The Journal of TRAUMA Injury, Infection, and Critical Care

Development and Implementation of a Clinical Pathway for

Severe Traumatic Brain Injury
Todd W. Vitaz, MD, Laura McIlvoy, RN, MSN, George H. Raque, MD, David Spain, MD, and
Christopher B. Shields, MD
Background: Clinical pathways
(CPs) have been shown to be beneficial in
optimizing patient care and resource use.
Methods: A multidisciplinary CP for
the treatment of severe traumatic brain
injury (Glasgow Coma Scale score of 37)
was developed. Data from these patients
(group I) were collected prospectively and
compared with a retrospective database
(group II).

Results: There were a total of 119

patients managed in conjunction with the
CP and 43 patients in the control group.
No statistical differences were found between the groups in age, Glasgow Coma
Scale score at 24 hours, or Injury Severity
Scores. There was a significant decrease in
the length of hospital stay, intensive care
unit stay, and length of ventilator support
in the study group (group I: 22.5, 16.8,

and 11.5 days, respectively; group II: 31.0,

21.2, and 14.4 days, respectively; p <
0.03).
Conclusion: The use of this CP
helped to standardize and improve patient
care with fewer complications and a potential cost savings of approximately
$14,000 per patient.
Key Words: Clinical pathway, Severe
traumatic brain injury.
J Trauma. 2001;51:369 375.

he care of patients with severe traumatic brain injuries

(TBIs) is complex and demanding, requiring the integration of skills from numerous different specialties. These
patients often have prolonged hospitalizations, which may be
marked by numerous complications. In an attempt to provide
a consistently high standard of care as well as maintaining
cost efficiency to all severely impaired TBI patients, we
developed a multidisciplinary clinical pathway (CP) to help
guide their care.1 4
Clinical pathways are a recent trend in medicine through
which multidisciplinary guidelines are developed for use in a
specific patient population. The pathway establishes a framework from which further treatment decisions are derived.
Such pathways do not limit physician decision-making capabilities, but help to coordinate patient care, thus improving
overall quality of care.2,3
Our traumatic brain injury clinical pathway (TBI-CP)
was developed by a team of physicians, nurses, and members
of ancillary services who are frequently involved in the care
of patients with TBI. The goal of the pathway was to stan-

Submitted for publication August 8, 2000.

Accepted for publication March 29, 2001.
Copyright 2001 by Lippincott Williams & Wilkins, Inc.
From the Departments of Neurological Surgery (T.W.V., G.H.R.,
C.B.S.) and General Surgery (D.S.), University of Louisville School of
Medicine, and University Health Care (T.W.V., L.M., G.H.R., D.S., C.B.S.),
Louisville, Kentucky.
Presented at the Kentucky Chapter of the American College of Surgeons, Trauma Resident paper competition, October 20, 1998, Louisville,
Kentucky; American College of Surgeons, Kentucky Chapter Annual Meeting, May 7, 1999, Louisville, Kentucky (2nd place resident research competition); and the American Association of Neurological Surgeons Annual
Meeting, April 24, 1999, New Orleans, Louisiana.
Address for reprints: Todd W. Vitaz, MD, 210 E. Gray Street, Suite
1102, Louisville, KY 40202; email: tvitaz@niky.com.

Volume 51 Number 2

dardize patient care and use of ancillary services to provide

all patients with the highest standard of care.

PATIENTS AND METHODS

The pathway was designed to standardize the routine
treatment options of patients with severe TBI (Glasgow
Coma Scale [GCS] score of 37). Secondary goals of the
pathway were to improve the care that these patients received, provide early mobilization, decrease the rate of complications, decrease the length of stay, and limit costs. All
patients were treated at the University of Louisville Hospital,
Louisville, Kentucky. Patients treated with the assistance of
the pathway since its inception in October 1995 were included in the pathway group (n 119).
Standard CPs are time dependant, with patients progressing from one stage to the next at a given time in their hospital
course. However, this TBI-CP was time independent, allowing patients to progress from one stage to the next at individual rates once predetermined goals have been met. This
prevents patients from deviating from the pathway when the
time spent in one stage is prolonged. The TBI-CP was designed to facilitate patient care and improve communication
between services without controlling physician decision
making.
Data for the pathway group (group I) were collected
prospectively and compared with a historical control group
(group II) that included retrospective data on 43 patients with
similar injuries treated over the previous year (October 1994
to October 1995). Only patients who survived their injuries
were included in either group. The GCS score at 24 hours
after injury was used to assess the level of neurologic dysfunction. This time course was chosen to ensure that adequate
postresuscitation evaluations were used and to exclude any
cases of transient neurologic dysfunction. The patients over369

The Journal of TRAUMA Injury, Infection, and Critical Care

all systemic injuries were graded using the Injury Severity
Score (ISS).5 All data were analyzed using Students t test.

Pathway Development
The initial step in designing the pathway involved the
formation of a multidisciplinary task force to review the
current practice patterns and to identify any areas where
improvement may have been necessary. This group also suggested possible changes or modifications in treatment plans
that would help correct these problems. Special emphasis was
given toward changes that had been proven to be effective in
other clinical scenarios or reported in the literature. This team
consisted of at least one member from each of the various
clinical services including trauma surgery, neurosurgery, orthopedic surgery, and members from ancillary services such
as nutrition, physical therapy, occupational therapy, and pharmacy. In addition, there were also representatives from hospital administration and both clinical and administrative nursing services.
At the beginning of the process, major areas of interest
were studied independently in an attempt to isolate areas
where the greatest fluctuation or variation existed. However,
as this process proceeded, all of the issues were combined
into the CP. Each department was then asked to approve the
final version of the CP. In an attempt to ensure the continued
participation of all of the specialties, the task force continued
meeting on a regular basis to review the progress of the CP
and to make any necessary modifications to the plan.

Severe Traumatic Brain Injury Clinical Pathway

The CP is broken down into four different treatment
phases (Fig. 1). All emergency care and resuscitation was
performed in conjunction with members of the trauma surgery, emergency medicine, and neurologic surgery services
according to Advanced Trauma Life Support and the Brain
Trauma Foundations recommendations.6,7
Phase 1
Phase 1 consists of admission to the intensive care unit
(ICU), where a standardized order sheet is used to ensure that
all the appropriate laboratory, radiographic studies, and consultations are ordered (Fig. 2). This helps to coordinate follow-up computed axial tomographic scans with other radiographic procedures, thus minimizing the number of times that
each patient needs to be transported from the ICU to the
radiology department. In addition, standards for the stepwise
treatment of intracranial hypertension, ventilator management (target PCO2 ranges), and volume management (determined by pulmonary capillary wedge pressure) are established. Suggestions for the use of antiepileptics, and
medications to control pain and agitation are also included in
an attempt to limit variations between physicians and on-call
coverage, standardize patient care, and provide clear guidelines for the nursing staff. This process also ensures early
participation of ancillary services such as physical and occu370

pational therapy, dietary services, and social services. Discharge planning is initiated at the time of admission so that
the appropriate paperwork for insurance coverage or government aide may be initiated if necessary. This facilitates placement in rehabilitation facilities as soon as the patient is
medically stable, thus avoiding long delays while posthospital placement is being arranged.
Phase 2
Phase 2 encompasses the acute critical care stage. Patients undergo continued treatment and stabilization of their
injuries. In addition, other nonlife-threatening injuries are
addressed at this time, with the appropriate treatment measures being instituted. Spine radiographs are completed with
a target of postinjury day (PID) 1 to prevent secondary
injuries and allow early mobilization. Patients who fail to
show improvement in their neurologic status (GCS score
8) undergo ventriculostomy placement. Intracranial hypertension was treated with a combination of ventricular drainage,
mannitol, and/or hyperventilation in accordance with the
Brain Trauma Foundation treatment guidelines.7 The actual
management changed throughout the course of the study.
Initially, patients were treated with fluid restriction and mild
hyperventilation (PCO2, 30 35 mm Hg), but these standards
were eventually abandoned for management with euvolemia
and normocapnia (PCO2, 35 40 mm Hg). In the first 2 years
of the study, patients were treated on the basis of intracranial
pressure (ICP) recordings; however, in the third year, there
was a change toward treatment on the basis of both ICP and
cerebral perfusion pressure results.
Patients who are ventilator dependent over the preceding
several days undergo tracheostomy (target PID 4) to provide
a secure airway and facilitate management of pulmonary
secretions and ventilator weaning. The diagnosis of pneumonia was determined by the treating physician on the basis of
a combination of the following criteria: hyperthermia, hypoxia, tachypnea, leukocytosis, changes on chest radiograph,
and culture results from bronchoalveolar lavage or tracheal
secretions, which required antibiotic administration for at
least 5 days. Finally, enteral feeding is usually started at the
end of the second phase, with a target of PID 3. Whenever
possible, percutaneous endoscopic or open gastrostomy tubes
were inserted in conjunction with the tracheostomy procedure
in an attempt to limit the number of anesthetic procedures.
Phase 3
Phase 3 focuses on ventilator weaning, which is handled
through a standardized slow wean protocol that is instituted
once the patients intracranial pressure and any pulmonary
injuries are stabilized. Under this protocol, the ventilator
settings are adjusted by the respiratory therapist according to
predetermined guidelines and monitored by the treating physician. In addition, rehabilitation efforts are intensified, with
mobilization being maximized and patients gotten out of bed
into cardiac chairs at least two times per day. Any outstandAugust 2001

Clinical Pathway for Severe TBI

Fig. 1. Severe traumatic brain injury clinical pathway.

ing medical or surgical issues are addressed; invasive lines

and Foley catheters are removed as soon as possible, and
patients are usually transferred out of the intensive care unit
to a step-down unit or regular floor. Patients who are not
receiving at least 50% of their nutritional goal through tube
feedings by PID 5 are considered candidates for total parenteral nutrition.
Phase 4
Phase 4 is the prerehabilitation phase. During this period,
physical therapy/rehabilitation is continued. Transfer plans
are finalized and the pathway concludes with transfer of the
Volume 51 Number 2

patient to a rehabilitation facility, nursing home, or home

with family members.

RESULTS
There were a total of 119 patients (group I) managed in
conjunction with the TBI-CP compared with 43 patients in
group II. There were no complications that could be directly
related to the use of the CP. No statistical difference was
found between the groups in age, GSC score at 24 hours, or
ISS (Table 1).
To evaluate the overall effectiveness of the CP, we
documented the length of treatment for three categories:
371

The Journal of TRAUMA Injury, Infection, and Critical Care

Fig. 2. Standardized order sheet.

total hospital length of stay, intensive care unit length of

stay, and number of ventilator days. There was a signifi-

Table 1 Patient Demographics

Age (y)
ISS
GCS at 24 h
GCS at discharge
Change in GCS

372

Group I (n 119)

Group II (n 43)

34.1 16.2
24.5 8.2
6.4 1.1
11.3 3.0
5.0 3.1

32.9 14.9
26.8 8.6
6.3 1.0
11.5 3.2
5.2 2.8

cant decrease in each of these parameters in the group

treated in conjunction with the pathway guidelines (group
I) (Table 2).
Ninety-seven percent (115 of 119) of the patients in
group I underwent tracheostomy versus 88% in group II. The
average PID ( SD) that the procedure was performed on
was 4.3 2.1 (median, 4) and 6.3 3.4 (median, 6) for
groups I and II, respectively. The overall incidence of pneumonia was remarkably high in both groups, but there were
fewer total episodes of pneumonia in patients managed in
conjunction with the CP (Table 3). The pathway also helped
August 2001

Clinical Pathway for Severe TBI

Table 2 Outcome: Length of Stay

Group II (n 43)

p Value

22.5 8.5
20
16.8 7.4
16
11.5 5.8
10

31.0 19.7
27
21.2 9.3
19
14.4 7.5
12

0.009

Hospital days
Median
ICU daysa
Median
Ventilator daysa
Median
a

Table 4 Mortality Statistics

Group I (n 119)

0.007
0.026

Average SD.

improve the time for initiation of enteral nutritional support

(group I, 3.8; group II, 4.7; p 0.03).
Finally, a series of subset analysis was completed. First,
group I was subdivided on the basis of the type of treatment
for intracranial hypertension. Group A (83 patients) consisted
of patients in the first portion of the study who were treated
on the basis of ICP recordings with routine use of hypovolemia and hyperventilation. Group B (36 patients) contained
those whose treatment was determined on the basis of both
ICP and cerebral perfusion pressure recordings with normocapnia and euvolemia. There was no significant difference in
patient demographics or length of stay parameters between
these two subsets; however, both were still significantly less
than group II. In addition, there was no significant difference
between groups A and B regarding change in GCS score from
24 hours after admission to discharge (group A, 5.2 3.3;
group B, 4.6 2.6).
The second subset analysis compared patients with isolated head injuries (group C; n 57) with patients with
closed head injury associated with multisystem trauma (group
D; n 62). Once again there was no significant difference in
the demographics, outcome parameters, or changes in GCS
score (group C, 5.0 3.6; group D, 5.0 2.6) between these
groups.

Mortality Statistics
Patients who died as a result of their injuries were not
included in the above analysis and are listed here separately.
The mortality rates were 47% (107 of 226) and 39% (27 of
70), respectively, for groups I and II (p 0.05). The demographics from those patients not surviving their injuries and
the entire population (deceased and survivors) are listed in
Tables 4 and 5, respectively. The average age for patients
who did not survive their injuries was 47 years (group I) and

Table 3 Outcome: Pneumonia

Pneumonia (%)
Total episodes
Number patients with
pneumonia
PID 03 (%)
PID 47 (%)
PID 7 (%)

Volume 51 Number 2

Group II (n 43)

p Value

97 (82)
152 (1.3)

42 (98)
88 (2.0)

0.0008

19 (44)
20 (47)
30 (70)

Group II (%)

107 (47)
47
35 (33)
65 (61)
28 (26)
14 (13)
79 (35)

27 (39)
38
5 (19)
21 (78)
4 (15)
2 (7)
23 (34)

DNR, do not resuscitate.

a
Delayed mortality: patients who received maximal care and
lived at least 48 h from admission but died prior to hospital discharge.
b
Adjusted mortality: overall mortality rate minus patients who
were DNR/comfort care.

38 years (group II), which was higher than the average age of
patients who survived (34 years for group I; 33 years for
group II). When cases where the family desired only comfort
care (group I, 28; group II, 4) were excluded, the adjusted
mortality rates were 35% for group I and 34% for group II (p
0.05) (Table 4).

DISCUSSION
The development of CPs has been a recent advancement
in the medical field. Previous pathways have shown that a
standardized routine of patient care not only improves patient
outcome, but also helps contain costs.1 4 We have replicated
those results with our TBI-CP. This pathway creates a systematic time-independent framework to ensure that all aspects of patient care are addressed. In addition, CPs have
been shown to help avoid inconsistency of treatment and
duplication of efforts by different caregivers.4 Use of CPs has
also been found helpful in assisting with nursing education as
well as providing an additional communication channel with
patients families. This helps keep family members informed
about anticipated therapies or interventions and potential
complications.5
Unlike other pathways, our TBI-CP is derived from a
time-independent framework. This accounts for the extreme
variability that exists in patients experiencing TBIs and allows for progression from one phase to the next at individual
rates. The CP acts as a facilitator to ensure that the appropriate ancillary services and diagnostic studies that are common
to most patients in this population, such as radiographic
clearance of the cervical, thoracic, and lumbar spine, are

Table 5 Demographics of All Patients (Survivors and

Group I (n 119)

41 (35)
41 (35)
50 (42)

Overall mortality
Average age of dead patients (y)
Dead patients 60 years old
Death in first 48 h
DNR/comfort care only
Delayed mortalitya
Adjusted mortalityb

Group I (%)

Dead)
Age
ISS
GCSa

Group I (n 226)

Group II (n 70)

p Value

40.2 20.1
24.5 8.9
5.4 1.6

34.7 18.2
27.7 8.1
6.2 1.1

0.001
0.004
0.002

a
GCS at 24 h unless patient died during first 24 h; then, admission GCS was used.

373

The Journal of TRAUMA Injury, Infection, and Critical Care

Table 6 Estimated Charge Reduction Analysisa

ICU days saved per patient
TCUa days saved per patient
Ventilator days saved per patient
Rehabilitation services days saved per patient
Total charge reduction per patient
Total charge reduction over entire study (119 patients)

No. of Days

Cost per Day ($)

Total ($)

4.4
4.1
2.9
8.5

1,484
1,017
510
279

6,530
4,170
1,479
2,372
14,551
1,731,569

TCU, transitional care unit.

a
On the basis of 1999 charges.

performed in a timely fashion. In addition to helping to

coordinate patient care and improve communication between
the numerous clinical services, this concept also allows physicians to maintain autonomy and individualize patient care.
The main focus of our TBI-CP was early involvement of
all the appropriate services and early mobilization. However,
a majority of these patients have significant respiratory compromise or require prolonged mechanical ventilation secondary to a decreased level of consciousness. We believe that
early tracheostomy and enteral feeding are extremely important components of this treatment strategy. Tracheostomy
ensures a safe airway, and facilitates pulmonary toilet, ventilator weaning, and patient mobilization. Previous studies
have shown that patients with similar injuries who require
mechanical ventilation and do not show evidence of weaning
in the first several days benefit from early tracheostomy.8 10
Early enteral feeding provides nutritional support and helps
counteract and prevent the negative affects of malnutrition
such as immune suppression.1114
Our severe TBI-CP has been exceptionally useful as
illustrated by our decreased length of stay and decreased total
episodes of pneumonia. This improved outcome was unrelated to method of treatment of intracranial hypertension or
association with other injuries, as was shown by the subset
analysis. However, it is difficult to decipher the exact causeand-effect relationship when performing this type of analysis.
We feel that the decreased length of mechanical ventilation,
decreased ICU length of stay, and decreased pneumonia incidence were attributed to the goals that were incorporated
into the CP. In addition, increased coordination of patient
care alone should diminish wide fluctuations in patient care
that would eliminate any extreme outliers and thus further
improve these parameters. On the other hand, it is impossible
to rule out the effect from other factors such as changes in
ventilator management with possible differences in the rates
of barotrauma, aspiration, or variations in pulmonary and
chest wall injury patterns.
One factor that may have affected the overall change in
hospital stay would have been a change in the availability of
rehabilitation beds in our community. This may have slightly
decreased overall length of stay throughout the study period,
but should not affect the length of ICU stay or number of
ventilator days. In addition, many of these patients are uninsured, which may also delay their transfer to an appropriate
374

rehabilitation facility. We have found that the early involvement of social services and discharge planning helps to minimize some of these delays; however, there are still instances
where discharge may be significantly prolonged.
Unfortunately, a change in hospital management precluded us from performing an actual cost comparison. Therefore, we conducted a projected charge analysis comparing
these two groups and found a $14,551 per patient charge
reduction in the CP patients (group I) (Table 6). This simple
analysis only includes standard charges and does not take into
account other variable charges such as medications, laboratory or radiographic studies, or any type of interventions.
We recognize the limitations inherent to the comparison
between historical controls and the prospective study group
given the different time periods in which each group was
treated as well as the multifactorial nature of such injuries.
However, the comparison was useful to show the general
trend toward improved outcomes with the use of this CP. In
addition, these comparisons only included patients who survived their injuries. There was a slightly higher overall incidence of death after the initiation of the CP. This was attributed to a substantial increase in deaths in elderly patients that
were the result of withdrawal or limitation of care. If these
types of deaths are excluded, the mortality rates between the
two time periods are comparable.
Another advantage of such pathways was that they in
essence standardize the general aspects of patient care, thus
preventing any major fluctuations in treatment. Therefore,
studies that incorporate CPs into their design may prevent
wide variations in patient care, therefore limiting the number
of variables that are being manipulated.
Finally, we found that through the current database
design and continuous reevaluation of our treatment patterns we were able to isolate other areas of ineffective or
costly care. An example of this was the development of our
phenytoin policy. Originally, patients who were started on
phenytoin were placed on doses determined by the admitting physician, with monitoring of blood levels at random
times as determined by that physician. However, we converted to a weight-based dosing regimen that was monitored at predetermined times by one of our clinical pharmacists. This change resulted not only in fewer phenytoin
(Dilantin) levels being drawn but also in a higher rate of
initially therapeutic levels. The physician retained the auAugust 2001

Clinical Pathway for Severe TBI

tonomy to determine which patients were treated and for
what duration, but the micromanagement of this decision
(drug levels) was predetermined and monitored by the
clinical pharmacist, thus freeing the physician of the associated hassles and burdens.
In conclusion, we feel that use of such CPs are important in
the treatment of patients with TBI. The CP facilitates patient
progression and communication between the various caregivers.
This limits duplication of workloads and also decreases length of
stay parameters and complications, therefore limiting costs.

10.

REFERENCES

11.

1.

2.
3.
4.
5.

Aronson B, Maljanian R. Critical path education: necessary

components and effective strategies. J Contin Educ Nurs. 1996;
27:215219.
Ibarra VA. Spine update: clinical pathways. Spine. 1997;22:352357.
Kitchiner D, Bundred P. Integrated care pathways. Arch Dis Child.
1996;75:166 168.
Zevola DR, Raffa M, Brown K, et al. Clinical pathways and
coronary artery bypass surgery. Crit Care Nurse. 1997;17:20 33.
Baker SP, ONeill B, Haddon W, et al. The Injury Severity Score: a
method for describing patients with multiple injuries and evaluating
emergency care. J Trauma. 1974;14:187296.

Volume 51 Number 2

6.

7.
8.

9.

12.

13.

14.

American College of Surgeons. Committee on Trauma. Advanced

Trauma Life Support Course. Chicago: American College of
Surgeons; 1998.
Florin R, Jagoda A, Kelly JP, et al. Guidelines for the Management
of Severe Head Injury. Chicago: Brain Trauma Foundation; 1995.
Armstrong PA, McCarthy MC, Peoples JB. Reduced use of
resources by early tracheostomy in ventilator-dependent patients with
blunt trauma. Surgery. 1998;124:763766.
Johnson SB, Kearney PA, Barker DE. Early criteria predictive of
prolonged mechanical ventilation. J Trauma. 1992;33:95100.
Ross BJ, Barker DE, Russell WL, Burns RP. Prediction of long-term
ventilatory support in trauma patients. Am Surg. 1996;62:19 25.
Borzotta AP, Pennings J, Papasadero B, et al. Enteral versus
parenteral nutrition following severe closed head injury. J Trauma.
1994;37:459 468.
DAmelio LF, Hammond JS, Spain DA, Sutyak JP. Tracheostomy
and percutaneous endoscopic gastrostomy in the management of the
head-injured patient. Am Surg. 1994;60:180 185.
Moore FA, Haenel JB, Moore EE, et al. Percutaneous tracheostomy/
gastrostomy in brain-injured patients: a minimally invasive
alternative. J Trauma. 1992;33:435 439.
Spain DA, DeWeese RC, Reynolds MA, Richardson JD.
Transpyloric passage of feeding tubes in patients with head injuries
does not decrease complications. J Trauma. 1995;39:1100 1102.

375