Discussion

Given the clinical and economic consequences, prevention of SSIs following abdominal hysterectomy surgery is being increasingly emphasized in hospitals. An important component of perioperative bundles is the timely administration of appropriate prophylactic antibiotics. Prior studies have prioritized β-lactam antibiotics (typically cefazolin) for prophylaxis for abdominal hysterectomy procedures. ^{Reference Uppal, Harris and Al-Niaimi17} Clindamycin with an aminoglycoside is recommended for patients undergoing hysterectomy who have a serious penicillin allergy. ^{Reference Bratzler, Patchen Dellinger and Olsen18} In our report, there were more patients with SSIs who received the recommended alternative prophylactic clindamycin with or without gentamicin compared to matched controls. This finding is likely multifactorial: increasing clindamycin resistance has been reported among staphylococci, streptococci, and anaerobic (eg, Bacteroides spp) bacteria. ^{Reference Carrel, Smith and Shi19–Reference Kierzkowska, Majewska and Mlynarczyk21} Also, the average BMI for our patients was >30 kg/m², suggesting suboptimal gentamicin dosing may be a contributing factor. Newer agents, such as tigecycline or eravacycline, should be explored as alternatives for penicillin-allergic individuals. Previous studies have shown that deviations from cefazolin-based prophylaxis may compromise infection prevention efforts, ^{Reference Anderson, Podgorny and Berríos-Torres1,Reference Andiman, Xu and Boyce6,Reference Calderwood, Anderson and Bratzler7,Reference Pop-Vicas, Musuuza and Schmitz22} underscoring the need for robust antimicrobial stewardship programs and accurate allergy documentation.

With the institution of widely accepted bundles, published rates of SSIs following hysterectomy have been varied, ranging from .4%–11%. ^{Reference Vurture, Mendelson, Grigorescu and Lazarou4,Reference Guo, Runge, Miller, Aaby and Milad5,Reference Agarwal, Sinha, Haripriya, Kokkayil and Singh8,Reference Lake, McPencow, Dick-Biascoechea, Martin and Erekson12} Over the final 5 years of our study period, the rate of SSIs was 2.94 ± .61%, and the corresponding standardized infection ratio was 1.260 ±.195 (range 1.035–1.454). We suspect the greater than predicted number of SSIs across our system involves factors beyond the CMS recognized risk factors of age, BMI, ASA score, and diabetes mellitus. Other variables, including some found in our report, that are associated with increased risk of infection include increased duration of surgery, ^{Reference Tserenpuntsag, Haley and Ann Hazamy9,Reference Morgan, Swenson and Streifel10} wound class, ^{Reference Tserenpuntsag, Haley and Ann Hazamy9} and gynecologic cancer. ^{Reference Tserenpuntsag, Haley and Ann Hazamy9,Reference Morgan, Swenson and Streifel10} Laparoscopic and vaginal hysterectomies have consistently lower rates of SSIs compared to open abdominal hysterectomies and are the preferred approaches when feasible. ^{Reference Agarwal, Sinha, Haripriya, Kokkayil and Singh8–Reference Lake, McPencow, Dick-Biascoechea, Martin and Erekson12}

Receiver operating curves demonstrated the current CMS risk adjustment variables were suboptimal for predicting SSIs in our population of patients. Adding other risk variables significantly improved the predictive index for all SSIs, but not for the CMS reported deep incisional and organ/space SSIs.

Nearly all the patients treated in the NYC Health + Hospitals system have Medicaid or Medicare as their health insurance, or are uninsured. A recent study documented posthysterectomy infection rates, for both deep incisional/organ space and superficial SSIs, that were nearly double for Medicaid patients when compared to patients with commercial insurance. ^{Reference Edmiston, Bond-Smith and Spencer23} In our report, 86% of infections were detected after hospital discharge for the hysterectomy surgery. This is remarkably similar to a recent study by our group, where 90% of infections following cesarean sections in our hospital safety net system were detected following hospital discharge; New York City borough infection rates were directly proportional to poverty level of the borough. ^{Reference Akter, Folks and Bakare24} As with cesarean section surgeries, we suspect outpatient wound care following hysterectomy may need to be improved for women of low socioeconomic status.

Several limitations of our study must be acknowledged. First, this was a retrospective analysis relying on surveillance data, which may be subject to underreporting or misclassification. Propensity score matching was used to identify additional factors for infection outside the known CMS risk variables. However, both the infection and control groups were derived from patients of low socioeconomic status attending our safety net hospitals; employing a control group of patients of higher socioeconomic status would likely identify additional factors.

It is apparent that estimating SSIs following hysterectomy is multifactorial and complicated. Current CMS risk assessments for SSIs following hysterectomy have been deemed “inadequate” and do not correlate with other quality summary scores. ^{Reference Minami, Dahlke and Barnard25} The implementation of the Hospital-Acquired Condition Reduction Program and Value-Based Purchasing program, that include hysterectomy SSIs in their measures, have disproportionately penalized safety net hospitals. ^{Reference Hsu, Wang and Broadwell26} The financial consequence of being labeled a “poor performer” is significant, as the CMS penalty of up to 2% of total inpatient charges can be devastating for safety net hospitals. For safety net hospitals, these inaccurate measures of quality may result in augmentation of healthcare disparities. ^{Reference Minami, Dahlke and Barnard25} Until more robust and comprehensive risk assessment criteria are developed, use of SSIs following hysterectomy as a financially-associated quality measure should be reconsidered.