizing intensive care unit discharge decisions with patient readmissions. Oper Res 60(6):1323–1341.

    Article  Google Scholar 

  10. 10.

    Cochran JK, Bharti A (2006) Stochastic bed balancing of an obstetrics hospital. Health Care Manag Sci 9(1):31–45.

    Article  Google Scholar 

  11. 11.

    Cochran JK, Roche K (2008) A queuing-based decision support methodology to estimate hospital inpatient bed demand. J Oper Res Soc 59(11):1471–1482.

    Article  Google Scholar 

  12. 12.

    Costa AX, Ridley SA, Shahani AK, Harper PR, De Senna V, Nielsen M (2003) Mathematical modelling and simulation for planning critical care capacity. Anaesthesia 58(4):320–327.

    Article  Google Scholar 

  13. 13.

    Creemers S, Lambrecht M (2008) Healthcare queueing models. Tech. rep.

  14. 14.

    Demir E, Lebcir R, Adeyemi S (2014) Modelling length of stay and patient flows: methodological case studies from the UK neonatal care services. J Oper Res Soc 65(4):532–545.

    Article  Google Scholar 

  15. 15.

    van Dijk N, Kortbeek N (2008) On dimensioning intensive care units. Springer, Berlin, pp 291–296.

    Google Scholar 

  16. 16.

    van Dijk NM, Kortbeek N (2009) Erlang loss bounds for OT–ICU systems. Queueing Systems 63(1-4):253–280.

    Article  Google Scholar 

  17. 17.

    Dobson G, Lee HH, Pinker E (2010) A model of ICU bumping. Oper Res 58(6):1564–1576.

    Article  Google Scholar 

  18. 18.

    Duraiswamy N, Welton R, Reisman A (1981) Using computer simulation to predict ICU staffing needs : journal of nursing administration. J Nurs Adm 11(1):45.

    Article  Google Scholar 

  19. 19.

    Erramilli A, Narayan O, Willinger W (1996) Experimental queueing analysis with long-range dependent packet traffic. IEEE/ACM Trans Networking 4(2):209–223.

    Article  Google Scholar 

  20. 20.

    Goldfrad C, Rowan K (2000) Consequences of discharges from intensive care at night. The Lancet 355 (9210):1138–1142.

    Article  Google Scholar 

  21. 21.

    Gray W, Wang P, Scott M (1992) An m/g/1-type queuing model with service times depending on queue length. Appl Math Model 16(12):652–658.

    Article  Google Scholar 

  22. 22.

    Green LV (2002) How many hospital beds? Inquiry: a Journal of Medical Care Organization, Provision and Financing 39(4):400–412

    Article  Google Scholar 

  23. 23.

    Griffiths JD, Price-Lloyd N, Smithies M, Williams JE (2005) Modelling the requirement for supplementary nurses in an intensive care unit. J Oper Res Soc 56(2):126–133.

    Article  Google Scholar 

  24. 24.

    Griffiths JD, Price-Lloyd N, Smithies M, Williams J (2006) A queueing model of activities in an intensive care unit. IMA J Manag Math 17(3):277–288.

    Article  Google Scholar 

  25. 25.

    Griffiths JD, Knight V, Komenda I (2012) Bed management in a Critical Care Unit. IMA J Manag Math 24(2):137–153.

    Article  Google Scholar 

  26. 26.

    Hagen M, Jopling J, Buchman TG, Lee EK (2013) Priority queuing models for hospital intensive care units and impacts to severe case patients. AMIA Ann Symp Proc 2013:841–850.

  27. 27.

    Harper PR, Shahani AK (2002) Modelling for the planning and management of bed capacities in hospitals. J Oper Res Soc 53(1):11–18.

    Article  Google Scholar 

  28. 28.

    Hohn N, Veitch D, Abry P (2002) Does fractal scaling at the IP level depend on TCP flow arrival processes? In: Proceedings of the 2nd ACM SIGCOMM…, pp 1–6.

  29. 29.

    Howard G (2005) A shocking admission: patient exclusion from intensive care therapy. Critical Care and Resuscitation : Journal of the Australasian Academy of Critical Care Medicine 7(2):97–101

    Google Scholar 

  30. 30.

    Kim SC, Horowitz I (2002) Scheduling hospital services: the eecacy of elective-surgery quotas. Omega 30:335–346.

    Article  Google Scholar 

  31. 31.

    Kim SC, Horowitz I, Young KK, Buckley TA (1999) Analysis of capacity management of the intensive care unit in a hospital. Eur J Oper Res 115(1):36–46.

    Article  Google Scholar 

  32. 32.

    Kolker A (2009) Process modeling of ICU patient flow: effect of daily load leveling of elective surgeries on ICU diversion. J Med Syst 33:27–40.

    Article  Google Scholar 

  33. 33.

    Komenda I (2013) Modelling critical care unit activities through queueing theory. PhD thesis, School of Mathematics, Cardiff University

  34. 34.

    Litvak N, van Rijsbergen M, Boucherie RJ, van Houdenhoven M (2008) Managing the overflow of intensive care patients. Eur J Oper Res 185(3):998–1010.

    Article  Google Scholar 

  35. 35.

    Lowery JC (1992) Simulation of a hospital’s surgical suite and critical care area. In: Proceedings of the 24th conference on winter simulation, ACM, Arlington, Virginia, USA, pp 1071–1078.

  36. 36.

    Lowery JC (1993) Multi-hospital validation of critical care simulation model. In: Proceedings of the 25th conference on winter simulation. ACM, Los Angeles, California, USA, pp 1207–1215.

  37. 37.

    Mallor F, Azcárate C (2014) Combining optimization with simulation to obtain credible models for intensive care units. Ann Oper Res 221(1):255–271.

    Article  Google Scholar 

  38. 38.

    Masterson BJ, Mihara TG, Miller G, Randolph SC, Forkner ME, Crouter AL (2004) Using models and data to support optimization of the military health system: a case study in an intensive care unit. Health Care Manag Sci 7(3):217–224.

    Article  Google Scholar 

  39. 39.

    McManus ML, Long MC, Cooper A, Litvak E (2004) Queuing theory accurately models the need for critical care resources. Anesthesiology 100(5):1271–1276.

    Article  Google Scholar 

  40. 40.

    Metcalfe MA, Sloggett A, McPherson K (1997) Mortality among appropriately referred patients refused admission to intensive-care units. Lancet 350(9070):7–11.

    Article  Google Scholar 

  41. 41.

    Pilcher DV, Duke GJ, George C, Bailey MJ, Hart G (2007) After-hours discharge from intensive care increases the risk of readmission and death. Anaesth Intensive Care 35(4):477–485.

    Article  Google Scholar 

  42. 42.

    Reinhardt U (2003) Does the aging of the population really drive the demand for health care? Health Aff 22(6):27–39.

    Article  Google Scholar 

  43. 43.

    Ridge J, Jones S, Nielsen M, Shahani AK (1998) Capacity planning for intensive care units. Eur J Oper Res 105(2):346–355.

    Article  Google Scholar 

  44. 44.

    Ridoux J, Nucci A, Veitch D (2006) Seeing the difference in IP traffic: wireless versus wireline. In: Proceedings - IEEE INFOCOM, pp 1–12.

  45. 45.

    Robinson S (2004) Simulation : the practice of model development and use. Wiley.

  46. 46.

    Shahani AK, Ridley SA, Nielsen MS (2008) Modelling patient flows as an aid to decision making for critical care capacities and organisation. Anaesthesia 63(10):1074–1080.

    Article  Google Scholar 

  47. 47.

    Shmueli A, Sprung CL, Kaplan EH (2003) Optimizing admissions to an intensive care unit. Health Care Manag Sci 6(3):131–136

    Article  Google Scholar 

  48. 48.

    Sinuff T, Kahnamoui K, Cook DJ, Luce JM, Levy MM (2004) Rationing critical care beds: a systematic review. Crit Care Med 32(7):1588–1597.

    Article  Google Scholar 

  49. 49.

    Troy PM, Rosenberg L (2009) Using simulation to determine the need for ICU beds for surgery patients. Surgery 146(4):608–620.

  50. 50.

    Wharton F (1996) On the risk of premature transfer from coronary care units. Pergamon Omega Int J Mgmt Sci 24 (4):413–423.

  51. 51.

    Williams SV (1983) How many intensive care beds are enough? Crit Care Med 11 (6):412–416.

    Article  Google Scholar 

  52. 52.

    Yang M, Fry MJ, Raikhelkar J, Chin C, Anyanwu A, Brand J, Scurlock C (2013) A model to create an efficient and equitable admission policy for patients arriving to the cardiothoracic ICU. Crit Care Med 41(2):414–422.

    Article  Google Scholar 

Download references


The authors would like to recognise the engagement of the ICU staff at the RAH. In particular, for providing practical insights into results, and for making the data accessible which was used to carry out this research. This publication is based on a project funded by the Premier’s International Research Fund. The international partners for the research grant are the international partners Cumberland Initiative and The AnyLogic Company. The views expressed in the paper are those of the authors and do not necessarily the views of the any other parties involved in this research.

Author information



Corresponding author

Correspondence to Josephine Varney.


Appendix A: ANZICs data

A summary of the ANZICS data is shown in Table 5.

Table 5 ANZICS data summary

Appendix B: Arrival rate and LOS analysis

Based on work in [13, 25, 33], discussions with clinicians, and our own observations we chose three likely classes. These classes are: 1. Emergency or elective patient, 2. Arrival day and shift, and 3. A patients’ APACHE III code.

B.1 Emergency and elective arrivals

Emergency and elective patients are commonly considered different classes in ICU modelling [25, 33], due to their differing arrival rates and differing LOSs. In this study ICU elective patients predominantly arrive on weekdays, during day and afternoon shifts (see Figs. 8 and 9). This is not surprising, since elective patients are scheduled in the normal business hours of the hospital. In comparison, emergency arrivals have a more even arrival pattern across the week and across shifts (see Figs. 8 and 9). This is expected, as emergency arrivals are unplanned/random arrivals.

Fig. 8

Boxplots of daily arrival rates and LOSs for emergency and elective arrivals. Note that the LOS graph has been cut-off at 16 days

Fig. 9

Bar plots showing counts of admissions and discharges per hour of day for the study ICU, split into emergency patients (left) and elective patients (right)

Emergency and elective patients also differ in their LOSs. As Fig. 8 shows, emergency patients tend to have longer LOSs than elective patients. Both the emergency and elective patient LOS distributions are long-tailed, as expected. However, the emergency patient LOS distribution has a much heavier tail than the elective patient LOS distribution. Note that Fig. 8 is cut-off at 16 days LOS, so the full length of the tails are not shown.

Based on this data it is clear that emergency and elective arrivals need to be considered as different classes of patients.

B.2 Time of arrival

In the previous section, we have seen graphically that there might be different LOS distributions depending on a patient’s day-of-week-of-arrival (see Fig. 8). Hence, we conducted one way ANOVA tests on LOS based on day-of-week-of-arrival. This analysis showed that the mean LOS for some day-of-week-of-arrival was different to other day-of-week-of-arrivals, with a p-value of 0.02. We then further subdivided the data into emergency and elective patients, and repeated the above ANOVA test on day-of-week-of-arrival, both of these ANOVA tests passed (with p-values of 0.87 and 0.36 respectively), which indicates that the LOS means based on day-of-week-of-arrival are not statistically different. Despite this last analysis, we still chose ‘day-of-week-of-arrival’ as a separate class for our data set. This is because, a count of the number of discharges, by day of week, shows a structure in the results (see Fig. 10). There are fewer discharges on weekends for emergency patients, and fewer discharges of elective patients on Sundays and Mondays. This structure indicates that there could be a correlation between the arrival process and LOSs based on day-of-week-of-arrival, so it is more conservative to choose each day-of-week-of-arrival as a class. Further, we would like this method to apply across hospitals, and so it is more robust to include each day-of-week-of-arrival as a class.

Fig. 10

Bar plots showing counts of discharges per day of the week for the study ICU, split into emergency patients and elective patients

Additionally, we have also just shown in Fig. 9, that there are significant differences in arrival and discharge counts based on the hour-of-day-of-arrival and hour-of-day-of-discharge for the study ICU. To investigate this idea, we constructed histograms of LOS, where each bin was set at 1/3 of a day. This bin-width was chosen because there are three shifts per day and because the arrival and discharge counts appear to be grouped into those three shifts. This plot, further divided into emergency and elective patients, with LOS truncated at 10 days, is shown in Fig. 2. It shows a significant stratification in the LOS profiles, indicating that LOS is indeed correlated with arrival time, specifically the arrival shift. The stratification can be explained as follows: if you arrive on day shift, you are likely to stay between (1 to 4/3) + n days (where n is a non-negative integer), because most discharges occur late on day shift and early afternoon shift. Similarly, if you arrive on afternoon shift your LOS is likely to be (2/3 to 4/3) + n days. Due to this correlation, we further class our data into arrival-shift: 1. day shift 2. afternoon shift, and 3. night shift.


Another class suggested by clinicians is the APACHE III code. The ICU provides care for a wide range of patients. We cannot create classes for each APACHE diagnosis in the ANZICS data, because this would make the size of the groups to be permuted too small to be meaningful. Therefore we have simplified, based on clinical advice and operational considerations, the APAHCE III codes to the following four diagnostic codes (DCs): 1. Cardiovascular related codes (Cardio), 2. Respiratory and gastrointestinal related codes (RespGastro), 3. Neurological and trauma related codes (NT), and 4. all other codes (other).

Note that these diagnostics codes were chosen for the study-ICU and may not be appropriate for all ICUs, because the patient cohort can vary significantly between ICUs. For example, the study-ICU treats emergency and elective patients, whilst other ICUs may only treat emergency patients, or only treat elective patients. Further, because the study-ICU is the main hospital in the area, it treats all presenting diagnostic codes, excluding paediatrics and obstetrics. In comparison, other ICUs may only, for example, treat paediatric patients or elective cardiac patients.

Boxplots for the LOS separated into 56 classes of: day-of-week-of-arrival (7 classes), diagnostic code (4 classes) and emergency/elective patients (2 classes) are shown in Fig. 11. Based on this figure we conducted a number of one way ANOVA tests. The first of these was on LOSs with diagnostic code, which showed that not all diagnostic codes have the same mean LOS. Then we subdivided each diagnostic code class into emergency and elective classes. We then conducted two one way ANOVA tests, one for elective patients and one for emergency patients, and in each ANOVA test we compared diagnostic code and LOS. The results from both ANOVA tests showed that not all diagnostic codes (for emergency or elective patients) have the same mean LOS.

Fig. 11

Boxplots for LoS separated into classes of: day-of-week-of-arrival, diagnostic code and emergency/elective patients, truncated at 16 days

Appendix C: Fitting arrival rate and LOS distributions

C.1 Classes

Based on our analysis of the data, arrival and LOS data was divided into the following classes:


When patients die in ICU, shortly after admission. Hence their LOS profile is quite different to other classes of patients. Additionally, our research sponsor was interested in scenarios involving increasing LOS to all patients who depart alive. Hence, constructing this class of patient was necessary for this modelling purpose.


Patients were divided according to whether they were emergency or elective patients.


Patients were divided according to the shift they arrived on, that is day shift, afternoon shift, or night shift. This class is in place due to the stratification in LOS based on arrival shift shown in Fig. 2. Note that there were very few elective arrivals on night-shift, hence not large enough for modelling purposes, so this group was set to zero arrivals.


This class is in place due to elective patients who have significantly different arrival patterns on weekdays in comparison to the weekends. Hence, elective arrivals were grouped as Monday to Friday and Saturday to Sunday. However, the elective Saturday to Sunday group was almost empty, hence not large enough for modelling purposes, so this group was set to zero arrivals. Emergency arrivals were not grouped according to day-of-week, as we found no significant change in arrival rate or LOS based on the day-of-week.

C.2 Fitting arrival rates

The key parameters for arrival rates were estimated using maximum-likelihood fitting. As expected, the arrival distributions were all found to be Poisson distributions (see Table 6).

Table 6 Admission processes for each class in D0

C.2 Fitting LOS distributions

The LOS data was transformed into discrete time intervals, representing the number of ‘shifts’ each patient stayed in the ICU. For example, if the patient arrived and left during the same shift, their LOS was 0 shifts. If a patient arrived on long-day shift and left on the subsequent long-night shift, then their LOS was 1 shift.

Calculating LOS in this manner means that all the LOS distributions (apart from deaths) show significant stratification (see Fig. 12), creating a significant modelling challenge. Clearly, a distribution like this cannot be modelled simply - using a single theoretical distribution. We successfully (and simply) dealt with this stratification by noticing that the proportion of discharges on day-shift, afternoon-shift and night-shift for a single day, was similar for one whole distribution, and further, across many distributions. This meant that we could fit a a typical discrete negative binomial distribution to daily LOS data, then use these percentages to transform the LOS (by day) distribution, into a LOS (by shift) distributions. The procedure we used is outlined below (Fig. 13).

  1. 1.

    Determine the early shifts which are best modelled as point densities. Record the point densities and remove that data. The point density data is shown in Table 7.

  2. 2.

    Transform the LOS (by shift) distributions into LOS (by day) distributions by aggregating the count for day-shift, afternoon-shift and night-shift for each day.

  3. 3.

    Determine the best fit for these distributions. These are given in Table 7.

  4. 4.

    Using the best fit theoretical distributions determined in Step 3, the LOS (by day) distributions, were transformed back into LOS (by shift) distributions as follows: the density for each day was spread across day-shift, afternoon-shift and night-shift (for that day) according the percentages shown in Table 8.

Fig. 12

Density plot of original and simulated LOS data from (or based on) the D0 dataset for the emergency, day-shift class

Fig. 13

Density plot of original and simulated LOS data from (or based on) the D0 dataset for the elective, Monday to Friday, day-shift class

Table 7 Negative binomial parameters and point densities, for each class of LOS distribution, for D0
Table 8 LOS split per shift for D0

Note that, the deaths distributions did not need this procedure as they do not show any significant stratification. This is because the stratification in LOS is due to patients rarely leaving ICU on night-shift, less often on afternoon-shift etc. However, this logic does not affect deaths. Hence the death distributions were simply modelled using a discrete binomial distribution.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Varney, J., Bean, N. & Mackay, M. The self-regulating nature of occupancy in ICUs: stochastic homoeostasis. Health Care Manag Sci 22, 615–634 (2019).

Download citation


  • Hospitals
  • Queueing
  • Decision analysis
  • Health services
  • Bed capacity