bmj.com Rapid Responses for Bagust et al., 319 (7203) 155-158

Rapid Responses to:

PAPERS:
Adrian Bagust, Michael Place, and John W Posnett
Dynamics of bed use in accommodating emergency admissions: stochastic simulation model
BMJ 1999; 319: 155-158 [Abstract] [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

Occupancy probably already higher than thought: Ruth Boaden (17 July 1999)

Previous paper concurs value of this approach: David J R Hutchon (18 July 1999)

Simulating bed use for emergency admissions: M Pidd (19 July 1999)

A Large Rise in Occupancy Levels in Acute NHS Hospitals is unlikely: Matthew G Dunnigan, Myra M McMurdo (26 August 1999)

Relative effects of internal and external factors on emergency department efficiency: Oscar Miro (30 November 1999)

Occupancy probably already higher than thought 17 July 1999

Ruth Boaden,
Senior Lecturer in Operations Management
Manchester School of Management, UMIST
Send response to journal:
Re: Occupancy probably already higher than thought

The paper by Bagust et all indicates that overall the NHS may currently be operating at 85% occupancy currently. Work we have been carrying out with bed managers in Greater Manchester (Boaden et al, 1999) recently indicates that occupancy is much higher than that already (generally in excess of 90%). Measuring occupancy by a midnight bed count may not be the most effective way of indicating the true pressure on beds, since most trusts would need some beds empty at midnight to be able to accomodate emergencies admitted during the night. A midday count would give an even more dismal prediction of the number of bed crises likely to occur (is it possible to have occupancy in excess of 100%?). The bed crises that the paper refers to are, in our view, happening all too frequently in many trusts and are not due in many cases to poor bed management. If this paper serves to encourage bed managers in their difficult work by showing that the problems they face daily in reconciling supply with excessive demand are related to occupancy, not their own professional competence, then it is of value. It should also encourage senior managers to look again at how the bed resource is managed and at the pressures on their bed managers. Local support networks, and the sharing of best practice, as are operating in Greater Manchester, may go some way towards improving the management of this scarce resource.
REFERENCE
Boaden R J, Proudlove N and Wilson M (1999)"An Exploratory Study of Bed Management in 14 Trusts", Journal of Management in Medicine, forthcoming

Previous paper concurs value of this approach 18 July 1999

David J R Hutchon,
Consultant Obstetrician & Gynaecologist
Memorial Hospital Darlington
Send response to journal:
Re: Previous paper concurs value of this approach

18/7/99
The Editor, British Medical Journal, BMA House, Tavistock Square, London. WC1H 9JR
EDITOR - I agree that a simulation model as described by Bagust et al 1 using a spread sheet is a valuable way of investigating the dynamics of bed use. We have already described a similar system using a Smartware II spreadsheet that showed the dynamic use of beds in an obstetric and gynaecology unit. 2 Obstetrics inevitably has a high percentage of "emergency" admissions.
In 1992 following the Changing Childbirth report there was increasing pressure to reduce the size of the estate and the bed complement. The required bed complement estimated from historical bed occupancy rates, from other comparable units or calculated from simple formulae often lacks credibility with clinicians. The dynamic spreadsheet approach allows "what if" scenarios to be entered and can take into account variations in length of stay, Caesarean section rates and community visits etc, and the staff requirements for the level of activity. Too many beds leads to inefficient use of resources while too few requires finding beds in other parts of the hospital during periods of high admissions and the associated risks to patients that this involves. We showed that such a model was a valuable tool in change management, and it has proved a realistic simulation for bed requirements since the change in bed numbers has been effected.
David J R Hutchon Consultant Obstetrician & Gynaecologist
References
1 Bagust A, Place M, and Posnett J W. Dynamics of bed use in accommodating emergency admissions: stochastic simulation model. British Medical Journal 1999;319:155-8
2 Hutchon D J R. Using computer modelling for the rationalisation of hospital beds. British Journal of Health Care Management 1997;3(6):332-4

Simulating bed use for emergency admissions 19 July 1999

M Pidd
Send response to journal:
Re: Simulating bed use for emergency admissions

EDITOR - I was pleased to see Bagust et al [1] demonstrating the value of stochastic simulation in investigating the effects of different policies for emergency admissions. They are correct to say that such approaches provide a neutral way to develop better policies in practice, without the need for dangerous and expensive experiments on the real system. Using proper discrete simulation software it would not be difficult to take this work rather further by showing, for example, the knock-on effects of emergency admissions on the availability of beds for planned admissions. This might enable the better management of available resources. Perhaps, too, health policy makers could be persuaded that analyses of this type should become a routine way of investigating the effect of changes before they are introduced?
Professor M. Pidd Head of Department of Management Science Lancaster University Lancaster LA1 4YX

A Large Rise in Occupancy Levels in Acute NHS Hospitals is unlikely 26 August 1999

Matthew G Dunnigan,
Senior Research Fellow
Department of Human Nutrition, Royal Infirmary Glasgow G31 2ER,
Myra M McMurdo
Send response to journal:
Re: A Large Rise in Occupancy Levels in Acute NHS Hospitals is unlikely

Editor - Bagust et al�s simulation model suggests that failure to provide a bed for patients in acute NHS hospitals in England should be infrequent at current annual occupancy levels (79.5% in 1997-98)[1]. In the absence of published data, their model corresponds to most clinicians� experience.
Their projection of rapidly rising annual occupancy levels for all acute specialties �which could exceed 90% by 2002-03� is implausible. Average occupancy in acute specialties in Scottish NHS hospitals has risen slowly because rising numbers of emergency and elective admissions and falling bed capacity have been accompanied by falling average lengths of stay[2]. Average occupancy rose by only 4.6% between 1980-98 (Table). While occupancy statistics in English NHS hospitals are limited to 1996-97 and 1997-98 (79.5% in both years), long-term English and Scottish occupancy trends are likely be similar[3]. In this respect the authors� simulation model conflicts with past experience.
In the absence of total closure, the frequent disruption experienced by patients and staff in acute hospitals at periods of high occupancy results from long waits in Accident & Emergency Departments due to delays in finding scarce beds and from the frequent failure to accommodate patients in their consultants� parent units. This leads to their dispersal within the Medical Directorate (since most emergency admissions are medical), extending to their wider distribution in other directorates as occupancy rises. Long ward rounds to multiple wards and the cancellation and postponement of elective surgical procedures follow, contributing to rising stress levels in medical and nursing staff and rising waiting lists. Far from planning to increase reserve bed capacity as Bagust et al sensibly suggest, new hospitals being built by the Private Finance Initiative will contain about 30% fewer acute beds than those they replace[4].
In the absence of increased reserve bed capacity, our experience suggests that pressures on acute beds can be alleviated by the appointment of experienced senior nursing staff as bed managers with wide responsibilities for the direction of inpatient �traffic� and for the implementation of effective discharge planning policies.
TABLE
Changes in inpatient activity in acute specialities in Scottish NHS hospitals between 1979 - 80 and 1997 - 98.
Year Inpatient Average discharges staffed beds 1979 - 80 508,360 18,144 1997 - 98 751,543 12,136 Change % + 47.8 - 33.1 Year Mean stay Occupancy (days) (%) 1979 - 80 10.8 71.9 1997 - 98 5.1 75.2 Change % - 52.8 + 4.6

REFERENCES
1. Bagust A, Place M, Posnett J W. Dynamics of bed use in accommodating emergency admissions: stochastic simulation model. BMJ 1999; 319: 155-8.
2. Scottish Health Statistics. Edinburgh: NHS (Scotland) Information and Statistics
Division, 1980-98.
3. Department of Health. Bed availability and occupancy. England, financial year
1997-98. London: DOH 1998.
4. Pollock A M, Dunnigan M G, Gaffney, D, Price D, Shaoul J. The private finance
initiative. Planning the �new� NHS: downsizing for the 21st century. BMJ, 319: 179-84.
Matthew G. Dunnigan Senior Research Fellow Department of Human Nutrition, Royal Infirmary Glasgow G31 2ER
Myra M. McMurdo Discharge Co-ordinator Stobhill Hospital, Glasgow G21 3UW

Relative effects of internal and external factors on emergency department efficiency 30 November 1999

Oscar Miro,
Department of Internal Medicine, Hospital Cl�nic
Hospital Cl�nic, Villarroel 170, 08036 Barcelona, Catalonia, Spain
Send response to journal:
Re: Relative effects of internal and external factors on emergency department efficiency

To accomplish their function with sufficiency, Emergency Departments (EDs) require an appropriate balance between demand of care and ED response. Different external and internal factors exert pressure against EDs function that can lead to a decay in its quality, but the relative influence of each factor on ED quality has been poorly explored. In their theoretical assay, Bagust et al.1 underlined the necessity to spare hospital beds to counteract the typical peaks of demand of emergency care and avoid the risk of ED overcrowding during periods of increased external pressure. We herein offer real data coming from an urban tertiary-care university hospital (referral population: 500,000) which attends around 40,000 medical consults annually2,3.
During 3 consecutive weeks (from February 10 to March 2, 1999), we recorded at 3 hours interval two different markers of overall ED efficiency: "patients waiting for visit" and "waiting time for visit". To evaluate effects of external factors on such efficiency markers, we compiled patient arrivals during each interval; on the other hand, to assess effects of internal factors, we attributed the main cause for patient remaining in ED after medical assistance had begun to: 1) ED- related factors (patients being visited or waiting for emergency physician, obtaining complementary tests results, or clinical evolution); 2) ED-Hospital interrelation-related factors (patients waiting for explorations performed outside of ED, or assessment by hospital specialist); 3) Hospital-related factors (patients waiting for going to or finding an inhospital bed); and 4) Non-Hospital-related factors (patients waiting for family, social assistance, or ambulance). Results were expressed as means�SD (range). The relationships between efficiency markers and external/internal factors affecting ED dynamics were assessed through simple lineal regression analyses.
During the studied period, 2,060 patients were visited. Occupation rate of hospital was 89.9%, and beds daily offered for emergency admissions were 36�10 (27-69). As average, waiting time was 85.8�100.8 (0- 470) minutes, waiting patients 5.8�6.2 (0-25), patient arrivals 11.7�6.6 (1-35), and patients remaining in ED after their assistance had begun 34.5�9.5 (17-65). As the table illustrates, while fluctuation of quality markers was not associated with changes in external pressure, internal pressure showed a positive, significant correlation with both markers, being ED and H-related factors those causing such a correlation. Table Evolution of quality markers and external and internal pressure on ED along daytime (mean�SEM) .
____________________________________________________________________________________________________ 9h 12h 15h 18h 21h 24h 3h 6h p value* ____________________________________________________________________________________________________ Quality markers �Waiting patients (number) 2�1 6�1 7�1 9�2 8�2 7�1 4�1 3�1 - �Waiting time (minutes) 27�20 38�8 79�14 106�23 117�20 149�28 92�25 76�26 - External pressure �Patients arrival during previous 3 hours (number) 12�1 17�2 13�1 12�1 14�1 10�1 9�1 8�2 n.s./n.s. Internal pressure �Patients waiting in ED (number) 31�2 38�2 42�2 38�2 35�2 32�2 30�2 30�2 <0.0001/<0.0001 Analysis of internal factors (%) �Patients waiting in ED for ED-related factors 35�2 39�2 41�3 40�3 44�4 38�2 40�3 38�3 n.s./n.s �Patients waiting in ED for ED-H interrelation-related factors 17�2 14�2 11�1 11�1 12�2 10�2 3�1 3�1 <0.01/<0.01 �Patients waiting in ED for H-related factors 42�2 40�2 43�3 44�3 38�4 45�2 50�2 51�3 n.s./n.s. �Patients waiting in ED for non-H-related factors 5�1 7�1 5�1 5�1 5�1 7�1 6�1 8�1 n.s./<0.05 ______________________________________________________________________________________________________ *p values have been obtained by linear regression analysis between each studied factor and quality markers (waiting patients/waiting times). n.s.: not significant. ED: Emergency Department. H: Hospital.

Our results show that fluctuating patient arrivals to ED is not associated with a decay in ED efficiency. Accordingly, efficiency should be improved through the modification of ED internal dynamics. Among such internal factors susceptible to be corrected, optimisation of the ED processes themselves, disposition of beds all daytime, and spare enough beds for ED necessities seem to be the most urgent issues to be improved.
�scar MIR�,
Miquel S�NCHEZ,
Jos� MILL�.
Emergency Department, Hospital Cl�nic, Barcelona, Catalonia, Spain.
Address for correspondence: Dr. �scar MIR�, Department of Internal Medicine, Hospital Cl�nic. Villarroel 170, 08036 Barcelona, Catalonia, Spain.
Email: omiro@medicina.ub.es
REFERENCES
1.-Bagust A, Place M, Ponsett JW. Dynamics of bed use in accommodating emergency admissions: stochastic simulation model. BMJ 1999; 319:155-158.
2.- Mir� O, Jimenez S, Alsina C, Tovillas-Mor�n FJ, S�nchez M, Borr�s A, Mill� J. Revisitas no programadas en un servicio de urgencias de medicina hospitalario: incidencia y factores implicados. Med Clin (Barc) 1999; 112: 610-615.
3.-Mir� O, Antonio MT, Jim�nez S, de Dios A, S�nchez M, Borr�s A, et al. Decreased health care quality associated with emergency department overcrowding. Eur J Emerg Med 1999; 6:105-107.