(158) Identification of Valid Indicators for the Detection of Potentially Adverse Drug Event-Related Hospitalisations in Routine Electronic Data Sources: a RAND Survey

Background: Adverse drug events (ADEs) can cause significant harm to patients and, in severe cases, lead to emergency hospital admissions. It is estimated that at least half of these would be avoidable. Policy and research efforts have therefore increased over recent decades, but evaluations of their effectiveness have rarely measured their impact on potentially ADE-related hospitalisations (pADEH), at least partly due to a lack of efficient methods to identify such endpoints.

Objectives: To develop drug-event pairs as valid indicators for the detection of pADEH in routine electronic data sources.

Methods: We conducted a two-stage expert consensus process based on the RAND/UCLA Appropriateness Method (RAM), in order to prioritise ADEs at hospital admission (stage 1) and to identify drug-event pairs as valid indicators for the most important adverse drug reactions (ADRs) (stage 2). Each stage consisted of two assessment rounds, with the first-round ratings fed back to the panelists before the second-round ratings were placed. In contrast to the Delphi method, but consistent with the RAM, two moderated face-to-face meetings (one for each stage) were held in between rounds to enable the exchange of arguments between experts. ADEs prioritised in the first stage were considered for the second stage, where the experts were asked to assess on a 4-point Likert scale (1=Unlikely to 4=Certain) how likely the listed medication could be assumed to contribute significantly to the adverse event in an average patient. Drug-event pairs with a median rating of ≥ 3 without disagreement (i.e., less than 30 % of expert ratings were < 3) were defined as „valid indicators” of ADRs.

Results: In stage 1, a total of 6 physicians and 7 pharmacists with clinical and/or scientific background from 11 university sites in Germany assessed 65 ADEs, 38 of which were prioritised, including 17 with a median of 4. For the latter 17 ADEs, combinations of each ADE with potentially causative medication (drug-event pairs) were assessed by 5 physicians and 6 pharmacists from 10 German university sites in stage 2. After the second rating round, 44 out of 318 drug-event pairs were confirmed as valid indicators of pADEH. Examples are acute kidney injury in combination with non-steroidal anti-inflammatory drugs (NSAIDs) or aminoglycosides, falls with benzodiazepines, and bleeding outside the gastrointestinal tract with antiplatelet drugs or anticoagulants (e.g., heparins, vitamin K antagonists).

Conclusions: The first stage of our survey led to a set of prioritised ADEs and provided a basis for our second stage, resulting in systematically developed indicators of pADEH for application in clinical practice (e.g., decision support), clinical surveillance and as endpoints for interventional and observational research.