Practice and outcomes for radical urological cancer surgery in England: a study based on hospital episode statistics and a review of the literature

Complete, timely and accurate data are needed to monitor the quality of surgical care in the UK. However, large-scale national audits are labour-intensive and expensive. Furthermore, within surgery many studies have noted wide variation between post operative outcomes amongst hospitals. Studies have also tried to determine whether the number of cases performed at a hospital is one component that can influence surgical outcomes. With this in mind, the overall aims of this thesis were to describe the practice and outcomes for radical urological cancer surgery in England, and to determine whether the number (or 'volume') of procedures that a hospital carries out is a potential determinant of outcomes. The objectives of this thesis were fivefold. Firstly, to use administrative data from Hospital Episode Statistics (HES) of the English Department of Health between 1995 and 2002 to report activity and outcomes for patients recorded as having undergone a radical prostatectomy (RP), radical cystectomy (RC) or radical nephrectomy (RN). Secondly, to develop and validate a comorbidity index for the first time within the HES database. The obtained index, based on the Charlson score, increased with age, was higher in patients admitted as an emergency, and also predicted outcomes. Thirdly, to review the literature to establish to what extent volume could be an explanation for variation in outcomes between surgical providers. This review concluded that surgical outcomes, on average, should improve with increasing hospital or surgeon volume, although the evidence for RN was weakest. Fourthly, to establish, through conducting a national survey, the views of the surgeons performing these procedures regarding 'volume and outcome' relationships. This survey discovered that most consultants supported the principle of volume thresholds, but with wide variation as to where thresholds should be set. Finally, to determine through analysis of the HES data whether post-operative outcomes in England depend on the volume of procedures that a hospital performs. In brief, hospital volume was inversely related to in-hospital mortality following RC. No similar effect was demonstrable following RN. No effect of volume on length of stay after RC or RN was identified, but length of stay after RP was shorter in high volume hospitals. Data from this thesis provides some evidence in support of volume-based health policies for RC, but less support for such policies relating to RN or RP