SSI in Cesarean Delivery with CHG Data
Description
This process improvement (PI) project was designed to determine whether surgical irrigation with a 0.05% CHG solution would reduce the number of Surgical Site Infections (SSI) following Cesarean Delivery (CD) at a 150-bed community hospital in San Antonio, TX. The control period was defined as Dec 2013 through Nov 2014 and the treatment period was between Dec 2014 and Nov 2015. All CD were performed by one of eleven physicians practicing at the facility. All data collected during both the control and the treatments cohorts was collected prospectively. There were no exclusions. The primary outcome was the presence or absence of post-cesarean SSI. The intervention used as the treatment was irrigation with 0.05% CHG in sterile water followed by a normal saline rinse. Treatment period SSI numbers were compared with SSI numbers among women who underwent CD by the same group of physicians during the control period. During the control period, in which there was no official irrigation protocol, the practice, if performed, was done with normal saline. Prospective data from the control period, including all patient and procedure characteristics, were tracked during the treatment period and extracted from EMRs by the women services’ management team. Cross-checks were done to verify accuracy. All patients were evaluated for infection using the CDC criteria for superficial, deep, and organ space SSI. The treatment group SSI rate was slightly lower (1.5% [5/343]) than the control group (1.8% [7/399]), but this was not statistically significant (p=0.780). Multiple logistic regression analysis revealed race Hisp/Latino and non-elective/urgent/emergent status as positive predictors of SSI. A significant interaction effect between irrigation with 0.05% CHG and antibiotic administration time (AAT) existed such that SSI occurrence in the control group was dependent on AAT (each additional minute between administration and incision lowering odds for SSI) while SSI occurrence in the treatment group was not dependent on AAT. The model indicates that AAT explains a portion of the protective effect against SSI occurrence, but only in the control group. Thus, treatment (0.05% CHG irrigation) is the effective factor in mitigating SSI risk in the treatment group, regardless of AAT. These results are consistent with subset correlation and simple logistic regression analyses. AAT was shown to be negatively correlated with SSI occurrence (lower time related to SSI) in the control subset only (r=-0.131, p-value=0.010), while no correlation was shown between AAT and SSI occurrence in the treatment subset only (r=-0.009, p=0.863). In subset simple logistic regression analysis, AAT was the only significant predictor (p-value=0.010) of SSI in the control subset, while it held no significance in predicting SSI in the treatment subset (p-0.372). This suggests a role for 0.05% CHG irrigation in mitigating SSI risk whether antibiotic prophylaxis timing is suboptimal or ideal.
Files
Steps to reproduce
All CD patients received antibiotic prophylaxis using weight-based dosing of IV cefazolin, unless they had a beta lactam allergy, in which case they received 900mg of clindamycin and a weight-based dose of gentamicin. AAT varied based on the patient’s urgency, but departmental protocol was compliant with the CDC's and ACOG’s recommendation that antibiotic prophylaxis be administered within one hour prior to incision. Almost all cases followed the established departmental protocol for skin prep which included the use of a single applicator containing a 2% CHG and 70% isopropyl alcohol solution (ChloraPrep) after hair removal (if necessary) with surgical clippers. 13 of the 742 participants were prepped with a 10% povidone iodine (PVP-I) “splash” in an unmeasured volume on the patient’s skin. None of these “splash” cases resulted in SSI. Vaginal preparation was not performed in either the control or treatment study periods. Incisional irrigation, if performed in the control period, was done with only normal saline at the physician’s discretion, but without a departmental protocol. Treatment period irrigation protocol occurred post-infant extraction and after closure of the fascia. The incision was irrigated with the medical device that was warmed in a fluid warmer designed not to exceed 40 degrees Celsius, and applied under low pressure. The solution was allowed to dwell for 60 seconds, monitored by OR techs. The solution was then suctioned out and the incision was irrigated with 0.9% sterile n All analyses were performed with R statistical software version 3.4.2. Continuous data were summarized using mean, standard deviation, median, and range. Categorical data were summarized using frequency distributions (percentage). Comparisons between control and treatment groups were made using t-tests for continuous variables and using both Pearson’s Chi-squared tests and Fisher’s exact tests for categorical data. In all tests, significance was assumed at p < .05. Pearson’s correlation coefficient (r) was calculated between each factor and SSI occurrence and t-tests were used to test the null hypothesis H0: r=0 with a p-value < .05 considered significant. Logistic regression analysis was performed to determine significant predictors of SSI following CD in the study population. Given the low occurrence of SSI in the sample, a penalized likelihood approach (Firth's method) was used. For multiple logistic regression, both forward and backward stepwise selection procedures were used. Factors considered for inclusion in the model included all variables and all interaction effects with treatment level. Criteria of p<.20 to enter and p≥.20 to be removed were used. After each procedure resulted in an initial model, factors were eliminated one by one from the initial model, according to highest p-value, until only factors significant at p<.10 were left in the resulting final model. Both methods resulted in the same final model.