Making useful clinical guidelines: the ESGAR perspective

Clinical guidelines are important and influential; they can improve processes involved in patient care, thereby also improving patient outcomes [1]. They are both commonly downloaded from journal websites and highly cited [2], helping clinical decision making and service commissioning. Yet, the quality of such guidelines is highly variable; a review of 279 guidelines published between 1985 and 1997 found that overall adherence to high-quality methodological standards was less than 50% [3]. Just as Altman [4] has argued that the misuse of statistics is unethical for primary research, it is equally inappropriate for guideline documents to recommend specific practices unless developed robustly and transparently. To do otherwise risks erroneous care, and, ultimately, patient harm. Readers of guidelines (clinicians, patients and policymakers) require reassurance that these authoritative documents have identified, appraised and considered the available evidence, or draw attention to weaknesses in the literature if appropriate. Accordingly, many organisations and professional societies issuing clinical guidelines have formalised their development process. This has several advantages, but perhaps most significantly, it permits methodological rigour to be embedded into the guideline development process in a consistent fashion. The organisation is able to stipulate, in advance, minimum standards for literature searching, evidence synthesis, construction and ratification of guideline statements; all of which would otherwise be at the whim of individual guideline development committees or groups. Yet, what makes a high-quality guideline? Fortunately, this question has been the subject of many years’ work by several organisations, notably the AGREE consortium, whose most recent iteration (AGREE II) encompasses 23 items that can be used to assess the quality of a particular clinical guideline [5]. As a by-product of this quality assessment tool, it provides a framework to permit highquality guideline development and reporting [6]. The European Society of Gastrointestinal and Abdominal Radiology (ESGAR) has developed its own guideline development process, aiming to standardise and improve the quality and transparency of guidelines issued under its aegis. This was initiated in the mid-2017, and guidelines published from October 2018 onwards will adhere to its structure. The key components of this process, which adhere to the principles of AGREE II, are summarised in Fig. 1 and below; the full document and supporting information are available on the ESGAR website (https://www.esgar.org/guidelinespublications/research-committee/).

organisation is able to stipulate, in advance, minimum standards for literature searching, evidence synthesis, construction and ratification of guideline statements; all of which would otherwise be at the whim of individual guideline development committees or groups. Yet, what makes a high-quality guideline? Fortunately, this question has been the subject of many years' work by several organisations, notably the AGREE consortium, whose most recent iteration (AGREE II) encompasses 23 items that can be used to assess the quality of a particular clinical guideline [5]. As a by-product of this quality assessment tool, it provides a framework to permit highquality guideline development and reporting [6].
The European Society of Gastrointestinal and Abdominal Radiology (ESGAR) has developed its own guideline development process, aiming to standardise and improve the quality and transparency of guidelines issued under its aegis. This was initiated in the mid-2017, and guidelines published from October 2018 onwards will adhere to its structure. The key components of this process, which adhere to the principles of AGREE II, are summarised in Fig. 1 and below; the full document and supporting information are available on the ESGAR website (https://www.esgar.org/guidelinespublications/research-committee/).

Guideline prioritisation
Potential topics for guidelines will be sought from the ESGAR membership, members and leaders of other societies and members of the public and journal editors. A form is available on the ESGAR website to assist this (at the same URL as listed above). Topics will be prioritised based on specific criteria, namely the burden of the relevant disease (or radiological technique), the impact that this has on radiological services, the extent of uncertainty in current clinical practice (including availability of other high-quality-related guidelines) and the availability of published evidence upon which a guideline could be based. The feasibility of developing a guideline using the resources and expertise of the ESGAR membership will Electronic supplementary material The online version of this article (https://doi.org/10.1007/s00330-019-6002-9) contains supplementary material, which is available to authorized users. also be considered. These criteria will be assessed by the ESGAR research committee before a new guideline is commissioned.
Step by step consensus process

Evidence selection and synthesis
The guideline development group will construct and circulate an initial questionnaire in order to identify the scope and aims of the guideline, and prioritise topics for detailed literature review. Thereafter, where possible these topics will be converted into specific clinical questions adopting the PICO format (patients/participants, intervention, control/comparators, outcomes). These clinical questions will then serve as the basis for the literature search. The literature search, including the databases used, search terms, inclusion dates and language restrictions, will be documented clearly and used uniformly by the guideline group. Relevant articles will be retrieved and summarised into evidence tables with accompanying explanatory text, including an assessment of the literature quality.

Consensus statements and committee voting
Using the evidence synthesis tables and explanatory text, the guideline group will construct draft consensus statements for the final guideline. Where the evidence is contradictory or incomplete, relevant statements can still be constructed, but may need to be based on the group's opinion with anonymous voting, using a modified Delphi process. Each voting member will be asked regarding their agreement with a given statement on a five-point scale (strongly disagree, disagree, neither agree nor disagree, agree and strongly agree). For a statement to achieve consensus, 80% of guideline group members must agree or strongly agree with that statement. Statements not achieving consensus initially may be modified and subjected to repeat voting; a maximum of three iterations is suggested as the upper limit before a statement is defined as failing to reach consensus. Statements that do not achieve consensus may be included in the Discussion of the final guideline document, but should not be recommended by the group.

Guideline reporting, appraisal and updating
The final output of the process will be a guideline document, which will include its background, target audience and endorsing societies. The consensus statements themselves, with an associated level of evidence quality and strength of recommendation, will be provided, with the evidence summary tables and explanatory text made available as supplementary information. Discussion points should include statements that failed to reach consensus, recommendations for how the guideline should be implemented in clinical practice, where the literature review uncovered gaps and opportunities for further research, and a timeframe for updating the guideline. During the process of constructing the guideline document, the group will conduct an internal appraisal of methodological quality by using the AGREE II instrument and use this to guide their manuscript preparation. For example, target users of the guideline must be explicitly identified; statement wording should be clear and unambiguous; and the key recommendations must be easily identifiable. The key aspects of ESGAR guideline reporting are summarised in Table 1. Reports will be published in the peer-reviewed literature and made Open Access, to enhance visibility and likelihood of implementation.

Summary
Well-constructed clinical guidelines have the potential to substantially improve patient care across multiple institutions in different countries. To maximise their benefit, they must be prioritised, developed, disseminated and implemented using methodologically sound and transparent techniques. The ESGAR is committed to ensuring consistency and excellence in its guidelines, and all future ESGAR guidance will follow the principles outlined here.
Plumb is funded by the National Institute for Health Research (NIHR) under its Fellowships Programme, PDF-2017-10-081. Research time for Stuart Taylor is partly funded by the NIHR under the Senior Investigator scheme. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care.

Compliance with ethical standards
Guarantor The scientific guarantor of this publication is Andrew Plumb.