Transparent Guideline Methodology Needed

As part of learning at the Nordic Workshop of Evidence-based Medicine, we have read with interest the practice guidelines for central venous access, published in your Journal in 2012.¹  We appraised the quality of this guideline using the checklist developed by The Evidence-Based Medicine Working Group.²  Similar criteria for guideline quality have been suggested elsewhere.³  Our conclusion was that this much needed guideline is currently unclear about several aspects of the methodology used in developing the recommendations. This means potential users cannot be certain that the recommendations are based on best currently available evidence.

Our concerns are in two main categories: the rigor of development, including methodology of searching, evaluating, and combining the evidence; and editorial independence, including funding and possible conflicts of interest. The methodological issues that we would like to see clarified in the guideline are:

1. Searches for literature: details of search strategies and databases used, including the search terms and strategies used in different databases;

2. Criteria for selecting the evidence: inclusion and exclusion criteria for publications found through the searches;

3. Process of selecting the evidence: number of persons independently applying the criteria at each step of selection and ways of solving disagreement;

4. Process and criteria of appraising the quality of studies for inclusion;

5. Methods used for synthesizing the evidence;

6. Results of meta-analyses (including tests for heterogeneity);

7. A statement of conflicts of interest for the authors; and

8. Sources of financial and other support for developing the guideline.

The searches may have missed relevant information. The authors report their searches covered a period from 1968 to 2011, identifying over 2,000 citations. When we made a PubMed search until the end of 2011 using “central venous catheter” as the search term, we found 12,453 references; adding “infection” retained 4,729 of these; and adding “coated” resulted in 128 studies, including a cohort study of catheters coated with antibiotics published in 2011.⁴  This study shows such catheters significantly decreased infections “in a manner that was independent and complementary to the infection control precautions.” In the paragraph on Recommendations for Use of Catheters Containing Antimicrobial Agents, the guideline has no references. A clear description of the search strategies could explain such a discrepancy.

The concept of “evidence linkage” is rather central for the methodology but not quite fully explained in the methods. Does it mean “a statement regarding potential relationships between clinical interventions and outcomes” (as in appendix 5)?

In reporting the results, we would appreciate separating the recommendations based on published scientific evidence and opinion-based evidence. Reading the guideline, it was first unclear to us whether the letters (e.g., Category B2 evidence) referred to scientific evidence or opinion-based evidence. Getting to the footnote of appendix 5, it seems the categories refer to scientific evidence only. It would be interesting to learn whether the guideline group for the next version could consider using a more widely used grading system, such as the GRADE (The Grading of Recommendations Assessment, Development and Evaluation).⁵ 

It may be that we have missed where the authors have referred to the methods used, although we did our best to study the publication and the Supplemental Digital Content carefully. We also searched the website of this Journal^† to make sure we did not miss any information on guideline methodology. In many guidelines, for example in a recent one on the prevention of intravascular catheter-related infections,⁶  the methodology is provided on an external site.^‡

Finally, we would have appreciated seeing the conflict of interest statements for all authors, and a declaration of how this undoubtedly work-intensive process has been funded.

Making the evidence more transparent would be helpful to guideline users. Our suggestions are anchored in principles of evidence-based medicine. Our aim with this letter is to encourage the authors to provide information which could increase our confidence in the guideline and thus promote its application in practice.

We hope that this important guideline will be updated within a couple of years, and that the American Society of Anesthesiologists Task Force on Central Venous Access will consider adding the necessary methodological information in their guideline.