Methodological Quality of Guidelines for Management of Lyme Neuroborreliosis

Résumé du livre

Abstract: Background

Many aspects of clinical management of Lyme neuroborreliosis are subject to intense debates. Guidelines show considerable variability in their recommendations, leading to divergent treatment regimes. The most pronounced differences in recommendations exist between guidelines from scientific societies and from patient advocacy groups. Assessment of the methodological quality of these contradictory guideline recommendations can be helpful for healthcare professionals.

Methods

Systematic searches were conducted in MEDLINE and databases of four international and national guideline organizations for guidelines on Lyme neuroborreliosis published from 1999-2014. Characteristics (e.g., year of publication, sponsoring organization) and key recommendations were extracted from each guideline. Two independent reviewers assessed the methodological quality of each guideline according to the Appraisal of Guidelines for Research and Evaluation II (AGREE II) tool. AGREE II scores from guidelines developed by scientific societies and from patient advocacy groups were compared across domains.

Results

We identified eight eligible guidelines of which n = 6 were developed by scientific societies and n = 2 by patient advocacy groups. Agreement on AGREE II scores was good (Cohen's weighted kappa = 0.87, 95 % CI 0.83-0.92). Three guidelines, all from scientific societies, had an overall quality score of ≥ 50 %. Two of them were recommended for use according to the AGREE II criteria. Across all guidelines, the AGREE II domain with the highest scores was "Clarity of Presentation" (65, SD 19 %); all other domains had scores 50 % with the domain "Applicability" having the lowest scores (4, SD 4 %). Guidelines developed by scientific societies had statistically significantly higher scores regarding clarity of presentation than guidelines from patient advocacy groups (p = 0.0151). No statistically significant differences were found in other domains.brbrConclusionsbr