Background and importance Medication discrepancies (MDs), defined as unexplained differences among medication regimens, cause important public health problems with clinical and economic consequences. Medication reconciliation (MR) reduces the risk of MDs but is time consuming and its success relies on the quality of different information sources. Online personalised health records (PHRs) may overcome these drawbacks, but the correctness of the identified MDs with a PHR compared with traditional MR is unclear.

Aim and objectives The aim of this study was to determine the level of agreement of identified MDs between traditional MR and an online PHR and the correctness of the identified MDs with an online PHR.

Material and methods Two weeks prior to a planned admission to the cardiology, neurology, internal medicine or pulmonary department, patients received an invitation from a PHR to update their medication file derived from the Nationwide Medication Record System (NMRS). At admission, MR was performed by a pharmacy technician, who created the best possible medication history (BPMH) based on the NMRS data and an interview. MDs were determined as discrepancies between the available information from the NMRS and the input and alterations patients or pharmacy technicians made. The number, correctness of patients’ alterations, type and severity of identified MDs were analysed.

Results Of 488 patients approached, 155 (31.8%) were included. The mean number of MDs identified with MR and PHR was 6.2 (SD 4.3) and 4.7 (SD 3.7), respectively. 82.1% of the drug information noted by the patient in the PHR was correct compared with the BPMH, and 98.6% had no clinically relevant differences between the lists.

Conclusion and relevance Patients who used an online PHR had the ability to correctly identify clinically relevant MDs in a manner that resembled traditional MR. Online PHRs may have the potential to replace MR in detecting MDs.

Conflict of interest No conflict of interest