Efficient clinical evaluation of guideline quality: development and testing of a new tool
1 International Centre for Allied Health Evidence (iCAHE), University of South Australia, City East Campus, School of Health Sciences, Centenary, GPO box 2471, Adelaide 5001, Australia
2 College of Rehabilitation Sciences, University of Santo Tomas, St. Martin de Porres Building University of Santo Tomas Espana, Manila 1083, Philippines
3 Florey Institute of Neurosciences & Mental Health, University of Melbourne, NHMRC Research Fellow, Melbourne, Victoria, Australia
BMC Medical Research Methodology 2014, 14:63 doi:10.1186/1471-2288-14-63Published: 10 May 2014
Evaluating the methodological quality of clinical practice guidelines is essential before deciding which ones which could best inform policy or practice. One current method of evaluating clinical guideline quality is the research-focused AGREE II instrument. This uses 23 questions scored 1–7, arranged in six domains, which requires at least two independent testers, and uses a formulaic weighted domain scoring system. Following feedback from time-poor clinicians, policy-makers and managers that this instrument did not suit clinical need, we developed and tested a simpler, shorter, binary scored instrument (the iCAHE Guideline Quality Checklist) designed for single users.
Content and construct validity, inter-tester reliability and clinical utility were tested by comparing the new iCAHE Guideline Quality Checklist with the AGREE II instrument. Firstly the questions and domains in both instruments were compared. Six randomly-selected guidelines on a similar theme were then assessed by three independent testers with different experience in guideline quality assessment, using both instruments. Per guideline, weighted domain and total AGREE II scores were calculated, using the scoring rubric for three testers. Total iCAHE scores were calculated per guideline, per tester. The linear relationship between iCAHE and AGREE II scores was assessed using Pearson r correlation coefficients. Score differences between testers were assessed for the iCAHE Guideline Quality Checklist.
There were congruent questions in each instrument in four domains (Scope & Purpose, Stakeholder involvement, Underlying evidence/Rigour, Clarity). The iCAHE and AGREE II scores were moderate to strongly correlated for the six guidelines. There was generally good agreement between testers for iCAHE scores, irrespective of their experience. The iCAHE instrument was preferred by all testers, and took significantly less time to administer than the AGREE II instrument. However, the use of only three testers and six guidelines compromised study power, rendering this research as pilot investigations of the psychometric properties of the iCAHE instrument.
The iCAHE Guideline Quality Checklist has promising psychometric properties and clinical utility.