Improved outcomes with routine respiratory therapist evaluation of non-intensive-care-unit surgery patients.

Abstract

Respiratory therapist (RT) driven protocols decrease ventilator days and resource utilization in the intensive care unit (ICU). Protocols have been studied in non-ICU settings, but their effect on mortality has been incompletely studied.
In our neurosurgery step-down, trauma/surgery step-down, and trauma/surgery general units we initiated an RT-driven evaluate-and-treat protocol that included a standardized, quantitative, RT-driven patient-assessment scale and protocolized interventions. Before and after initiation of the protocol we collected data on non-ICU patients at risk for pulmonary complications.
The patient groups before (n = 2,230) and after (n = 2,805) protocol initiation were well matched in age, sex, Charlson score, and admitting service. Most of the patients, whether assessed by a physician or an RT, were deemed to have low risk of pulmonary complications and did not require any respiratory treatments. The number of respiratory treatments increased after protocol initiation, but the patients who received respiratory treatments after protocol initiation had shorter ICU stay and hospital stay, and lower total hospital costs than those who received respiratory treatments before protocol initiation. There was a nonsignificant trend toward lower mortality after protocol initiation.
Our RT-evaluate-and-treat protocol for non-ICU surgery patients was associated with more patients receiving respiratory treatments but decreased ICU and hospital stay and lower total hospital costs. Routine RT-driven assessment of non-ICU patients may reduce pulmonary complications in high-risk patients.

Full text

Page 1

Improved Outcomes With Routine Respiratory Therapist Evaluation
of Non-Intensive-Care-Unit Surgery Patients
Brian G Harbrecht MD, Edgar Delgado RRT, Raymond P Tuttle RRT,
Mark H Cohen-Melamed RRT, Melissa I Saul MSc, and Cynthia A Valenta RN MSN CNRN
BACKGROUND: Respiratory therapist (RT) driven protocols decrease ventilator days and re-
source utilization in the intensive care unit (ICU). Protocols have been studied in non-ICU settings,
but their effect on mortality has been incompletely studied. METHODS: In our neurosurgery
step-down, trauma/surgery step-down, and trauma/surgery general units we initiated an RT-driven
evaluate-and-treat protocol that included a standardized, quantitative, RT-driven patient-assess-
ment scale and protocolized interventions. Before and after initiation of the protocol we collected
data on non-ICU patients at risk for pulmonary complications. RESULTS: The patient groups
before (n ? 2,230) and after (n ? 2,805) protocol initiation were well matched in age, sex, Charlson
score, and admitting service. Most of the patients, whether assessed by a physician or an RT, were
deemed to have low risk of pulmonary complications and did not require any respiratory treat-
ments. The number of respiratory treatments increased after protocol initiation, but the patients
who received respiratory treatments after protocol initiation had shorter ICU stay and hospital
stay, and lower total hospital costs than those who received respiratory treatments before protocol
initiation. There was a nonsignificant trend toward lower mortality after protocol initiation. CON-
CLUSIONS: Our RT-evaluate-and-treat protocol for non-ICU surgery patients was associated with
more patients receiving respiratory treatments but decreased ICU and hospital stay and lower total
hospital costs. Routine RT-driven assessment of non-ICU patients may reduce pulmonary compli-
cations in high-risk patients. Key words: atelectasis, outcomes, respiratory care, pulmonary diseases,
postoperative care, hypoxia. [Respir Care 2009;54(7):861–867. © 2009 Daedalus Enterprises]
Introduction
Pulmonary therapy is important for patients at risk for
atelectasis and pneumonia from immobility or inadequate
mobilization of pulmonary secretions. Aggressive pulmo-
nary care is particularly important for surgery patients,
because postoperative incisional pain can decrease inspira-
tory effort, alter normal respiratory mechanics, and in-
crease the potential for postoperative pulmonary compli-
cations. Many post-surgery patients require medications
that can depress the respiratory drive and further interfere
with pulmonary function. Atelectasis from inadequate in-
spiratory effort is a common complication following any
operation that requires general anesthesia, and is a com-
mon cause of postoperative hypoxemia.1Post-surgery pul-
monary complications consume substantial health-care re-
sources, can produce substantial morbidity, and can even
lead to death.
The rising interest in evidence-based medicine has re-
sulted in efforts to standardize care through the develop-
ment of clinical protocols. Protocols decrease the time
required to wean patients from mechanical ventilation, re-
duce the risk of ventilator-associated pneumonia, and im-
prove glucose control in the intensive care unit (ICU).2-6
At the time of this study Brian G Harbrecht MD was affiliated with the
Department of Surgery, University of Pittsburgh Medical Center, Pitts-
burgh, Pennsylvania. Edgar Delgado RRT, Raymond P Tuttle RRT, and
Mark H Cohen-Melamed RRT are affiliated with the Department of
Respiratory Care; Melissa I Saul MSc is affiliated with the Department
of Biomedical Informatics; and Cynthia A Valenta RN MSN CNRN is
affiliated with the Department of Nursing, University of Pittsburgh Med-
ical Center, Pittsburgh, Pennsylvania.
The authors have disclosed no conflicts of interest.
Correspondence: Brian G Harbrecht MD, Department of Surgery, Uni-
versity of Louisville School of Medicine, 550 South Jackson Street,
Louisville KY 40292. E-mail: briang.harbrecht@louisville.edu.
RESPIRATORY CARE • JULY 2009 VOL 54 NO 7861

Page 2

ProtocolizedcarehasalsobeenstudiedoutsidetheICU.7-15
Many of the studies on protocols in non-ICU settings eval-
uated protocol compliance or the number of respiratory
treatments provided, whereas relatively few have rigor-
ously analyzed the impact of protocols on patient-specific
outcomes such as admission to the ICU or mortality.7-15
Though it is clear that protocolized respiratory care can
decrease inappropriate application of respiratory services
and improve utilization,16the effect on patient mortality is
difficult to measure.7,12,15We hypothesized that a respira-
tory therapist (RT) driven evaluate-and-treat protocol for
non-ICU surgery patients would reduce respiratory com-
plications and improve patient outcomes.
Methods
This investigation was approved by the University of
Pittsburgh Medical Center-Presbyterian quality-improve-
ment committee as a performance-improvement project,
and was exempt from the University of Pittsburgh institu-
tional review board.
We implemented the protocol in July 2003. The proto-
col focused on early identification of patients at risk for
pulmonary complications and early provision of pulmo-
nary therapy and bronchodilator, if needed, to high-risk
patients, to decrease pulmonary complications and ICU
admissions. The protocol was implemented as a trial in the
neurosurgerystep-downunit,thetrauma/surgerystep-down
unit, and the trauma/surgery general ward, where non-ICU
patients at greatest risk for pulmonary complications are
concentrated. The control group consisted of the patients
admitted to those units during the 8 months prior to the
initiation of the protocol (November 2002 through June
2003). The treatment group consisted of the patients ad-
mitted to those units during the 8 months after the initia-
tion of the protocol (July 2003 through February 2004).
Prior to the initiation of the protocol, respiratory treat-
mentswereperformedasdirectedbyphysicianorder,based
on the physician’s assessment of the patient’s risk of pul-
monary complications and need for specific respiratory
treatments. RTs could recommend treatments, but only
treatments with specific physician orders were imple-
mented.
The protocol involved RT evaluation (without the need
for a physician order) of every patient admitted to the
neurosurgery step-down, trauma/surgery step-down, and
trauma/surgery general units. The RT used a standardized
patient-assessment tool to quantify the severity of pulmo-
nary risk factors and guide the frequency of intervention
(Fig. 1).9,16The patient assessment included pulmonary
history, surgical history, chest radiograph, respiratory pat-
tern, breath sounds, cough, mental status, activity level,
laboratory data, and pulse-oximetry readings. In each as-
sessment category the RT assigned a score from zero (nor-
mal) to 4 (marked abnormality). Based on the patient-
assessment cumulative score the RT selected one of the
protocol treatment groups (Fig. 2) and developed and im-
plemented an individualized care plan that specified the
respiratory treatments and the frequency of the treatments.
Treatments that could be implemented by RTs included
positive expiratory pressure (PEP), oscillatory PEP, incen-
tive spirometry, chest percussion, bi-level positive airway
pressure, airway suctioning, intermittent positive-pressure
breathing, and aerosol bronchodilator (via metered-dose
Fig. 1. Assessment form for our respiratory-therapist-driven evaluate-and-treat protocol. All categories are assessed by the respiratory
therapist, and an objective score is generated. RR ? respiratory rate. CXR ? chest radiograph.
RESPIRATORY THERAPIST EVALUATION OF NON-INTENSIVE-CARE-UNIT SURGERY PATIENTS
862RESPIRATORY CARE • JULY 2009 VOL 54 NO 7

Page 3

inhaler or nebulizer). The RT evaluated the effectiveness
of the first treatment, and subsequent treatments, if needed,
were performed by the “house” RT, who provided respi-
ratory care in several units. The house RT could modify
the care plan as necessary, per the protocol, the patient’s
response to therapy, and the patient’s general status, and
the physician reviewed important changes to the care plan
and patient status. Respiratory treatments were stopped
when deemed no longer beneficial. The RT met daily with
the unit charge nurse to discuss patient needs and com-
municate assessment findings and the care plan. The phy-
sician could write orders for respiratory care, but such
orders were not required prior to RT evaluation and im-
plementation of the care plan. If medications were indi-
cated, the relevant information was conveyed to the treat-
ing physician and a physician order was obtained.
We collected data on demographics, admitting service,
number and type of respiratory treatments, stay, and in-
hospital mortality. We tracked ICU admissions from the
neurosurgery step-down, trauma/surgery step-down, and
trauma/surgery general units and cross-checked that data
with a database maintained by the institution’s rapid-
response team. From the medical records we recorded the
primary indication for ICU admission, in a manner blinded
to the presence or absence of the RT-evaluate-and-treat
protocol. We assessed medical comorbidities with the
Charlson index.17We obtained total hospital charges from
the hospital’s financial database, and calculated hospital
costs as previously described.18We tabulated charge codes
to determine the number and type of respiratory treat-
ments delivered. Data are presented as mean ? SD.
We performed the analysis with statistics software
(StatView 5.0, SAS Institute, Cary, North Carolina), and
we compared continuous variables with analysis of vari-
ance, with Scheffe ´’s test, and categorical variables with
the chi-square test. Differences were considered signifi-
cant when P ? .05.
Results
The neurosurgery step-down, trauma/surgery step-
down, and trauma/surgery general units admitted 2,230
patients in the control period, and 2,805 patients in the
8 months following protocol initiation. The patients ad-
mittedafterprotocolinitiationwereolderandhadaslightly,
but statistically significantly, higher mean Charlson score
(Table 1).
There were no statistically significant differences in sex,
ICU stay, or hospital stay, although after protocol initia-
tion a slightly higher proportion of patients were admitted
to a general or specialty medical service (16% versus 20%,
see Table 1). There was no significant difference in overall
mortality. Before protocol initiation 59 patients (3%) re-
quired ICU admission, compared to 93 patients (3%, dif-
ference not significant) after protocol initiation (Table 2).
There was no significant difference in the overall pro-
portion of patients who required ICU admission for respi-
ratory reasons (27% before, 33% after). Most of the rea-
sons for ICU transfer were not primarily respiratory (see
Table 2).
After protocol initiation a greater number and propor-
tion of patients received respiratory treatments (662 [30%]
vs 1,341 [48%], P ? .01 via analysis of variance). There
was no significant difference in age, sex, or Charlson score
between the patients who did and did not receive respira-
tory treatments (data not shown). In both the control and
protocolperiods,thepatientswhoreceivedrespiratorytreat-
ments were predominantly trauma/general surgery patients
(data not shown), whereas those who did not receive re-
spiratory treatments were more frequently medical patients
Fig. 2. Protocol care plan tool.
RESPIRATORY THERAPIST EVALUATION OF NON-INTENSIVE-CARE-UNIT SURGERY PATIENTS
RESPIRATORY CARE • JULY 2009 VOL 54 NO 7863

Page 4

or surgical specialty patients (data not shown). Compared
to patients who received no respiratory treatments, pa-
tients who received respiratory treatments had longer ICU
stay (2.7 ? 6.0 d vs 0.5 ? 1.4 d, P ? .001), longer total
hospital stay (7.1 ? 8.2 d vs 3.4 ? 3.5 d, P ? .001), and
higher total hospital costs ($18.4 ? 22.9 thousand vs
$8.0 ? 8.0 thousand, P ? .001). These data support the
hypothesis that patients who received respiratory treat-
ments had more severe illness.
One would anticipate that the benefit of objective pa-
tient assessment and protocolized respiratory care would
be greatest in patients at highest risk for pulmonary com-
plications. We tested that hypothesis by comparing the
patients who received respiratory treatments before and
after protocol initiation. The patients who received respi-
ratory treatments after protocol initiation were older, but
their sex and Charlson score were similar (Table 3). ICU
stay and total hospital stay were shorter and total hospital
costs were lower in the patients who received respiratory
treatments after protocol initiation. They also had nonsig-
nificantlyfewerICUadmissionsandnonsignificantlylower
overall mortality (see Table 3).
Table 4 summarizes the respiratory treatments delivered
before and after protocol initiation. Significantly more pa-
Table 2.Primary Indications for Intensive-Care-Unit Admission
Before Protocol
Initiation
(Nov 2002-
Jun 2003)
(n)
After Protocol
Initiation
(Jul 2003-
Feb 2004)
(n)
Respiratory (hypoxia, secretions,
respiratory distress)
Postoperative monitoring
Volume (bleeding, hypotension,
dehydration)
Cardiac (arrhythmia,
hypertension)
Change in neurologic status
Other
1631
17
11
30
11
43
7
4
16
2
Total number of patients
transferred to intensive care
5993
Table 4.Respiratory Treatments Delivered
Before
Protocol
Initiation
(Nov 2002-
Jun 2003)
n (%)*
After
Protocol
Initiation
(Jul 2003-
Feb 2004)
n (%)*
Patients
Incentive spirometry
Positive-pressure treatment
Pharmacologic treatment
Tracheostomy secretion management
662 1,341
777 (58)†
310 (23)‡
358 (27)†
268 (20)†
270 (41)
192 (29)
234 (35)
230 (35)
Treatments
Incentive spirometry
Positive-pressure treatment
Pharmacologic treatment
Tracheostomy secretion management
20,192
1,915 (10)
2,462 (12)
5,009 (25)
10,806 (54)
23,157
3,685 (16)‡
6,680 (29)‡
5,769 (25)
7,023 (30)‡
*Because of rounding, percentage values do not necessarily sum to 100.
†P ? 0.001
‡P ? 0.005
Table 1.Patient Demographics
Before Protocol
Initiation
(Nov 2002-
Jun 2003)
After Protocol
Initiation
(Jul 2003-
Feb 2004)
Number of patients
Age (mean ? SD y)
Male (%)
Charlson score (mean ? SD)
Stay (mean ? SD d)
Intensive care unit
Hospital
Mortality (%)
Admit service (%)
Trauma
General surgery
Specialty surgery
General medicine
Specialty medicine
2,2302,805
49.2 ? 19.0
56
0.7 ? 1.4
51.3 ? 19.8*
57
0.8 ? 1.4*
1.3 ? 4.3
4.8 ? 6.1
0.3
1.4 ? 3.9
5.0 ? 6.2
0.4
46
19
20
43
19
18
47
1213
* P ? 0.01 via analysis of variance
Table 3. Outcomes of Patients at High Risk for Pulmonary
Complications
Before
Protocol
Initiation
(Nov 2002-
Jun 2003)
After
Protocol
Initiation
(Jul 2003-
Feb 2004)
Number of patients
Age (mean ? SD y)
Male (%)
Charlson score (mean ? SD)
ICU stay (mean ? SD d)
Hospital stay (mean ? SD d)
ICU admissions (%)
Mortality (%)
Hospital cost (thousand $)
Hospital charges (thousand $)
6621,341
48.3 ? 18.7
54
0.8 ? 1.3
3.6 ? 7.2
7.8 ? 9.1
6.3
0.9
20.3 ? 25.2
105.7 ? 126.3
51.3 ? 20.0*
56
0.8 ? 1.4
2.3 ? 5.3*
6.8 ? 7.7†
5.5
0.5
17.4 ? 21.5‡
88.8 ? 111.5‡
*P ? 0.002
†P ? 0.02
‡P ? 0.01
RESPIRATORY THERAPIST EVALUATION OF NON-INTENSIVE-CARE-UNIT SURGERY PATIENTS
864RESPIRATORY CARE • JULY 2009 VOL 54 NO 7

Page 5

tients received incentive spirometry after protocol initia-
tion. The percentage of patients who received PEP treat-
ments, pharmacologic treatments, and tracheostomy
secretion-managementcarealldecreasedafterprotocolini-
tiation. In the patients who received respiratory treatments,
the mean and median number of treatments per patient
decreased after protocol initiation, although the standard
deviation was high (pre-protocol median 7 treatments/pa-
tient, pre-protocol mean ? SD 21.4 ? 50.6 treatments/
patient, protocol median 4 treatments/patient, protocol
mean ? SD 14.7 ? 35.5 treatments/patient, P ? .001 via
analysis of variance).
The total number of respiratory treatments increased
after protocol initiation (see Table 4). The proportion of
total pharmacologic treatments was unchanged, but trache-
ostomy secretion-management treatments significantly de-
creased, and the proportion of incentive spirometry and
PEP treatments increased after protocol initiation.
We also compared the patients who did not receive
respiratory treatments before and after protocol initiation
(1,568 patients and 1,464 patients, respectively). After pro-
tocol initiation, the patients who did not receive respira-
tory treatments were older (51.2 ? 19.6 y vs 49.6 ? 19.1 y)
and had a higher Charlson score (0.8 ? 1.4 vs 0.7 ? 1.4,
P ? .02), but overall mortality was low and statistically
similar (0.21% before vs 0.06% after protocol initiation,
P ? .29). The groups had similar total hospital stay
(3.3 ? 3.6 d vs 3.5 ? 3.4 d, difference not significant).
ICU stay was slightly longer (0.5 ? 1.5 d vs 0.4 ? 1.2 d,
P ? .001), and total hospital costs were slightly higher
($8.3 ? 8.7 thousand vs $7.7 ? 7.2 thousand, P ? .05)
after protocol initiation. Those differences were statisti-
cally significant but are not clinically important.
After protocol initiation there was no significant change
in the percentage of patients admitted to an ICU (1.1%
before vs 1.3% after, difference not significant) or the
percentage of patients admitted to an ICU for respiratory
reasons (3/17 admissions [18%] before vs 1/19 admissions
[5%] during the protocol) from this lower-risk group.
Discussion
RT-driven protocols decrease mechanical-ventilation
days, decrease ventilator-weaning time, promote noninva-
sive ventilation in the ICU, and are used to diagnose/treat
hypoxia.2-5,19,20Previous studies found improved appro-
priateness of respiratory treatments and better adherence
to respiratory protocols when respiratory treatments were
directed by RTs, compared to by physicians.7-15Deter-
mining whether those effects decrease morbidity or mor-
tality is difficult. Previous studies7-15either did not assess
morbidity/mortality or were limited by small numbers of
patients.
We instituted the protocol to facilitate early identifica-
tion and treatment of patients who would benefit from
respiratory treatments. In the patients who received respi-
ratory treatments, the protocol was associated with de-
creased total hospital costs and stay, and there was a trend
toward fewer ICU admissions and lower hospital mortal-
ity. The mean and median number of respiratory treat-
ments per patient decreased after protocol initiation, but
the total number of respiratory treatments increased. After
protocol initiation there was more use of incentive spirom-
etry and PEP, and fewer patients received tracheostomy
secretion-management treatments. Those data support our
hypothesis that routine assessment and treatment by RTs
in non-ICU, surgery-ward settings is associated with better
patient outcomes.
Limitations
The patient groups before and after protocol initiation
were generally comparable in age, sex, and medical co-
morbidities, but we cannot exclude the possibility of co-
morbidities or differences that might not be identified by
the Charlson score and might have impacted our results.
This was not a prospective randomized trial. We insti-
tuted the protocol to decrease ICU admissions for respi-
ratory problems in hospital units where patients at risk for
pulmonary complications are generally admitted. How-
ever, less than half the patients in the studied units, either
before or after protocol initiation, were deemed by physi-
cians or RTs to need any form of respiratory treatment.
There was no significant reduction in ICU admissions, but
that finding may be in part due to the small number of ICU
admissions from the studied units, and the small percent-
age of ICU admissions due to respiratory problems. We
began the protocol in the neurosurgery step-down, trauma/
surgery step-down, and trauma/surgery general units be-
cause they have many patients at high risk for pulmonary
complications. Patients admitted to these wards, however,
were not restricted to those specific high-risk services, as
is demonstrated by the number of admitting services listed
in Table 1. Restricting the implementation of this protocol
to specific services would be difficult, given RT staffing
by hospital location rather than by service line. Therefore,
geographical implementation of the protocol appeared to
be the most feasible.
Implementing the protocol in selected units with a fo-
cused group of dedicated RTs, nurses, and physicians al-
lowed an objective assessment of the protocol’s benefits
and disadvantages but may have limited the number of
involved patients and our ability to detect differences be-
tweengroups.Evenwiththisimplementationstrategy,how-
ever, most patients did not require respiratory treatments,
and the total number of adverse outcomes (death or ICU
admission) was low, which makes it difficult to detect
RESPIRATORY THERAPIST EVALUATION OF NON-INTENSIVE-CARE-UNIT SURGERY PATIENTS
RESPIRATORY CARE • JULY 2009 VOL 54 NO 7 865