Baylor University Medical Center Proceedings April 2014, Volume 27, Number 2 | Page 10

obtained on a patient from any source. The term “positive screen”
is used for patients who had a positive nasal PCR screen denoting
MRSA colonization, but not an active MRSA infection.
All inpatient hospital admissions from October 1st to
September 30th for 2007 to 2008 (before cohort) and 2010
to 2011 (after cohort) were analyzed. The 12-month interval
between cohorts was included to allow full implementation
across all units of the hospital. All patients in the after cohort underwent nasal MRSA PCR screening under the criteria
mentioned above. Patients with a previous MRSA infection,
as well as those readmitted with a previous positive screen for
MRSA, were excluded. Patients who initially screened negative
and subsequently developed an MRSA infection were considered to have HA-MRSA-I. A subset of patients undergoing a
wide array of surgical procedures across multiple specialties in
the before and after cohort were analyzed separately. These included general surgical, gynecologic, orthopedic, cardiothoracic,

transplant, oral-maxillofacial, plastics, and urologic procedures.
International Classification of Diseases, Clinical Modification,
ninth revision (ICD-9) procedure codes were used to identify
procedures by those specialties. Procedures were included in
the analysis if at least 50 were performed during the study period. Logistic regression analysis was performed to determine
which factors contributed to predicting HA-MRSA-I and how
HA-MRSA-I affected patient mortality. P values <0.05 were
considered significant.
Our primary endpoint was the incidence of HA-MRSA-I.
The secondary endpoint was the impact of HA-MRSA-I on
patient mortality. Multivariate regression techniques were used
to elucidate predictors of these outcomes.

RESULTS
There were 36,244 patients in the before cohort, of whom
16,740 underwent a surgical procedure, and 36,068 in the
after cohort, of whom 15,044 underwent a
surgical procedure. Patient characteristics are
Table 1. Patient characteristics for all patients and patients in the surgical group
shown in Tables 1 and 2.
The overall MRSA infection rate was
Combined
Before
After
1.3% (453 patients) in the before cohort and
All patients
3.2% (1136 patients) in the after cohort (P <
N
72,312
36,244
36,068
0.001). When patients with previous MRSA
infection or colonization were excluded, those
Age at admit, mean ± SD (years)
54.2 ± 19.2
54.1 ± 19.1
54.2 ± 19.3
with HA-MRSA-I decreased from 1.2% to
Male
29,465 (41%)
15,020 (41%)
14,445 (40%)
0.87% (Table 3). Similarly, there was an inFemale
42,837 (59%)
21,216 (59%)
21,621 (60%)
crease in the rate of overall MRSA infection
Race
in the surgical group, from 1.4% to 2.3%,
White
43,215 (60%)
22,609 (63%)
20,606 (57%)
along with a decrease in HA-MRSA-I, from
African American
18,473 (26%)
8,712 (24%)
9,761 (27%)
1.4% to 1.0% (Table 3).
Hispanic
8,324 (12%)
3,770 (10.5)
4,554 (12.7%)
For the total patient population, including both before and after cohorts, patients in
Asian
705 (1%)
289 (0.8%)
416 (1.2%)
the surgical group, African Americans, and
Other
1047 (1.5%)
431 (1.2%)
616 (1.7%)
patients with an increased length of stay had
Length of stay, median (IQR) (days)
3.8 (2.3, 6.8)
3.8 (2.2, 6.8)
3.8 (2.3, 6.9)
an increased likelihood of developing HADischarged alive
69.761 (97%)
34,917 (96%)
34,844 (97%)
MRSA-I. Female patients and patients in the
Died
2,551 (4%)
1,327 (4%)
1,224 (3%)
after cohort were less likely to develop HASurgical group
MRSA-I (Table 4). When only patients in the
after cohort were analyzed, the results were
N
31,784
16,740
15,044
similar, with surgery, African American race,
Age at admit, mean ± SD (years)
50.7 ± 18.5
51.1 ± 18.2
50.3 ± 18.8
and length of stay associated with an increased
Male
11,667 (37%)
6,416 (38%)
5,251 (35%)
risk of HA-MRSA-I, and female sex associFemale
20,109 (63%)
10,317 (62%)
9,792 (65%)
ated with a decreased risk (Table 4).
Race
With regards to the likelihood of screenWhite
21,021 (67%)
11,460 (70%)
9,561 (64%)
ing positive for MRSA colonization, African
American race and increased age were assoAfrican American
5,820 (19%)
2,901 (18%)
2,919 (20%)
ciated with a higher likelihood of screening
Hispanic
3,705 (12%)
1,727 (11%)
1,978 (13%)
positive, while patients in the surgical group,
Asian
373 (1.2%)
151 (1%)
222 (1.5%)
female patients, and Hispanic patients were
Other
473 (1.5%)
183 (1.1%)
290 (2%)
less likely to screen positive (Table 5). HALength of stay, median (IQR) (days)
3.6 (2.3, 6.8)
3.7 (2.3, 6.7)
3.5 (2.3, 6.9)
MRSA-I was associated with a higher morDischarged alive
31,239 (98%)
16,430 (98%)
14,809 (98%)
tality for both the entire patient population
Died
545 (1.7%)
310 (2%)
235 (2%)
and surgical group in both cohorts (Table 5).
There was no significant difference in the overSD indicates standard deviation; IQR, interquartile range.
all mortality of patients with HA-MRSA-I

84

Baylor University Medical Center Proceedings

Volume 27, Number 2