Using more than one systematic review as evidence base for a guideline recommendation: Updating the National Disease Management Guideline (NDMG) Asthma

Article type
Authors
Vader I1, Schorr SG1, Prien P1, Schüler S1, Eissing L1, Schaefer C1, Härter M2
1Agency for Quality in Medicine
2Agency for Quality in Medicine, Universitätsklinikum Hamburg-Eppendorf UKE
Abstract
Background: The update of the NDMG Asthma is based on systematic reviews (SR). Methodological standards for SR require a documentation of excluded studies and reason for exclusion. This information is essential for the comparison of different SR for one research question.

Objectives: We aimed to compare SR on therapy of asthma with macrolides. Furthermore, we analysed the reporting quality on 'study exclusion' of all included systematic reviews on the diagnosis and treatment of asthma.

Methods: SR on the diagnosis and treatment of asthma between 01/01/2013 and 02/26/2016 were systematically searched in MEDLINE via Pubmed and in the Cochrane Database of Systematic Reviews. Identified reviews were screened in a 2-step procedure and only those scoring at least 6 points at the AMSTAR measurement tool were included.
The SR were allocated to the research questions of the guideline. If >1 SR was allocated to one research question, such as for macrolide therapy, results were analySed in detail.

Results: We identified 3 SR on macrolide therapy in patients with asthma. For the outcome 'quality of life', 4 of altogether 6 studies were used in all 3 SR. Two SR found no difference in quality of life, while 1 found a statistically significant improvement. Comparing the body of evidence to explain the contradiction was only partly possible as only 1 SR reported on excluded studies and reasons for exclusion.
Hitherto a total of 124 SR were included as the evidence base for the diagnosis and treatment of asthma. 42% (n=52) reported on excluded studies and reasons for exclusion, 77% (n=40) of those were Cochrane reviews.

Conclusions: Reporting excluded studies and reasons for exclusion is essential for the comparison of different SR. Unfortunately, less than half of the included SR reported on excluded studies. A better reporting of SR would help to increase their value for guideline developing groups and thus strengthen the link between evidence synthesis and guideline development.