Vol. 26 No.

5 INFECTION CONTROL AND HOSPITAL EPIDEMIOLOGY 449

SURGICAL-SITE INFECTIONS FOLLOWING

CESAREAN SECTION IN AN ESTONIAN UNIVERSITY
HOSPITAL: POSTDISCHARGE SURVEILLANCE AND
ANALYSIS OF RISK FACTORS

Piret Mitt, MD; Katrin Lang, MD, MPH; Aira Peri, MD; Matti Maimets, MD, PhD

ABSTRACT
OBJECTIVES: To evaluate a multi-method approach to RESULTS: The multi-method approach gave a follow-up
postdischarge sur veillance of surgical-site infections (SSIs) and rate of 94.8%. Of 305 patients, 19 (6.2%; 95% confidence inter val
to identify infection rates and risk factors associated with SSI [CI95], 3.8–9.6) had SSIs. Forty-two percent of these SSIs were
following cesarean section. detected during postdischarge sur veillance. We found three
DESIGN: Cross-sectional sur vey. variables associated with increased risk for developing SSI: in-
SETTING: Academic tertiar y-care obstetric and gynecol- ternal fetal monitoring (odds ratio [OR], 16.6; CI95, 2.2–125.8;
ogy center with 54 beds. P = .007), chorioamnionitis (OR, 8.8; CI95, 1.1–69.6; P = .04),
PATIENTS: All women who delivered by cesarean sec- and surgical wound classes III and IV (OR, 3.8; CI95, 1.2–11.8;
tion in Tartu University Women’s Clinic during 2002. P = .02).
METHODS: Infections were identified during hospital CONCLUSIONS: The high response rate validated the
stay or by postdischarge sur vey using a combination of tele- effectiveness of this kind of sur veillance method and was most
phone calls, healthcare worker questionnaire, and outpatient suitable in current circumstances. A challenge exists to de-
medical records review. SSI was diagnosed according to the cri- crease the frequency of internal fetal monitoring and to treat
teria of the Centers for Disease Control and Prevention National chorioamnionitis as soon as possible (Infect Control Hosp Epide-
Nosocomial Infections Sur veillance System. miol 2005;26:449-454).

The single most important risk factor for postpartum decrease, the increasing number of SSIs is not detected
maternal infection is delivery by cesarean section.1 Mater- through the standard surveillance method; therefore,
nal morbidity related to infections has been shown to be postdischarge surveillance has become increasingly im-
eightfold higher after cesarean section than after vaginal portant to obtain accurate rates of SSI. Several methods
delivery.2 Reducing the number of deliveries by cesarean for postdischarge surveillance of SSI have been evaluated
section and identifying risk factors for post-cesarean surgi- for efficiency, including direct examination of patients’
cal-site infections (SSIs) could contribute to a decrease in wounds during follow-up visits to either surgery clinics or
maternal morbidity. physicians’ offices, review of medical records of surgery
The reported incidence of SSI following cesarean clinic patients, and patient and healthcare worker surveys
section varies widely, ranging from 0.3% in Turkey3 to 17% by mail or telephone.4,7 Automated data, especially from
in Australia.4 Among hospitals reporting to the National pharmacy and administrative claims, might substantially
Nosocomial Infections Surveillance (NNIS) System, the improve both inpatient and postdischarge surveillance for
rate of SSI after cesarean section was 2.8% to 6.7% de- SSI while reducing the resources required8; however, this
pending on the risk index category.5 The incidence rate method is not available everywhere. Nevertheless, there
depends on the following: the definition of SSI adopted, is no universally accepted strategy for monitoring these
the intensity of surveillance, the prevalence of risk factors infections.6,7,9,10 When choosing a surveillance method, in-
for SSI in the patient group being audited, and whether fection control personnel must consider not only the sen-
the survey contains postdischarge data.6 SSI may not be sitivity and the sources for data, but also the human and
detected for several weeks after discharge and may not financial resources allotted for SSI surveillance.10
require admission to the operating hospital. Because Various factors affect infection rates in different set-
the length of postoperative hospitalization continues to tings; those most frequently cited in the literature include

Drs. Mitt and Maimets are from the Departments of Internal Medicine and Infection Control, and Dr. Peri is from the Women’s Clinic, Tartu
University Hospital, Tartu, Estonia. Dr. Lang is from the Department of Public Health, University of Tartu, Tartu, Estonia.
Address reprint requests to Piret Mitt, MD, Department of Infection Control, Tartu University Hospital, Lina 6, 51004 Tartu, Estonia.
[email protected]
The authors thank Fred Kirss, MD, head of the Department of Obstetrics, Women’s Clinic, Tartu University Hospital, for supporting the study and
Kristiina Rull, MD, Women’s Clinic, Tartu University Hospital, for critical review of the manuscript.
450 INFECTION CONTROL AND HOSPITAL EPIDEMIOLOGY May 2005

which accounts for approximately one-third of the Estonian

population, and has an average of 2,000 deliveries per year.

Patients
The study population consisted of women who de-
livered by cesarean section in Tartu University Women’s
Clinic during 2002. The purpose of the study was explained
to patients, and their verbal consent to participate was ob-
tained.

Data Collection
Infections were identified during hospital stay or
within 30 days following cesarean section by readmission
to the hospital or by postdischarge survey using the cri-
teria of the Centers for Disease Control and Prevention
(CDC) NNIS System.7 The postdischarge survey was per-
formed according to the modified methodology developed
by Stockley et al.19 The patient received a questionnaire
(Figure) to be given to the physician (ie, obstetrician or
general practitioner) to complete if problems developed
regarding the wound or if endometritis developed after
discharge. All study subjects were contacted at home by
telephone 30 to 35 days after surgery (including those
patients who had been diagnosed as having SSI during
admission). The patients were asked about their general
health and the state of their surgical wound using a stan-
dard format based on the physician’s questionnaire. If a
patient’s complaints referred to a possible infection and
the questionnaire had not been received, the investigator
contacted the physician for verbal confirmation of SSI or
FIGURE. Questionnaire given to patients to deliver to their physician if
wound or endometritis problems occurred after discharge. the medical chart was reviewed to determine whether the
patient had attended the outpatient department for the
treatment of SSI or had been admitted to the hospital. If it
was not possible to contact the patient, the outpatient med-
extremes of maternal weight (being underweight or obese), ical records of those patients who were known to have re-
prolonged labor or rupture of membranes, long duration of turned to Tartu University Women’s Clinic were reviewed.
surgery, multiple procedures, manual removal of the pla- When the completed questionnaire arrived and the CDC
centa, young maternal age, maternal preoperative condi- criteria were met, no further contact with the physician
tion, procedure-related blood loss, and absence of antibiotic was established.
prophylaxis.11-17 It is important to identify these factors to All surveillance data were collected by a single inves-
target high-risk patients who need specific prevention mea- tigator. Demographic information, potential risk factors,
sures. and surgical indications were recorded. Host-related vari-
Despite recent progress in developing a medical in- ables included age, nationality, parity, existing comorbidi-
frastructure, some of its components, such as a national ties (eg, diabetes, preeclampsia, anemia, or chorioamnion-
system for nosocomial infections surveillance, have not yet itis), bacterial vaginosis during pregnancy, a preoperative
been established in Estonia. In Estonia, only a few studies condition assessed by American Society of Anesthesiolo-
have been conducted on healthcare-associated infections.18 gists (ASA) score, number of prenatal care visits, duration
The aims of this descriptive study were to evaluate a multi- of ruptured membranes, duration of labor, preoperative
method approach to postdischarge surveillance of SSI and stay, length of hospital stay, use of internal fetal monitor-
to identify infection rates and risk factors associated with ing, and number of vaginal examinations prior to cesarean
SSI following cesarean section by comparing patients with section at the hospital. Surgery-related variables included
and without SSI. emergency nature of the operation, indications for cesar-
ean section, duration of the operation, manual removal of
METHODS the placenta, volume of blood loss, and antibiotic prophy-
Setting laxis. The study subjects were postoperatively monitored
Tartu University Women’s Clinic is a 54-bed (2002 for temperature, SSI, wound and endocervical culture, and
data) academic, tertiary-care obstetric and gynecology cen- antibiotic treatment.
ter that serves mainly the population of southern Estonia, Inpatient surveillance (including gathering surveil-
Vol. 26 No. 5 SSI FOLLOWING CESAREAN SECTION IN ESTONIA 451

lance data and explaining the purpose of the study) took TABLE 1
approximately 20 minutes per patient; telephone follow-up INDICATIONS FOR CESAREAN SECTION
took an average of 3 to 5 minutes per patient. No. of
Indication Patients (%)
Definitions Fetal distress 72 (23.6)
According to the modified wound classification, ce- Breech and malpresentation 63 (20.7)
sarean section deliveries were classified as follows: class I Cephalopelvic disproportion 41 (13.4)
if there was no rupture of membranes or labor, class II if
Previous delivery by cesarean section 39 (12.8)
there was less than 2 hours of membrane rupture without
Preeclampsia 25 (8.2)
labor or labor of any length with intact membranes, class III
Dystocia 13 (4.3)
for rupture of membranes greater than 2 hours, and class
IV for purulent amniotic fluid.11,12 Currently practiced hos- Placenta previa, placenta increta, or placental 12 (3.9)
pital policy for surgical antibiotic prophylaxis in Tartu Uni- abruption
versity Women’s Clinic recommends intravenous adminis- Other* 40 (13.1)
tration of 1 g of cefazolin as soon as the umbilical cord is Total 305 (100.0)
clamped for all patients undergoing non-elective cesarean *Included maternal diseases, twin pregnancy, and pelvic trauma.
section and for patients undergoing elective cesarean sec-
tion for rupture of membranes more than 4 hours or with
bacterial vaginosis diagnosed during pregnancy. Elective
cesarean section was defined as a planned procedure and were contacted by telephone. Fifteen of them described a
performed when scheduled (patient did not have active la- possible infection but only 6 cases were finally confirmed
bor or rupture of membranes prior to surgery), whereas all by the physician (4 with the questionnaire and 2 with
other instances of cesarean section were defined as emer- personal contact). Information was obtained on chart
gency procedures. review for 9 study subjects who could not be contacted
by telephone nor from whom a completed questionnaire
Statistics was received. No SSI was detected. In this study, 16 of
All comparisons were unpaired, and all tests of signifi- the 305 eligible patients were not available for any infor-
cance were two-tailed. For all categorical variables, Fisher’s mation during the postdischarge period. A combination
exact test or chi-square was used. For continuous variables, of healthcare worker questionnaires, telephone calls,
Student’s t test was performed. Odds ratios (ORs) and 95% and chart reviews gave a postdischarge follow-up rate of
confidence intervals (CI95) were calculated using standard 94.8%.
methods. A P value of .05 or less was considered to indicate Two patients were excluded from the study during
statistical significance. Multiple logistic regression analysis risk factor analysis because SSI developed more than 30
was performed to obtain adjusted estimates of OR and to days after their cesarean sections. The characteristics of
identify independent risk factors. Data were analyzed using the sample are given in Table 2. Variables such as national-
Stata statistical software (version 8.0; StataCorp, College ity, diabetes, preeclampsia, anemia, bacterial vaginosis dur-
Station, TX). ing pregnancy, manual removal of the placenta, and volume
of blood loss did not show significant association between
RESULTS patients with and without SSI (data not shown). Univariate
There were 310 cesarean sections performed among analysis identified that chorioamnionitis, duration of labor,
2,092 deliveries (14.8%; CI95, 13.3 to 16.4) during the study internal fetal monitoring, and surgical wound classes III
period. Three hundred five patients were enrolled in the and IV were associated with SSI (Table 2). Multiple logistic
study (4 patients refused to participate, and 1 patient died regression analysis found three variables that were inde-
during the operation due to a complication of underlying pendently associated with increased risk for developing
disease, rupture of an aortic aneurysm). Among the study SSI: internal fetal monitoring (OR, 16.6; CI95, 2.2 to 125.8;
patients, there were 192 (63%) emergency and 113 (37%) P = .007), chorioamnionitis (OR, 8.8; CI95, 1.1 to 69.6; P =
elective cesarean sections performed. Indications for cesar- .04), and surgical wound classes III and IV (OR, 3.8; CI95,
ean section are provided in Table 1. 1.2 to 11.8; P = .02).
During the study period, 19 SSIs were identified: 14 Analysis of the use of antibiotic prophylaxis re-
patients developed incisional (2 deep and 12 superficial) vealed that 163 (84.9%) of the emergency cesarean sec-
infections, 4 developed endometritis, and 1 developed in- tion deliveries and 37 (32.7%) of the elective cesarean
tra-abdominal abscess. The overall infection rate was 6.2% section deliveries received prophylaxis. Two hundred
(CI95, 3.8 to 9.6). Of the 19 SSIs identified, 11 (57.9%) were twenty-three (73.1%) of the patients had followed the
diagnosed before and 8 (42.1%) after discharge. Of the lat- guidelines according to the current policy for hospital
ter, 2 patients were readmitted to the hospital and 6 had antibiotic prophylaxis. Seventy-five patients received an-
SSIs that were detected by multi-method postdischarge tibacterial treatment without any confirmed diagnosis of
surveillance. infection after the operation.
During the postdischarge sur veillance, 280 patients Patients with SSI had a longer mean hospitalization
452 INFECTION CONTROL AND HOSPITAL EPIDEMIOLOGY May 2005

TABLE 2
ANALYSIS OF RISK FACTORS FOR SURGICAL-SITE INFECTIONS FOLLOWING CESAREAN SECTION
Mean ± SD or No. (%)

Risk Factor SSI (n = 19) No SSI (n = 284) OR CI95 P

Age, y 27.7 ± 6.5 28.5 ± 5.9 .59

Gestational age, wk 39.2 ± 2.6 38.6 ± 2.6 .38

Prenatal care visits 9.4 ± 2.2 9.6 ± 2.7 .75

Preoperative stay, d 1.2 ± 1.2 1.1 ± 1.7 .08

Vaginal examination 2.6 ± 1.7 1.9 ± 1.8 .06

Duration of labor, h 6.0 ± 8.5 2.8 ± 4.7 .008

Duration of ruptured membranes, h 10.5 ± 18.2 5.4 ± 21.2 .31

ASA score 1.3 ± 0.5 1.5 ± 0.8 .18

Duration of surgery, min 40.7 ± 6.8 40.6 ± 15.0 .97

Repeat cesarean section 4 (21) 62 (21) 0.95 0.32–2.85 1.00

Emergency 14 (74) 178 (63) 1.67 0.61–4.57 .46

Chorioamnionitis 2 (11) 2 (1) 16.58 2.75–100.29 .02

Nulliparity 11 (58) 133 (47) 1.56 0.63–3.89 .35

Internal fetal monitoring 2 (11) 3 (1) 11.02 1.72–70.43 .03

Absence of prophylaxis 7 (37) 96 (34) 1.14 0.44–2.99 .79

Surgical wound class* 12 (63) 84 (30) 4.08 1.55–10.73 .002

SD = standard deviation; SSI = surgical-site infection; OR = odds ratio; CI95 = 95% confidence interval; ASA score = American Society of Anesthesiologists preoperative assessment score.
*Contaminated or dirty.

time than did noninfected patients (5.8 ± 0.3 vs 7.9 ± 1.5 than the number of the total sample, should be used. As in-
days; P < .03). creasingly more infections emerge after discharge and in
cases of a low response rate, inclusion of nonresponders
DISCUSSION may lower the infection rate and produce inaccuracies
To our knowledge, this is the first extensive SSI when compared with other healthcare facilities. Thus,
surveillance study performed in Estonia. The main goals when these comparisons are being made, the denomina-
of our study were to evaluate a multi-method approach to tor has to be taken into consideration. In our study, the re-
postdischarge surveillance of SSI and to identify infection sponse rate was high and the exclusion of nonresponders
rates after cesarean section. from the denominator would not have changed the results
The overall SSI rate of 6.2% in our hospital is low- significantly (6.2% vs 6.6%).
er than rates from other studies that have used postdis- However, the best way to conduct postdischarge
charge surveillance. Rates have varied from 9.6% in Bra- surveillance is still a matter of dispute according to the
zil9 to 17% in Australia.4 The comparison of our SSI rates literature. The ideal methodology should have a high fol-
with NNIS System benchmarks is not meaningful because low-up rate, be cost-effective, and have high sensitivity
postdischarge surveillance is not required by the NNIS and specificity.4,6-8,10 The follow-up rate in our study was
System, but any comparison of SSI rates must take into 94.8%; in a study conducted by Stockley et al.,19 a combi-
account whether case-finding included the detection of nation of different postdischarge surveillance methods
SSI after discharge. Twelve percent to 84% of SSIs are gave a follow-up rate of 92.7%. In the study performed by
detected after patients are discharged from the hospital.7 Taylor et al.,20 patients were contacted only by telephone.
During postdischarge surveillance in our study, 42.1% of The compliance rate was 93.3%, and it was concluded that
SSIs were detected; however, the numbers were relatively this method of contact is feasible and effective. Stockley
small and may therefore have been affected by a random et al.19 found that the combination of different methods is
error. The data suggest the necessity to perform postdis- relatively simple to use and causes minimal inconvenience
charge surveillance to obtain more accurate SSI rates. Noy to patients and healthcare workers. We agree with this be-
and Creedy4 recommended in their study that when the cause most of the patients were interested in participating
rate is being calculated, the number of responders, rather in the surveillance and because given that approximately
Vol. 26 No. 5 SSI FOLLOWING CESAREAN SECTION IN ESTONIA 453

65% of the population owns a mobile phone, the telephone method allows efficient identification of postpartum infec-
questionnaire is a suitable method.21 The method we used tions not detected by conventional surveillance. Computer-
is not very time consuming. It is also more acceptable ized systems could reduce the time and costs required to
to patients because only those who have a problematic perform surveillance, but electronic medical records do not
wound area need to go to the hospital for checkups. The exist yet on a large scale in Estonia.
data collection was simplified by the fact that most of the Studies have indicated that antibiotic exposure is a
women returned to our clinic for postdischarge care. The sensitive indicator of an infection because relatively few
same applies to the physicians; only physicians whose serious infections are managed without antibiotics. Poor
patients have problems should take part in the study. specificity (too many false-positive results) has been a ma-
In the study by Stockley et al.,19 the medical records of jor problem because antibiotics are so widely used after sur-
patients who could not be contacted by phone were not gery for extended prophylaxis, empiric therapy of suspect-
reviewed. In contrast, we performed chart reviews and re- ed infection, and treatment of infections other than SSI.10,25
ceived information regarding nine more patients, a meth- In our study, 75 patients without any confirmed infection
od also practiced by Noy and Creedy.4 The high response diagnosis received antimicrobial treatment after cesarean
rate validates the effectiveness of this kind of surveillance section. Therefore, we cannot use therapy as an indicator of
method. SSI. Inappropriate use of antimicrobial agents not only adds
The literature suggests that direct observation of to the cost of medical care, but also needlessly exposes the
surgical sites by trained professionals (eg, infection con- patient to potential toxicity and risks that promote the de-
trol practitioners) is the most accurate method to detect velopment and spread of antimicrobial resistance in health-
SSI.6,7 However, in our study this was not possible because care facilities.26
we had to consider human and financial resources allotted Our postdischarge surveillance system was most
for SSI surveillance. It could be one of the reasons why suitable in the current circumstances.
our infection rate is lower than those from other studies. The second aim of this study was to identify the risk
The validity of the information obtained from patients and factors that contribute to SSI following cesarean section. In
physicians and whether the diagnosis of SSI can be based contrast to the NNIS System results, age, ASA score, dura-
on this is a matter of dispute in several research studies. tion of labor, and duration of surgery were not significant
Seaman and Lammers22 found that patients, despite using risk factors in our study sample.17 The average age of the
verbal or printed instructions, were unable to recognize patients in the infected group was younger than that of the
infections. They reported that patients correctly identi- noninfected group, but not significantly younger. However,
fied their infections in only 11 cases, whereas medical our rate estimates could have been affected by a random er-
examiners diagnosed infection in 21 wounds, and called ror caused by the small number of infections observed. The
into question the validity of data obtained using patient- multiple logistic regression revealed three variables inde-
returned questionnaires or telephone surveys. A recent pendently associated with post-cesarean SSI. Other studies
study,23 however, demonstrated that patients can accurate- have also reported a contaminated or dirty operation as a
ly diagnose the absence of a wound complication but are risk factor.17,27 Internal fetal monitoring and chorioamnion-
less accurate in diagnosing the presence of an infection. itis, although they occurred in only a few cases, appeared to
Patients frequently confuse serous discharge with pus predispose women strongly to SSI. A challenge exists to de-
and, therefore, this marker may overestimate infection crease the frequency of internal fetal monitoring and treat
rates. The results of the current study also support the chorioamnionitis as soon as possible.
latter because nine of the patients self-reported the SSIs, There were no significant differences regarding infec-
which were not confirmed by a physician. Confidence in tions between elective and non-elective groups or between
the results should be improved by gathering information those receiving and not receiving appropriate antibiotic pro-
from patient records, microbiologic findings, and discus- phylaxis. A Cochrane Review from 2002 recommends pro-
sions with the physician. The current data were collected phylactic antibiotics to all women undergoing cesarean sec-
from several sources, and SSIs were always confirmed by tion.1 According to our current hospital policy, antibiotics
the physician. All of the SSI diagnoses were determined should be given to only high-risk groups. Because for 73%
by the investigator to have met the CDC criteria. of the patients our current hospital antibiotic prophylaxis
Multi-method postdischarge surveillance has been policy was followed, we cannot conclude whether selective
described as cost-effective in several studies.4,19,20 The cur- prophylactics is a better alternative than routine prophylac-
rent study included the costs of labor, postage, and tele- tics. The results of this study indicate the need for interven-
phone calls. Sands et al.24 evaluated the use of automated tion to improve the rational use of antibiotic prophylaxis in
ambulatory diagnosis, testing, and pharmacy code screen- consonance with the hospital guidelines.
ing combined with discharge diagnosis to identify SSI in It has been proved in several studies that a significant
non-obstetric patients undergoing surgery. They found that increase in hospital stay occurs when a patient acquires
ambulatory code screening was a sensitive method for de- SSI.13,19 A significantly longer hospital stay also occurred in
tecting patients with SSI. Yokoe et al.8 screened automated the current study once SSI was identified.
ambulatory medical records, hospital and emergency de- Healthcare-associated infection control is a relatively
partment claims, and pharmacy records and found that this new field in Estonia, and we have to develop a surveillance
454 INFECTION CONTROL AND HOSPITAL EPIDEMIOLOGY May 2005

system and propose a control strategy for nosocomial infec- wound infections among women undergoing cesarean section. Obstet
Gynecol 1988;72:559-564.
tions. The baseline quantification of nosocomial infection 13. Killian CA, Graffunder EM, Vinciguerra TJ, Venezia RA. Risk factors
rates and their comparisons with external rates enable hos- for surgical-site infections following cesarean section. Infect Control
pitals to direct infection surveillance and control programs. Hosp Epidemiol 2001;22:613-617.
14. Myles TD, Gooch J, Santolaya J. Obesity as an independent risk fac-
The next step in our hospital would be to inform physicians tor for infectious morbidity in patients who undergo cesarean deliv-
about the results and set up further follow-up of SSIs. As er y. Obstet Gynecol 2002;100:959-964.
15. Lasley DS, Eblen A, Yancey MK, Duff P. The effect of placental re-
greater efforts are made to quantify and describe the char- moval method on the incidence of postcesarean infections. Am J Ob-
acteristics of healthcare-associated infections in Estonia, stet Gynecol 1997;176:1250-1254.
active surveillance, application of prevention interventions, 16. Mah MW, Pyper AM, Oni GA, Memish ZA. Impact of antibiotic pro-
phylaxis on wound infection after cesarean section in a situation of
and judicious antimicrobial use should greatly improve expected higher risk. Am J Infect Control 2001;29:85-88.
patient outcomes. 17. Horan TC, Edwards JR, Culver DH, Gaynes RP. Risk factors for in-
cisional surgical site infections after cesarean section: results of a 5-
year multicenter study. Infect Control Hosp Epidemiol 2000;21:145.
REFERENCES 18. Karki T, Truusalu K, Vainumae I, Mikelsaar M. Antibiotic suscepti-
1. Smaill F, Hofmeyr GJ. Antibiotic prophylaxis for cesarean section. Co- bility patterns of community- and hospital-acquired Staphylococcus
chrane Database Syst Rev 2002;CD000933. aureus and Escherichia coli in Estonia. Scand J Infect Dis 2001;33:333-
2. Ott WJ. Primar y cesarean section: factors related to postpartum infec- 338.
tion. Obstet Gynecol 1981;57:171-176. 19. Stockley JM, Allen RM, Thomlinson DF, Constantine CE. A district
3. Yalcin AN, Bakir M, Bakici Z, Dokmetas I, Sabir N. Postoperative general hospital’s method of post-operative infection sur veillance in-
wound infections. J Hosp Infect 1995;29:305-309. cluding post-discharge follow-up, developed over a five-year period. J
4. Noy D, Creedy D. Postdischarge sur veillance of surgical site infec- Hosp Infect 2001;49:48-54.
tions: a multi-method approach to data collection. Am J Infect Control 20. Taylor EW, Duffy K, Lee K, et al. Telephone call contact for post-dis-
2002;30:417-424. charge sur veillance of surgical site infections: a pilot, methodological
5. National Nosocomial Infections Sur veillance System. National Noso- study. J Hosp Infect 2003;55:8-13.
comial Infections Sur veillance (NNIS) System report: data summar y 21. European Sur vey of Information Society. Estonia: Innovative IT Solu-
from Januar y 1992 to June 2002, issued August 2002. Am J Infect Con- tions. Tartu, Estonia: Foundation Archimedes, Inc.; 2004. Available at
trol 2002;30:458-475. www.esis.ee/ist2004/403.html. Accessed March 1, 2004.
6. Smyth ET, Emmerson AM. Surgical site infection sur veillance. J Hosp 22. Seaman M, Lammers R. Inability of patients to self-diagnose wound
Infect 2000;45:173-184. infections. J Emerg Med 1991;9:215-219.
7. Mangram AJ, Horan TC, Pearson ML, Silver LC, Jar vis WR. Guideline 23. Whitby M, McLaws ML, Collopy B, et al. Postdischarge sur veillance:
for prevention of surgical site infection, 1999. Infect Control Hosp Epi- can patients reliably diagnose surgical wound infections? J Hosp Infect
demiol 1999;20:250-278. 2002;52:155-160.
8. Yokoe DS, Christiansen CL, Johnson R, et al. Epidemiology of and sur- 24. Sands K, Vineyard G, Platt R. Surgical site infections occurring after
veillance for postpartum infections. Emerg Infect Dis 2001;7:837-841. hospital discharge. J Infect Dis 1996;173:963-970.
9. Couto RC, Pedrosa TM, Nogueira JM, Gomes DL, Neto MF, Rezende 25. Platt R, Yokoe DS, Sands KE. Automated methods for sur veillance of
NA. Post-discharge sur veillance and infection rates in obstetric surgical site infections. Emerg Infect Dis 2001;7:212-216.
patients. Int J Gynaecol Obstet 1998;61:227-231. 26. Martone WJ, Nichols RL. Recognition, prevention, sur veillance, and
10. Roy MC, Perl TM. Basics of surgical-site infection sur veillance. Infect management of surgical site infections: introduction to the problem
Control Hosp Epidemiol 1997;18:659-668. and symposium over view. Clin Infect Dis 2001;33(suppl 2):S67-S68.
11. Tran TS, Jamulitrat S, Chongsuvivatwong V, Geater A. Risk factors for 27. Eriksen HM, Chugulu S, Kondo S, Lingaas E. Surgical-site infections
postcesarean surgical site infection. Obstet Gynecol 2000;95:367-371. at Kilimanjaro Christian Medical Center. J Hosp Infect 2003;55:14-20.
12. Emmons SL, Krohn M, Jackson M, Eschenbach DA. Development of