Background
Background
Major progress in the accuracy and efficacy of diagnostic and therapeutic colonoscopy has been achieved over the last 30 years. Patient safety and comfort has increased with the advent of new endoscopic techniques [1 2 3 4 5 6 7 8 4 9 10 11 12 13 14 15 16 17
Maximizing the appropriateness of healthcare interventions is a key component of quality care. The indication for an intervention is deemed appropriate when the expected benefits are greater than the expected risks or inconvenience to the patient by a sufficiently large margin that the intervention is worth performing. Assessment of appropriateness should ideally be based on high-quality evidence; unfortunately, however, in many domains and for multiple clinical situations, the evidence base is poor or nonexistent. In such situations, an explicit multidisciplinary expert panel approach (the RAND/UCLA Appropriateness Method) has been used as a complement to evidence-based data [18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 37 41 42 43 44 45 46 47 48 49 50
Methods
Methods
The RAND/UCLA Appropriateness Method
The same methodological approach as for the 1998 EPAGE I panel was used [18 38 Table e1 ): eight gastroenterologists, three primary care physicians and three surgeons. Four out of these 14 panelists were already familiar with the process as they had participated as panelists or organizers in EPAGE I.
A series of clinical indications was identified and categorized into 11 clinically relevant chapters corresponding to customary use of diagnostic colonoscopy. These chapters were: iron-deficiency anemia, hematochezia, nonspecific abdominal symptoms (uncomplicated lower abdominal pain and/or constipation and/or bloating), uncomplicated chronic diarrhea, evaluation of known inflammatory bowel disease (IBD), screening for colorectal cancer (CRC) in asymptomatic individuals and in IBD patients, surveillance after colonic polypectomy or after curative intent resection of CRC, and miscellaneous indications. A detailed description, including situations that were explicitly excluded from assessment, is shown in [Table 2
Table 2 EPAGE II: main clinical categories (chapters).
Chapter Title Situations explicitly excluded from assessment Number of clinical scenarios
Round 1 Round 2
1 Iron-deficiency anemia Malabsorption syndrome 57 48
2 Hematochezia IBD 24 54
3 Lower abdominal symptoms (chronic constipation/lower abdominal pain/bloating) Duration < 3 months 24 12
4 Uncomplicated diarrhea Acute diarrhea (< 4 weeks) 22 6
5 Evaluation of ulcerative colitis Prior colonoscopy 26 27
6 Evaluation of Crohn’s disease Prior colonoscopy 28 23
7 Screening for CRC in known IBD Other specific colitides (or ileitis): 27 43
8 Surveillance (follow-up colonoscopy) after colonic polypectomy Symptomatic individuals 84 77
9 Surveillance after curative-intent resection of colorectal cancer Palliation 17 16
10 Screening for CRC Symptomatic individuals 64 97
11 Miscellaneous indications (lesion detected at recent barium enema or sigmoidoscopy, preoperative colonoscopy, FOBT-positive stools, fulminant colitis, stenosis in IBD, acute diverticulitis, endometriosis, unexplained weight loss, melena, massive hematochezia, iron-deficiency without anemia, hyperplastic polyps at sigmoidoscopy) All situations covered by other chapters 40 60
Total 413 463
EPAGE II, 2008 European Panel on the Appropriateness of Gastrointestinal Endoscopy; CRC, colorectal cancer; FAP, familial adenomatous polyposis; FOBT, fecal occult blood test; HNPCC, hereditary nonpolyposis colorectal cancer (Lynch syndrome); IBD, inflammatory bowel disease.
Literature review
An update of the literature review was established, covering from 1997 to February 2008 (the previous literature review covered publications up to 1997), based on a three-step search strategy. First, a systematic search of Medline (1997 to February 2008) and the Cochrane Database was conducted to identify selected published guidelines, systematic reviews and/or meta-analyses, as well as primary studies assessing the use of colonoscopy in adults in the various clinical categories. Details of these OVID searches using Mesh words, keywords and specific limits are available online (Table e3 ). A comprehensive search of clinical and medical scientific websites issuing guidelines and recommendations for endoscopy was also performed. Details of websites that contributed useful information are available in Table e4 . Finally, a complementary manual search was carried out based on the reference lists of retrieved guidelines, reviews and articles.
Only articles in English, French or German, and dealing with adults, were considered for this literature review. One search strategy aimed to collect information about risks and complications of colonoscopy and the results are reported here.
Rating of clinical indications for appropriateness and necessity
About 3 months before the panel meeting, the panel experts received the list of clinical indications, the literature review, and detailed instructions on how to rate the level of appropriateness of each indication. Given the very low risk of complications related to the procedure, as shown below, and complementing the usual definition of appropriateness, as indicated above, a modification of the definition of appropriateness was used: an indication was considered appropriate if there is a reasonable likelihood of a significant finding, if such a finding would alter therapy or prognosis, and if therapy would be beneficial. In rating the appropriateness of colonoscopy, panelists were asked to imagine a “typical” patient with the described set of clinical features receiving care from a “typical” physician performing the procedure. Financial considerations were not included in the judgment process at this stage. Costs, social and other related health technology evaluation features may be introduced at a later stage, i. e. when developing guidelines from appropriateness criteria. The experts used a scale graded from 1 to 9 to rate the appropriateness of each of the several hundred indications for which colonoscopy could, in practice or in theory, be considered, using the following scores: 1 = extremely inappropriate, 5 = uncertain, 9 = extremely appropriate. The clinical indications were presented as a rating matrix which allowed a tabular presentation of the various scenarios by chapter. An indication for colonoscopy was considered appropriate if the median of the panelists’ ratings was between 7 and 9, without disagreement, and inappropriate if the median was between 1 and 3, without disagreement. Scenarios with a median rating of 4 to 6, or those revealing disagreement among the panelists, were considered “uncertain” as to the appropriateness of colonoscopy in such cases. Disagreement was defined as occurring when at least four panellists rated an indication from 1 to 3 and four others from 7 to 9. [Fig. 1
Fig. 1 Rating matrix for the first round (example). Example for iron-deficiency anemia (extract showing scenarios 1 – 30). The symptoms evaluated are represented in columns. The scenarios are further described in detail, with color coding for greater clarity. In each box, values are shown on a 9-point scale. Appropriateness scale: 1 = extremely inappropriate; 5 = uncertain; 9 = extremely appropriate.
The panel meeting took place in Montreux, Switzerland, over 2 days (April 17 – 19 2008). All chapters were discussed in the light of existing evidence that was summarized before the discussion and emphasized by the moderators, focusing on those indications for which there was disagreement between experts or uncertainty. A second round of ratings was then carried out during the panel meeting, followed by a third necessity rating round for those indications which were considered appropriate. An indication that the procedure was considered necessary fulfilled all of the following criteria [51
The procedure is appropriate (that is, it must have a median rating of 7, 8 or 9 without disagreement at the second appropriateness rating round).
It would be considered improper care or negligence not to provide this procedure.
There is a reasonable chance that this procedure will benefit the patient. (A procedure could be deemed appropriate even if it had a low likelihood of benefit but there were few risks; such procedures should not be considered necessary.)
The benefit to the patient is not small. (A procedure could be considered appropriate if it had a minor, but almost certain benefit to the patient, but would not, however, be considered necessary.)
A general description of the results of the second round of ratings, including the proportion of appropriate, uncertain and inappropriate ratings, globally and by chapter, is reported here. Additional information about the variation in agreement in panel ratings between the first and second round of ratings was established.
Results
Results
Appropriateness and necessity ratings
A total of 413 indications were submitted to the panelists for the first round of ratings. The analyses of the votes, experts’ comments, and new evidence published during the process led to some slight modifications of the scenarios submitted for discussion by the panel and to the second round of ratings: 463 scenarios were rated by all experts, an increase of 11 % on average. These indications were considered appropriate, uncertain, or inappropriate, in 255 (55 %), 73 (16 %), and 135 (29 %) of the clinical indications, respectively. [Table 5
Table 5 EPAGE II appropriateness results by main clinical categories (chapters).
Chapter* Scenarios Appropriateness, % of scenarios
Number % of total Appropriate Uncertain Inappropriate
1 48 10 58 27 15
2 54 11 83 13 4
3 12 3 42 25 33
4 6 1 67 0 33
5 27 6 41 4 55
6 23 5 30 30 40
7 43 9 47 19 35
8 77 17 56 7 38
9 16 4 63 25 12
10 97 21 51 20 29
11 60 13 53 10 37
Total 463 100.0 55 16 29
EPAGE II, 2008 European Panel on the Appropriateness of Gastrointestinal EndoscopyTable 2
There was an increase in the proportion of agreement between panelists between the first and second round of ratings. Disagreement between experts and the percentage of uncertain indications decreased between round 1 and round 2, from 15 % to 9 % and 23 % to 16 %, respectively. Detailed results by category of indication are presented in the companion articles. A total of 255 appropriate ratings were evaluated for necessity; in 160 (63 %) a colonoscopy was considered necessary.
Risks and complications of colonoscopy
A total of 39 articles, published between 1997 and February 2008 and describing the complication rates of colonoscopy, were identified (Table e6 ) [1 2 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88
Perforation and hemorrhage rates were generally < 0.1 % and < 0.3 %, respectively. The highest rates reported were 0.25 % for perforation [73 88
Cardiovascular events were the most frequent complication occurring in between < 0.001 % and < 2 % of colonoscopies [64 54 55 60 62 63 68 74 75 76 77 78 81 84 85 88 55 62 70 74 88 69 78 87 1 5 81 89 90
Conclusion
Conclusion
An update of the EPAGE appropriateness criteria for the use of colonoscopy was undertaken in the spring of 2008. The majority of the indications rated were considered appropriate, whereas about a third were considered inappropriate. The panel meeting led to enhanced agreement between panelists and, indeed, disagreement decreased to less than 10 % and the proportion of uncertain indications was reduced by a third. For almost two-thirds of the indications rated as appropriate, the experts considered that colonoscopy was necessary. Given the large number of indications considered by the panel, tabular presentation is not feasible. A presentation by chapter, according to the constituent characteristics of the clinical indications, is, however, available in a convenient internet presentation. Answering a small number of questions allows the physician to determine, in a few simple steps, the degree of appropriateness of colonoscopy for any clinical indication (www.epage.ch).
It is difficult to assess complications of colonoscopy precisely, since there is no consensus on the definition of what constitutes a complication. While accurate data may be, and often are, obtained for severe and acute complications, the extent of delayed and/or minor complications is probably underestimated because of underreporting and the difficulties entailed in data collection. There are also minimal data for minor adverse events, and indirect negative outcomes such as work days lost secondary to colonoscopy, are rarely considered and recorded.
Acknowledgments
Acknowledgments
The authors gratefully acknowledge the selfless commitment and invaluable contribution of the expert panel members, who made this project possible: Lars Agréus (SE), Christoph Beglinger (CH), Peter Bytzer (DK), Michel Delvaux (FR), Volker F. Eckardt (DE), Peter D. Fairclough (UK), François Lacaine (FR), Olivier Le Moine (BE), Vicente Lorenzo-Zúñiga Garcia (ES), Giorgio Minoli (IT), Mattijs E. Numans (NL), Daniel Oertli (CH), John O’Malley (UK), Alastair Windsor (UK). The authors warmly thank Susan Giddons for her valuable assistance in the administration of the expert panel process, as well as in the meticulous preparation of the manuscripts.
This work was supported by a grant from the Loterie Romande, Switzerland (CH).
Competing interests: None
Appendix: The EPAGE II Study Group
Appendix: The EPAGE II Study Group
See page 205.
