Dr Mark Cottrill explains how adoption of the EPAGE guidelines could improve access to endoscopy for patients and make best use of limited resources

There has long been a need to identify which patients would benefit from an endoscopic procedure (gastroscopy or colonoscopy) and to standardise such indications.

A European Panel of experts met in Switzerland in November 1998 and, based on the best available clinical evidence, considered several hundred patient situations to determine whether such an investigation was appropriate in each case.

The panel results were published in October 1999.1 A novel approach to their dissemination was to make them available on the internet (European Panel on the Appropriateness of Gastrointestinal Endoscopy; EPAGE).

Primary or secondary care physicians (and even patients) can access the website (www.epage.ch). By going through a series of on-screen prompts, the appropriateness of an endoscopy in an individual patient can be determined.

The benefits of such guidelines are numerous:

Inappropriate or unnecessary referrals can be reduced, which should, in turn, lead to a fall in waiting times for endoscopy
GPs, armed with best expert advice, can determine when to refer a patient for endoscopy
A trained worker can screen open access referrals for appropriateness.

In the last 30 years, endoscopy has gradually replaced radiology as the Iost common procedure in the investigation of gastrointestinal disease.

As a principle, any test should only be performed when it may impact on the patient's management. Favourable outcomes include relief of symptoms, reduction in morbidity or mortality, exclusion of serious disease, or improvement in the quality of life.

The use of endoscopy as an investigation should be balanced against the potential risk of complications, poor acceptance by patients, the relatively high cost and, in many parts of the UK, unacceptably long waiting lists.

The best evidence for any guideline is obtained from well conducted clinical trials or even high quality observational studies. In considering the efficiency of endoscopy there are surprisingly few well conducted clinical trials. Randomised trials are rarely used to validate diagnostic tests.

The goal of the project is to improve understanding of clinical situations in which expert opinion is still needed. Expert opinion for colonoscopy is usually determined by the consultant gastroenterologist after careful clinical evaluation.

For gastroscopy the decision for the investigation often rests with the GP, especially if he or she has open access to the hospital endoscopy list. This opinion derives from the skill and experience of the practitioner and possibly local guidelines or advice from the local gastroenterologist.

Despite this there is still considerable uncertainty about which patients to refer. With long waiting lists and diminishing resources in the health service it is important that referrals are at least appropriate.

When evidence from clinical trials or observational studies is lacking, the next best evidence is expert opinion. Such opinion is derived from an understanding of the clinical situation.

For the opinions expressed by the panel, the RAND appropriateness method was employed. This involved a detailed literature search to identify studies evaluating the efficacy, safety, side-effects and possible complications of digestive endoscopy. It also included clinical situations where endoscopy was relevant.

The literature review included MEDLINE (up to 1997), EMBASE, and the Cochrane Library, as well as several internet sites. A synthesis of the literature was distributed to each member of the panel before the meeting. They, in turn, rated the appropriateness of each of several hundred clinical situations on a scale of 1–9.

The panellists comprised 14 'experts' I8 gastroenterologists, 2 surgeons, and 4 GPs from 9 European countries.

When the panellists met, the results of individual ratings were anonymously distributed to allow discussion of each situation. The ratings were categorised as being appropriate, uncertain or inappropriate.

Appropriate was defined as the expected health benefit exceeding the risk of performing the procedure. A 'necessary' indication would fulfil the following criteria:

The procedure is appropriate
It would be negligent not to propose it to the patient
The potential benefit for the patient is high

The probability that this benefit will occur is also high.

The final round of ratings was 601 (gastroscopy 292: colonoscopy 309). This is the number of clinical indications. Overall, 37% of the 601 situations were rated appropriate, 19% uncertain and 44% inappropriate. For the 292 indications for gastroscopy 11% were deemed necessary: for colonoscopy the figure was 17%.

The guidelines were funded by a grant from the European Community, supplemented by additional funding from the Swiss National Science Foundation and the Swiss Federal Office of Education and Science. There were no conflicts of interest among any of the participants.

Unlike guidelines in other disease areas,uthese guidelines cover more than 600 clinical situations, so it is not possible to list them in any meaningful form on paper. Use of the internet, however, provides ready access to all the situations within a few keystrokes. This is best explained with an example.

A patient with no known coronary artery disease presents with atypical chest pain. This is defined as 'chest pain with no objective evidence of coronary artery disease'.

In the USA, about 30% of patients undergoing coronary angiography each year have no abnormality of the coronary arteries and are referred to a gastroenterologist for oesophageal evaluation.

In determining the appropriateness of gastroscopy, there are a number of variables (see Figure 1, below). The results of the panel votes in these situations are also shown in Figure 1.

Figure 1: Results of the EPAGE guideline for determining the appropriateness of gastroscopy for a patient with no known coronary artery disease presenting with atypical chest pain.

Possible variables:

If the patient had already had cardiac investigations (the panel defined this as one or more of the following: exercise ECG, stress thallium, stress echocardiogram, angiography)

If the patient had had gastro-oesophageal reflux disease (GORD) treatment (defined as more than 7 days of a proton pump inhibitor) and if the patient had responded to such treatment.

The results of the panel votes in these situations were as follows:

No cardiac investigation no GORD treatment inappropriate (gastroscopy)
No cardiac investigation GORD treatment with no response inappropriate
No cardiac investigation GORD treatment with response inappropriate
Cardiac investigation normal no GORD treatment uncertain
Cardiac investigation normal GORD treatment without response appropriate
Cardiac investigation normal GORD treatment with response uncertain

The EPAGE website summary of the characteristics for the same patient is shown in Figure 2.

Figure 2: Frame from the EPAGE website summarising the patient characteristics for atypical chest pain
epage website

These conclusions are based on evidence that only gastroscopy can accurately establish the diagnosis of oesophagitis, but that the majority of patients with reflux disease have a normal gastroscopy.

In addition, acid-lowering therapy benefits patients with chest pain resulting from acid reflux, and therefore therapeutic interventions may be a better option than endoscopy.

Standardisation of the indications for endoscopy has obvious benefits. The expert panel determined that there are a number of clinical situations where the appropriateness of the investigation is uncertain. Such a decision will depend on the circumstances of the individual patient.

Open access gastroscopy is widely available to GPs in the UK. Use of the EPAGE guidelines should help to reduce the number of inappropriate referrals.

Long waiting lists for gastroscopy and colonoscopy are common within the NHS. The early detection of neo-plastic disease is cost-effective and prompt diagnosis is clearly desirable. Should referrals for endoscopy be rationalised, there is the potential for a reduction in referrals, with a resulting reduction in waiting lists.

Endoscopy is not without risk and not well tolerated by some patients. Adoption of the EPAGE guidelines could improve access for patients and ensure best use of the limited resources in secondary care.

There is still considerable uncertainly about the appropriateness of endoscopic procedures at all levels of clinical practice. Although there is also some uncertainty about indications, even among the expert panel, the guidelines provide an excellent consensus in a variety of clinical situations.

This will be of great value to GPs in determining the value of investigation in an individual patient, especially when the GP is part of an open access referral scheme. Also, the opinions of the expert panel could be invaluable to both primary and secondary care physicians.

If the guidelines were to be adopted by a primary care group, referrals within the group could be standardised and a referral form could be devised to include (or exclude) certain patient groups. A trained worker with access to the internet could filter referrals and streamline procedures.

In our own department, patients waiting for gastroscopy are often treated with acid-lowering drugs; by the time of gastroscopy, many are asymptomatic. The reason for investigation then changes from diagnosis to exclusion of more serious disease.

Guidelines are based on the clinical evidence available at the time of their inception. The EPAGE experts remain available for any updating in the light of new clinical evidence. The novel approach to publishing the guidelines on a website provides the ideal opportunity to upgrade as appropriate.

If you have access to the internet I would strongly urge you to visit the EPAGE site. You may find that it will change your clinical practice!

  1. Endoscopy No. 8, Vol 31, October 1999, pp 571-696.

Guidelines in Practice, April 2000, Volume 3
© 2000 MGP Ltd
further information | subscribe