Detection of adverse drug events in e-prescribing and administrative health data: a validation study

Table 2 Adverse Drug Event (ADE) Rates, by Definition and Data Source

Data Source	ADE in all Study Patients (N = 688)	ADE in Patients Prescribed an Antidepressant (N = 302)	ADE in Patients Prescribed an Antihypertensive (N = 386)
Gold Standard^a	114 (16.6%)	74 (24.5%)	40 (10.4%)
Electronic Prescribing Data, Treatment Change Orders^b	40 (5.8%)	20 (6.6%)	20 (5.2%)
Administrative Health Data, Expanded ICD Code Set^c	62 (9.0%)	11 (3.6%)	51 (13.2%)
Administrative Health Data, Standard ICD Code set^d	19 (2.8%)	3 (1.0%)	16 (4.1%)

^aGold standard assessment of ADEs was based on patient interview data, adjudicated by a clinical expert to assess causality based on the Naranjo criteria. ^bPotential ADEs in electronic prescribing data were defined as study drug discontinuations or dose changes due to safety or effectiveness reasons. ^cPotential ADEs in administrative health data using the expanded ICD code set were defined as a physician visit, ED visit, or hospital admission during follow-up for which the recorded ICD code was (1) a relevant external cause code or (2) an adverse effect of the study drug. ^dPotential ADEs in administrative health data using the standard ICD code set were defined as a physician visit, ED visit, or hospital admission during follow-up for which the recorded ICD code was (1) a relevant external cause code or (2) an adverse effect of the study drug and (3) was included in previously validated code sets.

ISSN: 1472-6963