Use of unlicensed drugs in a Swiss Pediatric University Hospital and associated prescribing error rates - a retrospective observational study.

AIMS OF THE STUDY: Unlicensed drugs are frequently used in paediatric care. To what extent they are prescribed in hospital care in Switzerland is unclear. Because prescribing errors seem to occur more frequently with unlicensed drugs, we aimed to assess the prevalence of unlicensed drug prescriptions in two study periods (2018 and 2019) at the University Children's Hospital Zurich, compare these periods and investigate whether unlicensed drugs were more prone to prescribing errors than licensed drugs. METHODS: We conducted a sub-analysis of a retrospective single-centre observational study and analysed 5,022 prescriptions for a total of 1,000 patients from 2018 and 2019 in paediatric general wards. The rate of unlicensed drugs, consisting of imported or formula drugs, was investigated. The prescriptions from 2019 were further analysed on prescribing errors to see whether errors occurred more often in unlicensed or licensed drug use. RESULTS: Of all prescriptions, 10.8% were unlicensed drugs, with around half each being imported and formula drugs. Among all patients, 34% were prescribed at least one unlicensed drug. Younger paediatric patients were prescribed more unlicensed drugs than older paediatric patients (newborns: 15.8% of prescriptions, infants: 13.4%, children: 10.6%, adolescents: 7.1%). Ibuprofen suppositories, midazolam oral solution and gentamicin i.v. solution were the most frequently prescribed imported drugs. Macrogol powder, lisinopril oral suspension and potassium chloride i.v. solution were the most frequently prescribed formula drugs. The most common drug forms in unlicensed use were oral liquid forms and i.v. SOLUTIONS: Unlicensed drugs had a significantly higher rate of prescribing errors than licensed drugs (31.6 errors per 100 prescriptions [95% CI: 26.1-37.0] versus 24.3 errors per 100 prescriptions [95% CI: 22.3-26.2], p = 0.024). In particular, formula drugs carried a higher risk (36.4 errors per 100 prescriptions, p = 0.012). CONCLUSIONS: Unlicensed drugs are frequently prescribed in this paediatric hospital setting in Switzerland. Around every tenth prescription is an unlicensed drug. Because unlicensed drugs showed a significantly higher rate of prescribing errors, licensed drugs are favourable in terms of medication safety and should be prescribed whenever possible. If no licensed drug is available, imported drugs should be favoured over formula drugs due to lower prescribing error rates. To increase medication safety in paediatrics in Switzerland, efforts are necessary to increase the number of suitable licensed drug formulations for paediatric patients, including developing new innovative drug formulations for children.

Prescribing Patterns in Pediatric General Wards and Their Association with Prescribing Errors: A Retrospective Observational Study.

PURPOSE: There are only limited data on drug utilization patterns in pediatric inpatients, especially on general wards. The aim of the study was to describe prescribing patterns and their associations with prescribing errors in a university children's hospital in the German-speaking part of Switzerland. METHOD: This was a subanalysis of a retrospective single-center observational study. Patient characteristics and drug use of 489 patients with 2693 drug prescriptions were associated with prescribing errors. Drugs were categorized by the Anatomic Therapeutic Chemical Classification System (ATC), patients were categorized by age group according to European Medicines Agency guidelines, and prescribing errors were analyzed by type [Pharmaceutical Care Network Europe (PCNE) classification] and severity of error [adapted National Coordinating Council for Medication Error Reporting (NCC MERP) index]. RESULTS: The most frequently prescribed ATC classes were nervous system (N) (42.6%), alimentary system (A) (15.6%), and anti-infective drugs (J) (10.7%). Eighty-two percent of patients were prescribed an analgesic. Most drugs were prescribed for oral (47%) or intravenous (32%) administration, but the rectal route was also frequent (10%). The most frequently prescribed drugs were paracetamol, metamizole, and ibuprofen. The high number of metamizole prescriptions (37% of patients were prescribed metamizole) is typical for German-speaking countries. Older pediatric patients were prescribed more drugs than younger patients. A statistically significant difference was found in the rate of potentially harmful errors across age groups and for gender; children between 2 and 11 years had a higher rate of potentially harmful errors than infants under 2 years (p = 0.029) and female patients had a higher rate of potentially harmful errors than male patients (p = 0.023). Recurring errors were encountered with certain drugs (nalbuphine, cefazolin). CONCLUSIONS: Our study provides insight into prescribing patterns on pediatric general wards in a university children's hospital in Switzerland and highlights some areas for future research. Especially, the higher risk for prescribing errors among female pediatric patients needs further investigation.

Prescribing errors in children: what is the impact of a computerized physician order entry?

Prescribing errors represent a safety risk for hospitalized patients, especially in pediatrics. Computerized physician order entry (CPOE) might reduce prescribing errors, although its effect has not yet been thoroughly studied on pediatric general wards. This study investigated the impact of a CPOE on prescribing errors in children on general wards at the University Children's Hospital Zurich. We performed medication reviews on a total of 1000 patients before and after the implementation of a CPOE. The CPOE included limited clinical decision support (CDS) such as drug-drug interaction check and checks for duplicates. Prescribing errors, their type according to the PCNE classification, their severity (adapted NCC MERP index), as well as the interrater reliability (Cohen's kappa), were analyzed. Potentially harmful errors were significantly reduced from 18 errors/100 prescriptions (95% CI: 17-20) to 11 errors/100 prescriptions (95% CI: 9-12) after CPOE implementation. A large number of errors with low potential for harm (e.g., "missing information") was reduced after the introduction of the CPOE, and consequently, the overall severity of potential harm increased post-CPOE. Despite general error rate reduction, medication reconciliation problems (PCNE error 8), such as drugs prescribed on paper as well as electronically, significantly increased after the introduction of the CPOE. The most common pediatric prescribing errors, the dosing errors (PCNE errors 3), were not altered on a statistically significant level after the introduction of the CPOE. Interrater reliability showed moderate agreement (Κ = 0.48).  Conclusion: Patient safety increased by reducing the rate of prescribing errors after CPOE implementation. The reason for the observed increase in medication reconciliation problems might be the hybrid system with remaining paper prescriptions for special medication. The lacking effect on dosing errors could be explained by the fact that a web application CDS covering dosing recommendations (PEDeDose) was already in use before the implementation of the CPOE. Further investigations should focus on eliminating hybrid systems, interventions to increase the usability of the CPOE, and full integration of CDS tools such as automated dose checks into the CPOE. What is Known: • Prescribing errors, especially dosing errors, are a common safety threat for pediatric inpatients. •The introduction of a CPOE may reduce prescribing errors, though pediatric general wards are poorly studied. What is New: •To our knowledge, this is the first study on prescribing errors in pediatric general wards in Switzerland investigating the impact of a CPOE. •We found that the overall error rate was significantly reduced after the implementation of the CPOE. The severity of potential harm was higher in the post-CPOE period, which implies that low-severity errors were substantially reduced after CPOE implementation. Dosing errors were not reduced, but missing information errors and drug selection errors were reduced. On the other hand, medication reconciliation problems increased.