THE PROBLEM OF DELAYED DISCHARGE
IN LABOR AND DELIVERY
Jeff W. Trailer, Ph.D.
Professor of Management
College of Business
The California State University, Chico
Chico, CA 95929-0031
Office: 530-898-5663
Fax: 530-898-5501
jtrailer@ csuchico.edu
Matthew B. McLaughlin, M.B.A.
Financial Analyst
Enloe Medical Center
Chico, CA 95929
Matthew.McLaughlin@ enloe.org
Abstract
Enloe Medical Center is a non-profit community hospital in Chico, California. Among
the many services they provide is a Labor and Delivery Department. While mothers are
routinely admitted from 1:00pm to 1:00am, they are generally discharged between 10:00am and
5:00pm. This results in a generic bell curve behavior pattern for patient occupancy during the
daytime. Hospitals are reimbursed for inpatient services in two major ways: either on a per
diem basis, or by diagnosis related groups (DRG). Either way, the revenue to the hospital
remains the same, regardless if the patient is discharged at 4:00am or 4:00pm. In California,
state mandated nurse to patient ratios require hospitals to maintain a minimum level of nurse
staffing for inpatient services. Thus, as the patient census rises during the day, so must the
number of nurses on staff. This is the problem studied; costs expended for patient discharge
Contact author: Jeff Trailer, Ph.D.
The Problem of Delayed Discharge in Labor and Delivery
Enloe Medical Center is a non-profit community hospital in Chico, Califommia. Among
the many services they provide is a labor and delivery department for expectant mothers, which
serves over 1,500 mothers and babies each year. While mothers are routinely admitted anywhere
from 1:00pm to 1:00am, they are generally discharged between 10:00am and 5:00pm. This
results in a generic bell curve behavior pattern that reflects extremely high patient census, or
patient occupancy, during daytime business hours.
While this may appear to be an inevitable cycle for the hospital, it is in fact a major target
area for financial improvement. Hospitals are reimbursed for inpatient services in two major
ways: either on a per diem basis, or by diagnosis related groups (DRG). Either way, the revenue
to the hospital remains the same, regardless if the patient is discharged at 4:00am or 4:00pm.
Thus, there is created strong pressure for efficiency and corresponding cost control. This
pressure has resulted in lowering staffing levels to bare minimum levels. In some cases, it
appears that the burden of work on nurses is so great, the quality of care is degraded. Thus,
hospitals must now comply with AB394 (California Assembly Bill 394), which imposes
minimum nurse to patient staffing levels.
In 1999, the California State Legislature passed A B394, adding section 1276.4 to the
Health and Safety Code (HSC), later amended by AB1760. Basically, this requires the California
Department of Health Services to develop minimum, specific, numerical nurse to patient ratios
for specified units of general acute care hospitals. These ratios are imposed to ensure minimum
necessary nurse staffing levels for protecting public health and safety. This means that as the
patient census rises during the day, so must the number of nurses on staff. With the recent
nursing shortage this can become a problem both financially and logistically. In addition to
salaries, other costs such as food, linens, utilities, etc are expended for each hour that a patient
remains in a bed that could be discharged.
Healthcare reimbursements are in decline while salaries, pharmaceuticals, and other
related costs are steadily rising. This double squeeze on hospitals makes every improvement
opportunity essential to maintaining a healthy financial position. Carol Butler, the VP of
Nursing Services at Enloe Medical Center, had this to say about this issue.
“The efficient movement of patients through the inpatient areas is essential to the
management of hospital resources. Staffing levels can be managed more easily to meet
the patient needs. The quality of care given is improved when patients are admitted and
discharged in a controlled environment. Unanticipated and delayed discharges result in
unreimbursed expenditure of resources, and chaotic work practices. The financial
constraints of the organization can be more easily met by reducing overtime hours
incurred by overlap of patients waiting for discharge and patients requiring admission.
Bringing control to discharge practices and timeliness improves the satisfaction of
patients, staff and hospital management.”
The purpose of this study is to understand some of these unanticipated and delayed
discharges to help better comply with AB394 and manage this highly important healthcare
problem. This study is executed in the following steps (Richardson and Pugh, 1981; Vennix,
1996; Forrester, 1999): (1) Problem Identification and Model Purpose. As described above in
this section, the problem of delayed discharge in Labor and Delivery was identified and refined
in interviews with Enloe Medical Staff, and based on expert experience of the co-author. (2)
System conceptualization. In this section, we describe the process for identifying the model
boundaries and the reference mode. (3) Model formulation and parameter estimation. Here we
present the model and explain the basis for the parameter estimates. (4) Analysis of model
validity and behavior. In this section, we will discuss validity checks and show the baseline
performance pattern, comparing this to the reference mode; the actual performance data from
Enloe Medical Center, for the past three years. (5) Policy Analysis. Here we will show the tests
of alternative policies to improve Labor and Delivery system performance. (6) Model use for
policy implementation. In this final section, we will discuss feedback we receive from Enloe on
management use of the model and on implementation of policy changes.
System conceptualization
We are currently working in this phase of the project. By the time of the SDS
conference, we anticipate completion through policy analysis. For this system
conceptualization, rather than starting with system archetype causal loop diagrams to explain a
reference mode behavior pattern, we adopted the Forrester (2000) process of starting with
identifying the levels. Based on interviews at EMC, and expert knowledge of the co-author, we
identified levels for two required subsystems, patients and nurses. Patients accumulate in three
different stages in the life cycle of their treatment, as necessary to explain the reference mode of
behavior; patients waiting for discharge. We next focused on the rates (policies) that move
patients (patient flow) through the Delivery and Labor process. Nurses, and as a consequence
nurse staffing policies, are the resource identified that most strongly limit the rates at which
patients can flow through the system. The accumulation of Nurses on duty is thus required.
Correspondingly, we then specified the rates (policies) that govern arrival and departure of the
nurses. In our interviews, there was discussion of the different professional categories of nurses.
This motivated the creation of several stocks of nurses in the model, to control the availability of
the different types of nurses on duty. However, this added complexity that tended to obfuscate,
rather than clarify, the overall process. Thus, the model has been simplified to include only a
generic “average” nurse. At this point we feel the model still retains the validity, if not the
precision, of the more complex specification.
Other required resources identified were physicians, beds, equipment, and medical
supplies. Of these, only physicians were identified as a significant constraint. This constraint
was not associated with the accumulation of physicians, but with the timing of the physician
visits. Thus, at this point in the modeling process, the accumulation of physicians is considered
exogenous, however the timing of physician visits is endogenous. This may change when we
have feedback from the model validity assessment.
Model formulation and parameter estimation.
The model, to this point in the project, is presented in Figure 1. We have historical data
with which to specify the parameters for the stocks, flows and the lookup tables. We have
interview notes that allow us to set realistic parameter estimates for the remaining variables.
Figure 1.
Modeling the flow of labor and delivery patients
—.
“Time available for
discharging
—~ Average time needed
Time needed for peg
Average time admits
needed for admits f ~~ ___ Time meded for.
‘Time avalable for
admits
Patins wating Pateais i |
fai to be admit a Treatment | fordsctare | patos
a x complete |
Patent anival table ‘ aa
Physician rival
Nurse to Patient
be Ratio Desired
| "ine rene for \
en Ral
a Nurse to Patient
Comerter _ Nurse to Patient —r Ratio gap
Calladditionala— Ratio
muses
Average time needed
for treatment
3 =
Nurse arival table_ ~ E
Average time
available per muse
Analysis of model validity and behavior.
We currently have three years of historical data with which to conduct checks of
simulation behavior with reference mode behavior. We are currently in the process of testing
consistency in dimensions.
Policy Analysis.
Based on interviews with the staff, we anticipate testing the following policies:
Policy 1. Use of an admit/discharge specialty team of nurses
Policy 2 Use ofa physician “ discharge visit” schedule
Model use for policy implementation.
Acknowledgements
References
Forrester ].W. 1961. Industrial Dynamics . MIT Press: Cambridge, MA.
Forrester ].W. 1971. Counterintuitive behavior of social systems. Technol Rev. 1971; 73:53-68.
Forrester ].W. and Senge P.M. 1980. Tests for building confidence in system dynamics models.
In: Legasto AA, Forrester JW and Lyneis JM (eds). System Dynamics. TIMS Studies in
the Management Sciences , Vol. 14 . North-Holland: New Y ork, pp 209-228.
Forrester, J.W. 2000. Forrester Ph.D. Seminar in System Dynamics. Session F - Group
Model Building. Albany, NY. System Dynamics Society.
HomerJ. 1997. Structure, data and compelling conclusions: notes from the field. System
Dynamics Review 13 : 293-309.
Homer J, Hirsch G, Minniti M, Pierson M. 2004. Models for collaboration: how system
dynamics helped a community organize cost-effective care for chronic illness. System
Dynamics Review. 20:199-222.
Lane DC, Monefeldt C, Rosenhead JV. 2000. Looking in the wrong place for healthcare
improvements: a system dynamics study of an accident and emergency department. J
Operational Res Soc. 51: 518-531.
Lattimer V. 2004. Reviewing emergency care systems I: insights from system dynamics
modelling. Emergency Medicine J ournal, Nov; 21 (6): 685-91.
Morecroft JDW. 1985. Rationality in the analysis of behavioral simulation models.
Management Science. 31:900-916.
Oliva, R., andJ. Sterman 2001 “Cutting Corners and Working Overtime: Quality Erosion in the
Service Industry.” Management Science, 47: 894 - 914.
Repenning N, Sterman J. 2001. Nobody ever gets credit for flying problems that never
happened: creating and sustaining process improvement. California Management
Review. 2001;43:64-88.
Repenning, N.and Sterman, J. 2002. Capability Traps and Self-Confirming A ttribution Errors in
the Dynamics of Process Improvement. Administrative Science Quarterly, 47: 265-295.
Richardson, G.P. and Pugh, A.L. 1981. Introduction to system dynamics modeling with
DYNAMO. Cambridge (MA). M.LT. Press.
Royston G, Dost A, Townshend J, Turner H. 1999. Using system dynamics to help develop and
implement policies and programmes in health care in England. System Dynamics
Review.; 15: 293-313.
Sterman J.D. 1984. Appropriate summary statistics for evaluating the historical fit of system
dynamics models. Dynamica 10 : 51-66.
Sterman, J.D. 1989. Modeling Managerial Behavior: Misperceptions of Feedback in a Dynamic
Decision Making Experiment. Management Science, 35: 321-339.
Sterman J.D. 2001. System dynamics modeling: tools for learning in a complex world.
California Management Review. 2001; 43:8-25.
Sterman J.D. 2000. Business Dynamics . Irwin McGraw Hill: New Y ork.
Taylor K and Lane D. 1998. Simulation applied to health services: opportunities for applying
the system dynamics approach. J Health Serv Res Policy 3 : 226-232.
Tucker, A and Edmondson, A. 2003. Why hospitals don’t learn from failures: Organizational
and psychological dynamics that inhibit system change. California Management
Review, 45, 2: 55-72.
Vennix J. 1996. Group Model-building: F acilitating Team Learning Using System Dynamics.
Chichester, UK: Wiley.
Walker B; Haslett T. 2001. System dynamics and action research in aged care. Australian
Health Review, 24(1): 183-91.
Wolstenholme EF. 1999. A patient flow perspective of UK Health Services: exploring the case
for new "intermediate care" initiatives. System Dynamics Review. 15: 253-271.
Zimmerman, R. 2004. Hospital Capacity, productivity and patient safety - it all flows together.
Frontiers of Health Services Management, Summer, 20,4: 33-38.
