Clinical risk management in Dutch community pharmacies: the case of drug-drug interactions.

BACKGROUND: The prevention of drug-drug interactions requires a systematic approach for which the concept of clinical risk management can be used. The objective of our study was to measure the frequency, nature and management of drug-drug interaction alerts as these occur in daily practice of Dutch community pharmacies.
METHODS: In total, 63 Dutch pharmacies collected all drug-drug interaction alerts during 153 research days (on average 2.4 days/pharmacy), as well as variables related to these alerts, such as involved medicines, first time or recurrent drug-drug interaction, same or different prescribers, patient data (age, sex) and information about the management of drug-drug interactions by the pharmacy. The latter was discriminated into internal procedures only and external action, such as communication with the patient, the prescriber or the anticoagulation clinic and prescription modification. All drug-drug interactions were classified into categories of clinical relevance (A-F) and available evidence (0-4).
RESULTS: A total of 43,129 prescription-only medicines were dispensed during the study period. On average, 16.8 interaction alerts per day per pharmacy were collected. Approximately 6% of all prescriptions generated a drug-drug interaction alert. Of all alerts (n = 2572), 31.1% occurred for the first time and with 21% two different prescribers were involved. The 20 most frequently occurring drug-drug interaction alerts accounted for approximately 76% of all alerts. Cardiovascular drugs, NSAIDs, oral contraceptives and antibacterials were most frequently involved. External action was taken in response to 27.3% of the alerts, meaning either a modification of one of the concerned prescriptions (n = 65; 9.3%), communication with the prescriber or anticoagulation clinic (n = 90; 12.8%) or communication with the patient or a relative (n = 547; 77.9%). Where there was no external action (n = 1860; 72.3%), pharmacists concluded in about two-thirds of cases that the drug-drug interaction had been managed in the past. Other reasons not to intervene externally were for instance: incorrect alert; acceptable drug-drug interaction; or outcome of the interaction considered irrelevant. Adjusted for several variables, a first alert was found to be a main determinant for external action. After stratifying for first alerts no other significant determinants were found.
CONCLUSIONS: A high frequency of drug-drug interaction alerts was found. Most concerned recurrent alerts, which were the main reason not to act externally. Concerning the assessment phase in the risk-management process, drug-drug interactions with no or low evidence/relevance should be reconsidered. Concerning the management of drug-drug interactions in pharmacies, the opportunity to actively suppress alerts for a certain period of time should be studied in more detail. There are indicators that the management of patient-orientated advice could be improved and a greater degree of consistency developed for the management of first and recurrent interaction alerts.