Association of Coloproctology of Great Britain & Ireland (ACPGBI): Guidelines for the Management of Cancer of the Colon, Rectum and Anus (2017) - Audit and Outcome Reporting.

Abstract

Defining quality in colorectal cancer surgery is complex. Single metrics that describe safety or performance fail to adequately describe the complexity of colorectal care (Almoudaris et al., 2013). Meaningful performance measurement is dependent upon a) case-mix adjustment, b) an understanding of volume-outcome effects and c) reliable data sources. Old age and patient co-morbidity significantly increase the likelihood of peri-operative events and mortality. To enable comparison of surgical performance, adjustment must be made for case-mix differences between providers. Scoring systems, such as Charlson (Charlson et al., 1987) and Elixhauser (Elixhauser et al., 1998) are used widely to case-mix adjust models derived from large clinical datasets. Currently risk stratification models used for comparison of outcomes have limitations as they do not adjust for case mix, including BMI (body mass index) and socio-economic deprivation (Harris et al., 2009; Haydon et al., 2006). The main data sources in the United Kingdom are; The National Bowel Cancer Audit (NBOCA) is commissioned by the Healthcare Quality Improvement Partnership (HQIP) delivered by the Health and Social Care Information Centre working in collaboration with the ACPGBI. The data are submitted from various NHS Trusts in the UK and the Republic of Ireland via the Open Exeter system. These data are used to examine variation in surgical practice, to describe outcome following colorectal resections for cancer and to derive predictive models of mortality (Tan et al., 2007; Tekkis et al., 2003; Tilney et al., 2009). Findings are disseminated in an annual report. These datasets are primarily collected for administrative purposes but are increasingly used for clinical performance benchmarking. The Hospital Episode Statistics (HES) represents the dataset for NHS England [www.hesonline.nhs.uk] and Information Services Division Scotland (ISD) for Scotland [www.isdscotland.org]. Such databases comprise of clinical codes that relate to diagnoses, procedures and demographic information. These data can be used to examine procedure volumes and basic clinical outcomes, including inpatient mortality rates, readmission rates and length of hospital stay. HES data have been linked to other national datasets such as the Office of National Statistics and the Cancer Registry for additional clinical and outcome information, such as out-of-hospital survival and tumour staging. The latter National Cancer Data Repository (NCDR) has recently been used in research studies to compare institutional mortality rates following colorectal surgery in England (Morris et al., 2011). The NCDR can also been linked to the national Radiotherapy Dataset (RTDS) to investigate patterns of management of rectal cancer across the English NHS (Morris et al., 2016). Surgeon level reporting of colorectal cancer surgery outcomes to the UK public was introduced in 2013. Capture of information on patients undergoing planned colorectal resection for cancer (excluding appendiceal neoplasms) are included in analyses. Surgeon level reporting has been spearheaded by Cardiac Surgeons but there are concerns amongst many surgeons that unit level reporting may be more appropriate. The ACPGBI agreed to participate in the process as the Department of Health stated that it would proceed with, or without the involvement of the ACPGBI. There are concerns regarding the quality of data and limitations of their interpretation. There are discrepancies in both information technology infrastructure and the availability of data capture personnel. As of 2013, NHS England (through the Healthcare Quality Improvement Partnership) has published 90-day risk-adjusted mortality figures for colorectal cancer patients undergoing planned surgery on a hospital and surgeon basis. This is based on data submitted to NBOCA. Monitoring of outcomes is required to define acceptable quality of care in elective colorectal cancer surgery. For the reasons already cited comparison can only be made when funnel plots are used to adjust for operative volumes, when data are reliable and when case-mix has been adjusted for. Outcomes for emergency surgery are generally less favourable. The following figures represent some published outcomes following elective colorectal cancer surgery: Surgeons and trusts must make provisions for the prospective collection of accurate clinical data for submission to the National Bowel Cancer Audit (NBOCA). Recommendation grade B Omar Faiz has no conflicts of interest to declare.