Article Text

Download PDFPDF
5PSQ-093 Assessment tool for hospital admissions related to medications, 10 questions (at-harm10): a validation study
  1. Y Al-Sammak1,
  2. U Gillespie2,
  3. T Kempen3
  1. 1Uppsala University, Pharmacotherapy, Uppsala, Sweden
  2. 2Patientsäkerhetsavdelningen, Clinical Pharmacy, Uppsala, Sweden
  3. 3Patient Safety, Uppsala University Hospital, Uppsala, Sweden


Background The MedBridge study, with the aim of evaluating the effects of comprehensive medication reviews performed by ward-based clinical pharmacists on elderly patients’ healthcare consumption, started in February 2017. A secondary outcome measure in the study will be drug-related admissions (DRAs) during the follow-up year. As the identification of DRAs inevitably has a degree of subjectivity, a reliable and standardizsed method is required. The involvement of senior clinicians is often deemed necessary, making the assessment relatively expensive. We therefore developed an Assessment Tool for Hospital Admissions Related to Medications, consisting of 10 questions (AT-HARM10), which could be used by less experienced clinical pharmacists and advanced pharmacy students instead.

Purpose To validate the final version of the AT-HARM10.

Material and methods The contents and lay-out of the previous version of AT-HARM10 was discussed between the investigator, the supervisors and a group of clinical pharmacists to obtain a new version with good face-validity. This was followed by a validation process where the inter-rater reliability (IRR) and criterion-related validity (CRV) of AT-HARM10 was determined. Five pharmacy students and two clinical pharmacists separately applied the tool to 100 hospital admissions that had previously undergone assessment by one senior clinical pharmacist and one experienced geriatrician, which was regarded as the ‘gold standard’. The results were analysed using Cohen’s kappa and Fleiss’ kappa for IRR, and sensitivity and specificity for CRV.

Results The final version of AT-HARM10 had good face-validity according to the assessors. The tool showed a moderate to substantial IRR with Cohen’s kappa values ranging between 0.45 to 0.75 and Fleiss’ kappa values of 0.46 and 0.58. For the CRV, the sensitivity and specificity ranged between 68% and 82% and 64% and 89% respectively.

Conclusion AT-HARM10 seems to have sufficient capacity to determine whether a hospital admission is medication-related when used by clinical pharmacists or advanced pharmacy students after a half-day training in the use of the tool.

References and/or Acknowledgements Thanks to all pharmacists and students participating in the study and to the experts making up the gold standard.

No conflict of interest

Statistics from

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.