Article Text

PDF
PS-117 Data quality analysis of adverse drug events in a voluntary reporting system
  1. M Aznar Saliente,
  2. L Roca Aznar,
  3. P Herraiz Robles,
  4. M Bonete Sánchez,
  5. L Pons Martínez,
  6. S Ruiz Darbonnens
  1. Hospital Universitario de Sant Joan, Pharmacy, Alicante, Spain

Abstract

Background In 2009, SINEA, a voluntary reporting system for adverse events (AEs) in healthcare was implemented, designed for direct online reporting. It cannot ensure the consistency of the information, nor the quality of the reports.

Purpose To determine the number and type of errors found in the SINEA database reports of drug adverse events (DAE); to propose improvements to reduce them and to note the differences in the results of the raw and refined databases, in order to skip the refining process if possible.

Material and methods AEs reported between 1 January 2014 and 30 August 2014 were extracted and revised by a pharmacist to refine the database considering the field “describe_what_happened” as the gold standard. Percent of medicines errors (MEs), adverse drug reactions (ADRs), potential (PME) and real (RME) medicines errors, description of the effect on the patient, the impact on assistance and the most frequently reported drugs (MFD) were compared in both raw and refined databases. Cohen’s kappa (k) statistic defining concordance was calculated.

Results 364 AEs were reported, of which 66.7% were classified as MEs, 2.7% as ADRs (2 wrongly classified as both, thus total percent > 100%) and 31% as other events. After refining, MEs totalled 69.5%; ADRs, 5.8% and events not related to medicines, 24.7% (k = 0.85 CI95% [0.80–0.90]). Before refining, 73.6% of MEs were considered PMEs versus 82.3% after refining (k = 0.65 CI95% [0.54–0.76]). With refined data, the MFD was trastuzumab (20.9%), due to exhaustive notification in oncology (all PMEs). The “active_ingredient” field was empty in 133 reports in the raw database. A mean of 1.8 ± 1.9 errors per report were detected.

Conclusion Although concordance is good, the tough refining process cannot be skipped as it provides quality information so that improvements in pharmacotherapy can be implemented. Data quality could be improved by reducing the number of type-in text fields and using checkboxes or drop-down lists and by increasing the staff’s knowledge of DAEs.

References and/or acknowledgements No conflict of interest.

Statistics from Altmetric.com

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.