Besides the fact that most SSIs surface only after patients are long gone from your facility, surgical site infections in ambulatory surgery patients haven't been extensively studied. The reason is simple: Traditional surveillance methods don't let us comprehensively detect surgical site infections in ambulatory surgery. You can probably guess why.

• Patient observation times are shortened;
• cultures aren't done or are sent elsewhere;
• patients aren't readmitted, or they go to the ER if they experience complications; and
• empiric antibiotic treatment is usually given in physicians' offices.

But your infection control department must be responsible for these tasks, and for structuring them in a way that is efficient, gets results and meets accreditation requirements. JCAHO, for one, requires post-op surveillance in the ambulatory setting. If that's not incentive enough, active surveillance and regular feedback of SSI rates to surgeons has been shown to reduce post-op infection rates.

An end in mind
I conducted a study to determine the frequency of SSIs in ambulatory surgery patients at Bon Secours Cottage Health Services in Grosse Pointe, Mich. Here's how we ran our study, its results and how you can apply them to developing your own SSI surveillance program.

We entered the surveillance with three aims:

• that data analysis would establish endemic rates of infection to monitor clusters, outbreaks and areas for process improvement;
• that we'd use SSI rates for intra-hospital comparison, as benchmark data for this patient population did not exist; and
• that wound class rates would be reported for all clean and clean-contaminated procedures, and would be calculated and reported separately from the inpatient population.

Of course, we kept all surveillance data strictly confidential and non-discoverable.

How we collected our data
During the two-and-a-half-year study period (May 1999 to October 2001), BSCHS performed 21,654 ambulatory surgical procedures. We enrolled 8,226 (38 percent) of these patients, spanning abdominoplasty, appendectomy, arthroscopy, breast biopsy, breast reconstruction, breast reduction, breast augmentation, mastectomy, carpal tunnel release, lumbar discectomy, facial rhytidectomy, bunionectomy, hammertoe correction, hernia repair, laparoscopic cholecystectomy, laparoscopy and open orthopedic procedures. We excluded contaminated and dirty/infected surgical cases from the study.

Using the CDC definitions of nosocomial SSIs to standardize reporting (see "Defining Surgical Site Infections" on page 68), we monitored patients for signs and symptoms of infection for 30 days post-operatively or up to one year if an implant was present. Here's how we collected our data:

• Each month we sent a line-listing of all ambulatory surgery patients, generated by the OR computer system, to the infection control practitioner.
• At the end of the month after the date of surgery, we mailed to our physicians a surgeon-specific questionnaire that listed patients by surgery date and procedure, and included a yes/no column regarding the development of an SSI. Along with the questionnaire, we included a cover letter and summary of the CDC definition of SSI. We asked physicians to return questionnaires to the ICP by the 15th of the month. The ICP followed up all yes responses with a phone call to the surgeon's office to confirm infections.
• We reviewed admission diagnoses daily to screen for patients admitted with potential post-op complications.
• We reviewed all culture reports daily for possible association with ambulatory surgery.
• For those cases associated with an SSI, we collected the following data through medical record review and interview of the surgeon or office staff: OR room number, ASA score, wound class, implant data, duration of surgery, antibiotic therapy, culture reports, date of onset, signs or symptoms of infection, type of SSI, other infections, hospital admission and method of notification by established monitors.

When docs document
While JCAHO's requirement of SSI surveillance for ambulatory surgery should serve as an incentive for healthcare facilities to develop a surveillance method that's easy to perform and enhances case finding of post-operative infections, literature shows that passive reporting involving patient questionnaires or telephone surveys has a low sensitivity for detecting SSIs. The ideal situation is one in which a limited number of validated observers see all post-op patients at preset intervals and use standardized objective criteria. Unfortunately, meeting all of these conditions is seldom possible.

But our monthly surgeon-specific questionnaires contributed significantly to the detection of SSIs that would otherwise have gone undetected (see "SSI Rates Identified by Surgeon Questionnaire vs. Traditional Surveillance" on page 67). Traditional surveillance uncovered 24 infections - which would have been an overall SSI rate of 0.29 percent, compared with the observed rate of 0.75 percent (38 total infections). That means surgeon questionnaires identified an additional 14 (37 percent) infections. Using a two sample T-test, this difference is statistically significant (p=0.003). This was despite the fact that we had a slightly lower response rate for returned surveys than that reported in the literature. The mean response rate for the surgeon-specific questionnaire was 66 percent, representing 61 percent of the initial study population (see "Rating the Responses" on page 69). That means we followed 5,039 procedures by monthly survey, culture report review and screening of hospital admission data, while we lost 3,187 (39 percent) procedures.

Finding problem areas
Of the procedures followed up, 4,148 (82 percent) were classified as clean surgery and 891 (18 percent) as clean-contaminated surgery. The overall surgical site infection rate was 0.82 percent for clean and 0.45 percent for clean-contaminated procedures (see "SSI Rate by Wound Class" on page 69).

Superficial SSI represented 27 (71 percent) of the observed infections; the remaining 11 (29 percent) were deep SSI. Hospital admission was necessary in 14 (37 percent) of the cases (nine deep and five superficial SSI). We obtained cultures in 28 (74 percent) of the infected cases. The average length of time from surgery until infection diagnosis was 15.5 days, though diagnosis ranged from three days to 94 days post-op (see "SSI Breakdown" on page 69).

Our highest infection rates were in breast biopsy and orthopedics (see "Distribution of Infections by Surgical Service" and "SSI Rate by Surgical Service" on page 69). Breast biopsy accounted for 12 (31.6 percent) of the infections identified - an SSI rate of 1.48 percent, nearly double our average overall rate.

The percentage of breast biopsy procedures covered by returned surveys was slightly higher than the overall return rate at 66 percent (monthly range: 23 percent to 98 percent), representing 809 procedures. All the patients were women between the ages of 36 and 80 years with a median age of 53 years. Superficial SSI was documented in 10 (83 percent) of the infected breast biopsy cases; the remaining two were classified as deep SSI. We performed cultures in nine of these cases, and four patients required hospital admission for treatment. The average length of time from surgery until infection diagnosis was 7.3 days (range: three days to 15 days).

We identified 10 infections for orthopedic cases, three of which involved arthroscopy; the remaining seven were various open procedures.

The overall 0.75 percent SSI rate for ambulatory surgery is comparable to rates of previously published studies using similar case-finding methods. Our lower response rate may have been due in part to the fact that we didn't follow up with individual surgeons for delinquent surveys. But even without follow-up, the questionnaires, used along with traditional surveillance methods, are effective and efficient for increasing the accuracy of SSIs in the ambulatory patient population.

Streamlined, but continuing
In May 2004, I joined St. John Hospital and Medical Center in Detroit, where I continue to conduct surveillance for ambulatory surgery. The scope of conducting post-discharge surveillance is limited in comparison to the study from BSCHS, but it suits our needs. At SJHMC, I report surgical site infection rates for all clean, clean-contaminated and contaminated procedures performed at our free-standing ambulatory surgery center. I accomplish this through such traditional surveillance methods as reviewing microbiology reports and readmission reports. We limit post-discharge surveillance to one high-risk/high-volume procedure category per year, which we choose based on a review of previous year's surveillance data.