A colorectal perspective on voluntary submission of outcome data to clinical registries

Abstract

The aim of the study was to identify outcome differences amongst patients undergoing resection of colorectal cancer at English National Health Service trusts using Hospital Episode Statistics (HES). A comparison was undertaken of trusts that submitted and those that did not submit, or submitted only poorly, voluntarily to a colorectal clinical registry, the National Bowel Cancer Audit Programme (NBOCAP). The NBOCAP data set was used to classify trusts according to submitter status. HES data were used for outcome analysis. Data for major resections of colorectal cancer performed between 1 August 2007 and 31 July 2008 were obtained from HES. Trusts not submitting data to NBOCAP and those submitting less than 10 per cent of their total workload were termed 'non-submitters'. HES data for 30-day mortality, length of stay and readmission rates were compared according to submitter and non-submitter status in multifactorial analyses. A total of 17,722 patients were identified from HES for inclusion. Unadjusted 30-day in-hospital mortality rates were higher in non-submitting than in submitting trusts (5·2 versus 4·0 per cent; P = 0·005). Submitter status was independently associated with reduced 30-day mortality (odds ratio 0·76, 95 per cent confidence interval 0·61 to 0·96; P = 0·021) in regression analysis. A higher postoperative mortality rate following resection of colorectal cancer was found in trusts that do not voluntarily report data to NBOCAP. Implications regarding the voluntary nature of submission to such registries should be reviewed if they are to be used for outcome benchmarking.