El-Dhuwaib, Yesar (2021) Evaluation of surgical outcomes using administrative hospital data. Doctoral thesis, Keele University.

[thumbnail of El-DhuwaibPhD2021.pdf]

Download (2MB) | Preview


Hospital Episode Statistics (HES) is the main hospital administrative dataset in England. Since the early 1990s HES has been used for research. The aim of the thesis is to evaluate the use of HES data to measure surgical outcomes.

1. Can HES data be used to measure rare surgical complications when there is a code?
2. Can HES data be used to measure rare surgical complications when there is no code?
3. Can HES data be used to compare different types of surgery for the same condition?
4. Can HES data be used to measure changing trends in surgical practice?
5. Can HES data be used as a national audit tool?

Methods: Hospital Episode Statistics from @1, @2, @3. @4 and @5 were used to address each individual objective. HES data were obtained from the National Health Service (NHS) Information Centre and stored into Microsoft SQL server for analysis. HES data were cleaned and converted into clinical database. Patients' data were identified using operative and diagnostic codes. All duplicates were removed and patients were followed across time and place to identify complication. For each study a different Statistical approach was planned to answer each question.

Objective 1: Venous Thromboembolism (VTE) following colorectal resection was used to answer this objective. A total of 35,997 underwent colorectal resection during the period of study. The VTE rate was 2.3%. Two hundred and one (0.56%) patients developed VTE during the index admission and 571 (1.72%) were readmitted with VTE. Following discharge from the index admission, the risk of VTE in patients with cancer remained elevated for six months compared with two months in patients with benign disease. Age, postoperative stay, cancer, emergency admission, and emergency surgery for patients with inflammatory bowel disease (IBD) were all independent risk factors associated with an increased risk of VTE. Patients with ischaemic heart disease and those having elective Minimal Access Surgery (MAS) appear to have lower levels of VTE.
Objective 2: Bile duct reconstruction following Laparoscopic Cholecystectomy (LC) in England was used for this objective. Five hundred seventy-two thousand two hundred and twenty three LC and attempted LC were performed in England between April 2001 and March 2013. Five hundred (0.09%) of these patients underwent Bile Duct Reconstruction (BDR). The risk of BDR for Admission with other causes is significantly lower than acute cholecystitis (Odds ratio OR0.48 (95%CI 0.30—0.76). The regular use of On Table Cholangiogram (OTC) (OR 0.69 (0.54—0.88) and high consultant caseload >80 LC/year (OR 0.56 (0.39—0.54) reduced the risk of BDR. Patients who underwent BDR were 10 times more likely to die within a year than those who did not require further surgery (6% vs. 0.6%).
Objective 3: Laparoscopic versus open repair of inguinal hernia: a longitudinal cohort study was used for this objective. Between April 2002 and April 2004 125,342 patients who underwent inguinal hernia repair were included in the analysis and they were followed until April 2009. There were no differences in postoperative stay between the laparoscopic and open groups except for the laparoscopic bilateral hernia repair patients who had a shorter stay than the open group. Infection and bleeding were more common following open repair, whilst urinary retention and injury to an organ were more frequent after laparoscopic repair. Reoperation for another inguinal hernia was more common after laparoscopic (4.0%) than after open repair of primary inguinal hernia (2.1%). There was no difference in reoperation rate following repair of a recurrent inguinal hernia. Consultant caseload was strongly inversely correlated with reoperation following laparoscopic but not open repair of primary inguinal hernia.
Objective 4: Epidemiological trends in surgery for rectal prolapse in England was used for this objective. During the study period, a total of 25,238 adults underwent a total of 29,379 operations for rectal prolapse [median age 73 years (IQR 58-83) years; female to male ratio: 7:1]. Median LOS was 3 days (IQR 1-7) with an overall in-hospital mortality rate of 0.9%. Numbers of total admissions (2001: 4,950 vs. 2012: 8,927) and of patients undergoing prolapse surgery (2001: 2,230 vs. 2012: 2,808) significantly increased (P < 0.001 for trends) throughout the study period. The overall increase in surgery (up about 1/3rd overall and 44% for elective) was dwarfed by an increase in popularity of laparoscopic surgery (increased 15-fold during the period). Overall prolapse reoperation rate (as a surrogate of recurrence) was 12.7%. The lowest recurrence rate was observed for elective open resection (9.1%) but this had the highest mortality (1.9%) Laparoscopic and perineal fixations were also associated with low reoperation rates (<11%) but lower mortality rates of approx. 0.3% for elective surgery. The data refute a trend toward subspecialisation (by surgeon or hospital) during the study period.
Objective 5: Definitive management of Gallstone pancreatitis in England was used for this objective. A total of 5,454 patients were admitted with GSP between April 2007 and April 2008, of whom 1,866 (34.2%) underwent definitive treatment according to BSG guidelines, 1,471 on the index admission. Patients who underwent a cholecystectomy during the index admission were less likely to be readmitted with a further bout of GSP (1.7%) than those who underwent endoscopic sphincterotomy alone (5.3%) or those who did not undergo any form of definitive treatment (13.2%). Of those patients who did not undergo definitive treatment before discharge, 2,239 received definitive treatment following discharge but only 395 (17.6%) of these had this within 2 weeks. Of the 505 patients who did not undergo definitive treatment on the index admission and who were readmitted as an emergency with GSP, 154 (30.5%) were admitted during the 2 weeks immediately following discharge.

Hospital Episode Statistics can be used to measure surgical outcome in a number of useful and reproducible ways. HES can be used to measure mortality, complications, compare different surgical approaches, assess the effect of changes in practice, and assess caseload outcome association. Those metrics can be used to inform health care planners, develop guidelines, inform patients, and reward hospitals for improved outcome. The use of HES, however, has weaknesses which to a certain extent could be overcome easily with minor alteration in the way that diagnostic, consultant/operator and procedure fields are recorded.

Item Type: Thesis (Doctoral)
Subjects: R Medicine > R Medicine (General)
Divisions: Faculty of Medicine and Health Sciences > School of Medicine
Contributors: Pandyan, A (Thesis advisor)
Depositing User: Lisa Bailey
Date Deposited: 23 Mar 2021 17:08
Last Modified: 23 Mar 2021 17:08
URI: https://eprints.keele.ac.uk/id/eprint/9294

Actions (login required)

View Item
View Item