Understanding unplanned return to theatre in rural South Australia general surgery: review of four major hospitals over a six‐year period

Unplanned return to theatre (URTT) is associated with longer hospital stay and higher mortality rates, placing extra burden on hospital resources. There is a lack of literature analysing causes of URTT in a rural general surgery department. This knowledge may be important to help identify patients at risk of URTT. This study aims to identify causes of URTT in rural general surgical patients.


Introduction
Surgery is a common cause of adverse events in hospital admitted patients. 1,2 These adverse events include complications after surgery that result in patients undergoing unplanned return to theatre (URTT) for reoperation. URTT are associated with longer length of hospital stay and higher mortality rate. 3,4 These URTT not only result in poorer patient outcomes but potentially put a strain on a hospital's resources. This is particularly relevant for rural hospitals with reduced resources, and limited availability for operating theatre space and staffing.
Being aware of the causes of URTT is important to help prepare aspiring rural general surgeons to identify unexpected outcomes and manage them appropriately. Additionally, this knowledge may help identify patients at risk of URTT and allow steps to be taken to mitigate against URTT. Information regarding rates of URTT and associated complication is also essential for obtaining informed operative consent. To the best of our knowledge, there is currently no literature available regarding all causes of rural general surgical URTT and associated risk factors. By auditing a six-year experience, this multicenter study aims to analyse all-causes of URTT in rural general surgical patients to identify potential areas of improvement. A secondary outcome of this study is to identify if there are any differences in URTT for elective versus emergency admissions.

Methods
This is a retrospective multicenter cohort study of operative patients from February 2014 to March 2020, involving four rural South Australian (SA) hospitals: Mount Gambier (MGH), Whyalla (WH), Port Augusta (PAH), and Port Lincoln (PLH). According to the Rural, Remote and Metropolitan Area (RRMA) these hospitals are classified as rural and remote. 5 Hospital setting, population coverage, and specialist availability were described in our previous papers. 6,7 All four hospitals have 24-h access to operating theatres and none of the hospitals have a formal intensive care unit (ICU) on site.
Data was obtained from six-monthly departmental audits which were peer reviewed at The Queen Elizabeth Hospital, Adelaide, and South Australia with the surgeons from respective rural hospitals. Data was sourced from the Operating Room Management Information System (ORMIS), a clinical administration system used in theatres for patient management and billing. Every operation performed has to be registered into ORMIS prior to proceeding. Data extracted includes demographics (age and gender), index surgery performed, indication for URTT, length of stay (LOS), and mortality. Index surgery was defined as the primary operation which the patient underwent resulting in the need for URTT. Based on the index surgery, cases were categorized into their surgical subspecialties. Cases were categorized as general surgery cases if they were common conditions that all general surgeons can manage (e.g., hernia repair, abscess drainage, and carpal tunnel syndrome repair). If patients underwent more than one URTT, the number of days to the first URTT was used. Admissions were classified either as emergency or elective. Emergency admissions were defined as cases where the patient needed to be admitted within 24 h for management. 8 To capture all causes of URTT, patients who had any URTT related to a short-term complication from the index surgery were included in this study regardless of the timeframe. Patients who were transferred to a metropolitan hospital for their URTT were also included. Patients with appropriate planned escalation of treatment were excluded from this study (e.g., operative repair of bleeding peptic ulcer disease after unsuccessful endoscopic management or abscesses with planned relook and washout). Categorical variables are presented as frequency and percentage, and continuous variables are presented as median and range. Data with outliers were expressed as median with interquartile range (IQR). Days until URTT and LOS are displayed as median and interquartile ranges (IQR). Further subgroup analysis was done to identify different characteristics between elective and emergency cases. Univariate analyses were performed using the Mann-Whitney U test or t test for continuous variables, the chi-squared and Fisher's exact tests for categorical variables. Statistical analysis was performed using IBM SPSS Statistics Version 27 (SPSS Inc). A statistically significant P-value was defined as ≤0.05.
As this project fell under audit and quality assurances, formal ethical review was not required. All data have been managed appropriately under the Australian code of the Responsible Conduct of Research.

Results of all admissions
From February 2014 to March 2020, 44 191 surgical procedures were performed and 67 (0.15%) of these cases had URTT. 64% of all patients were male (three of the patient's gender data were unavailable for analysis). Median age was 59 years (IQR = 30.0, 75.8). Emergency admissions accounted for 24% of these URTT cases. Of the index surgery, 19 (28.4%) were for malignancy of which majority of these cases were related to colorectal cancer (n = 14(20.9%)) followed by cutaneous malignancy (n = 3(15.8%)).
The median number of days between index surgery and URTT was 3 days (IQR 6, range between zero (same day as index surgery) and 33 days). Only one case of URTT was outside a 30 day range which was a delayed presentation of retrocaecal abscess post laparoscopic appendectomy (represented on day 33). The remaining 66 cases of URTT occurred within the 30 day range. Three (4.5%) of the cases required more than one URTT. Two of these cases were patients who developed early adhesional bowel obstruction post colorectal cancer resection requiring repeated adhesiolysis. The last case needing more than one URTT was a reversal of Hartmann's surgery which sequentially developed an intraperitoneal haemorrhage after an anastomotic leak.
Of the 67 cases of URTT, 2 (3%) had negative laparotomies. One of the negative laparotomies was a patient who underwent open peptic ulcer disease (PUD) repair and post-operatively developed worsening abdominal pain, high bilious drain output, and had a systemic inflammatory response syndrome (SIRS). Upon URTT, no cause of bilious drain output was found, and the PUD repair site was intact. The other negative laparotomy was a patient who had an open repair of sigmoid mesocolon tear from a motor vehicle accident. Postoperatively the patient had worsening abdominal pain and SIRS, however, no additional injury was found during URTT. It is worth nothing this was the only URTT associated with trauma (1.5%).
The median LOS for patients having an URTT was 8 days (IQR 9). 12 (17.9%) of cases needed to be transferred to a metropolitan hospital. Seven (10.4%) were due to need for intensive care unit (ICU) support, 2 (3%) were transferred for quaternary specialist input, and 1 (1.5%) case needed (ERCP) and ICU support. Two URTT (3%) cases resulted in mortality. Both patients were in their early eighties and died from aspiration pneumonia around 22 days from index surgery.

Results of elective versus emergency admissions
Comparing elective and emergency surgical patients, there were no statistical differences between gender, median age, speciality type, types of surgery performed, median number of days until URTT, and transfer to metropolitan hospital (see Table 2). URTT cases after elective surgery have statistically more malignancy related to index surgery (P = 0.0277), and more elective index surgeries were performed open (P = 0.0048).

Discussion
The rate of URTT is regarded as an indicator of quality of care in general surgery. 9 Unplanned reoperations have multiple detrimental sequalae such as draining limited hospital resources, causing psychological and social distress in patients, and loss of patient confidence in the health care system. 10 The lack of existing rural URTT studies make comparison of our data difficult, however, previous American studies quote URTT rates in metropolitan general surgery patients ranging between 3.5% and 5.6%. 9,11 Of the 44 191 surgical procedures (excluding endoscopy cases) performed during the study period, our URTT rate was 0.15%. A possible reason for the low rates of URTT in rural hospitals is that consultant surgeons often more involved as primary surgeon or first assistant when compared to metropolitan hospitals where some procedures are managed by fellows and senior registrars. Another possible reason for the low rates of URTT could be partly attributed to the fact that the majority of surgery performed in the four rural hospitals were minor procedures. As described in our previous paper, the most common surgeries performed in our hospitals were skin lesion excision (6349 (14.4%)), cholecystectomies (2210 (5%)), and drainage of abscess/debridement of wounds (1729 (3.9%)). 7 These statistics substantiates that minor procedures can be performed safely in our rural centres with low URTT rates and low associated mortality rates (3%). The two mortalities following URTT during the study period were due to aspiration pneumonia in octogenarians. In our previous paper exploring causes of rural general surgical mortality, aspiration pneumonia was shown to be the most common potentially preventable cause of rural general surgical mortality. 6 Methods to prevent aspiration pneumonia were also explored in our previous paper. In this study, URTT does not appear more likely to be associated with emergency surgery (24%), trauma (1.5%), or cancer related surgery (28.4%).
The top two operations done during URTT were: 22 (32.8%) washouts, and 11 (16.4%) haemostasis AE ligation of bleeding vessels. The most common indication for washouts was post-operative haematoma (8, 12%) without active bleeding. The retrospective nature of this study makes it impossible to determine if the URTT for post-operative haematoma and bleeding were preventable. However, it highlights the importance of careful reevaluation of haemostasis at the end of surgery before closure such as checking for pooling of blood, slow ooze, and appropriate placement of surgical clips. If there are ongoing concerns regarding bleeding risk, topical haemostatic agents such as thrombin-based agents or sealants should be considered intra-operatively. The third most common surgery performed during URTT was bowel resections 9 (13.4%). Colorectal was also the most common surgical subspeciality encountered 36 (54%), followed by 19 (28%) general surgical, 5 (7.5%) plastic surgery, 3 (4.5%) upper gastrointestinal. Some general surgery trainees do not have the opportunity to undertake rotations in the aforementioned subspecialties. Aspiring rural general surgeons should have a tailored curriculum to include subspecialities such as colorectal surgery, and plastic surgery. This is further supported by our previous paper analysing the rural general surgical caseload which found that almost 40% of procedures performed were considered non-general surgical operations. 7 There was no significant difference between gender, age, surgical speciality, nor types of surgery performed during URTT when comparing between elective and emergency index surgeries needing URTT. The only statistically significant difference between the two groups was that elective index surgery that need URTT tend to be related to malignancy (35.3% versus 6.3%, P = 0.0277) and performed open (82.3% versus 56.3%, P = 0.0048). This is most likely due to the nature of cancer that tended to be diagnosed early in a primary care or elective setting (screening or surveillance). The earlier presentation allowed for planning of cancer surgeries with fewer requiring emergency surgeries.
We recognize that there are limitations to this study such as its retrospective nature, making some factors unavailable for analysis such as comorbidities, American Society of Anaesthesiologist (ASA) score, and if the URTT were preventable. Additionally, the small sample size may make comparison of elective versus emergency statistically inaccurate. Lastly, we recognize that this study may not be representative as all rural hospitals have different infrastructure, resources, and speciality availability. To the best of our knowledge, this is the first study exploring URTT in the rural general surgical population.
In conclusion, rates of URTT are low in rural South Australian general surgery departments, with low associated mortality rates. Only 28% of index surgery were general surgical cases. This further supports the need for general surgical trainees to have a tailored curriculum to include surgical subspecialities.