The art and science of surgery: Do the data support the banning of surgical skull caps?

Recommendations of the Joint Commission discourage the use of surgical skull caps in favor of bouffant or helmet headwear; however, data supporting such recommendations are limited and have been questioned in recent studies, as well as by our departmental and hospital leadership. At the end of December 2015, surgical caps were removed from our institution with the theoretic goal of decreasing surgical site infections. We aimed to assess the impact of this intervention on surgical site infection occurrence at our institution. Using our institutional American College of Surgeons National Surgical Quality Improvement Program General and Vascular procedure-targeted data, we identified patients undergoing any surgical procedure classified as clean or clean-contaminated during a 12-month period before and after implementation of the surgical headwear policy. Patients without complete 30-day follow-up were excluded. Cases with active infection at the time of operation were excluded. Vascular surgery operations were excluded because of the implementation of a separate intervention to decrease surgical site infections during the study period. Patients were grouped according to timing of the operation in relation to the policy change (12 months before or after). Descriptive statistics focused on proportions and adjusted logistic regression models were used to investigate the association of alternative headwear use with any type of surgical site infection. Models were adjusted for potential confounders that included demographics and clinical characteristics (age, sex, race or ethnicity, obesity, diabetes, steroid use, smoking status, cancer, urgency of the operation, and wound classification). A total of 1,901 patients underwent 1,950 procedures during the study period, with 767 (40%) before and 1,183 (60%) after the headwear policy measure was adopted. The most common procedures overall were colectomy (18%), pancreatectomy (13.5%), and ventral hernia repair (8.9%). The overall rate of any surgical site infection was 5.4%, with no difference before and after policy implementation (5.3% versus 5.5%; P = .81). Multivariate analysis controlling for age, sex, race or ethnicity, obesity, diabetes, smoking status, steroid use, cancer diagnosis, and type of wound classification showed no association between implementation of this new policy and surgical site infections occurrence (odds ratio 1.12 [95% confidence interval 0.73–1.71]; P = .59). In our institution, the strict