Background:
Differences between what medications patients think they should be taking and what is ordered after hospital discharge (i.e., medication errors) place patients at risk for harmful adverse drug events and hospital readmission. We examined patient‐ and medication‐related factors associated with medication errors after discharge.
Methods:
We analyzed data from a prospective cohort study of patients hospitalized with cardiovascular disease. Medication errors were determined by comparing discharge medication lists to what patients reported taking during a phone interview in the week following discharge. We utilized logistic regression for binary end points (i.e., presence/absence of an error) and binomial logistic regression for count end points (i.e., number of errors) to examine the association between prespecified risk factors and types of medication errors (discordance between the presence of a medication on the discharge list and the patient‐reported list; omission and commission errors; and for cardiac medications, discrepancies in indication, dose, and frequency). Risk factors included demographics, health literacy, subjective numeracy, marital status, cognitive function, social support, education, income, depression, global health status, and medication adherence.
Results:
Among 473 patients, 51% had at least 1 discordant medication (i.e., it did not appear on both the discharge list and the patient‐reported list). More than a quarter of patients (27%) were not taking a medication that they were supposed to be taking per the discharge list (an omission), whereas more than one‐third (36%) were taking a medication not listed on the discharge list (a commission error). Nearly 60% of patients reported a discrepancy in indication, dose, or frequency for at least 1 cardiac medication on the discharge list. In adjusted analyses, the higher the number of medications taken, the higher the odds were for having a medication being discordant (OR, 1.10; 95% CI, 1.05–1.15), having a commission error (OR, 1.09; 95% CI, 1.04–1.14), or having a discrepancy in indication, dose, or frequency (OR, 1.38; 95% CI, 1.17–1.62). Older age (OR, 1.215; 95% CI, 1.07–1.38) and more cardiac medication changes between admission and discharge (OR, 1.05 per medication change; 95% CI, 1.00–1.10) were associated with higher odds of discrepancies in indication, whereas female sex (OR, 0.683; 95% CI, 0.52–0.89), higher health literacy (OR, 0.972; 95% CI, 0.96–0.99), and higher subjective numeracy (OR, 0.863; 95% CI, 0.78–0.96) were protective. Worse cognitive function was associated with higher odds of discrepancies in frequency (OR, 1.407; 95% CI, 1.08–1.84), whereas female sex was protective for discrepancies in dose (OR, 0.649; 95% CI, 0.47–0.91).
Conclusions:
Patients with advanced age, lower health literacy and numeracy, and cognitive impairment and who take more medications are at risk for experiencing more postdischarge medication errors and subsequent potential harm because of these errors.