Patient boarding in the emergency department as a symptom of complexity-induced risks

Full text

Chapter x: Patient Boarding in the
Emergency Department as an Indicator of
Complexity-Induced Risks
Robert J. Stephens, David D. Woods and Emily S. Patterson
Patient Boarding in the Emergency Department as a
Symptom of Complexity
The emergency medicine system absorbs much of the burden of the
chronic problems facing healthcare in the United States today: an
increasingly unhealthy populace, variability in access to primary care,
and severe resource shortages (Committee on the Future of Emergency
Care in the US, 2006). Studies of patient boarding, overcrowding, and
limits on surge capacity often examine how these factors lead to poor
care processes and poor outcomes for patients (e.g., Chalfin, et al.,
2007; Pines et al., 2007; Viccellio, et al., 2009). As a result, emergency
medicine, and in particular the emergency department (ED), is an
excellent natural laboratory in which to study how health care personnel
in different units adapt to complexity related factors (Wears, et al., 2008;
Wears and Woods, 2007).
“Boarding” is a label used to refer to situations when admitted patients
are held in the ED for long time periods after their treatment there has
concluded and while awaiting transfer to another service -- situations
where patients flow into and then get stuck in an ED (Hoot and
Aronsky, 2008; AECP, 2008). Observations of how different types of
patients move from the ED to other units such as an intensive care unit
(ICU) show that there is a gray area in between “completing” treatment
in the ED and when other units are willing to accept and take
responsibility for the patient. A variety of factors regarding current
and future demands on the different units operate in this gray area and
Chapter x 1

limited coordination between the units combine to determine the risk
of boarding for different classes of patients.
This chapter relates the findings from a study of ED boarding
(Stephens, 2010; Stephens, Cudnik and Patterson, 2011) to factors
about human complex adaptive systems (Woods and Branlat, 2011).
Patient boarding can be interpreted as a symptom of complexity, and
represents a gap or shortfall in adaptive capacity (Cook et al., 2000;
Woods, 2006; Woods, in preparation). Studying how different parts of
the hospital adapt in the face of this shortfall reveals the operation of a
key concept in resilience - capacity for maneuver, or CfM (Woods and
Branlat, 2011; Woods, in preparation). The shortfall also indicates the
system is at risk of complexity-induced forms of breakdown and the
adaptations by the different units are, in effect, attempts to manage
these risks by regulating their CfM relative to interactions with other
parts of the health care system. Analyzing patient boarding as a
symptom of complexity provides new explanations of how boarding
occurs and points to new approaches to resolve the risk for patients.
Patient Transfers from the ED
Patient boarding reflects processes associated with moving patients
from the ED to other units. Stephens et al. (2011) tracked two classes
of patients to understand how boarding occurred. One class was
patients who needed to be transfered to an Intensive Care Unit (ICU),
and the second group were mental health patients who needed to be
transferred to another facility.
The study observed what happens in the gray area between
“completing” treatment in the ED and when other units are willing to
accept and take responsibility for the patient. One activity the study
observed is that ED providers expend considerable effort to build a
case for admission to another unit, then negotiate for that patient to be
admitted. Breakdowns and delays in the negotiation process increase
patient length of stay in the ED (LOS) and increase the risk of being
“stuck” waiting.
2 Resilience Engineering in Health Care

The negotiation process builds slowly as an asynchronous sequence of
interactions across multiple parties with differing priorities. Part of the
build-up to a transfer includes the steps to making a patient stable and
transportable but also, importantly, to making the patient look like a
patient who fits the criteria for treatment in the other unit. The ED
performs cognitive work and physical actions to build proof that the
patient should be admitted to the other unit, and this work can result in
delays to patient disposition.
On the other hand, the receiving unit has its own resource limits,
current patient load, and future expected patient load, all of which
constrain its capacity for receiving patients from the ED. For example,
the medical ICU (MICU) was sometimes observed to refuse to receive
a patient until the ED completed additional tasks such as central line
placement. This is an example of workload shifting from the ICU to
the ED. The viability of this shift depends on the answer to the
question of what is the potential for workload bottlenecks for each unit
-- which unit is more workload constrained now or in the near future?
Shifting workload to another unit frees or preserves resources to reduce
the risk of future bottlenecks and the potential degradation of
indicators of the performance for unit shifting workload. But this gain
for one unit is a cost on the unit receiving the additional workload. The
cost for the receiving unit depends on its ability to absorb that
workload relative to other demands that are ongoing, building, or
anticipated. The example of workload shifting from an ICU to an ED
also points to another relevant factor -- differential authority between
the units.
These observations highlight that patient LOS represents the results of
an interplay between the capacity of the receiving unit (e.g., ICU) and
the capacity of the ED to prepare patients and to continue to provide
appropriate care for patients. Because hospitals track performance
indicators such as ED LOS and are aware of potential difficulties in
transfer across units, they adapt in various ways to control LOS and
improve transfers. In the hospital Stephens (2010) observed, a transfer
center had been set up to facilitate admission to the ICU resulting in
interactions across three units. Even with this extra unit to smooth the
transfer process, Stephens (2010) observed, when crunched, ED
Chapter x 3

attending physicians going to the medical ICU to assess directly that
unit’s capacity to handle ED patients ready to transfer. The reverse was
also observed where, when crunched, ICU attending physicians went to
the ED to judge the ED’s capacity to act as a buffer for the ICU and to
continue to care for critically ill patients. The study also noted cases of
direct face-to-face negotiation between ED personnel and ICU
personnel over how to balance capacity.
ED providers begin by calling the transfer center, which initiates a
complex and asynchronous chain of interactions that may or may not
result in smooth transfer for the patient. The nominal interplay
observed across the three units is: the ED passes responsibility for
negotiation to the transfer center, who then notifies the MICU, who
then calls the ED back completing the loop. It is not until the third step
of the sequence that the ED has a synchronous communication with
the MICU, and this communication primarily consists of one-way
transfer of information. The MICU may or may not confirm during the
call that the patient has been accepted. If the MICU decides not to
accept the patient, the transfer center once again becomes involved,
sending an update to the ED, who must then decide whether to push
for admission to the MICU harder or negotiate for admission
elsewhere. The complexity of this sequence creates additional delays for
patients (in one case this factor added about an hour delay).
ED attending physicians sometimes used phrases such as, “I could
admit this patient now but I’d get in trouble,” in reference to those
patients who they believed to be candidates for MICU admission, but
for which they did not yet have sufficient basis to initiate the transfer
process. Further lab tests and imaging studies were ordered first, and in
some cases, the patient’s condition would need to worsen (e.g. being
placed on a ventilator) before an attempt at transfer would be made. In
the observed cases getting computed tomography study results resulted
in additional wait times of 1 and 6 hours depending on the load on the
imaging centers.
In all observed cases, when the ED attending spoke with the MICU to
negotiate for admission, more lab values were given than when
speaking to the transfer center. In addition, the attending increased the
4 Resilience Engineering in Health Care

amount of data communicated to both the MICU and the transfer
center following an unsuccessful attempt at admission. For all of the
cases observed, interpretation of laboratory data and imaging studies
were intricately woven into the process of presenting a patient for
admission.
Because the different units are physically separate, each relies on
updates to computer displays for status information and for
communication with other units (e.g., Fairbanks et al., 2007). Computer
status displays are extremely poorly designed so that it is easy to miss
updates such as new lab results or changes in transfer status, especially
in the multi-task environment of an ED. Direct communication by one
physician in one unit to personnel in another unit can overcome the
limits of the computer systems and aid the negotiation process in other
ways, but it is also a significant additional task that must be integrated
into ED providers work load and workflow.
These observations illustrate how neither unit, even with the
interposition of the transfer center, have a clear picture of the load on
the other unit relative to its capacity trends and how that relates to their
own capacity to meet current and anticipated future loads. This lack of
common ground (Klein et al., 2004) undermines the ability to
coordinate across units and leaves room for one unit to prioritize its
own workload and risk regardless of the state of the other unit. For
example, bed hiding, maintaining a reservation for an ICU bed when it is
not immediately needed (Cook, 2006), can occur as a defensive strategy
to handle possible future load, but this occurs at a cost to the ED if it
is experiencing crowding or its surge capacity is being challenged.
Stephens (2010) also observed boarding risk for mental health patients
in the ED. Because mental health facilities tend to have limited capacity,
patients with mental health issues can be stuck in the ED for long
periods after ED treatment is complete while awaiting transfer to in-
patient or to other facilities. The ED in the US context often becomes a
de facto buffer for mental health patients given the general low
resources available for these patients. The study observed how the ED
adapted over time to compensate. The observed ED developed a ‘Flex
unit’ as an internal buffer within its physical and team structure, and
Chapter x 5

almost all mental health patients spent some of their ED stay in the
Flex unit.
Previously, the ED at the study site had a finite number of
undifferentiated ED beds, and the option to use hallway space when
crunches arose. At the time of observation, the ED developed a set of
different mini-units with a central pool of “regular” beds coupled with
observation beds, trauma beds, Flex unit beds, and was developing the
possibility to mobilize extra space when overcrowding was near.
The Flex unit provided a number of adaptations to reduce the
workload and resource demands of managing mental health patients
including: reduced impact on the central ED’s physical bed space (Flex
Unit per-bed space is about one half of other ED beds), partially
reduced the impact of agitated or disruptive patients on other ED
patients, and a single staff person could monitor up to six patients in
the Flex unit.
The adaptation of the Flex unit produced unintended new difficulties
as well. The physical separation also impacted on number and
timeliness of physician interactions. The patients in the Flex unit easily
slipped into the background for ED practitioners as they dealt with
patients in the other parts of the ED. The ability to get support from
other personnel declined should one patient’s agitation effect others
and produce a cascade of demands for the staff in the Flex unit (and
risk to the patients and to the staff).
Patterns of Adaptation in Managing Patient Transfers
and the Risk of Patient Boarding
The data from the Stephens et al. (2011) study can be analyzed in terms
of concepts from Resilience Engineering to reveal general patterns
about human complex adaptive systems at work.
First, the work to reduce LOS/avoid patient boarding, and the work to
move patients to other units, adds to the workload on the ED which
reduces its CfM. Should patient load on the ED increase, it risks falling
behind the pace of these new demands (Cook and Rasmussen, 2005).
6 Resilience Engineering in Health Care

This is the basic adaptive system breakdown of decompensation (Woods
and Branlat, 2011) where the unit exhausts its capacity to adapt -- CfM
-- as the unit falls behind the pace of incoming demands. In this
pattern, breakdown occurs when challenges grow and cascade faster
than responses can be decided on and deployed to effect.
Units are aware, at least implicitly, of the risk of the adaptive system
breakdown of decompensation. We know this because we can observe,
as in this study, units engage in locally adaptive responses to reduce the
risk of exhausting their CfM given their appraisal of upcoming or
potential events, demands, and challenges. The study of patient
boarding found hospital units making many different adaptations to
manage their CfM and reduce the risk of decompensation for that unit.
The second basic adaptive system breakdown is working at cross purposes
which addresses the ability to coordinate different units at the same or
different echelons in the face of goal and resource conflicts (Brown,
2005; Woods and Branlat, 2011). Each unit works hard to achieve local
goals as defined for their scope of responsibility. But the adaptations
made to achieve their goals turn out to make it more difficult for other
related units to meet the responsibilities of their roles -- the adaptations
of one unit also squeezes other interdependent units. As a result of the
mis-coordination between the units, each can be responding in ways
that reduce their risk of exhausting CfM while at the same time these
responses undermine the capacity of the system defined more broadly
to meet global or long term goals that cut across all units -- for
example, in the case of boarding, maximizing patient benefits. Thus,
this risk can be described as behavior that is locally adaptive, but
globally maladaptive (Woods and Branlat, 2011).
The study of patient boarding noted many different adaptations relative
to the risk of the adaptive system breakdown of working at cross
purposes (Table 1). Moving a patient from the ED to another unit, e.g.,
to an ICU or to transfer to another facility, depends on the other unit
having the capacity to accept the patient relative to other ongoing and
upcoming demands. Some of the observed adaptations were directed
at reducing the risk of breakdowns in transfers (reducing LOS/
instances of boarding). More adaptations were observed that
Chapter x 7

attempted to reduce or cope with the impact on a unit’s CfM when
coordination across units was limited.
Table 1 groups the set of observations about adaptations that occurred
around the issue of patient boarding into three groups:
• responses that defended a unit’s CfM from being reduced by activity of
other units;
• responses that re-adjusted a unit’s own CfM by reconfiguring resources,
roles, teams, and activities within that unit’s scope of authority;
• responses that attempted to improve coordination across units by
adjusting the match of resources to demands based on the impact on
both units’ CfM.
Table 1. Patterns of Adapting to Shortfall in Capacity for Maneuver
Pattern
Tactic for Managing Capacity for Maneuver
Defend local
adaptive capacity
from the impact
of other units
(Cross-Unit)
Use authority differential to shift work to other units
Use authority differential to delay work being
transferred from another unit
Hoard and/or hide resources from other units
Manipulate image of capacity and resources to appear
worse than the actual conditions
Avoid offering resources to other units
Re-adjust local
adaptive capacity
(Within Unit)
Reorganize to reduce resource consumption
Handle exceptions differently from standard workflow
Relax constraints on work quality
Add additional capacity to do work
Coordinate
adaptive capacities
across units
(Cross-Unit)
Build common ground to manage interactions
between CfM of different units
Create a unit to facilitate interactions across multiple
units
Create a centralized authority that considers CfM from
a broader perspective that includes multiple units
8 Resilience Engineering in Health Care

The findings in Table 1, first, highlight that units work to create,
manage and sustain their capacity for maneuver (CfM). Units are able
to recognize how events can reduce their CfM and respond by
defending their CfM from the impact of actions by other units. Units
also respond to recurring challenges by reconfiguring their internal
resources, structure, tactics, and functions to preserve CfM.
Both of these kinds of responses are rather limited. Local re-
adjustments are just that -- local and limited by the resources under
control of that unit. Defensive responses preserve one unit’s CfM, but
only as a result of moving the pressure on CfM to other parts or levels
of the network. When one shifts to a broad perspective, these two
classes of responses are notable for their restricted or even or
counterproductive impact.
One goal of work on managing complexity has been to identify
heuristics or principles for coordinating across multiple units in an
interdependent or polycentric network -- what layered network design
or polycentric governance principles reduce the risk of working at cross
purposes? (Alderson and Doyle, 2010; Ostrom, 1999; Dietz et al.,
2003). The tactics observed in the study of patient boarding (row 3 of
Table 1) fall short of the latest thinking on coordination in a network
of adaptive units (Dietz et al., 2003; Ostrom 2003; Bengtsson et al.,
2003; Andersson and Ostrom, 2008).
A few of the research findings on layered networks and polycentric
governance are particularly relevant to the case of patient boarding.
When faced with symptoms of the adaptive system breakdown of
working at cross purposes, many try to move to more centralized or
command architecture. As in this case, such a change has limited
impact (Dietz et al., 2003).
Adding new units to the network with the purpose of facilitating
interactions (the transfer unit) has some appeal but results are
disappointing, as in this case (such units seems to add new burdens as
they also smooth out other aspects of cross-unit coordination. The
design principles for such a unit have yet to be determined.
Chapter x 9

Mechanisms to directly support horizontal coordination using
computer mediated interactions are promising (Smith et al., 2010), but
the computer tools for building such common ground are quite weak in
the observed case.
Notable by its extensive absence in the study of patient boarding are
signs of reciprocity across units (Ostrom 2003). Ostrom’s work has
highlighted the critical role of reciprocity in effective polycentric
governance of networks. Reciprocity is a form of trust that is critical
to carry out joint, interdependent activity in multi-agent, networked
systems. The research shows that the parties involved in joint activity
enter into a “Basic Compact,” i.e., an agreement (often tacit) to
facilitate coordination, work toward shared goals, and prevent
coordination breakdowns (Klein et al., 2004). In a reciprocal
cooperative relationship, unit 1 shows “trust” for unit 2 by taking an
action that gives up some amount of immediate benefit in return for a
longer run benefit for both, but in doing so unit 1 relies on unit 2 to
“reciprocate “in the future by taking an action that gives up some
benefit to make both units better off. The observed low reciprocity in
the study of patient boarding indicates that there is a significant
opportunity for re-designing the relationships and computer mediated
communications to build reciprocity and coordination in the face of
demand/resource crunches.
Overall, recurring patient boarding and long lengths of stay (LOS) in an
ED is a symptom of complexity. The ED has been recognized as a
critical natural laboratory to study the dynamics of brittleness,
resilience, and adaptation (Wears and Perry, 2006; Wears et al., 2008).
Basic concepts in the emerging theory of adaptive systems such as
capacity for maneuver and the patterns of adaptive system breakdown
(decompensation and working at cross purposes) help explain how
boarding develops and how different units locally adapt to cope.
But boarding in the ED is a tangible and significant problem for patient
centered care. The value of modeling patient boarding as a symptom of
complexity depends on the degree to which the model generates and
measures the effectiveness of new solutions to resolve the risk for
patients.
10 Resilience Engineering in Health Care

References
Andersson, K. P. and Ostrom, E. (2008). Analyzing decentralized
resource regimes from a polycentric perspective. !Policy Science,
41, 71-93.
Anders, S., Woods, D., Wears, R., Perry, S., & Patterson, E. (2006).
Limits on adaptation: modeling resilience and brittleness in hospital
emergency departments. In E. Rigaud & E. Hollnagel (Eds.)
Second Symposium on Resilience Engineering. Juan-les-Pins,
France, Nov. 8-10, 2006.
ACEP Boarding Task Force (2008). Emergency Department
Crowding: High-Impact Solutions. April, 2008. Available at: http://
www.acep.org/content.aspx?id=32050. Accessed November 5,
2013.
Alderson, D. L. and Doyle, J. C. (2010). Contrasting views of
complexity and their implications for network-centric
infrastructures. IEEE Systems, Man and Cybernetics, Part A, 40(4),
839-852.
Bengtsson, J., Angelstam, P., Elmqvist, T., Emanuelsson, U., Folke, C.,
Ihse, M., Moberg, F. and Nyström, M. (2003). Reserves, Resilience
and Dynamic Landscapes. Ambio, 32(6), 389-396.
Brown, J. P. (2005). Key themes in healthcare safety dilemmas. In M. S.
Patankar, J. P. Brown, & M. D. Treadwell (Eds.), Safety Ethics: Cases
from Aviation, Healthcare, and Occupational and Environmental Health
(pp. 103-148). Adelshot, UK: Ashgate.
Chalfin, D. B., Trzeciak, S., Likourezos, A., Baumann, B. M., &
Dellinger, R. P. (2007). Impact of delayed transfer of critically ill
patients from the emergency department to the intensive care unit.
Critical Care Medicine, 35(6),1477-1483.
Committee on the Future of Emergency Care in the US (2006).
Hospital-based Emergency Care: At the Breaking Point. National
Academic Press, Washington, DC.
Cook, R. I. (2006). Being bumpable: consequences of resource
saturation and near-saturation for cognitive demands on ICU
practitioners. In D. D. Woods & E. Hollnagel (Eds.), Joint Cognitive
Systems: Patterns in Cognitive Systems Engineering. (pp. 23–35). Boca
Raton, FL: Taylor & Francis/CRC Press.
Chapter x 11

Cook, R. and Rasmussen, J. (2005). “Going Solid”: A model of system
dynamics and consequences for patient safety. Quality and Safety in
Health Care, 14, 130-134.
Cook, R. I., Woods, D. D. and Miller, C. (1998). A Tale of Two Stories:
Contrasting Views of Patient Safety. Chicago, National Patient Safety
Foundation. (available at http://csel.eng.ohio-state.edu/blog/
woods/archives/000030.html ).
Dietz, T., Ostrom, E. and Stern, P. C. (2003). The Struggle to Govern
the Commons. Science, 302, 1907-1912 (Dec 12).
Fairbanks, R. J., Bisantz, A. M., & Sunm, M. (2007). Emergency
department communication links and patterns. Annals of
Emergency Medicine, 50(4), 396–406.
Hoot NR, Aronsky D. (2008). Systematic review of emergency
department crowding: causes, effects, and solutions. Annals of
Emergency Medicine, 52, 126–136. [PubMed]
Klein
Ostrom, E. (1999). Coping with Tragedies of the Commons. Annual
Review of Political Science, 2, pp. 493—535.
Ostrom, E. (2003). Toward a behavioral theory linking trust, reciprocity,
and reputation. In E. Ostrom & J. Walker (Eds.), Trust and
reciprocity: Interdisciplinary lessons from experimental research.
New York: Russell Sage Foundation.
Perry S. J. and Wears, R. L. (2012). Underground adaptations: case
studies from health care. Cognition, Technology & Work, 14(3),
253-260.
Pines J. M., Localio, R., Hollander, J. E., Baxt, W. G., Lee, H., Phillips,
C. and Metlay, J. P. (2007). The impact of ED crowding measures
on time to antibiotics for patients with community-acquired
pneumonia. Annals of Emergency Medicine, 50(5): 510-516.
Smith M. W., Giardina, T. D., Murphy, D. R., Laxmisan, A. and Singh,
H. (2013). Resilient Actions in the Diagnostic Process and System
Performance. BMJ Quality & Safety. 2013 Jun 27. [Epub ahead of
print]. PMID: 23813210.
Smith, P.
Stephens, R. J. (2010). Managing the Margin: A Cognitive Systems
Engineering Analysis of Emergency Department Patient Boarding.
Doctoral Dissertation, The Ohio State University.
12 Resilience Engineering in Health Care

Stephens, R. J., Cudnik, M., & Patterson, E. (2011). Barriers and
Facilitators to Timely Admission and Transfer of Patients from an
Emergency Department to an Intensive Care Unit. In Proceedings
of the Human Factors and Ergonomics Society 55th Annual
Meeting, September, 55, 763-767.
Viccellio, A., Santora, C., Singer, A. J., Thode Jr, H. C., & Henry, M. C.
(2009). The Association Between Transfer of Emergency
Department Boarders to Inpatient Hallways and Mortality: A 4-
Year Experience. Annals of Emergency Medicine, 54(4), 487-491.
Wears, R. L. and Woods, D. D. (2007). Always Adapting. Annals of
Emergency Medicine, 50(5), 517-519.
Wears, R. L. and Perry, S. J. (2006). “Free fall” – a case study of
resilience, its degradation, and recovery in an emergency
department. In E. Rigaud & E. Hollnagel (Eds.) Second
Symposium on Resilience Engineering. Juan-les-Pins, France,
November 8-10, 2006.
Wears, R. L., Perry, S. J., Anders, S., & Woods, D. D. (2008). Resilience
in the emergency department. In E. Hollnagel, C. P. Nemeth, & S.
W. A. Dekker (Eds.), Resilience Engineering: Remaining Sensitive to the
Possibility of Failure. (pp. 143-158). Aldershot, UK: Ashgate.
Woods, D. D. (2006). Essential characteristics of resilience. In E.
Hollnagel, D. D. Woods, & N. Leveson (Eds.), Resilience Engineering:
Concepts And Precepts (pp. 19–30). Adelshot, UK: Ashgate.
Woods, D. D. (2009). Escaping Failures of Foresight. Safety Science,
47(4), 498-501.
Woods, D. D., & Branlat, M. (2011). Basic Patterns in How Adaptive
Systems Fail. In E. Hollnagel, J. Pariès, D. D. Woods & J. Wreathall
(Eds.), Resilience Engineering in Practice (pp. 127-144). Farnham, UK:
Ashgate.
Woods, D. D., & Hollnagel, E. (2006). Joint Cognitive Systems: Patterns in
Cognitive Systems Engineering. Boca Raton, FL: Taylor & Francis/CRC
Press.
Woods, D. D. and Wreathall, J. (2008). Stress-Strain Plot as a Basis for
Assessing System Resilience. In E. Hollnagel, C. Nemeth and S. W.
A. Dekker, eds., Resilience Engineering: Remaining sensitive to the
possibility of failure. Ashgate, Aldershot, UK, pp. 145-161.
Woods, D. D., Chan Y.-J., and Wreathall, J. (2013). The Stress-Strain
Model Of Resilience Operationalizes The Four Cornerstones Of
Chapter x 13

Resilience Engineering. In Proceedings of the Fifth Resilience
Engineering Symposium. Soesterberg, The Netherlands, 24-27 June
2013.
Woods, D. D. (2013). Presidential Talk. Fifth Resilience Engineering
Symposium. Soesterberg, The Netherlands, 25 June 2013.
14 Resilience Engineering in Health Care