SUPPLEMENT ARTICLE

Surgical Site Infection (SSI) Rates

in the United States, 1992–1998:
The National Nosocomial Infections

Downloaded from https://academic.oup.com/cid/article/33/Supplement_2/S69/527418 by guest on 02 April 2024

Surveillance System Basic SSI Risk Index
Robert P. Gaynes, David H. Culver, Teresa C. Horan, Jonathan R. Edwards, Chesley Richards, James S. Tolson,
and the National Nosocomial Infections Surveillance System
Hospital Infections Program, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta

By use of the National Nosocomial Infections Surveillance (NNIS) System’s surgical patient surveillance com-
ponent protocol, the NNIS basic risk index was examined to predict the risk of a surgical site infection (SSI).
The NNIS basic SSI risk index is composed of the following criteria: American Society of Anesthesiologists
score of 3, 4, or 5; wound class; and duration of surgery. The effect when a laparoscope was used was also
determined. Overall, for 34 of the 44 NNIS procedure categories, SSI rates increased significantly (P ! .05)
with the number of risk factors present. With regard to cholecystectomy and colon surgery, the SSI rate was
significantly lower when the procedure was done laparoscopically within each risk index category. With regard
to appendectomy and gastric surgery, use of a laparoscope affected SSI rates only when no other risk factors
were present. The NNIS basic SSI index is useful for risk adjustment for a wide variety of procedures. For 4
operations, the use of a laparoscope lowered SSI risk, requiring modification of the NNIS basic SSI risk index.

In recent years, increased attention has been given to SSI. A composite risk index that captures the joint in-
measuring clinical outcomes as a component of com- fluence of this and other risk factors is required before
prehensive quality assurance programs [1]. A significant meaningful comparisons of SSI rates can be made
impediment to developing meaningful hospital-ac- among surgeons, among institutions, or over time.
quired infection rates that can be used for intra- and A simple index was developed during the Study on
interhospital comparisons has been the lack of an ad- the Efficacy of Nosocomial Infection Control (SENIC)
equate means of adjusting for case mix. For surgical project [5]. In that study, highly trained data collectors
site infections (SSI), the traditional wound classification evaluated 1338,000 patient records from a probability
system, which stratifies each wound into 1 of 4 cate- sample of hospitals in the United States to calculate
gories—clean, clean-contaminated, contaminated, and infection rates. The hospitals’ surveillance and preven-
dirty-infected—has been available since 1964 [2–4].
tion/control programs were also evaluated. The SENIC
Limitations of this system of risk stratification are well
study found that hospitals with lowest nosocomial in-
recognized. One of the major problems is its failure to
fection rates had strong surveillance and prevention/
account for the intrinsic patient risk of developing an
control programs. In particular, the SENIC study de-
veloped the first risk index to aid in risk adjustment of
infections that occur after surgery and showed that the
Reprints or correspondence: Dr. Robert Gaynes, Hospital Infections Program,
Mailstop E-55, Centers for Disease Control and Prevention, 1600 Clifton Rd. NE, collection, calculation, and dissemination of surgeon-
Atlanta, GA 30333. specific SSI rates to surgeons lowered the SSI rates. The
Clinical Infectious Diseases 2001; 33(Suppl 2):S69–77
 2001 by the Infectious Diseases Society of America. All rights reserved.
SENIC study provided the strongest scientific evidence
1058-4838/2001/3305S2-0002$03.00 to date for the efficacy of the surveillance of SSIs. We

NNIS Basic SSI Risk Index • CID 2001:33 (Suppl 2) • S69

previously reported an adaptation of this risk index used by risk of developing an SSI, we calculated the Goodman-Kruskal
44 National Nosocomial Infections Surveillance (NNIS) system (G) statistic [15]. Ranging from ⫺1 to ⫹1, this nonparametric
hospitals that collected data from January 1987 through De- correlation coefficient is most useful for comparing the relative
cember 1990 under the surgical patient surveillance component predictive power of different risk factors or comparing a risk
protocol, which includes definitions of eligible patients, oper- factor with the composite index.
ations, and hospital-acquired infections [6]. The NNIS basic From 1992 through 1998, several changes were made in the
SSI risk index was a significantly better predictor of SSI risk data collection process of the NNIS system, and these changes
than was the traditional wound classification system, and it had an impact on the nature of the data available for analysis.
performed well across a broad range of operative procedures. In January 1992, patients undergoing coronary artery bypass
However, the 1990s witnessed changes in health care delivery graft procedures were categorized into those with 2 incisions
with regard to surgical procedures. Considerable numbers of (chest and donor vessel site) and those with only a chest incision

Downloaded from https://academic.oup.com/cid/article/33/Supplement_2/S69/527418 by guest on 02 April 2024

procedures are now done on an outpatient basis, and the sur- (e.g., internal mammary arteries used for grafting). In 1992,
gical patients admitted to hospitals tend to have higher intrinsic we also began to identify operative procedures done via a fiber-
risk and are often discharged earlier [7–9]. We examine the optic scope.
effect of these changes and the dramatic expansion of the NNIS In June 1994, we subdivided spinal fusion/laminectomy pro-
system on the ability of the NNIS basic SSI risk index to predict cedures into either fusion or laminectomy. In addition, ortho-
the risk of an SSI following an operative procedure. pedic procedures that involved prosthetic implants, which had
previously been combined, were divided into 3 groups: hip,
knee, or other prosthetic orthopedic procedures. In June 1994,
METHODS discharge date was required on all operative records. We also
began collecting a data field called “detected” for all SSIs. This
The surveillance protocols used by hospitals in the NNIS sys- field required NNIS hospitals to indicate whether the SSI had
tem, including the mapping of International Classification of been detected during the same hospital admission as the op-
Diseases, 9th revision, codes into the 44 NNIS operative pro- erative procedure, while performing postdischarge surveillance,
cedure categories, have previously been described [10]. Under or on readmission to the hospital. No formal postdischarge
the surgical patient surveillance component, all patients un- surveillance protocol was developed for the NNIS system, be-
dergoing operations in preselected operative procedure cate- cause no method has been shown to be optimal and the par-
gories are monitored, at least until discharge, for postoperative ticipating hospitals indicated that the most effective postdis-
infections. Standard definitions for hospital-acquired infections charge surveillance methods are likely to vary according to each
are used [11]. institution’s setting. The data we used herein were reported by
In the NNIS basic SSI risk index, each operation is scored NNIS hospitals, each of which followed the surgical patient
by counting the number of risk factors present among a patient surveillance component protocol for ⭓1 month from the pe-
having an American Society of Anesthesiologists (ASA) pre- riod from January 1992 through June 1998.
operative assessment score of 3, 4, or 5; an operation classified
as either contaminated or dirty-infected; and an operation with
duration of 1T h, where T depends on the operative procedure RESULTS
being done.
The ASA score assesses preoperatively the overall physical Data regarding 738,398 NNIS operative procedures performed
status of the patient and is itself a scoring system. Preopera- during January 1992 through June 1998, including 19,267 sub-
tively, patients are assessed and given a score that ranges from sequent SSIs, were reported from 225 NNIS hospitals. More
1 (for an otherwise healthy patient) to 5 (for a patient not than 63% of these procedures were done during the period of
expected to survive the next 24 h) [12–14]. The distribution 1995–1998.
of duration of operation for the different operative procedures The NNIS basic SSI risk index: duration of surgery. The
was determined. The 75th percentile of each distribution was cut point for duration of surgery ranged from 1 h, for appen-
identified, rounded to the nearest whole number of hours, and dectomy, limb amputation, and cesarean section, to 5 h, for
used as the cut point, T, for distinguishing between operations coronary artery bypass graft (chest and donor vessel site) and
of short and long duration. The NNIS basic SSI risk index has cardiac surgery, and 7 h, for organ transplantation (table 1).
values of 0, 1, 2, or 3. Risk categories were defined by combining From 1987–1991 to 1992–1998, the 75th percentiles for the
adjacent risk index values when no significant difference in SSI 44 NNIS operative procedure categories changed very little (!20
risk was found between them (table 1). min), except for head and neck surgery, for which the cut point
To summarize the strength of the association between a po- increased from 5 h to 7 h. The 8 procedures with changes in
tential risk factor, or the composite risk index, and a patient’s cut points were procedures for which the cut points had been

S70 • CID 2001:33 (Suppl 2) • Gaynes et al.

close to the half hour and, when rounded to the nearest integer has the potential for lowering the risk of SSI, we investigated
for the hour, had a greater chance of changing over time despite this possibility for each of the procedures identified in the
the small change in duration. previous section as having significantly lower overall rates when
Utility of the NNIS basic SSI risk index. The NNIS basic a laparoscope was used. For only 4 NNIS operative procedures
SSI index was a useful method for risk adjustment for a wide did we find it necessary to incorporate laparoscope use into
variety of procedures (table 1). Overall, for 34 of the 44 NNIS the risk index: cholecystectomy, colon surgery, appendectomy,
operative procedure categories, SSI rates increased significantly and gastric surgery. For cholecystectomy, within each of the
(P ! .05) as the number of risk factors increased, on the basis basic SSI risk index categories, the SSI rate was significantly
of Goodman-Kruskal coefficients (SEM, 0.33  0.006; P ! lower when a laparoscope was used (table 2). Moreover, as
.0001 for all procedures combined). The NNIS basic SSI risk table 2 suggests, the influence of the laparoscope could be cap-
index was a particularly useful method of risk adjustment for tured by simply subtracting 1 from the basic SSI risk index

Downloaded from https://academic.oup.com/cid/article/33/Supplement_2/S69/527418 by guest on 02 April 2024

28 of the 31 NNIS operative procedures that were not part of whenever the procedure was done laparoscopically. With this
a category designated as “other,” such as “other genitourinary modification, the risk index has values of M (or ⫺1), 0, 1, 2,
system procedures.” The 3 procedures that were not part of a or 3; the SSI rate in each category is significantly lower than
category designated as “other” for which there was no signif- the rate in the next category (table 3).
icant increase in SSI rates with increasing numbers of risk fac- Only 2.6% of colon surgery procedures were done laparos-
tors were nephrectomy, splenectomy, and limb amputation. copically. The influence of laparoscope use on SSI was the same
The NNIS basic SSI index was not as useful for NNIS op- as that for cholecystectomy. Hence, the NNIS modified SSI risk
erative procedures that combined a variety of operations. For index could be defined by subtracting 1 from the basic SSI risk
6 of the 13 combinations of procedures designated as “other,” index whenever colon surgery operations were done laparoscop-
SSI rates increased significantly (P ! .05 ) with increasing num- ically, once again yielding risk categories with values of M (or
bers of risk factors. These 6 were other cardiovascular system ⫺1), 0, 1, 2, or 3. All of the SSI rates were significantly different
procedures; other genitourinary system procedures; other ear, among the 5 risk categories for colon surgery (table 3).
nose, and throat procedures; other gastrointestinal procedures; For appendectomy, laparoscope use was high (19%). SSI rates
other musculoskeletal procedures; and other endocrine pro- were not significantly lower when the laparoscope was used
cedures. For 4 procedures, analysis suggested the need to in- within each of the basic SSI risk index categories, except for 0;
corporate an additional measure into the SSI index: the use of this was unlike the pattern for cholecystectomy and colon sur-
the laparoscope. gery. The use of a laparoscope did not call for subtracting 1
Use of the laparoscope. Since 1992, laparoscopes have been from the basic SSI risk index, except when the index was 0.
increasingly used in surgery. From 1992 through 1997 (the last Hence, we split the risk category 0 group into “0-No” and “0-
complete year for data collection), the most common proce- Yes” and otherwise ignored whether the procedure was done
dures done laparoscopically were cholecystectomy operations, laparoscopically. SSI rates with 2 or 3 of the other risk factors
with the laparoscope being used 64% of the time. Laparoscope (ASA score, wound class, or duration of surgery) did not differ
use for cholecystectomy operations increased steadily from 59% significantly, and the data were combined to form a single
in 1992 to 72% in 1997, and SSI rates were significantly lower category 2,3 (table 3). Therefore, there were 4 risk categories
when a laparoscope was used (0.6% vs. 1.8%; P ! .001). for appendectomy: 0-Yes, 0-No, 1, and 2,3, where “Yes” or “No”
Other procedures frequently done by means of a laparoscopic refers to the appendectomy having been done with or without
(or endoscopic) approach during the period of 1992–1998 were a laparoscope. For gastric surgery, a laparoscope was used
the following: appendectomy (19%), vaginal hysterectomy ∼8.5% of the time. As seen for appendectomy, the use of a
(15%), gastrectomy (8.5%), exploratory laparotomy (6.9%), laparoscope significantly reduced the SSI risk only when the
herniorrhaphy (4.5%), ventricular-peritoneal shunt (4.2%), risk category was 0, thereby yielding 4 risk categories for gastric
and colon surgery (2.6%). NNIS hospitals also reported the surgery: 0-Yes, 0-No, 1, and 2,3 (table 3).
use of a laparoscope in other procedures, but in general, the Postdischarge surveillance of SSIs. Of the 19,267 SSIs,
use was !2% of the procedures reported or the NNIS operative only 14,949 (78%) had a recorded value in the category “de-
procedure was a combination of procedures (such as “other tected,” because this variable was not collected until 1994. Of
ear, nose, or throat” surgery, with a reported laparoscope use these 14,949 SSIs, 46% were detected during the current ad-
of 14%), and the combination of procedures did not show a mission, 16% through postdischarge surveillance efforts, and
significant difference in SSI rates with laparoscope use. 38% on readmission (figure 1).
Modified NNIS basic SSI risk index for cholecystectomy, The more serious SSIs were detected before discharge or on
colon surgery, appendectomy, and gastric surgery: the impor- readmission. Of 2392 SSIs detected by infection control pro-
tance of the laparoscope. Because the use of a laparoscope fessionals at NNIS hospitals in the postdischarge outpatient

NNIS Basic SSI Risk Index • CID 2001:33 (Suppl 2) • S71

 Table 1. Surgical site infection rates, by operative procedure and risk index category, National Nosocomial Infections Surveillance system, 1992–1998.

Duration
cut point, Risk index Risk index Risk index Risk index
Operative procedure categorya hours G category n Rateb category n Rateb category n Rateb category n Rateb
CARD: Cardiac surgery 5 0.31 0 1021 0.59 1 13,285 1.69 2,3 4010 2.84

S72
CBGB: CABG–chest and donor site 5 0.28 0 1098 0.73 1 113,169 3.46 2 22,942 5.82 3 57 17.54
CBGC: CABG–chest only 4 0.22 0,1 6210 2.62 2,3 2420 4.05
OCVS: Other cardiovascular surgery 2 0.42 0,1 5313 0.77 2 1660 1.69 3 69 5.80
ORES: Other respiratory system 2 — 0,1,2,3 1352 2.74
THOR: Thoracic surgery 3 0.51 0 936 0.43 1 2876 1.29 2,3 1048 3.24
BILI: Liver/pancreas 4 0.39 0 309 3.24 1,2,3 1094 7.04
OGIT: Other digestive surgery 3 0.45 0,1 2290 3.23 2,3 432 8.10
SB: Small bowel surgery 3 0.21 0 823 5.59 1 1876 7.52 2 1010 9.80 3 183 14.75
XLAP: Laparotomy 2 0.37 0 3733 1.69 1 4125 3.15 2 2181 5.36 3 363 7.99
NEPH: Nephrectomy 4 — 0,1,2,3 2046 1.22
OGU: Other genitourinary surgery 2 0.54 0 8946 0.44 1 4016 1.17 2,3 983 2.95
PRST: Prostatectomy 4 0.50 0 1648 0.91 1,2,3 1306 2.68
HN: Head and neck 7 0.48 0 442 2.94 1 595 5.71 2,3 280 13.93
OENT: Other otorhinolaryngological 2 0.85 0,1 2474 0.24 2,3 272 2.94
HER: Herniorraphy 2 0.44 0 7251 0.79 1 3982 1.86 2,3 901 3.44
MAST: Mastectomy 3 0.32 0,1 11,178 2.07 2,3 403 3.97
CRAN: Craniotomy 4 0.50 0 2054 0.58 1,2,3 8112 1.75
ONS: Other nervous system 4 — 0,1,2,3 1648 1.76
VSHN: Ventricular shunt 2 0.17 0 1549 3.68 1,2,3 3573 5.12

Downloaded from https://academic.oup.com/cid/article/33/Supplement_2/S69/527418 by guest on 02 April 2024

 CSEC: Cesarean section 1 0.22 0 59,921 3.27 1 19,920 4.74 2,3 1641 8.65
OOB: Other obstetrical procedures 1 — 0,1,2,3 793 0.50
VHYS: Vaginal hysterectomy 2 0.16 0 7959 1.08 1,2,3 3937 1.47
AMP: Limb amputation 1 — 0,1,2,3 5991 4.29
FUSN: Spinal fusion 4 0.51 0 12,306 1.23 1 7206 3.07 2,3 1979 7.23
FX: Open reduction fracture 2 0.39 0 8474 0.64 1 12,709 1.33 2,3 2931 2.59
HPRO: Hip prosthesis 2 0.28 0 9841 0.78 1 17,638 1.55 2,3 5120 2.07
KPRO: Knee prosthesis 2 0.24 0 13,721 0.87 1 17,101 1.22 2,3 4928 2.03
LAM: Laminectomy 2 0.32 0 18,951 0.85 1 14,064 1.38 2,3 4122 2.57
OMS: Other musculoskeletal 3 0.31 0 9493 0.65 1 6680 0.93 2,3 1788 2.07
OPRO: Other prosthesis 3 — 0,1,2,3 1396 0.64
OBL: Other hematopoietic/lymphatic system 3 — 0,1,2,3 844 2.01
OES: Other endocrine system 3 0.74 0 1364 0.15 1,2,3 1046 0.96
OEYE: Other eye 2 — 0,1,2,3 437 0.69
OSKN: Other integumentary system 2 — 0,1,2,3 5501 1.38
SKGR: Skin graft 3 0.52 0,1 1872 1.44 2,3 806 4.47
SPLE: Splenectomy 3 — 0,1,2,3 1016 2.85
TP: Organ transplantation 7 0.56 0,1 2077 5.39 2,3 5711 6.99
VS: Vascular surgery 3 0.49 0 3579 0.98 1 30,595 1.79 2,3 12,515 5.05

S73
NOTE. CABG–chest and donor site, coronary artery bypass graft, chest and donor site incisions; CABG–chest only, coronary artery bypass graft, chest incision (e.g., internal mammary artery); G,
Goodman-Kruskal correlation coefficient, which was significant (P ! .05) for all procedures except those that resulted in only 1 combined risk category (0,1,2,3).
a
Does not include 4 procedures in which laparoscope use was incorporated into index: cholecystectomy, colon surgery, appendectomy, and gastric surgery (see tables 2 and 3).
b
Rate is per 100 operations.

Downloaded from https://academic.oup.com/cid/article/33/Supplement_2/S69/527418 by guest on 02 April 2024

 Table 2. Surgical site infection (SSI) rates for cholecystectomy, stratified by risk index and
laparoscope (“scope”) use.

SSI rate
Risk Overall Scope
index SSI rate use Scope p no Scope p yes P
0 127/23,891 p 0.53% 71.6% 44/6782 p 0.65% 83/17,095 p 0.49% .07
1 184/14,589 p 1.26% 59.6% 121/5892 p 2.05% 63/8689 p 0.73% !.0001
2 117/3916 p 2.99% 38.9% 87/2391 p 3.64% 30/1525 p 1.97% .001
3 21/419 p 5.01% 24.1% 21/318 p 6.60% 0/101 p 0.00 .003
All 449/42,815 p 1.05% 64.1% 273/15,383 p 1.77% 176/27,410 p 0.64% !.001

Downloaded from https://academic.oup.com/cid/article/33/Supplement_2/S69/527418 by guest on 02 April 2024

setting, 78% were skin infections, 13% were deep incisional diagnostic accuracy, and risk factor misclassification error, the
infections, and 9% were organ/space infections (figure 2). In NNIS basic SSI risk index proved useful in risk adjustment for
contrast, the distribution of 6876 SSIs detected before discharge most procedures. Furthermore, this analysis emphasized the
was 43% skin infections, 19% deep incisional infections, and need to incorporate risk factors other than the traditional
38% organ/space infections, and the distribution of 5681 SSIs wound classification into a composite index of SSI risk before
detected on readmission was 40% skin infections, 31% deep attempting to compare infection rates among surgeons, among
incisional infections, and 29% organ/space infections. institutions, or across time. The ASA score is a critical com-
ponent of the index, included in an attempt to measure intrinsic
host susceptibility. Somewhat analogous to the number of dis-
DISCUSSION
charge diagnoses used in the SENIC index, the ASA score has
A number of studies have reported a decrease in the incidence the advantage of being readily available at the time of surgery.
of SSIs when surveillance programs have been implemented The approximate 75th percentile of duration of operation pro-
that included the feedback of postoperative wound infection vides the index with additional discriminatory power when
rates to practicing surgeons [16–19]. Indeed, the SENIC project applied to specific operative procedures, such as coronary artery
showed such feedback to be an essential component of an bypass grafts. Of interest was the remarkable lack of change in
effective infection control program [19]. Warnings have been the duration of surgery cut points in this report compared with
sounded regarding an overly simplistic approach to the cal- those in our previous report [6]. Cardo et al. [21] found that
culation and comparison of surgeon-specific wound infection the accuracy of surgical team members in assessing wound
rates [20]. The results of applying the basic SSI risk index to classification for general and trauma surgery was 88% (95%
NNIS data reaffirm the general conclusions drawn from the CI, 82%–94%). The accuracy of recording the duration of op-
SENIC risk index and our previous report [6]. When applied eration (i.e., time from skin incision to skin closure) and the
to a more recently collected set of data, which was subject to ASA class has not been studied.
the normal interhospital variations in case-finding methods, Our report also demonstrates the value of including an ad-

Table 3. Surgical site infection (SSI) rates, by selected operative procedures and modified risk index category by laparoscope,
1992–1998.

Duration
cut point,
a a a a a
Procedure hours RI n Rate RI n Rate RI n Rate RI n Rate RI n Rate
CHOL: Cholecystectomy 2 M 17,095 0.49 0 15,471 0.69 1 7417 2.04 2 2492 3.49 3 318 6.60
COLO: Colon surgery 3 M 288 0.69 0 6812 4.32 1 11,856 6.24 2 5267 9.55 3 718 12.95
APPY: Appendectomy 1 0-Yes 583 0.56 0-No 3866 1.37 1 4957 3.17 2,3 2121 5.85
GAST: Gastric surgery 3 0-Yes 203 0.49 0-No 1144 2.71 1 2416 5.13 2,3 1184 10.73

NOTE. For cholecystectomy and colon surgery, influence of laparoscope was captured by subtracting 1 from basic SSI risk index (no. of risk factors present,
as described in text) whenever procedure was done laparoscopically; M indicates modified risk category in which no risk factors were present and procedure
was done with laparoscope. For appendectomy and gastric surgery, basic SSI index value of 0 (no risk factors) was split into 0-No (laparoscope not used) and
0-Yes (laparoscope used), and whether procedure was done laparoscopically was otherwise ignored because SSI rates did not vary depending on use of laparoscope
when other risk factors were present. SSI rates with 2 or 3 other risk factors (American Society of Anesthesiologists score, wound class, or duration of surgery)
did not differ significantly, and data were combined to form category 2,3. RI, risk index.
a
Infection rate is per 100 operations.

S74 • CID 2001:33 (Suppl 2) • Gaynes et al.

 In that event, one can calculate how many infections would
have been “expected” to occur among patients, taking into
account the type and number of procedures and the risk cat-
egories of the patients. The expected number of SSIs can be
obtained by multiplying the number of operations done by the
surgeon in each procedure risk category by the NNIS rate for
that same procedure risk category and then dividing by 100.
By summing the numbers of expected SSIs for the procedure
and risk categories in which surgery was done, we can compare
it with the number of observed SSIs for the surgeon. The ratio
of the observed number of SSIs (O) that occurred to the ex-

Downloaded from https://academic.oup.com/cid/article/33/Supplement_2/S69/527418 by guest on 02 April 2024

pected number (E) is called the standardized infection ratio or
SIR: SIR p O/E.
The SIR is deceptively simple. It is an easy way to interpret
summary measures of the SSI experience of an individual sur-
geon, service, or hospital. Values that exceed 1.0 indicate that
more SSIs occurred than were expected (and by how much),
whereas values of !1.0 indicate the opposite. In calculating the
expected number of SSIs, we account for the type of procedures
Figure 1. Surgical site infections (SSIs) by location of detection. performed and the distribution of patients by risk index, that
is, the case mix. Therefore, the SIR is a risk-adjusted summary
ditional variable in the index, namely the use of the laparoscope. measure and can be used for comparative purposes. To test
Although only 4 procedures in the index use the laparoscope whether the SIR differs significantly from its nominal value of
at this time, we expect that number to increase. 1.0, a Z statistic can be calculated by the following formula
The results in table 1 can be used by hospitals in several (valid as long as E ⭓ 1): Z p 2(冑O ⫺ 冑E).
ways. For each surgeon, procedure-specific SSI rates can be The SIRs for 2 surgeons or for the same surgeon during 2
calculated and compared against the corresponding rates in time periods can also be compared [22].
table 1 by means of a simple Z test or Fisher’s exact test [22]. The value of comparative SSI rate data depends largely on
In many hospitals, the number of procedures done by an in- the accuracy and consistency with which the data are collected.
dividual surgeon in some of the risk categories may be small. If SSI rates vary because of differences in postdischarge sur-

Figure 2. Detection of surgical site infections (SSI) by specific type and location of detection. Admission refers to during same hospital admission
as operative procedure. From the National Nosocomial Infections Surveillance system, 1994–1998.

NNIS Basic SSI Risk Index • CID 2001:33 (Suppl 2) • S75

veillance intensity, then the value of the comparison is dimin- tion. For a particular group of patients, the index may still not
ished. Our data suggest that SSI rates are not generally cor- have adequately adjusted for differences in case mix between
related with postdischarge surveillance intensity, as measured an individual hospital’s group and the comparison group. Also,
by the percentage of SSIs detected after discharge among pa- in the case of the NNIS basic SSI risk index, 2 of the risk factors
tients who are not readmitted to the hospital. Previous studies in the index, wound class and duration of operation, may in-
have shown that 12%–84% of SSIs are detected after patients directly reflect quality of care. Adjustment for these factors may
are discharged from the hospital [16, 23–41]. Postdischarge mask rather than elucidate a potential problem. A comparison
surveillance methods have been used with varying degrees of of the distribution of the operations among the risk categories
success for different procedures and among hospitals, and they with the distribution in a corresponding group, such as that
include direct examination of patients’ wounds during follow- in table 1, may also be useful.
up visits to either surgery clinics or physicians’ offices [8, 18, As we move forward in the development of measures of

Downloaded from https://academic.oup.com/cid/article/33/Supplement_2/S69/527418 by guest on 02 April 2024

23, 16, 30, 37, 42–44], review of medical records of surgery health care quality, a simple index to predict risk for SSIs will
clinic patients [29, 42, 45], administration of questionnaires to be less optimal for interhospital comparison once more risk
patients by mail or telephone [8, 25, 27, 28, 31, 32, 38, 46–48], factors are ascertained. Interhospital comparisons can be im-
or administration of questionnaires to surgeons by mail or proved by using the SIR and by enhancing the accuracy of the
telephone [8, 16, 24, 27–29, 33, 34, 36, 38, 45]. One study expected number of SSIs from multivariate models by use of
found that patients have difficulty assessing their wounds for aggregated NNIS data. This approach to comparison will allow
infection (specificity, 52%; positive predictive value, 26%) [49], inclusion of the full range of risk factors for operative proce-
suggesting that data obtained by use of the patient questionnaire dures. Until then, the NNIS basic SSI risk index remains the
may inaccurately represent SSI rates. A review suggested that best currently available method for benchmark comparisons of
there is no consensus for monitoring SSIs after discharge from SSI rates.
the hospital [50].
Recently, Sands et al. [36] performed a computerized search References
of 3 databases: ambulatory encounter records for diagnostic,
1. O’Leary DS. The Joint Commission looks to the future [editorial].
testing, and treatment codes; pharmacy records for specific an- JAMA 1987; 258:951.
timicrobial prescriptions; and administrative records for re- 2. Altemeir WA, Culbertson WR. Surgical infection. In: Moyer CA,
hospitalizations and emergency room visits. The purpose of the Rhoads JE, Allen JG, Harkins HN, eds. Surgery, principles and practice.
Philadelphia: JB Lippincott, 1965:51–77.
search was to determine which database best identified SSIs. 3. National Academy of Sciences/National Research Council. Postoper-
They found that pharmacy records indicating that a patient ative wound infections: the influence of ultraviolet irradiation of the
had received antimicrobial agents commonly used to treat soft operating room and of various other factors. Ann Surg 1964; 160(Suppl
2):1–132.
tissue infections had the highest sensitivity and positive pre- 4. Garner JS. Guideline for prevention of surgical wound infections, 1985.
dictive value. Their study also showed the low sensitivity of Infect Control 1986; 7:193–200.
many of the other case-finding methods after discharge. During 5. Haley RW, Culver DH, Morgan WM, et al. Identifying patients at high
risk of surgical wound infection: a simple multivariate index of patient
the period of 1992–1998, most NNIS hospitals did not have susceptibility and wound contamination. Am J Epidemiol 1985; 121:
access to pharmacy records for information on specific anti- 206–15.
microbial prescriptions after discharge. This suggests that post- 6. Culver DH, Horan TC, Gaynes RP, et al. Surgical wound infection
rates by wound class, operative procedure, and patient risk index. Am
discharge case finding by most NNIS hospitals may have been
J Med 1991; 91(Suppl 3B):152S–7S.
consistently insensitive. 7. Flanders E, Hinnant JR. Ambulatory surgery postoperative wound sur-
The general applicability of the NNIS basic SSI risk index veillance. Am J Infect Control 1990; 18:336–9.
8. Fanning C, Johnston BL, MacDonald S, et al. Postdischarge surgical
within a broad range of operative procedure categories, with
site infection surveillance. Can J Infect Control 1995; 10:75–9.
the inclusion of laparoscope use where needed, is encouraging, 9. Hecht AD. Creating greater efficiency in ambulatory surgery. J Clin
but there remains room for improvement. Almost certainly, Anesth 1995; 7:581–4.
10. Emori TG, Culver DH, Horan TC, et al. National nosocomial infections
additional important risk factors for specific procedures need
surveillance system (NNIS): description of surveillance methods. Am
to be identified and incorporated into such an index [51], such J Infect Control 1991; 19:19–35.
as antibiotic prophylaxis during procedures in which it has been 11. Horan TC, Gaynes RP, Martone WJ, et al. CDC definitions of noso-
shown to be effective but not universally adopted [52–57]. comial surgical site infections, 1992: a modification of CDC definitions
of surgical wound infections. Infect Control Hosp Epidemiol 1992; 13:
It is important to keep in mind the limitations of any risk 606–8.
index. A statistically significant difference between a risk in- 12. New classification of physical status. Anesthesiology 1963; 24:107–11.
dex–adjusted SSI rate for a cohort of patients and an appro- 13. Owens WD. ASA Physical Status Classification: a study on consistency
of ratings. Anesthesiology 1978; 49:239–43.
priate comparison group merely indicates the presence of a 14. Keats AS. The ASA classification of physical status: a recapitulation.
potential problem, one generally worthy of further investiga- Anesthesiology 1978; 49:233–6.

S76 • CID 2001:33 (Suppl 2) • Gaynes et al.

15. Goodman LA, Kruskal WH. Measures of association for cross classi- 36. Sands K, Vineyard G, Platt R. Surgical site infections occurring after
fications. J Am Stat Assoc 1954; 49:732–64. hospital discharge. J Infect Dis 1996; 173:963–70.
16. Cruse PJE, Foord R. A five-year prospective study of 23,649 surgical 37. Weigelt JA, Dryer D, Haley RW. Necessity and efficiency of wound
wounds. Arch Surg 1973; 107:206–10. surveillance after discharge. Arch Surg 1992; 127:77–82.
17. Cruse PJE, Foord R. The epidemiology of wound infection: a 10-year 38. Gravel-Tropper D, Oxley C, Memish Z, Garber GE. Underestimation
prospective study of 62,939 wounds. Surg Clin North Am 1980; 60: of surgical site infection rates in obstetrics and gynecology. Am J Infect
27–40. Control 1995; 23:22–6.
18. Condon RE, Schulte WJ, Malangoni MA, Anderson-Teschendorf MJ. 39. Taylor S, Pearce P, McKenzie M, Taylor GD. Wound infection in total
Effectiveness of a surgical wound surveillance program. Arch Surg joint arthroplasty: effect of extended wound surveillance on wound
1983; 118:303–7. infection rates. Can J Surg 1994; 37:217–20.
19. Haley RW, Culver DH, White JW, et al. The efficacy of infection sur- 40. Hulton LJ, Olmsted RN, Treston-Aurand J, Craig CP. Effect of post-
veillance and control programs in preventing nosocomial infections in discharge surveillance on rates of infectious complications after cesa-
US hospitals. Am J Epidemiol 1985; 121:182–205. rean section. Am J Infect Control 1992; 20:198–201.
20. Scheckler WE. Surgeon-specific wound infection rates: a potentially 41. Law DJW, Mishriki SF, Jeffery PJ. The importance of surveillance after
dangerous and misleading strategy. Infect Control Hosp Epidemiol discharge from hospital in the diagnosis of postoperative wound in-

Downloaded from https://academic.oup.com/cid/article/33/Supplement_2/S69/527418 by guest on 02 April 2024

1988; 9:145–6.
fection. Ann R Coll Surg Engl 1990; 72:207–9.
21. Cardo DM, Falk PS, Mayhall CG. Validation of surgical wound clas-
42. Weigelt JA. Risk of wound infections in trauma patients. Am J Surg
sification in the operating room. Infect Control Hosp Epidemiol
1985; 150:782–4.
1993; 14:255–9.
43. Henley MB, Jones RE, Wyatt RWB, et al. Prophylaxis with cefamandole
22. Horan T, Culver D. Comparing surgical site infection rates. In: Pfeiffer
nafate in elective orthopedic surgery. Clin Orthop 1986; 209:249–54.
J, ed. APIC text of infection control and epidemiology. Washington,
44. Garibaldi RA, Cushing D, Lerer T. Risk factors for postoperative in-
DC: Association for Professionals in Infection Control and Epidemi-
fection. Am J Med 1991; 91(Suppl 3B):158S–63S.
ology, 2000:14-1–14-7.
23. Olson MM, Lee JT. Continuous, 10 year wound infection surveillance: 45. Mertens R, Jans B, Kurz X. A computerized nationwide network for
results, advantages, and unanswered questions. Arch Surg 1990; 125: nosocomial infection surveillance in Belgium. Infect Control Hosp
794–803. Epidemiol 1994; 15:171–9.
24. Burns SJ, Dippe SE. Postoperative wound infections detected during 46. Manian FA, Meyer L. Comparison of patient telephone survey with
hospitalization and after discharge in a community hospital. Am J traditional surveillance and monthly physician questionnaires in mon-
Infect Control 1982; 10:60–5. itoring surgical wound infections. Infect Control Hosp Epidemiol
25. Polk BF, Shapiro M, Goldstein P, et al. Randomised clinical trial of 1993; 14:216–8.
perioperative cefazolin in preventing infection after hysterectomy. Lan- 47. Holbrook KF, Nottebart VF, Hameed SR, Platt R. Automated post-
cet 1980; 1:437–41. discharge surveillance for postpartum and neonatal nosocomial infec-
26. Salem RJ, Johnson J, Devitt P. Short term metronidazole therapy con- tions. Am J Med 1991; 91(Suppl 3B):125S–30S.
trasted with povidone-iodine spray in the prevention of wound infec- 48. Zoutman D, Pearce P, McKenzie M, Taylor G. Surgical wound infec-
tion after appendicectomy. Br J Surg 1979; 66:430–1. tions occurring in day surgery patients. Am J Infect Control 1990; 18:
27. Walsh A, Roberts FJ, Bryce EA. Post-discharge surveillance of surgical 277–82.
wound infections [letter]. Can J Infect Control 1996; 11:29. 49. Seaman M, Lammers R. Inability of patients to self-diagnose wound
28. Brown RB, Bradley S, Opitz E, et al. Surgical wound infections doc- infections. J Emerg Med 1991; 9:215–9.
umented after hospital discharge. Am J Infect Control 1987; 15:54–8. 50. Holtz TH, Wenzel RP. Postdischarge surveillance for nosocomial
29. Rosendorf LL, Octavio J, Estes JP. Effect of methods of postdischarge wound infection: a brief review and commentary. Am J Infect Contol
wound infection surveillance on reported infection rates. Am J Infect 1992; 20:206–13.
Control 1983; 11:226–9. 51. Hall JC. A casemix of patients undergoing abdominal surgery. Med J
30. Ferraz EM, Ferraz AAB, Coelho HSTD, et al. Postdischarge surveillance Aust 1992; 156:863–5.
for nosocomial wound infection: does judicious monitoring find cases? 52. Haines SJ, Walters BC. Antibiotic prophylaxis for cerebrospinal fluid
Am J Infect Control 1995; 23:290–4. shunts: a metaanalysis. Neurosurgery 1994; 34:87–93.
31. Andenaes K, Amland PF, Lingaas E, et al. A prospective, randomized
53. Conte JE, Cohen SN, Roe BB. Antibiotic prophylaxes and cardiac sur-
surveillance study of postoperative wound infections after plastic sur-
gery: a prospective double-blind comparison of single-dose versus mul-
gery: a study of incidence and surveillance methods. Plast Reconstr
tiple-dose regimens. Ann Intern Med 1972; 76:943–9.
Surg 1995; 96:948–56.
54. Gentry LO, Zeluff BJ, Cooley DA. Antibiotic prophylaxis in open-heart
32. Keeling NJ, Morgan MWE. Inpatient and post-discharge wound in-
surgery: a comparison of cefamandole, cefuroxime, and cefazolin. Ann
fections in general surgery. Ann R Coll Surg Engl 1995; 77:245–7.
33. Manian FA, Meyer L. Adjunctive use of monthly physician question- Thorac Surg 1988; 46:167–71.
naires for surveillance of surgical site infections after hospital discharge 55. VanEk B, Dijkmans BAC, VanDulken H, VanFurth R. Antibiotic pro-
and in ambulatory surgical patients: report of a seven-year experience. phylaxis in craniotomy: a prospective double-blind placebo-controlled
Am J Infect Control 1997; 25:390–4. study. Scand J Infect Dis 1988; 20:633–9.
34. Manian FA, Meyer L. Comprehensive surveillance of surgical wound 56. Ergina PL, Gold S, Meakins JL. Antibiotic prophylaxis for hernior-
infections in outpatient and inpatient surgery. Infect Control Hosp rhaphy and breast surgery. N Engl J Med 1990; 322:1884–6.
Epidemiol 1990; 11:515–20. 57. Page CP, Bohnen JMA, Fletcher JR, et al. Antimicrobial prophylaxis
35. Reimer K, Gleed C, Nicolle LE. The impact of postdischarge infection for surgical wounds: guidelines for clinical care. Arch Surg 1993; 128:
on surgical wound infection rates. Infect Control 1987; 8:237–40. 79–88.

NNIS Basic SSI Risk Index • CID 2001:33 (Suppl 2) • S77