MODELING DECISIONS FOR HOSPITAL BED MANAGEMENT

A Review

Joaquim Matos

1,2

and Pedro Pereira Rodrigues

1,3

Faculty of Medicine of the University of Porto, Al. Prof. Hernâni Monteiro, 4200-319 Porto, Portugal

Faculty of Sciences of the University of Porto, Rua do Campo Alegre, s/n, 4169-007 Porto, Portugal

LIAAD - INESC Porto, L.A. & CINTESIS - Center for Research in Health Technologies and Information Systems

Universidade do Porto, Porto, Portugal

Keywords: Hospital bed management, Hospital capacity planning, Decision support systems, Decision support models.

Abstract: With today’s hospital demands and financial constraints, hospital inpatient bed management is becoming

increasingly complex. The use of decision support systems could enable hospital staff and health decision

makers to perform more focused management of the hospital inpatient beds, thus potentially reducing costs

and inpatient length of stay. A literature review was carry out on both PubMed and ISI Web of Knowledge

in order to identify studies evaluating the use of decision support systems when applied to hospital inpatient

bed management. Two different approaches were identified: one approach based on the use of mathematical

models to support the planning and allocation of hospital inpatient beds and another approach consisting in

the utilization of information technologies to support timely inpatient placement. It was perceived that

mathematical models could be safely used to model annual patient arrival rates and bed occupancy, thus

forecasting hospital/department bed demand and underlying cost structures/revenues. It was also perceived

that the use of bed management information systems provides hospital staff (administrative clerk, clinicians

and housekeepers) with the necessary information to timely assess performance measures based on the

hospital/department activity thus increasing resource effectiveness, optimizing established clinical

pathways, reducing inpatient length of stay and associated costs.

1 RATIONALE

With today’s hospital demands and increasing

financial constraints, efficient hospital inpatient bed

planning and allocation is becoming increasingly

difficult. In recent years, hospitals have engaged in

various cost-cutting efforts that include department

downsizing, the consolidation of small services and

the decrease of the average inpatient length of stay

(ALoS). The number of hospital inpatient beds is

usually determined by the hospital or by associated

health authorities using methods based on ratios

and/or target bed occupancy rates (Green and

Nguyen, 2001; National Audit Office, 2000; Nguyen

et al., 2005; Kokangul, 2008; Mackay and Lee,

2005; Millard et al., 2000). The optimal number of

hospital/department inpatient beds can be defined as

the number for which the following three criteria are

met (Nguyen et al., 2005):

 The number of unoccupied beds is not excessive,

to avoid resource misuse thus levering the efficiency

and maximizing revenues (productivity);

 The number of patients transferred to other

departments or other hospitals because of lack of

available bed is not excessive (security). Transfers to

other hospitals result in a lost of revenues while

transfers to other departments may cause the

placement in less appropriate units, compromising

the quality of care and possibly resulting in a

increase of costs (e.g. need for additional staff);

 One or more beds are available for unscheduled

admissions (accessibility).

The introduction/development of decision

support information systems, either based on

mathematical models or based on information

systems, could allow bed managers (and other staff

that hold responsibilities in the inpatient bed

management process) to perform a more focused

management of the hospital inpatient beds, either by

providing new planning forecast tools and/or tools

capable to timely-inform about current bed

occupancy/utilization, short-term levels of planned

504

Matos J. and Pereira Rodrigues P..

MODELING DECISIONS FOR HOSPITAL BED MANAGEMENT - A Review.

DOI: 10.5220/0003135005040507

In Proceedings of the International Conference on Health Informatics (HEALTHINF-2011), pages 504-507

ISBN: 978-989-8425-34-8

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

elective admissions, likely emergency inpatient

admissions and likely inpatient discharges.

2 METHODS

A literature search was performed in December 2009

with the goal of identifying studies evaluating the

use of decision support systems when applied to

hospital inpatient bed management. The search was

conducted both on PubMed and ISI Web of

Knowledge resulting in a base data collection of 65

studies. Thirteen studies were excluded by lack of

available abstract or full text and 38 studies were

excluded after abstract and/or full text review (e.g.

studies that didn’t address explicitly hospital bed

capacity and/or hospital bed allocation management

subjects). The final size of the overall data collection

was of 14 studies. Ten studies (~71,4%) addressed

the use of mathematical models to forecast and/or

plan hospital bed capacity (Green and Nguyen,

2001; Nguyen et al., 2005, 2007; Kokangul, 2008;

Mackay and Lee, 2005; Millard et al., 2000;

Cochran and Roche, 2008; Gorunescu, McClean and

Millard, 2002; Belien and Demeulemeester, 2007;

Ridge et al., 1998). Four studies (~28,6%) addressed

the use of information technologies to support

inpatient bed management, currently under use

(Blair, 2005; Szabo, 2003; Reuille, 2004) or under

evaluation (Kannry et al., 2007). One literature

source with publication date in the current decade

(National Audit Office, 2000) was handpicked and

added to the overall data collection in order to

support theoretical and practical concepts

surrounding the bed management subject.

3 FINDINGS

From the analysis of the selected data collection

from both PubMed and ISI Web of Knowledge, it

was perceived that the subject under revision (the

use of decision support systems for hospital inpatient

bed management) can be undertaken by two

distinctive approaches. One approach is based on the

use of mathematical models to support the planning

and allocation of hospital inpatient beds. Another

approach consists on the utilization of information

technologies to support the timely-decision/timely-

action of the hospital staff (administrative clerk,

clinicians and housekeepers) in order to facilitate

inpatient placement thus optimizing the bed

management process. The main conclusions and

recommendations extracted from the reviewed

studies are sketched in table 1.

The studies that addressed the planning and

allocation of inpatient beds by means of

mathematical models covered such diverse

techniques as: queuing models, stochastic models,

flow models and other general proposed models.

Different distributions of patient

arrival/department service rates were analyzed for

different department/unit profiles: geriatric

department (Gorunescu, McClean and Millard,

2002), intensive care unit (Ridge et al., 1998;

Cochran and Roche, 2008), intermediate care,

obstetrics/gynaecology and surgical departments

(Cochran and Roche, 2008) and paediatric intensive

care unit (Kokangul, 2008).

Mathematical distributions were originated from

data collections as diverse as: department historical

patient arrival rate/LoS, midnight census by

department, midnight bed holds by department and

patient billed nights by level of care. All of the

analyzed studies point out advantages in the use of

mathematical models to forecast department

demands in terms of allocating the right number of

beds. It was concluded that the mathematical models

also provide better information concerning cost

structures and revenue characteristics and how these

affect capacity and resource allocation decisions.

Given accurate data about the actual costs of

occupied and empty beds, modelling could be used

to balance the cost of providing excess standby beds

(accessibility), with the costs associated with

rejecting patients. Thus, the modelling approach

facilitates the provision of cost efficient and cost

effective services.

Four studies addressed the use of information

technologies to support and facilitate inpatient

placement. Reuille (2004) proposed a centralized

bed management system (Bed Control Report) based

on five Excel spreadsheets, dispersed by four

departments, were information was inputted by the

target department staff. Szabo (2003) describes the

use of the Bed Tracking System, developed by Tele-

Tracking Technologies, to support the bed

placement workflow and centralize bed status

information (bed released, bed in maintenance and

bed available), that is entered into the system by the

hospital staff (administrative clerk, clinicians and

housekeepers) using specific coded calls. A very

similar approach is described by Blair (2005) were

the BedCentral system, developed by MediLogistics,

is also used to centralize bed status information

entered by the hospital staff. Unlike the Bed

Tracking System, the information is inputted on both

MODELING DECISIONS FOR HOSPITAL BED MANAGEMENT - A Review

505

patient admission and patient discharge. Mobile staff

members use PDAs to generate, transmit and receive

information about bed status while other members

use department workplaces and/or an electronic

dashboard. It was concluded that the use of Bed

Tracking System resulted in a door-to-bed

improvement by nearly 30 percent and that the use

of BedCentral resulted in a reduction of the time

needed to perform bed maintenance from nearly two

hours to about 45 minutes reducing FTEs by 25

percent. The last analyzed study evaluated the use of

RFID technology to accelerate identification of

empty beds by assigning to each discharged patient a

unique RFID tag number that was read at the patient

departure time (the effective physical discharge). It

was concluded that the proposed system enabled the

identification of empty beds within an average of 25

minutes earlier when compared with the pre-existing

process (manual information recording in the ADT

system).

4 CONCLUSIONS

The analyzed mathematical models were in general

complex and didn’t address numerous internal and

external factors that may affect the provision of

hospital/department inpatient services, such as

staffing levels, multi-profile beds, requirements for

isolated beds and clinical pathways that are cross-

departmental (e.g. multi-department beds). As such,

the models should be augmented by other

information to ensure a more comprehensive

understanding of the dynamics that arise from an

analyzed hospital/department. For general planning

and cost structures/revenue purposes, an annual

model of patient arrival rates and bed occupancy

(department service) may be sufficient. Models of

shorter duration, however, may be necessary to

reflect the changes in bed occupancy that occur

throughout the year.

Table 1: Main conclusions and recommendations extracted from the analyzed studies.

Information technologies Mathematical models

BedCentral has helped St. Luke's

Episcopal Hospital to maximize its

existing resources, e.g. ICU capacity

tuning, telemetry/high-dollar beds swift

filling and minimization of the time

required to clean rooms and transport

patients (Blair 2005)

The use of target occupancy levels, such as the

average length of stay (LoS), as the primary

determinant of bed capacity is inadequate due to

the unpredictable nature and distribution of

hospital admission rates and patient LoS over time

(Green & Nguyen 2001; Nguyen et al. 2005, 2007;

Kokangul 2008; Mackay & Lee 2005; Millard et

al. 2000; Ridge et al. 1998)

More sophisticated methodologies

should be considered to support

decisions that involve bed capacity,

e.g. methodologies that capture

population structure, organization

and resource use (Green & Nguyen

2001; Mackay & Lee 2005)

Bed Tracking System has improved

door-to-bed time and reduced the “hold-

time” by nearly 30%. It also lengthened

the time clinicians can spend with

patients

instead of trying to locate a empty bed

(Szabo 2003)

Stochastic and flow models could be used to

determine the optimum size of bed requirement,

the size of additional required resources (e.g.

workforce or facility planning) or to capture the

variation in bed occupancy (Kokangul 2008;

Mackay & Lee 2005)

Models should consider elective

patient scheduling, different levels of

care (e.g. ICU and HDU beds) and

semi-automated historical data

(Ridge et al. 1998)

The proposed RFID-based system

identified empty beds within an average

of 25 minutes earlier when compared

with the pre-existing process - manual

information recording in the ADT system

(Kannry et al. 2007)

Posterior financial/billing data, rather than the

census data commonly relied upon, yields the

true hospital bed demand as it estimates true

demand for service rather than merely the

service available to be offered

(Cochran & Roche 2008)

Seasonal bed demand patterns

should be considered when

evaluating bed management models

(Kokangul 2008, Cochran & Roche

2008)

There is a clear need to address both

the issue of capacity planning at each different

level of care within individual hospitals and also

relative levels of care across different hospitals

in a region (Ridge et al. 1998)

Additional parameters should be

included such as population rate,

staff/equipment requirements and

budget constraints, in order to

universalize stochastic models

(Kokangul 2008)

Conclusions Recommendations

HEALTHINF 2011 - International Conference on Health Informatics

506

Bed management systems based on information

technologies, supply hospital staff with the

information necessary to timely assess performance

measures based on the department activity, thus

increasing resource effectiveness, optimizing

established clinical pathways, reducing inpatient

length of stay and associated costs. However

performance measures could be biased because of

the implied cross reliance on hospital staff to timely

input patient/bed status information. The analyzed

systems rely too much on timely inputted

information in order to proper track

hospital/department bed status – current bed status

may not be the real bed status. In this sense, it is

recommended that these systems be integrated with

the hospital ADT (Admission/Discharge/Transfer)

information system in order to reduce human

interaction to the minimum. Yet, even after

performing this kind of integration, another problem

would still exist in the patient discharge process –

the time de-synchronization between the patient

department discharge and the manual recording of

that information in the ADT system. The suggested

RFID approach could provide a simple solution for

this problem.

REFERENCES

Kokangul, A. (2008). A combination of deterministic and

stochastic approaches to optimize bed capacity in a

hospital unit. Computer Methods and Programs in

Biomedicine, 90(1), 56-65.

Belien, J. and Demeulemeester, E. (2007). Building cyclic

master surgery schedules with levelled resulting bed

occupancy. European Journal of Operational

Research, 176(2), 1185-1204.

Blair, R. (2005). Capacity management: the bedrock of

efficiency. Health Management Technology, 26(9), 26,

30-31.

Cochran, J. K. and Roche, K. (2008). A queuing-based

decision support methodology to estimate hospital

inpatient bed demand. Journal of the Operational

Research Society, 59(11), 1471-1482.

Gorunescu, F., McClean, S. I. and Millard, P. H. (2002).

Using a Queuing Model to Help Plan Bed Allocation

in a Department of Geriatric Medicine. Health Care

Management Science, 5(4), 307-312.

Green, L. V. and Nguyen, V. (2001). Strategies for cutting

hospital beds: The impact on patient service. Health

Services Research, 36(2), 421-442.

Nguyen, J. M., Six, P., Antonioli, D., Glemain, P., Potel,

G., Lombrail, P. and Le Beux, P. (2005). A simple

method to optimize hospital beds capacity.

International Journal of Medical Informatics, 74(1),

39-49.

Kannry, J., Emro, S., Blount, M. and Ebling, M. (2007).

Small-scale Testing of RFID in a Hospital Setting:

RFID as Bed Trigger. AMIA Annual Symposium

Proceedings, 384-388.

Mackay, M. and Lee, M. (2005). Choice of Models for the

Analysis and Forecasting of Hospital Beds. Health

Care Management Science, 8(3), 221-230.

National Audit Office (2000). Inpatient Admissions and

Bed management in NHS acute hospitals.

Nguyen, J. M., Six, P., Chaussalet, T., Antonioli, D.,

Lombrail, P. and Le Beux, P. (2007). An objective

method for bed capacity planning in a hospital

department - A comparison with target ratio methods.

Methods of Information in Medicine, 46(4), 399-405.

Millard, P. H., Mackay, M., Vasilakis, C., Christodoulou,

G. (2000). Measuring and modelling surgical bed

usage. Annals of the Royal College of Surgeons of

England, 82(2), 75-82.

Szabo, P. (2003). Bed-der than ever. Pittsburgh hospital

uses automated bed tracking and control to speed

efficiency in its ED. Health Management Technology,

24(3), 58-59.

Reuille, R. (2004). Bed Control Report: A Computer-

based System to Track Patient Admissions Delayed or

Rescheduled Due to a Bed Shortage. The Journal of

Nursing Administration, 34(12), 539-542.

Ridge, J. C., Jones, S. K., Nielsen, M. S. and Shahani, A.

K. (1998). Capacity planning for intensive care units.

European Journal of Operational Research, 105(2),

346-355.

MODELING DECISIONS FOR HOSPITAL BED MANAGEMENT - A Review

507