Current guidelines poorly address multimorbidity: Pilot of the interaction matrix method

Christiane Muth*, Hanna Kirchner, Marjan Van Den Akker, Martin Scherer, Paul P. Glasziou

*Corresponding author for this work

Research output: Contribution to journalArticleResearchpeer-review

30 Citations (Scopus)


Objectives To develop a framework to identify and classify interactions within and among treatments and conditions and to test this framework with guidelines on chronic heart failure (CHF) and its frequent comorbidity.

Study Design and Setting Text analysis of evidence-based clinical practice guidelines on CHF and 18 conditions co-occurring in ≥5% of CHF patients (2-4 guidelines per disease). We extracted data on interactions between CHF and comorbidity and key recommendations on diagnostic and therapeutic management. From a subset of data, we derived 13 subcategories within disease-disease (Di-Di-I), disease-drug (Di-D-I), drug-drug interactions (DDI) and synergistic treatments. We classified the interactions and tested the interrater reliability, refined the framework, and agreed on the matrix of interactions.

Results We included 48 guidelines; two-thirds provided information about comorbidity. In total, we identified N = 247 interactions (on average, 14 per comorbidity): 68 were Di-Di-I, 115 were Di-D-I, 12 were DDI, and 52 were synergisms. All 18 comorbidities contributed at least one interaction.

Conclusion The interaction matrix provides a structure to present different types of interactions between an index disease and comorbidity. Guideline developers may consider the matrix to support clinical decision making in multimorbidity. Further research is needed to show its relevance to improve guidelines and health outcomes.

Original languageEnglish
Pages (from-to)1242-1250
Number of pages9
JournalJournal of Clinical Epidemiology
Issue number11
Publication statusPublished - 1 Nov 2014


Dive into the research topics of 'Current guidelines poorly address multimorbidity: Pilot of the interaction matrix method'. Together they form a unique fingerprint.

Cite this