Surgical Site Infections Occurring after Hospital Discharge

0 Channing Laboratory and Departments of Medicine and Surgery, Brigham and Women's Hospital, and Department ofAmbulatory Care and Prevention, Harvard Medical School and Harvard Pilgrim Health Care , Boston, Massachusetts 1 The Journal of Infectious Diseases 1996; 173:963-70 1996 by The University of Chicago. All rights reserved. 0022~ 1899/9617304~0025$Ol.OO 2 Received 27 July 1995; revised 27 November 1995. Presented in part: American Federation for Clinical Research , April 1993, Washington, DC (Clin Res 1993;41 :235A). This study was done in accordance with the human experimentation guide lines of Brigham and Women's Hospital and Harvard Pilgrim Health Care. Financial support: Agency for Healthcare Policy and Research, Department of Health and Human Services (l F32 HS00057 -01); Harvard Pilgrim Health Care Foundation. Hospital , 330 Brookline Ave., Boston, MA 02215 Surgical Site Infections Occurring after Hospital Discharge Although surgical site infections (SSIs) occurring after hospital discharge cause substantial morbidity, their epidemiology is not well understood, and methods for routine postdischarge surveillance have not been validated. Inpatient and outpatient surveillance followed 5572 nonobstetric procedures among ll)embers of a health maintenance organization with extensive automated records. Records were screened for coded diagnoses, tests, and prescriptions and, if positive, were reviewed by reading full text. Questionnaires regarding the occurrence of an SSI were sent to the same patients and their surgeons. One hundred thirty-two SSIs were documented, of which 84% occurred after hospital discharge and 63% were managed outside the surgical facility. Postdischarge SSIs led to an average of 4.6 additional ambulatory encounters. Patient and surgeon questionnaires had a sensitivity of 28% and 15%, respectively. These data suggest that most SSIs occur after discharge and are not detectable by conventional surveillance. Nonetheless, they cause substantial resource utilization. - Postoperative surgical site infections (SSIs) are among the leading nosocomial causes of morbidity and increased medical expense. An estimated 325,000 SSIs occur each year in the United States and generate additional medical costs in the range of $1-2 billion [1, 2]. Routine surveillance for SSIs is recommended by both the Centers for Disease Control and Prevention (CDC) [3] and the Surgical Infection Society [4] as a mechanism for reducing the rate of these infections, presumably by providing feedback to surgeons on their performance [4-6]. SSI surveillance has been associated with decreased infection rates of as much as 35%, [3, 7] but even if the true effect of surveillance is much less than this, it would be highly cost-effective: the cost of surveil lance has been estimated to be about one-fifth the cost of treating preventable infections [4, 8]. SSI rates are thus widely used by hospitals as a quality indicator and as the basis for quality improvement efforts. Decreasing hospital lengths of stay and increasing use of ambulatory surgery may compromise the accuracy of surveil lance data [9]. Traditionally, surveillance involved inpatient follow-up only, but previous studies have reported that as many as 71% of SSIs occur after hospital discharge [10-16]. For hospitals to report accurate absolute infection rates or to make Methods Study population. The study population was drawn from adult members of Harvard Pilgrim Health Care (HPHC) who underwent a nonobstetric operating room procedure at Brigham and Women's Hospital from 10 February 1992 through 7 March 1993. HPHC is a multimodel health maintenance organization that included a staff model division with ~ 300,000 members in the greater Boston area at the time of this study. Members pay a monthly fee, after which office visits, urgent care visits, and hospitalizations generate only nominal charges. At the time of the study, 92% of persons de scribed above received care at centers that used an automated medical record system for both daily charting and archiving; only members with automated records were included in the study. The automated medical record system uses standardized forms that are completed for every patient encounter at HPHC centers, including telephone calls, office visits (scheduled or unscheduled), urgent care visits, and hospitalizations. Information is recorded on forms that are customized for the type of encounter: The provider either writes in or selects from a list all coded diagnoses, tests, procedures, and prescriptions relevant to that encounter and enters additional comments as free text. All information, including free text, is entered into an automated medical encounter record. The results of diagnostic tests are entered directly into the automated record linked to the patient encounter during which they were ordered. Information about hospitalizations and emergency room visits appear in both encounter records and separate administrative records. HPHC pharmacies are also computerized and linked to the automated medical record. Ninety percent of HPHC members have prepaid coverage for pharmaceuticals and so are likely to use HPHC pharmacies for any prescriptions that have more than a nominal charge at outside pharmacies. Brigham and Women's Hospital is the most active surgical facil ity for greater Boston HPHC members. HPHC patients undergoing surgery were identified within 3 weeks of their procedure from the hospital's computerized operating room log, which records information on every operation in real time, including procedure date, surgeons involved, duration of the procedure, wound class, admission status of the patient, and up to three ICD-9 procedure codes. Identification of SSIs. Hospital-based information about SSIs was identified by review of data gathered by the hospital's infection control unit, which uses routine microbiology, nurse's notes, and surgical ward rounds for surveillance, and review of the inpatient charts of all patients receiving the ICD-9 code of 998.5 (postopera tive wound infection) before discharge. Outpatient information about SSIs was identified by a two-step procedure for reviewing outpatient records that used the automated record systems de scribed above. The first step was a computerized search of three automated data bases. Ambulatory encounter records were screened for any of 102 diagnostic, testing, or treatment codes that may have indicated an SSI. A partial list of these codes is given in figure 1. The pharmacy data base was screened for specific antibiotic prescriptions, and administrative records were searched for all rehospitalizations or emergency room visits and their corre sponding ICD-9 codes for those potentially indicative of an SSI. The second step was a physician review of full-text outpatient records and relevant hospital records for the 30-day postoperative period for all patients identified by the computerized search. All records with any (...truncated)