Article type
Abstract
"Background
Decision support systems (DSS) warning for potentially problematic drug interactions are widely used in clinical practice. A drug combination that is frequently alerted as a clinically significant interaction due to an increased risk of bleeding is selective serotonin inhibitors (SSRI) in combination with low-dose acetylsalicylic acid (ASA). However, this drug interaction could not be confirmed in a recent systematic review performed by our group.
Aim
To compare the evidence base for the alleged interaction between SSRI and ASA in well-renowned DSSs to articles identified using a systematic review approach.
Methods
Articles cited as supporting an interaction between SSRI and ASA were extracted from drug interaction alerts in four DSS: Stockley, Lexicomp, Micromedex and Janusmed. Cited articles were compared to articles identified in the recently published systematic review.
Results
A total of 28 articles published between 1999-2017 were cited in the DSS as support for an interaction between SSRI and ASA. One of these, published in 2003, was cited by all four DSS. In that study, the observed-expected ratio for hospital admission due to upper gastrointestinal bleeding was reported separately for SSRI+ASA and SSRI, with a reference population as denominator and without considering disease in the analyses. At the other end, 22 of the 28 articles were cited in one DSS only. The systematic literature search identified 1252 records, 1062 were screened, 241 assessed for eligibility, and 24 finally included in the review whereof seven did not have major risk of bias. One of these was cited in two DSS, whereas the other six, published 2011‒2023 and including one randomised controlled trial, were not cited in any DSS.
Conclusion
The cited literature supporting a drug combination alert for SSRI and ASA differed considerably between four DSS and most of the relevant articles identified in a systematic literature search were not cited by any DSS. This study provides an example of a drug interaction alert that is classified as clinically significant in DSS but cannot be confirmed by a systematic review with critical appraisal of the literature. We conclude that literature search- and selection-related aspects deserve increased attention in DSS."
Decision support systems (DSS) warning for potentially problematic drug interactions are widely used in clinical practice. A drug combination that is frequently alerted as a clinically significant interaction due to an increased risk of bleeding is selective serotonin inhibitors (SSRI) in combination with low-dose acetylsalicylic acid (ASA). However, this drug interaction could not be confirmed in a recent systematic review performed by our group.
Aim
To compare the evidence base for the alleged interaction between SSRI and ASA in well-renowned DSSs to articles identified using a systematic review approach.
Methods
Articles cited as supporting an interaction between SSRI and ASA were extracted from drug interaction alerts in four DSS: Stockley, Lexicomp, Micromedex and Janusmed. Cited articles were compared to articles identified in the recently published systematic review.
Results
A total of 28 articles published between 1999-2017 were cited in the DSS as support for an interaction between SSRI and ASA. One of these, published in 2003, was cited by all four DSS. In that study, the observed-expected ratio for hospital admission due to upper gastrointestinal bleeding was reported separately for SSRI+ASA and SSRI, with a reference population as denominator and without considering disease in the analyses. At the other end, 22 of the 28 articles were cited in one DSS only. The systematic literature search identified 1252 records, 1062 were screened, 241 assessed for eligibility, and 24 finally included in the review whereof seven did not have major risk of bias. One of these was cited in two DSS, whereas the other six, published 2011‒2023 and including one randomised controlled trial, were not cited in any DSS.
Conclusion
The cited literature supporting a drug combination alert for SSRI and ASA differed considerably between four DSS and most of the relevant articles identified in a systematic literature search were not cited by any DSS. This study provides an example of a drug interaction alert that is classified as clinically significant in DSS but cannot be confirmed by a systematic review with critical appraisal of the literature. We conclude that literature search- and selection-related aspects deserve increased attention in DSS."