Adverse drug reactions commonly occur during drug treatment and can lead to a deterioration in the patient's condition. In hospitals, ADRs account for about 5 percent of admissions. Furthermore, it is estimated that 5 percent of hospitalized patients experience ADRs. The use of electronic patient documentation and keyword search tools can simplify the identification of ADRs, ultimately improving clinical outcomes and reducing treatment costs.
The aim of the study was to determine the sensitivity, specificity and positive predictive value of the keywords used for identifying ADRs from electronic patient records. We manually reviewed the electronic documentation of 640 patients treated at one of seven different clinical departments of the University Medical Centre in Ljubljana during a two-month period, January and February 2023. 246 of these patients had at least one ADR. A total of 501 ADRs were identified, of which 154 were recorded at the time of the admission, 149 during the patient's hospitalisation, 138 were previously known allergies, and 60 had a history of previous ADRs. Identification of ADRs was then performed using a keyword search tool, where 51.3 percent of the ADRs were identified. The keywords generally had good specificity, with majority of keywords having a specificity of more than 80 percent. However, their sensitivity was quite low, meaning that a large proportion of ADRs remained unrecognized.
|
