MATERIALS AND METHODS

We used the Hospital Episodes Statistics (HES), a routinely collected administrative dataset which includes a record of all admissions to National Health Service (NHS) hospitals in England 9, 10. HES data are primarily collected to guide government payment of the activity undertaken by the hospitals and therefore often analysed and reported according to ‘financial years’ (1st April to 31st March).
The study population included all women aged 18 years and above who underwent an elective robotic gynaecological procedure as in-patients between 1st April 2006 and 31stMarch 2018.
Surgical procedures were defined according to the United Kingdom (UK) Office for Population Censuses and Surveys Classification version 4.8 (OPCS-4) codes 11-13: hysterectomy (Q07), adnexal surgery (Q22, Q23, Q24), and urogynaecological surgery, including sacrocolpopexy (P24.2, P24.3, P24.5), sacrohysteropexy (Q54.5), and colposuspension (M52.3). The code Y75.3 was used to identify robotic procedures. Detailed definitions of the OPCS-4 codes that were used are provided in the supplementary material (Appendix S1).
If multiple procedures were recorded during the same hospital admission, only one procedure was counted, first considering hysterectomy, followed by urogynaecological surgery and then adnexal surgery.
Procedure numbers were reported by year to identify time trends in the uptake of surgery. Procedure numbers were also mapped to the 44 geographical areas with an average population size of about 500,000 adult women, known as Sustainability and Transformation Partnership (STP) areas. These areas were set up to improve delivery of NHS services14, 15.
We determined length of hospital stay (LOS) in terms of nights spent in hospital, obtained as the difference between admission date and discharge date. Laparotomy including conversion during the primary procedure or after return to theatre for surgical management of complication was captured using the following OPCS-4 codes: Y50.2, T30; T30.1, T30.2, T30.3, T30.4, T30.8, T30.9. 30-day emergency re-admission was identified by monitoring readmissions to any NHS hospital in England within 30 days of the primary procedure regardless of the cause.
The relationship between categorical variables (laparotomy or readmissions, and year of surgery) was explored using the Pearson chi-square test following tabulations, and where the cell value was less than five, using Fisher’s exact test. The equality of median length of stay across years was tested using the non-parametric k-sample test on the equality of medians in STATA, with cases at the median split evenly between the above and below groups. P-values are reported from the chi-square statistic. P < 0.05 was considered significant.