Matukaitis

10.1177/1062860604274382
Decision Tree
et al
Approach to Appropriate Admissions

Appropriate Admissions
to the Appropriate Unit:
A Decision Tree Approach
Joanne Matukaitis, RN, MSN, CNA, BC
Paula Stillman, MD, MBA
Elizabeth Wykpisz, RN, MSN, MBA, CNAA, BC
Edward Ewen, MD

An intermediate care decision tree tool was devel-
oped to meet the demand for intermediate care beds.
Concurrently, a charging process was developed to
support the acuity adaptable model of care, allowing
the patient to remain in the same bed from admission
to discharge, regardless of level of care required, ad-
justing nurse-to-patient ratios as acuity changes.
Since beginning this pilot, 96% to 100% of the patients
admitted to intermediate care from the emergency de-
partment met the criteria. Wait time from request to
admission was reduced from 5.5 hours to 2.5 hours. A
reduction in nursing costs was noted. The average
number of patients waiting daily in the emergency de-
partment for an intermediate care bed has been re-
duced by approximately 80%. A significant difference
in length of stay was not noted. (Am J Med Qual 2005;
20:90-97)
Keywords: intermediate care; admissions; decision tree;
acuity adaptable; intensive care
Ms Matukaitis is the director, Patient Care Services, Critical
Care; Dr Stillman is the senior vice president for special projects;
Ms Wykpisz is the former vice president of cardiovascular ser-
vices; and Dr Ewen is the director of Clinical Informatics, Disease
Management, and Case Management at the Christiana Care
Health System, Newark, Delaware. The authors have no affilia-
tion with or financial interest in any product mentioned in this ar-
ticle. The authors’ research was not supported by any commercial
or corporate entity. Corresponding author: Joanne Matukaitis,
Christiana Care Health System, 4755 Ogletown-Stanton Road,
Room 2539B, Newark, DE 19713 (e-mail: jmatukaitis@
christianacare.org).
American Journal of Medical Quality, Vol. 20, No. 2, Mar/Apr 2005
DOI: 10.1177/1062860604274382
Copyright © 2005 by the American College of Medical Quality
Christiana Care Health Systems, located in New-
ark, Delaware, is the largest health system in the state.
Included in the system are 2 acute care hospitals:
Christiana Hospital, with a total of 650 beds, and
Wilmington Hospital, with a total of 150 beds. There
are a total of 94 medical intermediate care (IMC) beds
between both hospitals, which are designed to provide
an intermediate level of care that requires more skill
than the general floor beds and less than the intensive
care unit (ICU) beds. The levels of care are the
following:
• ICU: the highest acuity patient requiring a 1:1 or
1:2 nurse-to-patient ratio;
• IMC unit: the next level after ICU and requiring a
1:4 or 1:5 nurse-to-patient ratio;
• floor monitored: general medical and/or surgical
patients who are telemetry monitored, requiring
a 1:6 nurse-to-patient ratio; and
• floor nonmonitored: general medical and/or sur-
gical patients who are not telemetry monitored
and with a level of care that requires a 1:6 or 1:7
nurse-to-patient ratio.
PURPOSE
The IMC beds were in great demand but short sup-
ply. Most of the IMC patients were admitted from the
emergency department (ED). The average wait time
for a bed was 5.5 hours. An audit revealed that, on av-
erage, only 42% of the patients admitted to an IMC
bed from the ED actually met basic InterQual crite-
ria.1 Thus, our nursing resources were being overused
in terms of nurse-to-patient ratios. There was a need

90

Downloaded from ajm.sagepub.com at WAYNE STATE UNIVERSITY on April 6, 2015

AMERICAN JOURNAL OF MEDICAL QUALITY
to better match patient acuity to our nursing
resources.
In addition, there were patients who had to be
transferred because they were not initially placed in
the appropriate bed. The transfers occurred as early
as 1 hour after admission because the patient should
have been admitted to the ICU. We also transferred
patients to a floor level of care when they no longer
met IMC criteria. The cost to transfer patients from
the IMC to a medical/surgical floor was estimated to
be $200 per patient transfer. We averaged 77 trans-
fers per month at a cost of $15 400 per month. All pa-
tients in the IMC were automatically charged the
IMC rate regardless of the level of care required, as
this was the charge for a room on that unit.
Although IMC admission criteria had been written
years ago, they were kept on the units in a policy man-
ual and not widely used. This allowed physicians to
subjectively interpret the need for IMC admissions.
Also, some physicians favored the nursing staff on the
IMC and admitted to those units whether or not that
level of care was required. All of the nurses on the
IMC unit were certified in advanced cardiac life sup-
port and arrhythmia interpretation. They were
trained to administer specific intravenous medica-
tions as designated by hospital policies. The nurses on
the medical/surgical units did not have this level of
knowledge.
Our other challenge was finding a billing method-
ology that would allow us to modify the accommoda-
tion code, which is used to charge for the skill level re-
quired, without changing the patient’s physical
location. Our original charging methodology was by
the unit type, not by the bed. Specific units were iden-
tified as IMC and charged the IMC rate with no
methodology available to change the rate as the level
of care changed. For example, if a patient was admit-
ted to a double room as an IMC patient, but then the
level of his or her care changed to medical/surgical, we
could not modify the accommodation code unless the
patient was moved to a room on a unit designated as
medical/surgical. We needed a methodology that pro-
vided the capability to charge for the appropriate
level of intensity of care regardless of the bed the pa-
tient occupied. This would allow the patient to stay in
the same bed from admission to discharge. We agreed
to test this concept on the IMC units and then expand
it throughout the institution by training all nurses in
the advanced skills necessary to provide an IMC level
of care.
The purpose of our pilot was to
Downloaded from ajm.sagepub.com at WAYNE STATE UNIVERSITY on April 6, 2015
Decision Tree Approach to Appropriate Admissions 91
1. decrease the inappropriate use of the IMC beds so
they would be available for patients who required
that level of care,
2. promote greater clarity concerning those conditions
and circumstances that required IMC admission,
3. provide an easy tool for the ED and other caregivers
to use when considering the level of care required by
each patient, and
4. redesign our billing methodology to charge by indi-
vidual bed, not by unit type.
REVIEW OF THE LITERATURE
A review of the literature to search for other insti-
tutions that had developed a decision tree for appro-
priate admissions to various levels of patient care was
completed. The keywords used included admission,
step-down unit, step-down admission, and intermedi-
ate care admissions. The review revealed a paucity of
available literature.
Lenox and New did report a 13% Medicare denial
rate at a loss of approximately $20 000 per month
2
after opening a clinical observation unit. The loss was
due to missed opportunities for admissions, inappro-
priate admissions, and extended length of stay past
the normal 48 to 72 hours that is customary for an ob-
servation unit. The facility adopted measures to im-
prove the appropriateness of admissions to the beds
and implemented corrective measures, which led to a
0% denial rate and steady increases in revenue.
Clarke and Normile reported a study designed to
examine the length of stay in the ED prior to admis-
3
sion to a critical care unit. The purpose of the project
was to assess (1) if holding critical care patients in the
ED after admission was due to nursing shortages or
lack of resources in the ED and (2) if delays in admis-
sion orders or tests occurred. A Likert-type scale sur-
vey was sent to directors of critical care and emer-
gency services areas. A total of 109 responses were
received. The responses revealed a positive correla-
tion between an increased length of stay in the ED
and delay in treatment. In addition, 81% strongly
agreed or agreed that there is a lack of space, staff,
and available critical care expertise in the ED. They
reported that the critical care units lacked clear ad-
mission criteria and assigned physician coverage. The
inappropriateness of admissions resulted in use of ex-
pensive technology and a high use of nursing re-
sources, leading to unnecessary expenses that
increased health care costs without improving
quality.
92 Matukaitis et al
METHODOLOGY AND IMPLEMENTATION
A multidisciplinary team was formed to develop an
IMC decision tree to be used by the ED as a guide for
admissions to the IMC units. The team included phy-
sicians and nurses from the ED and IMC units, ad-
ministration, information services, the finance depart-
ment, and the utilization management department.
The decision tree was developed based on InterQual
criteria and “Guidelines for ICU Admission, Dis-
charge and Triage,” as published by the Society of
4
Critical Care Medicine (Figure 1). The draft was re-
viewed and approved by the Critical Care Committee,
Cardiology Section, IMC nursing leadership, ED phy-
sicians, and physician leaders. Clear definitions were
included. The guideline was formatted as a checklist
for ease of use.
A physician champion in the ED instructed other
ED physicians how to use the tool as a guide for ad-
mission decisions, with the understanding that the
admitting physician could override the guidelines
based on appropriate and reasonable clinical judg-
ment. The ED nurses also were trained, with particu-
lar emphasis on the triage nurses, so that they could
assist with the appropriate assignment of patient
transfers/admissions. The IMC charge nurses and
staff also were educated about the criteria. The IMC
staff performed weekly audits to monitor admission
appropriateness.
Working with Information Services, charges were
changed from being room based to bed based. Accom-
modation codes were assigned to provide a charge for
IMC, general floor, and general floor monitored.
Using the approved decision tree, each evening, the
charge nurse reviews patients’ charts to ensure the
codes are correct and makes changes as needed. The
revised form is then faxed to our bed board, where the
accommodation code is entered into our information
services network for charging.
Our pilot initially focused on 2 units at Christiana
Hospital, 2C and 5C. The 2C unit is a 41-bed IMC unit
with a primary focus on cardiovascular patients.
However, about 30% of the admissions are
noncardiac, medical patients with other disease enti-
ties such as gastrointestinal bleed, renal disease, and
diabetes. Cardiac telemetry is housed on the 5C unit.
Both of these IMC units were modified to function
as acuity adaptable units to reduce patient transfers
by keeping the patient in the same room from admis-
sion through discharge. The literature revealed sev-
eral institutions that use this model for patients ad-
Downloaded from ajm.sagepub.com at WAYNE STATE UNIVERSITY on April 6, 2015
AMERICAN JOURNAL OF MEDICAL QUALITY
mitted to the critical care unit. Hendrich et al de-
scribed how this was accomplished at Methodist Hos-
pital, part of Clarion Health Partners, Inc, in India-
5
napolis, Indiana. Patients admitted to the critical
care unit remained in the same room until discharge.
Adaptable equipment was changed as the patient’s
condition improved. Rather than moving the patient,
adjustments were made to the nurse-to-patient ratio
based on the patient’s condition. This resulted in re-
duced transfers, efficient use of the budgeted nursing
care hours, and a reduced length of stay.
We could not find literature describing institutions
that initiated this model on an IMC unit. Because
only 42% of the patients on the IMC unit actually met
criteria on admission, we decided to redesign the car-
diac telemetry unit (5C), which originally was en-
tirely IMC, to be a more flexible unit. We budgeted
nursing care hours for 8 IMC beds, with the remain-
ing being floor monitored or nonmonitored beds for
medical patients. This reduced our daily nursing
hours of care from 9.2 to 8.5. Unit 2C remained en-
tirely IMC. The IMC decision tree was initially pi-
loted on 5C, and revisions were made prior to expand-
ing the pilot to 2C. The charge nurses of both 2C and
5C were given guidelines to help them decide how to
adjust staffing, depending on the level of care re-
quired. Intermediate care patients generally required
a 1:4 nurse-to-patient ratio; medical patients re-
quired a ratio of 1:6. This approach was taken to re-
duce nursing care hours required by allocating staff
based on the needs of the patients, rather than
traditional nurse-to-patient ratios based on the type
of unit.
Outcome Measurements
The following outcome measures were used to eval-
uate the success of this pilot:
1. appropriateness of admission from the ED to IMC
beds,
2. average time (in hours) from request for an ED pa-
tient to be admitted to an IMC bed compared to the
actual time the patient arrived on the IMC unit (we
felt this was a good measure as delays were docu-
mented by the coordinators daily as “patients wait-
ing in the ED for an IMC bed”; we knew delays were
not related to transport as the ED has dedicated staff
for transfers),
3. average length of stay on the IMC unit,
4. nursing care hours and dollars on units 2C and 5C,
and
AMERICAN JOURNAL OF MEDICAL QUALITY Decision Tree Approach to Appropriate Admissions 93

Start Nursing Interventions at least

every 2 hours for > 8 hours
(Examples)
Yes
Ÿ continuous pulse oximetry or

Intermediate Care
Yes
Ÿ arterial catheter
Ÿ neuro checks
Frequent Nursing Ÿ complex wound care
interventions ? Ÿ glucose measurement

No

vital signs at least hourly for up or

No Yes
to 4 hours

Hemodynamic hemodynamic
Yes Yes
problems? instability or lability?

No

level B medication or level C

Intermediate Care
No medication approved for or
Yes
Intermediate Care use
(see reverse side for list)

Ÿ dysrhythmia that is uncontrolled or

Cardiac Rhythm
Yes Ÿ requires IV drug suppression or
problems?
Ÿ cardioversion or or
Yes

ICU
Ÿ epicardial temporary pacing or
Ÿ is non-sustained VT or
Ÿ other potentially fatal dysrhythmia

No
No
transvenous temporary pacing Yes

Ÿ new S-T T wave changes or

or
AMI ? Yes Ÿ elevated troponin or Yes
Ÿ unstable coronary syndrome

No

Ÿ SaO2 < 92% and FiO2 > 50% or

Respiratory or
Yes Ÿ continuous CPAP, BiPAP or Yes
Intermediate Care
Insufficiency ?
Ÿ hypercapnia with pH < 7.35

No
Ÿ GCS < 15 or
Ÿ neurological deficit or
Ÿ liver or spleen laceration or or
Acute Trauma ? Yes Yes
Ÿ complex pelvic fracture or
Ÿ r/o blunt myocardial injury with
dysrhythmia and/or elevated troponin or
Ÿ pneumothorax without chest tube
No

Ÿ bleeding controlled through intervention or

Acute GI bleeding ? Yes and Yes
Ÿ hemodynamically stabile

No
No

Floor

Figure 1. Guidelines for intensive care unit admission, discharge and triage.

Downloaded from ajm.sagepub.com at WAYNE STATE UNIVERSITY on April 6, 2015

94 Matukaitis et al AMERICAN JOURNAL OF MEDICAL QUALITY

100%

90%

80%

70%

60%
Percentage Met

50%

40%

30%

20%

10%

0%
3

4
03

03

03

03

03

04
00

00

00

00

00
20

20

20

20

20

20
/2

/2

/2

/2

/2
6/

4/

5/

2/

2/

9/
19

/5

26

25

29
/1

/1

/2

/1

/2

3/
12
9/

1/

3/

3/
10

11

11

12

12

of
of

of

of

of

of
of

of

of

of

of

k
ee
k

k
ee

ee

ee

ee

ee
k

k
ee

ee

ee

ee

ee

W
W

W
W

% Met IMC Criteria

Figure 2. Appropriate admissions to intermediate care (IMC) from the emergency department.

5. number of patients waiting daily in the ED for an weeks after initiating the pilot, this wait time was re-
IMC bed. duced to 2.5 hours. Although we had a slight increase
in wait time in early 2004, we reeducated the staff and
RESULTS immediately saw a reduction as evidenced in Figure
3. We have continued to see a decline in wait time
The pilot was begun in October 2003. Data were (Figure 3).
collected from May 2003 through September 2003 There was an overall reduction of 2025 hours of
and revealed that only 42% of the patients admitted care from October to November for both units com-
to either of the 2 IMC units met IMC criteria. This im- bined after initiating the pilot (Figure 4). This repre-
proved immediately and has continued (Figure 2). sents a cost reduction in regularly paid hours of ap-
Since beginning this pilot, 96% to 100% of the patients proximately $56 862 (based on an average hourly pay
admitted to IMC from the ED meet the criteria. rate of $28.08; Figure 5). These reductions were not
The appropriateness of level of care also was moni- sustained but remained within budget. This can be
tored 48 hours after admission. Initially, there was explained by the fact that prior to initiating the
poor compliance with changing the accommodation project, we staffed for an IMC level of care regardless
status; approximately 60% of the patients were inap- of patient needs. Immediately after we initiated the
propriate for IMC 2 days after admission, and the sta- project, our hours of care dropped significantly be-
tus had not been changed. Through education of the cause the percentage of patients that actually met
nursing staff, there has been a gradual improvement IMC criteria was low. This was not sustained be-
to 96% appropriateness at the 48 hour postadmission cause, rapidly, 96% to 100% of the patients were true
mark. IMC patients requiring higher nurse-to-patient
The average time from bed requested to the patient ratios.
being admitted to the IMC unit from the ED was 5.5 We initially saw a significant drop in the number of
hours in October 2003. In November, approximately 3 patients waiting daily in the ED for a bed in the IMC

Downloaded from ajm.sagepub.com at WAYNE STATE UNIVERSITY on April 6, 2015

AMERICAN JOURNAL OF MEDICAL QUALITY Decision Tree Approach to Appropriate Admissions 95

5
Hours

0
June 2003 Aug 2003 Oct 2003 Nov 2003 Jan 2004 Feb 2004 Mar 2004 May 2004 June 2004 July 2004
Date

Avg. Time (Hours)

Figure 3. Average wait from emergency department bed request to admission to intermediate care.
Nursing Care Hours

12 000
10 000
8000
6000
4000
2000
0
Se 003
Au 03

Ja 03
ov 3
03

ec 3

04
ar 4
Fe 04

Ap 04

Ju 004
ay 4
0

00
20

20
20
20

20

20
20
20

20
2

2
r2
g
ly

ct
pt

ne
b
n
Ju

M
D
N

Month

Figure 4. Average nursing care hours on a monthly basis for units 5C and 2C.

after an admission decision was made from a total of Barriers

97 in October 2003 to 39 in November 2003. This did
fluctuate temporarily in December 2003; however, as As with any new project, we experienced some ini-
indicated in Figure 6, it has continued to decrease. We tial challenges. Our first barrier was building a com-
feel this was a positive result of having the appropri- fort level for the physicians and nurses so they did not
ate admission criteria available. feel we were trying to control their practice. Sharing
We did not see a significant change in the length of the guidelines with the ED staff was an essential part
stay on either unit (Figures 7 and 8). We do feel this is of the process. During the first few weeks, there were
an opportunity to pursue in the future. daily morning meetings, reviewing patients admitted

Downloaded from ajm.sagepub.com at WAYNE STATE UNIVERSITY on April 6, 2015

96 Matukaitis et al AMERICAN JOURNAL OF MEDICAL QUALITY

$500 000
$400 000

Dollars
$300 000
$200 000
$100 000
$0

03
03

Ja 03
N 003
ct 3

ec 3

04
ar 4
Fe 04

A p 04

Ju 004
M 004
0

0
20
20

20
20

20

20
20
20

20
2

2
r2
g
y

pt

ov

ne
b
n

ay
l
Au
Ju

O
Se

M
D
Month

Figure 5. Average nursing care dollars on a monthly basis for units 5C and 2C.

120

100

80
Census

60 Census

40

20

0
Oct 2003 Nov 2003 Dec 2003 Jan 2004 Feb 2004 Mar 2004 April 2004
Date

Figure 6. Intermediate care patients waiting daily in the emergency department for an appropriate bed.

to the IMC from the ED the previous day to discuss
appropriateness. Within several weeks, IMC admis-
sion criteria were met 96% to 100% of the time. In ad-
dition, the beds were more readily available when
needed as demonstrated by the reduction in wait time
for IMC beds. Anecdotally, the IMC nurses told us
they felt more satisfied because the patients admitted
to the unit belonged in the unit and because they now
had a concrete tool to use when discussing reasons a
patient should not be admitted to the unit or should
have a change in the level of care.
Our greatest challenge has been changing the way
we charge for level of care. The current system has
been in place since 1991. Our institution’s reimburse-
ment source is approximately 65% Medicare and 35%
commercial payor or self-pay. The finance depart-
ment initially felt that we could not change from
charging by room to charging for a bed because it may

Downloaded from ajm.sagepub.com at WAYNE STATE UNIVERSITY on April 6, 2015

AMERICAN JOURNAL OF MEDICAL QUALITY
Figure 7. Average length of stay, unit 5C.
IMC = intermediate care.
Figure 8. Average length of stay, unit 2C.
IMC = intermediate care.
lead to denials or investigations if the codes were in-
correct. We tested our new methodology prior to final-
izing it to ensure that all patients were coded appro-
priately by working with the utilization management
and the unit nurse managers. Our test revealed our
daily codes were correct. There also were concerns
that appropriate charging would affect revenue or in-
Downloaded from ajm.sagepub.com at WAYNE STATE UNIVERSITY on April 6, 2015
Decision Tree Approach to Appropriate Admissions 97
crease denials. Our initial pilot revealed we were
charging appropriately 100% of the time.
An unexpected outcome was the difficulty in imple-
menting an acuity adaptable model. We quickly found
that this would not work well unless a majority of the
floors used this model. We reduced transfers, but
were not able to eliminate them because of the need to
vacate beds on IMC to admit true IMC patients. If all
the units in the hospital (except ICU) provide the
same level of skilled nursing, we could admit IMC pa-
tients throughout the hospital. We discovered that we
needed to train the entire cadre of medical and surgi-
cal unit nurses to be able to care for IMC level of care
patients. Then the patient could remain in the same
room regardless of level of care required.
Future Implications
As we move forward, our goal is to expand the deci-
sion tree model for admission criteria to all levels of
care. We are planning to take the tool to the physician
offices so that direct admissions will be assigned to
the correct unit and level of care. We are developing a
plan that will empower the nurse to change the level
of care from IMC to floor or floor monitored based on
the criteria. We will continue our goal of developing
an acuity adaptable model throughout the
institution.
REFERENCES
1. McKesson Health Solutions, LLC. InterQual Level of
Care–Adult. Available at: www.interqual.com.
2. Lenox AC, New H. Clinical observation units help manage costs
and care. Health Finance Manage. 1997;51(4):88-89.
3. Clarke K, Normile LB. Delays in implementing admission or-
ders for critical care patients associated with length of stay in
emergency departments in six mid-Atlantic states. J Emerg
Nurs. 2002;28:489-495.
4. Society of Critical Care Medicine. Guidelines for ICU admis-
sions, discharge and triage. Critical Care Medicine.
1999;27:633-638.
5. Hendrich AL, Fay J, Sorrells AK. Effects of acuity-adaptable
rooms on flow of patients and delivery of care. Am J Crit Care.
2004;13(1):35-45.