Manual for ACC/AHA Guideline Writing Committees

Methodologies and Policies from the ACC/AHA Task Force on Practice Guidelines

Section II: Tools and Methods for Creating Guidelines

• Step Four: Synthesize and Interpret the Evidence

Guideline Authoring Template
To improve the consistency of guideline content, both within and between guidelines, the Task Force has created a guideline authoring template. Guideline writers will receive one template for each clinical objective they are responsible for writing. The following is a brief introduction to using the template:

• The fields for guideline, author, section name, and clinical objective will be completed by staff.
• The evidence base field will consist of the articles gathered through the literature search and sorting (Steps Two and Three). Writers whose clinical objectives have a large evidence base may also receive full text articles or an evidence table along with the template.
• There are four fields for recommendation, including a checklist for classification and level of evidence for each recommendation (more than four recommendations can be written). After reviewing the evidence base, the writer should create recommendations that answer the clinical objective. See Steps Five and Six for details on writing recommendations.
• The text field is used for placing recommendations in context (see "Narrative synthesis of evidence" below).
• The additional references field can be completed by a writer who references other sources than those provided in the evidence base field.
• If the clinical objective has a diagram, table, or graphic associated with it, it can be added or referenced in this field (see Step Seven).

Assessing how much heterogeneity exists among trials included in a review and assessing whether they should be combined make up one of the crucial steps in a systematic review. - Jadad, 1998

Narrative Synthesis of Evidence
Summaries of evidence should generally be in tabular for, and not in the text of the guideline. Text should be reserved for qualifying or clarifying the recommendations. The Task Force prefers that clinical trial data and other evidence be displayed in an evidence table or included in meta-analysis. When multiple trials have yielded similar, non-controversial results (e.g., the use of aspirin post myocardial infarction), a single sentence with appropriate references may suffice. Long, descriptive paragraphs of the methodology and findings of individual trials are discouraged.

Visual Synthesis of Evidence
Preparing an evidence table involves identifying and extracting the key data from the relevant studies. The Cochrane Collaboration recommends beginning by deciding what comparisons need to be made, then identifying the data elements necessary to make those comparisons. Salient data elements may include, but are not limited to, number of patients, morbidity, mortality, dose-response, sensitivity, specificity, p-values, confidence intervals, positive predictive value, negative predictive value, and relative risk.

The next step is to prepare visual summaries of the results of the studies included in each comparison. The data are usually displayed in a table that allows the studies' designs and results to be easily compared. However, sometimes the data are better summarized in a bar chart or other graphic summary. Information presented graphically can replace the need for "text-heavy" sections of the guideline. Examples of visual synthesis of evidence from published ACC/AHA guidelines include:

• Acute Myocardial Infarction Guideline, Figure 7
• Atrial Fibrillation Guideline, Figure 16
• Chronic Stable Angina Guideline, Table 21

Analytical Synthesis of Evidence
Sometimes recommendations can confidently be written based on the organization of evidence in tables or graphs. Other times, a further step is necessary; analyzing the data statistically to get an estimate of the heterogeneity of the individual effect sizes, an estimate of the summary effect size, and a measure of its variance. Guideline writers generally rely upon meta-analytic methods to conduct such analyses.

A detailed guide to the methods of meta-analysis is beyond the scope of this manual. However, ACC staff is available to assist writing committees in conducting meta-analysis for guidelines.

Also recommended is the Cochrane Collaboration, which offers free on its web site (www.cochrane.org) software called Review Manager 4.1. RevMan allows for entry of the characteristics of studies and their findings, and the creation of comparison tables. It can perform meta-analysis of the data entered, and present the results graphically. A comprehensive handbook for conducting systematic reviews that can be printed from the same web site accompanies the software.

Use of Other Guidelines/Authorities
Guideline text, recommendations, and evidence tables may be replicated from previous ACC/AHA guidelines and statements endorsed by both organizations (e.g., National Cholesterol Education Program). Consensus statements or guidelines developed by others and not endorsed by the AHA and ACC should not be cited or referenced unless absolutely necessary, as this implies endorsement on the part of the organizations.

Discussing Pharmacotherapy in Guidelines
The Task Force has provided a detailed list of policies on discussing pharmacotherapy in guidelines as Checklist 3. In addition, a pharmacologist will either be assigned to each guideline or will be used in a consulting role to review the guideline's pharmacotherapy discussions before publication.

Investigational treatments or drugs that are not available for general use may be mentioned, but should be clearly described as such and not given Class I, IIa, or IIb recommendations. The writing committee should decide whether to list them as Class III, or not to list them at all. The presence or absence of FDA or CMS approval of a drug or device for a specific purpose should generally not be mentioned. The criteria used by regulatory authorities are frequently different, and the ACC/AHA process should be independent of these regulatory issues.

In the case of international guidelines co-sponsored by the ACC/AHA/ESC, it is understandable that rare occasions may require a discussion of international availability of certain medications. However, such content should be addressed from the perspective of the patient or indication, and not from a policy (i.e. drug approval) perspective.

 

Checklist 3. Discussing Pharmacotherapy in Guidelines Use generic or chemical name not trade name e.g., simvastatin, not Zocor Use broadest and most generic name of class appropriate e.g., cholesterol-lowering not "statins" List classes of drugs or drugs within classes according to evidence-based rationale and state rationale e.g., first-line, second-line or side effects or cost-effectiveness If no evidence-based rationale, list alphabetically List all drugs (or none) within class Indicate whether each is approved for the indication(s) under discussion e.g., statins for primary prevention Indicate whether each has evidence for the indication(s) under discussion e.g., IIb/IIIa's Discuss evidence for or against "class effect" e.g., issue raised by ramipril in HOPE study When so-called "alternative medicines" are known to be widely used, discuss the evidence about them and the issues raised by their use e.g., possible interactions Avoid the use of symbols and abbreviations when discussing drug dosing and timing. e.g., use "micrograms" or "mcg" instead of "µg" The Institute for Safe Medication Practices has issued a drug error alert regarding some commonly used abbreviations Whenever a guideline includes specific drug information, such sections of the guideline should be reviewed by a pharmacologist during peer review.