National audit of the sensitivity of double-contrast barium enema for colorectal carcinoma, using control charts: For the Royal College of Radiologists Clinical Radiology Audit Sub-Committee
Introduction
Double-contrast barium enema (DCBE) is a standard technique for investigating colonic disease, and is widely used in the diagnosis of colorectal cancer.1, 2, 3, 4, 5, 6 Ideally, 100% of such lesions would be demonstrated on DCBE and diagnosed (interpreted) correctly. Unfortunately, 100% rates are seldom, if ever, attainable2, 3, 4, 5, 6 because some lesions cannot be seen, because it is sometimes not possible to be sure that the lesion is malignant, and because human error in perceiving and reporting the lesion cannot be ruled out. Less than 100% performance can have potentially adverse consequences for patient outcomes and satisfaction, stemming from delayed diagnosis. In 1995, the Wessex Regional Audit1 found the average demonstration rate of 10 radiology departments was 92.7% (96.8% after correction for errors of perception) and the diagnosis rate was 84.6%. The authors proposed targets of 97% demonstration rates and 94% diagnosis rates.
We undertook a national audit of DCBE demonstration and diagnosis rates to determine if any improvement had occurred since the Wessex Regional Audit (1995), and also to provide guidance on how current performance could be improved. For the latter purpose, we adopted Shewhart's theory of variation,7, 8 which seeks to provide guidance on the action required to deliver continual improvement. After some debate, we adopted the targets set out in Table 1. These targets are higher than those reported by the Wessex Audit and lower than those it proposed for re-audit, but were considered by the Audit Sub-Committee as attainable procedural and personal standards.
Section snippets
Methods
All 302 UK NHS radiology departments on the Royal College of Radiology (RCR) audit database were, through their nominated contact, invited to participate; 131 did so. A consultant radiologist from each department was asked to undertake this retrospective study and to obtain, from the pathology department, a list of colorectal cancer cases with a histologically confirmed diagnosis recorded during the year 2001. This list was then used to identify all such patients who had had a DCBE at any time
Results
Fig. 1 displays on a control chart the diagnosis rate for all departments. The overall diagnosis rate, horizontal line, was 85.9% (4687/5454), range 50% to 100%. The control chart divides the departments into three distinct groups. The 7 departments on or above the upper control limit indicate that special-cause variation rather than chance has produced these high diagnosis rates. The 3 departments below the lower control limit again imply special-cause variation, this time producing low
Discussion
The overall diagnosis rate (85.9%) in this study is comparable with that of the Wessex Audit (84.6%), but the range in our study was wide (50% to 100%), as was that in Wessex (36% to 95%). To improve the diagnosis rate, we need to identify strategies for shifting the mean and reducing the tail. The control-charts methodology incorporates such guidance. The 7 departments that had particularly high diagnosis rates should be studied to learn from their apparently superior process, although the
Conclusion
When set against the diagnosis rates and demonstration rates reported in the Wessex Audit 1995,1 it appears that a similar level of performance is currently being achieved in the NHS today, suggesting that the basic process for undertaking and reporting DCBE has remain relatively unchanged over the last few years. This indicates that improvement in the future requires fundamental changes to the process of reporting DCBE, in order to minimize the perception failure rate and accurately to
Acknowledgements
The authors would like to thank all the radiologists of the 131 participating departments for their contribution to this audit, particularly the audit leads, who were responsible for collating and submitting the data. Without their hard work, national audits such as this would not be possible. We would also wish to thank our colleagues from the Clinical Radiology Audit Sub-Committee for their input into the preparation of this paper.
References (14)
- et al.
Wessex regional audit: barium enema in colo-rectal carcinoma
Clin Radiol
(1995) - et al.
Bristol, Shipman and clinical governance: Shewhart's forgotten lessons
Lancet
(2001) - et al.
Barium enema in the diagnosis of colonic carcinoma
Austrian Clin Rev
(1984) - et al.
Colonoscopically detected colo-rectal cancer missed on barium enema
Gastrointest Radiol
(1991) Radiographic findings in overlooked colon carcinomas: a retrospective analysis
Acta Radiol
(1998)- et al.
Colo-rectal carcinoma missed on double contrast barium enema study: a problem in perception
Am J Roentgenol
(1981) - et al.
Carcinoma of the colon and rectum: tumours missed by radiologic examination in 61 patients
Cancer
(1988)