|Home | About | Journals | Submit | Contact Us | Français|
Provisional reports from the Intercollegiate British Society of Gastroenterology National Colonoscopy audit show completion rates of 57-77% for the procedure and poor levels of training and supervision. We prospectively audited all aspects of colonoscopy performed at a combined district general hospital and specialist endoscopy unit. Details of referral, examination, endoscopist, complications and follow-up were recorded and patients were sent questionnaires for long-term follow-up.
505 patients (246 male) underwent colonoscopy by 27 different endoscopists. Their median age was 57 years (range 13-92) and 93% were outpatients. 64% patients were symptomatic and 36% were having surveillance or follow-up colonoscopy. The overall caecal intubation rate was 93%, with little difference between surgeons, physicians and experienced trainees (89%, 92%, 94%) and specialist endoscopists (98%). In only one case was an inexperienced trainee (<100 procedures) unsupervised. Pain scores estimated by the endoscopist were well matched with those given by the patient—medians 29 and 26 (maximum 100) respectively. Median satisfaction score was 96 (maximum 100). Polyp pick-up rate was 26.9% and there were 11 new cancers. 16 (3%) minor immediate complications were recorded—5 oversedation, 6 vasovagal attacks, 3 polypectomy haemorrhages and 2 mucosal injuries (neither requiring treatment). 3 patients died within 6 months of follow-up but no death was colonoscopy related.
Completion rates in this setting were adequate for all endoscopists studied. Patient satisfaction with the procedure was high and very few immediate or long-term complications were encountered.
Examination of the colonic surface by flexible video-endoscopy (colonoscopy) is the most accurate method for diagnosis of colonic disease and for surveillance of patients at high-risk of developing colon cancer. Colonoscopy is a technically demanding procedure with potential for harm if technique is poor. It is highly operator-dependent and standards vary greatly1. In the recent National Colonoscopy audit2, caecal intubation was reported in 77.1%; however, if identification of the ileocaecal valve or ileal intubation were taken as the only reliable indicators of caecal intubation, the rate fell to 57.1%. Current guidelines suggest that all trained colonoscopists should achieve caecal intubation rates of at least 90%3. We audited the practice in our centre, which comprises a district general and a specialist endoscopy unit combined.
All patients referred for colonoscopy and having their examination in June or July 2000 were included in the audit. A questionnaire was designed with advice from the National Audit's chief investigator and the local departments of Continuous Improvement and Risk Management. The questionnaire was piloted and amended within the unit before the start of the audit. Among the information recorded for each colonoscopy were referral pattern, examination details, immediate and long-term follow-up and details of the endoscopist and his or her supervision where relevant.
Referral details were recorded from the investigation request form by the audit investigator. Patient demographics were noted as well as the referring team, date and urgency of referral and the indication for the procedure. The investigator, without knowledge of the examination findings, made an assessment of the appropriateness of the referral. The total time the patient was in the department and the time in the procedure room were recorded. All patients were prescribed standard bowel preparation consisting of senna granules (13 g) and 2 sachets of magnesium citrate (29.5 g).
The endoscopist recorded details of the procedure and these together with information from the report form were entered onto the database. The following was recorded regarding the procedure: availability of notes and X-rays (where relevant), drugs given and their dosage, insufflating gas used, quality of bowel preparation, macroscopic findings and maximum depth of insertion (if not the caecum the reason for failure and plans made after incomplete examination). The endoscopist was asked to photo-document the caecum and to state any identifying landmarks used. The caecal intubation rate was calculated after elimination of cases with impassable strictures or very inadequate bowel preparation. The endoscopist recorded any immediate complications.
Details of the endoscopist performing the procedure were recorded. This included the grade of the endoscopist (and of the supervisor where appropriate), the number of procedures the endoscopist had previously performed and whether a supervised trainee was observed or had assistance during the procedure.
The endoscopist was asked to estimate how much pain the patient had experienced by completing a 100 point visual analogue scale (VAS). After the procedure, just before leaving the department following the recovery period, the patient completed two further VAS questions regarding pain experienced during the procedure and level of satisfaction with the entire procedure. The patient was able to make free-text comments on the same form.
All patients were sent a follow-up questionnaire 6 months after their colonoscopy unless they had stated that they did not want to be contacted or hospital records indicated that they had died. The follow-up questionnaire recorded details of symptoms arising after colonoscopy and any action taken as a result. The time taken to recover from the procedure and any time off work was documented. Any hospital admissions within a month of the procedure were also recorded. Data were entered onto a Microsoft Access database and analysed with the Microsoft Excel statistics package.
505 patients (246 male) underwent colonoscopy. The median age was 57 years (13-92), 58 for men and 57 for women. 85% of the referrals were from either the gastrointestinal physicians or the colorectal surgeons at the hospital. Only 7 cases (1.3%) were referred from the geriatric firms and 9 (1.7%) directly from general practitioners. The priority was stated on the request form in 374 cases (74%) and was felt to be appropriate in 300 cases (80%) and inappropriate in 39 (10%); in the remaining 35 cases no judgment could be made. 468 (93%) were outpatients; 408 had conscious sedation, 1 had general anaesthesia and 59 outpatients and 6 inpatients had no sedation. Of the 65 patients who had no sedation 48 were male. The median outpatient colonoscopy waiting-list times were as follows: ‘urgent’ 4 weeks (0-13), ‘soon’ 7.5 weeks (1-33) and ‘routine’ 10 weeks (0-66). The indications for the procedures are shown in Table 1.
The patients' case-notes were unavailable at the time of the procedure in at least 20 cases, and X-rays, where applicable, were unavailable in 66 out of 85 cases. Patients waited a median of 60 minutes (0-210) from their appointment time to the time they went into the procedure room and were in the procedure room for a median of 40 minutes (10-120). 70% of the procedures were diagnostic only and in 30% of cases polypectomy was performed. 2 patients underwent dilatation. In 41% of all cases the colon was normal. The symptoms or indications most likely to be associated with macroscopic disease were suspected inflammatory bowel disease (66%), bleeding (64%) and anaemia (52%). The symptoms pain and diarrhoea alone were much less likely to result in macroscopic disease being found (27% and 25%, respectively). No abnormality was found in 41% of cases. The commonest findings were polyps (26.9%), inflammatory bowel disease (17%), diverticular disease (9%) and cancer (2.1%).
95 requests (18.8%) were stated as being very urgent or urgent. In 61 of these disease was found. In 8 of the 11 patients in whom definite cancers were found the procedure had been requested as very urgent or urgent.
For conscious sedation a combination of midazolam and pethidine was used, in median doses of 2.5 mg (0-6) and 50 mg (0-100), respectively. Buscopan was used in most cases, at a median dose of 20 mg (0-60). Air and carbon dioxide were used either alone or in combination. The quality of the bowel preparation was judged ‘good’ in 66% (clear fluid/small amounts of fluid residue easily suctioned, good views), ‘fair’ in 28% (liquid stool requiring repeated suctioning for adequate views) or ‘poor’ in 6% (solid/semi-formed stool preventing adequate examination, repeat procedure required). 27% of inpatients had poor bowel preparation.
The overall caecal intubation rate was 93% (72-100%). This is the adjusted rate following exclusion of cases with impassable strictures, previous bowel resection and therefore no caecum, and abandonment due to poor bowel preparation. Surgeons, physicians and unsupervised specialist registrars had similar success rates (caecal intubation rates of 89%, 92% and 94%, respectively); specialist endoscopists had an intubation rate of 98%. The landmarks to be used for identification of the caecum were not specified in the protocol and the following features were stated as having been used: ileocaecal valve (58%), terminal ileal mucosa (48%), appendix orifice (38%), tri-radiate fold (11%), right iliac fossa transillumination (7%) and indentation (5%). Photodocumentation was achieved in 85% of claimed caecal intubations. The photographs were reviewed independently by two experienced endoscopists and judged as being definite evidence of the caecum in 88%. Excluding cases in which there had been a colonic resection or there was stricturing or poor bowel preparation not allowing caecal intubation, the reasons for failure to intubate the caecum (33 cases) were uncontrollable looping (13), diverticular disease (9), pain (3), other (7) and unstated (1). The endoscopist was asked what plans were made for those patients in whom total colonoscopy was not achieved. In 16 of the 33 cases, no further plans were made immediately in 9, a barium enema was booked in 2, a repeat examination was to be done by one of the unit's specialist endoscopists and in 6 another definitive plan was made. The results of the barium enemas were severe diverticular disease in 4 and a long redundant colon in 3—accounting for the difficulties with colonoscopy. In 2 cases no barium enema was done for other reasons. In the 2 cases rebooked with a specialist endoscopist, the caecum was reached.
A total of 27 different endoscopists performed 505 colonoscopies—2 specialist endoscopists 108 cases; 6 gastroenterology consultants 93 cases; 4 colorectal surgeons 67 cases; 13 specialist registrars, 5 of whom were experienced and performing colonoscopy unsupervised, 230 cases; 1 nurse practitioner 4 cases; and 1 visiting consultant having specialist training 3 cases. Of the 234 cases performed by non-consultant endoscopists, 79 were supervised (58 by consultants and 21 by competent specialist registrars). Of the 155 cases performed by unsupervised endoscopists 145 were performed by competent specialist registrars. In 7 of the 155 the trainee had had moderate previous experience (> 100 previous procedures) and in only one case was an inexperienced trainee (< 100 previous cases) left unsupervised.
355 patients completed the VAS sheet for pain on a scale of 0-100. Overall the endoscopists' perception of how painful the patient found the procedure (29 [0-100]) agreed well with the patients' score (26 [0-100]); all groups of endoscopists except the surgeons tended to over-estimate the patient's pain. The patients' median satisfaction score was 96 (0-100). The pain perceived by patients was somewhat less if CO2 alone was used (n=86), median score 24.5 (0-100) rather than air alone (n=202), median score 27.5 (0-100). Pain scores for women were higher than those for men (median 33 [0-100] and 18 [0-100] respectively) but median satisfaction score was 96 (0-100) for both groups. Pain scores did not necessarily correlate with the indication for the procedure. The highest pain scores were found in those patients with inflammatory bowel disease (median 40 [0-95]) and the lowest in those having cancer follow-up (median 10 [0-91]). 3 patients scored pain at the maximum; one, however, we feel may have incorrectly completed the VAS as satisfaction was also scored at a maximum. 4 patients scored satisfaction at zero but again looking at their comments and pain scores we suspect they misunderstood the VAS scoring system. We have included the scores for all of these patients to avoid bias.
There were no major complications. 16 minor complications arose during the study period. There were 6 vasovagal episodes and 5 episodes of over-sedation (all of which responded to reversal therapy). 3 minor postpolypectomy haemorrhages were identified during the procedure and treated with endoclips in 2 cases and with adrenaline injection in 1. There were 2 ‘mucosal injuries’. One was a minor probable suction injury; the patient had no symptoms. The second was in a patient with ulcerative colitis. A tear down to the muscle layer in the sigmoid colon was noted. A radiograph showed gas in the bowel wall but no free air. The patient was admitted and treated conservatively with antibiotics and later readmitted for elective proctocolectomy and pouch procedure for extensive poorly controlled disease.
328 (69%) of 474 who were sent the follow-up questionnaire replied. Patients reported a median requirement of 1 day off work (range 1-28) and, if not working, a median of 1 day unwell at home (0-60) after their colonoscopy. 32 (9.7%) patients reported some bleeding (3 attended hospital) but none required any active treatment. 8 of these had had a polypectomy or ‘hot’ biopsy but 24 had only had a diagnostic procedure. 69 patients reported abdominal pain, of whom 21 described it as ‘a lot’ of pain and 2 attended hospital but required no specific treatment. 2 patients had a high temperature (one of whom was the patient with a mucosal tear down to muscle); neither required emergency treatment.
14 patients in total attended hospital within a month of their colonoscopy; 1 was admitted after colonoscopy but settled and later underwent elective colectomy for extensive ulcerative colitis; 8 were admitted electively for surgery (colorectal or other); 5 attended as outpatients (2 complaining of abdominal pain). 3 patients died during the 6-month follow-up period, 2 of these within a month of the procedure. One of these patients was 86 and had coexistent ischaemic heart disease and oesophageal cancer; the other had extensive colonic carcinoma. Neither death was related to the colonoscopy.
Early data from the Intercollegiate-BSG National Colonoscopy (IBNC) audit in the United Kingdom2 suggest that the standard of colonoscopy in the UK is generally far short of the British and American guidelines3,4. As in many other areas in medicine, standards in colonoscopy are being evaluated world wide5,6,7 and we wished to see how our centre performed. The unit is atypical for the UK in serving both a tertiary referral centre and a district general hospital, with specialist endoscopists, gastroenterology physicians and colorectal surgeons performing 3000-4000 colonoscopies a year. At present all referrals are accepted without being vetted by medical personnel. The high proportion of appropriate referrals may be explained by the fact that most referrals come directly from gastroenterology or colorectal teams. However, there are still a substantial number of inappropriate referrals and referrals in which inadequate information is provided; radiographs were commonly unavailable when needed. The quality of bowel preparation was ‘poor’ (requiring a repeat attempt) in 6% of patients overall but in 27% of inpatients. Part of the reason, perhaps, is that inpatients are generally sicker and less able to tolerate purging.
The areas in which this audit has highlighted particularly good results are caecal intubation and supervision of trainees. In this unit there are, on average, over 20 different practitioners performing colonoscopy and yet the rate of total colonoscopy is as high as 93%, with photodocumentary evidence in 85% of claimed caecal intubations. Admittedly, photodocumentation is not an entirely accurate method of confirming caecal intubation8,9. The landmarks most commonly used for successful caecal intubation were ileocaecal valve and terminal ileal mucosa and these are probably the most reliable; but some colonoscopists also used transillumination and finger indentation in the right iliac fossa, which are unreliable. In the IBNC audit the caecal intubation rate fell from 77.1% to only 57.1% when only the two most reliable indicators were used2.
General gastroenterologists, colorectal surgeons and specialist registrars performed only slightly less well than the unit's expert colonoscopists. Specialist registrars and other trainees were very well supervised, with only one inexperienced trainee (< 100 procedures) performing a single colonoscopy unsupervised. This compares well with the provisional data from the IBNC audit (Bowles CJA, unpublished) where only 16% of trainees were supervised for their first 100 colonoscopies and 33% of colonoscopists described themselves as ‘self-taught’. Dafnis looked at the impact of experience and learning curves on colonoscopy completion rates and found that some endoscopists achieve a high completion rate early in their career and further experience has little impact. Some endoscopists never achieve a 90% completion rate, with rates levelling off from around 250 cases6. Innate ability, intensity of practice and formalized training are probably important factors. We found that some trainees with relatively small experience of colonoscopy (in terms of total number of procedures performed) had caecal intubation rates at least as good as those of consultant-grade practitioners with much more experience. It must be said, however, that many of the trainees have a special interest in the procedure; St Mark's is not one of the basic training centres in the RCS/BSG scheme1.
Might the prospective audit have favourably influenced results during the study period? This may be so; but audit is intended to improve standards and make practitioners aware of the need for continual improvement. Colonoscopy, like many other areas of medical practice, cries out for better methods of training, assessment and accreditation.