Background: Medication errors are the cause of common and dangerous adverse effects on patients. They occur as a result of preventable failures in the prescribing (ordering), transcribing, dispensing and administration of medications.
Objective: The aim of this study was to determine the relationship between patient characteristics and prescribing and transcribing medication errors during acute hospitalization of elderly patients in an internal medicine ward.
Methods: A cohort case-control study was conducted in a 37-bed medical ward at a tertiary medical centre in Israel. The study included 137 patients in whom potentially harmful medication errors were detected, and 137 sex- and age-matched controls. Clinical data were collected and Charlson Comorbidity Index scores were calculated. Conditional logistic regression was used to identify factors associated with medication errors.
Results: Independent factors associated with any type of medication error included Charlson Comorbidity Index score ≥8 (odds ratio [OR] 2.97 [95% CI 1.16, 7.61], p = 0.023), number of medications ≥ 9 (OR 2.02 [95% CI 1.00, 4.05], p = 0.049) [both upon admission] and length of hospital stay ≥ 13 days (OR 4.41 [95% CI 2.25, 8.62], p < 0.0001). Independent factors associated with prescribing errors included Charlson Comorbidity Index score ≥ 8 (OR 6.34 [95% CI 1.63, 24.71], p = 0.008) and length of hospital stay ≥13 days (OR 3.19 [95% CI 1.23, 8.26], p = 0.017), while independent factors associated with transcribing errors included number of medications ≥9 (OR 2.58 [95% CI 1.02, 6.51], p = 0.04) and length of hospital stay ≥ 13 days (OR 6.90 [95% CI 2.76, 17.23], p < 0.0001). The median time to an error was 3 days, and was half as long for prescribing errors as for transcribing errors (2 and 4 days, respectively, p = 0.017).
Conclusions: The risk of medication errors among elderly patients during acute hospitalization in an internal medicine ward is associated with Charlson Comorbidity Index score (for prescribing errors), number of medications (for transcribing errors) and length of hospital stay (for both types of errors). Further study will determine whether these factors can be used to identify patients at risk and to prevent prescribing and transcribing medication errors.