Background Electronic prescribing systems in outpatient care have been implemented widely in our country. The pharmacotherapeutic information it contains is used in both primary and hospital healthcare. In daily clinical practice, systematic errors are observed in this information, even in narrow therapeutic index drugs, which could reach the patient, especially in transitions.

Purpose To quantify the frequency of errors that occur in narrow therapeutic index drugs monitored in the service of pharmacokinetics.

To assess whether these errors influence plasma drug concentrations (Cp).

To determine whether follow-up queries to hospital or outpatient care reduces errors.

Material and methods Prospective observational study.

Period: 5 months.

Population All patients receiving carbamazepine (CBZ), phenytoin (PHE) and valproic acid (VPA) were selected.

Information sources: pharmacotherapeutic electronic information/prescription (IANUS), pharmacokinetic history (Openlab).

Cp determination: Architect 1200SR.

Variables collected: age, monitored drug, Cp, error (mismatch between prescribed dose and actual patient dose), physician follow-up (outpatient or hospital).

Statistical evaluation: Stata 12. Descriptive statistics. Mean comparison using Student’s t test. Proportions with χ².

Results Population variables: 103 patients (34 CBZ, 27 PHE, 41 VPA). Values are mean±SD. Age (years) (45.8 ± 24.5). Error (%) (30.1 ± 46.1). Error effect on Cp (mg/mL) by drug: without error versus with error CBZ (11.5 ± 17.85 vs 7.17 ± 2.75; p = 0.395), PHE (8.83 ± 3.48 vs 6.70 ± 4.73; p = 0.215) and VPA (67.17 ± 22.92 vs 61.8 ± 21.55; p = 0.502).

Hospital follow-up (%) (70.59 ± 46.79). Follow-up effect on errors: hospital versus outpatient errors (hospital without/with error vs outpatient without/with error) (47/25 vs 24/6; p = 0.141).

Conclusion We have shown that this information is unreliable as it has a very large amount of errors (30.1%). The hospital follow-up was not related to fewer errors than outpatient care. These errors were not associated with a different Cp. This may be related to the narrow therapeutic index of these drugs and the small sample size of the study. Future studies should assess the frequency of adverse effects with greater numbers of patients. The pharmacist should review this information to communicate and correct errors and to prevent them from reaching patients.

No conflict of interest.