The Guideline Language and Format Instrument (GLAFI): development process and international needs assessment survey

Abstract

Background: Successful guideline implementation depends both on factors extrinsic to guidelines and their intrinsic features. In the Guideline Implementability for Decision Excellence Model (GUIDE-M), "communicating" content (language and format) is one of three core determinants of intrinsic implementability, but is seldom addressed. Our aims were to develop a tool that could be used by guideline developers to optimize language and format during development; identify gaps in this type of guidance in existing resources; and evaluate the perceived need for and usefulness of such a tool among guideline developers.

Methods: Our mixed-methods design consisted of (1) content development (selection and organization of evidence-based constructs from the GUIDE-M into a prototype Guideline Language and Format Instrument (GLAFI), followed by face validation with guideline developers); (2) document analysis (duplicate) of seven existing guideline tools to measure coverage of GLAFI items and identify new items; and (3) an international survey of guideline developers (corresponding authors of recent Canadian Medical Association or Guidelines International Network database guidelines) to measure perceived importance of language and format, quality of existing resources, and usefulness of a language and format tool.

Results: GLAFI items were organized into 4 language and 4 format subdomains. In face validation with guideline developers (17 clinicians, 1 methodologist), all agreed that the tool would improve guideline implementability and 93% indicated a desire for regular use. In the existing guideline tool document analysis, only 14/44 (31.8%) GLAFI items were operationalized in at least one tool. We received survey responses from 148/674 (22.0%) contacted guideline authors representing 45 organizations (9 countries). Language was rated as "extremely important" or "important" in determining uptake by 94% of respondents, and format by 84%. Correspondingly, 72% and 70% indicated that their organization would likely use such a tool.

Conclusions: Optimal language and format are fundamental to guideline implementability but often overlooked. The GLAFI tool operationalizes evidence-based constructs, most of which are absent in existing guideline tools. Guideline developers perceive these concepts to be important and express a willingness to use such a tool. The GLAFI should be further tested and refined with guideline developers and its impact on end-users measured.

Keywords: Clinical practice guidelines; Guideline implementation; Implementation science.

Fig. 1

Language and format tool organizational structure into domains and main subdomains

Fig. 2

Language and format action item coverage in existing guidance tools. Constructs meeting inclusion criteria were organized into the following items: domains (global categories), subdomains (sub-categories within each domain), and action items (individual actionable recommendations with explanatory operational definitions and examples). Domains are capitalized; sub-domains are underlined; and action items are italicized (note that some sub-domains were also considered action items). Action items that were operationalized in at least 1 tool are shaded green, those that were either mentioned or described in at least 1 tool are shaded yellow, and those that were neither mentioned, described, nor operationalized in at least 1 tool are shaded red (items are ordered green/yellow/red where applicable, within each category). M denotes mentioned; D denotes described (implies that the item was also mentioned); O denotes operationalized (implies that the item was also mentioned and described)

Fig. 3

Example of construct organization into domains, subdomains, and action items in the GLAFI. Under the global “LANGUAGE” category and “Simple” domain, a main subdomain was called “Succint and uncomplicated.” Under this subdomain were 4 action items, including “Avoid recommendations requiring many steps … ” and the following distinct items under that category: “Limit the number of distinct elements … ”; “Use conditional statements … ” and “Limit any checklists to 5 to 7 items … ”

Fig. 4

Flowchart of survey respondents

Fig. 5

Survey respondent (guideline developer) ratings of the importance of main language (a) and format (b) subdomains for recommendation uptake. Guideline developer ratings of the importance of main language (a) and format (b) subdomains for recommendation uptake, in the GLAFI. The mean Likert scale response (out of 5) for each question is represented by the length of the bar and stipulated numerically within the bar. The proportion with each response type is represented by corresponding colors within each the bar