The impact of comorbidities on postoperative outcomes of ventral hernia repair: the patients' perspective

Introduction Ventral hernia repair (VHR) outcomes can be adversely affected by modifiable patient co-morbidities, such as diabetes, obesity, and smoking. Although this concept is well accepted among surgeons, the extent to which patients understand the significance of their co-morbidities is unknown, and a few studies have sought to determine patient perspectives regarding the impact of their modifiable co-morbidities on their post-operative outcomes. We attempted to determine how accurately patients predict their surgical outcomes after VHR compared to a surgical risk calculator while considering their modifiable co-morbidities. Methods This is a prospective, single-center, survey-based study evaluating patients' perceptions of how their modifiable risk factors affect outcomes after elective ventral hernia repair. Pre-operatively, after surgeon counseling, patients predicted the percentage of impact that they believed their modifiable co-morbidities (diabetes, obesity, and smoking) had on 30-day surgical site infections (SSI) and hospital readmissions. Their predictions were compared to the Outcomes Reporting App for CLinicians and Patient Engagement (ORACLE) surgical risk calculator. Results were analyzed using demographic information. Results 222 surveys were administered and 157 were included in the analysis after excluding for incomplete data. 21% had diabetes, 85% were either overweight with body mass index (BMI) 25-29.9 or obese (BMI >= 30), and 22% were smokers. The overall mean SSI rate was 10.8%, SSOPI rate was 12.7%, and 30-day readmission rate was 10.2%. ORACLE predictions correlated with observed SSI rates (OR 1.31, 95% CI 1.12-1.54, p < 0.001), but patient predictions did not (OR 1.00, 95% CI 0.98-1.03, p = 0.868). The correlation between patient predictions and ORACLE calculations was weak (p = 0.17). Patient predictions were on average 10.1 +/- 18.0% different than ORACLE, and 65% overestimated their SSI probability. Similarly, ORACLE predictions correlated with observed 30-day readmission rates (OR 1.10, 95% CI 1.00-1.21, p = 0.0459), but patient predictions did not (OR 1.00, 95% CI 0.975-1.03, p = 0.784). The correlation between patient predictions and ORACLE calculations for readmissions was weak ( p = 0.27). Patient predictions were on average 2.4 +/- 14.6% different than ORACLE, and 56% underestimated their readmission probability. Additionally, a substantial proportion of the cohort believed that they had a 0% risk of SSI ( 28%) and a 0% risk of readmission (43%). Education, income and healthcare employment did not affect the accuracy of patient predictions. Conclusions Despite surgeon counseling, patients do not accurately estimate their risks after VHR when compared to ORACLE. Most patients overestimate their SSI risk and underestimate their 30-day readmission risk. Furthermore, several patients believed that they had a 0% risk of SSI and readmission. These findings persisted regardless of level of education, income level, or healthcare employment. Additional attention should be directed toward setting expectations prior to surgery and using applications such as ORACLE to assist in this process.