Prolonged Hospital Stay is Associated with Increase Surgical Site Infection, Gastrointestinal and Hepatobiliary Surgery. – A propensity Matched Analysis Dr.Bhavin Vasavada. Consultant hepatobiliary and liver transplant surgeon, Shalby Hospitals, Ahmedabad

Dr. Hardik Patel. Consultant hepatobiliary and liver transplant surgeon, Shalby Hospitals, Ahmedabad.

following surgery each year, including at least 1 million deaths, a magnitude that exceeds both maternal and AIDS-related mortality. As many as 50 percent of these deaths and complications are preventable. [1] These estimated deaths are more than projected global deaths due to HIV/AIDS by 2030 which is estimated to about 6.5 million. [2] Prolong hospital stay is associated with increased morbidity and mortality in many diseases. [3] However very few studies examined effect of prolonged hospital stay on actual morbidity and mortality. Postoperative hospital stay also found to be associated with surgical site infections. [4]. However, there is very limited data of association of length of hospital stay before diagnosis of morbid event/surgical site infection or occurrence of mortality.

Aim of Study:
Primary aim of the study was to evaluate effect of prolonged hospital stay on Surgica l site infections We also evaluated effect of prolonged hospital stay morbidity and mortality in Gastrointestinal and Hepatobiliary Surgery as secondary outcome.

Study Design:
All gastrointestinal and hepatobiliary procedures performed in last 3 years have been evaluated retrospectively. It is very well known and logical that after complications or wound infection has happened subsequent hospital stay increases for managing those complications. To know cause effect relationship, we considered hospital stay as hospital stay before diagnosis of complications, surgica l site infections or any morbidity or in case of sudden deaths hospital stay before mortality in the study group. For control group consisting of patients with uneventful recovery total hospital stay was considered as hospital stay. To evaluate cutoff for hospital stay we evaluated ROC curves for 90 days morbidity and surgica l site infections.
Surgical site infections were defined as any culture positive discharge from the wound either superficial or deep within 30 days post operatively. [5] Statistical analysis: We used ROC curve analysis with coordinates for each factor to select optimal hospital stay beyond which it was associated with SSI, morbidity and mortality and we found that hospital stay beyond 4 days was associated with SSI, morbidity and 30 days mortality with highest sensitivity and specificity.
To avoid selection bias in attempt to evaluate effect of prolonged hospital stay less than or greater than 4 days on surgical Site infections we did 1:1 propensity score matching analysis with surgical site infections as dependent factor. Propensity scores were calculated using logistic regression. for nonparametric data and student t test for parametric data, after checking for skewness and standard error of skewness and also kurtosis and standard error of kurtosis. Medians were used for nonparametric data. Two-sided p value less than 0.05 was considered as significant. Statistical analysis was done using SPSS version 23.

RESULTS:
We included 348 patients operated between April 2017 and March 2020 in our analysis. After 1:1 propensity score matching 58 patients included in study arm (prevent hospital stay more than 4 days) and 56 patients in control arm. majority of immediate postoperative deaths suggested by median hospital stay in patient who died was 2 days vs 5 days in group with no post-operative mortality.

Discussion:
Prolonged Hospital stay has been shown to be associated with perioperative morbidity and mortality in various surgical disciplines. [6,7]. However very few studies evaluated increase hospital stay before diagnosis of surgical site infectio n, morbidity or occurrence of mortality as a causative factor for increase surgical site infection or morbidity.
Surgical science has progressed to a great extent in last century. Despite such a great progress Surgical site infection remains a major challenge and its incidence rates still remains high due to prevalence of wide range of protocols and practices. [8].
Causes of Surgical site infection can be multifactorial and include variety of patient related, hospital related and procedural related factors and it includes use of variety of protocols and procedures to prevent them. [9] Mortality and morbidity are also related to various perioperative factors like ASA score, intraoperative blood product use, intra operative hypotension and preoperative patients' physiology also contribute to morbidity and mortality.
So, to avoid selection bias and various confounding factors we did propensity score matching analysis with various pre-operative and intraoperative factors as mentioned in study design as cofactors.
After 1:1 propensity score matching we found that increase post-operative hospital stay was associated with increase rate of surgical site infections and 90-day morbidities.
We decided 4 th day as a cutoff based on ROC curve analysis for each factor like SSI, Morbidity and Mortality. [ Figure 3]. All showed cutoff of 3.5-4.5 days showing maximum sensitivity and specificity based on that we decided to take 4 days as optimum cutoff.
In our analysis mortality was significantly higher in patients whose hospital stay was lesser than 4 days. This was due to most of the mortalities were in emergency and sicker patients who died in immediate post-operative period. Williams et al.
also did not find any association between hospital stay and mortality. [11] There are studies about effect of night times and weekend admission on morbidity and mortality. [12,13]. However, we could not find effect of prolonged hospital stay on over all morbidity and mortality.
Mujagic et al [14] showed that increase hospital stay was associated with increase rate of surgical site infections, similar to our findings.
Our studies have certain limitation like low numbers of patients, also SSI, morbidity and mortalities depend on many factors which might have been missed in our analysis.
In conclusion, prolonged hospital stay is associated with increased 90 day morbidity