Abstract
OBJECTIVE
The study aims to identify the rate of inappropriate prescribing per the Key Potentially Inappropriate Drugs in Pediatrics (KIDs) List versus total prescribing in patients at University of Missouri Health Care hospitals.
METHODS
This retrospective study evaluated orders for patients treated at University of Missouri Health Care inpatient units or emergency departments with a KIDs List medication between September 1, 2019, and September 1, 2020, or a reported adverse event to one of these medications between September 1, 2015, and September 1, 2020. Patients were excluded if the patient safety report was related to a medication error rather than an adverse event. Safety measures assessed included age and weight filtering, dose-range checking, clinical decision support, and override availability.
RESULTS
There were 39 inappropriate orders and 4 possible adverse events identified. A total of 8 of 33 medications (24%) had age and weight filtering in place for at least 1 order sentence, 1 of 38 (2.6%) had dose-range checking, no medications had an active clinical decision support alert, and 33 of 38 (87%) had availability on automated dispensing cabinet override.
CONCLUSIONS
Use of KIDs List medications is appropriately low, but low levels of safety measure implementation leave pediatric patients vulnerable.
Keywords: electronic health records, medical order entry systems, patient safety, pediatrics, potentially inappropriate medication list
Introduction
Although recommendations for use of medication in adults are generally guided by significant high-quality research, such recommendations are often not available in pediatrics or are based on expert opinion because this population is often excluded from clinical trials. Evidence-based medicine is considered the standard of care, but it is limited for special populations. The Beers Criteria,1 developed in 1991, as well as the Screening Tool to Alert to Right Treatment and Screening Tool of Older Persons' Prescriptions criteria,2 developed in 2003, have guided the potentially appropriate and inappropriate use of medication in the elderly population, but the pediatric community has not had such guidance. This leaves a greater margin for error in select pediatric populations which may have a reduced capacity to compensate for adverse drug events.3 Adverse drug events are defined as any noxious and unintended response to a drug that occurs in humans at doses normally used for prophylaxis, treatment, or diagnosis of disease, or for the modification of physiologic function.1
The publication of the KIDs List seeks to provide guidance to practitioners caring for pediatric patients by evaluating the literature that is available to assess the appropriateness of medications and medication excipients in this population. Many recommendations are considered weak because of the lack of primary literature and the reliance on clinical experience and knowledge of pharmacokinetics and pharmacology in pediatrics.3 The list is not intended to provide contraindications, but rather to assist in the decision-making and selection of appropriate medication regimens in children. The KIDs List includes lists of medications and excipients that are potentially harmful to pediatric patients, with rationale for inclusion of each.3 The included products are recommended to either avoid if evidence was strong or the potential adverse event was life-threatening or life-altering, or use with caution if evidence was less strong and the potential adverse event was less serious.3
Medication errors and adverse drug events disproportionately affect patients at extremes of age, and exact rates of errors are difficult to identify. A study of medication errors considered to be malpractice found that two thirds of errors were in pediatric or geriatric patients.4 Children may be particularly at risk because of the additional steps involved in identifying appropriate drugs and doses. A total of 69% of calculation errors involve pediatric patients, and 34% of errors in patient weight are associated with medication errors.5 Errors in dose, however, are not the only type of errors with particularly concerning effects on children. A study of errors in pediatric intensive care unit patients found a 6.6% rate of errors in dose selection and a 2.1% rate of errors in drug selection.6
This study was designed to evaluate the use of and safety measures in place for medications on the KIDs List recommended to be avoided that are available for use within the University of Missouri Health Care system to identify opportunities for improvement. Implementation of safety measures related to medications on the KIDs List is expected to reduce errors in drug selection.
Materials and Methods
The primary objective was to identify the rate of inappropriate prescribing per the KIDs List versus total prescribing in patients at University Hospital and Women's and Children's Hospital of selected KIDs List medications. Secondary objectives were to identify adverse drug reactions (ADRs) to KIDs List medications in patients in whom use was recommended against per the KIDs List, and safety measures in place for these medications. Safety measures assessed were age and weight filtering, clinical decision support alerts, dose-range checking, and automated dispensing cabinet (ADC) override availability.
Age and weight filtering of order sentences guides the provider toward appropriate dosing by hiding orderables (medication orders with dose and frequency information) for patients outside the recommended age group or weight range and capping orders at the adult or maximum dose once the weight cutoff is reached. For the purpose of our data collection, this end point was only evaluated for formulary medications because there are no order sentences (medication, dose, route, frequency, duration) viewable to providers for nonformulary medications. Clinical decision support alerts are pop-ups that provide information upon placing an order. These alerts can be a hard stop in which the user cannot proceed with their intended action, or soft stops in which action is recommended but not required to continue. Dose-range checking is similar to clinical decision support alerts, but the information included alerts specifically based on the dose of the medication being higher or lower than expected for normal or abnormal renal or hepatic function. Alerts can also be added to notify providers of inappropriate use for specified age groups. The ADC override allows removal of a medication from an automated dispensing cabinet prior to an order being placed in the electronic medical record.
Medications recommended to be avoided, and which are also available at University Hospital and Women's and Children's Hospital, were evaluated in this study (Table 1). Orders for medications with recommendations to be avoided in specific indications (i.e., topical corticosteroids, benzocaine, lidocaine 2% viscous, loperamide, and naloxone) were not evaluated for appropriateness of order given the challenges in retrospectively in determining indication; however, they were considered for safety measures. Excipients were not evaluated in this study.
Table 1.
Medications Evaluated
| Drug3 | Avoid Criteria3* | Risk3 |
|---|---|---|
| Azithromycin | Neonates unless Bordetella pertussis, Chlamydia trachomatis pneumonia, ureaplasma | Hypertrophic pyloric stenosis |
| Benzocaine | Infants for teething or pharyngitis | Methemoglobinemia |
| Chlorpromazine | Infants | Acute dystonia (dyskinesia); increased risk of respiratory depression, extravasation, death with intravenous use |
| Codeine | Children unless pharmacogenetic testing is used | Respiratory depression, death |
| Corticosteroid topical medium to very high potency | <1 yr for diaper dermatitis | Adrenal suppression, higher rates of absorption |
| Darunavir | <3 yr or <10 kg | Seizures, death |
| Desipramine | Children | Sudden cardiac death |
| Dicyclomine | <6 mo | Apnea |
| Diphenoxylate atropine | <6 yr | Respiratory depression, death |
| Divalproex | Infants | Pancreatitis, fatal hepatotoxicity |
| Erythromycin | Neonates unless chlamydia trachomatis pneumonia | Hypertrophic pyloric stenosis |
| Fluphenazine | Infants | Acute dystonia (dyskinesia); increased risk of respiratory depression, extravasation, death with intravenous use |
| Gentamicin ophthalmic ointment | Neonates | Severe ocular reactions |
| Haloperidol | Infants | Acute dystonia (dyskinesia); increased risk of respiratory depression, extravasation, death with intravenous use |
| Hexachlorophene | Neonates | Neurotoxicity |
| Ivermectin | <1 yr | Encephalopathy |
| Lidocaine 2% viscous | Infants for teething | Seizures, arrhythmia, death |
| Linaclotide | <6 yr | Death from dehydration |
| Loperamide | Infants for acute infectious diarrhea | Ileus, lethargy |
| Meperidine | Neonates | Respiratory depression |
| Metoclopramide | Infants | Acute dystonia (dyskinesia); increased risk of respiratory depression, extravasation, death with intravenous use |
| Midazolam | Very low birth weight neonates | intraventricular hemorrhage, periventricular leukomalacia, death |
| Mineral oil oral | <1 yr | Lipid pneumonitis |
| Naloxone | Neonates for postpartum resuscitation | Seizure |
| Nitrofurantoin | Neonates | Hemolytic anemia |
Table 1.
Medications Evaluated (cont.)
| Drug3 | Avoid Criteria3* | Risk3 |
|---|---|---|
| Opium tincture | Neonates | Respiratory depression |
| Perphenazine | Infants | Acute dystonia (dyskinesia); increased risk of respiratory depression, extravasation, death with intravenous use |
| Pimozide | Infants | Acute dystonia (dyskinesia); increased risk of respiratory depression, extravasation, death with intravenous use |
| Prochlorperazine | Infants | Acute dystonia (dyskinesia); increased risk of respiratory depression, extravasation, death with intravenous use |
| Promethazine | Infants | Acute dystonia (dyskinesia); increased risk of respiratory depression, extravasation, death with intravenous use |
| Propofol | >4 mg/kg/hr for >48 hr | Propofol-related infusion syndrome |
| Silver sulfadiazine | Neonates | Kernicterus |
| Sodium phosphate rectal | Infants | Electrolyte abnormalities, acute kidney injury, arrhythmia, death |
| Sodium polystyrene sulfonate | Very low birth weight neonates | Colonic perforation |
| Sulfamethoxazole | Neonates | Kernicterus |
| Trifluoperazine | Infants | Acute dystonia (dyskinesia); increased risk of respiratory depression, extravasation, death with intravenous use |
| Valproic acid | Infants | Pancreatitis, fatal hepatotoxicity |
| Verapamil | <1 yr | Asystole |
* Infants defined as <24 mo of age and children as <18 yr of age per KIDs List criteria.3
Patients were included if they were treated at University Hospital or Women's and Children's Hospital inpatient or emergency departments with a reported ADR to a KIDs List medication listed in Table 1 between September 1, 2015, and September 30, 2020, or an inpatient order for a KIDs List medication between September 1, 2019, and September 30, 2020. Time frame was extended for adverse event reporting versus chart review given the expected lower volume of potential adverse event reports compared with the number of charts to review, because adverse events are reported via the Patient Safety Network system outside of the medical record and must be recognized as a potential unexpected event in order to be reported. If patients had an order for a KIDs List medication that was not appropriate per the KIDs List but was administered, the chart was evaluated for presence of the adverse event the KIDs List is trying to avoid. Patients were excluded if the patient safety report was related to a medical error rather than an adverse reaction.
The study was conducted via retrospective chart review and evaluation of the electronic medical record. Data were analyzed using descriptive statistics.
Results
Prescribing rates in patients for whom medications are recommended to be avoided were low. Most of the identified medications had no orders during the study period (Table 2). Of the medications with orders in the targeted population during the study period, these orders were greatly outnumbered by appropriate medication orders. Of the medications ordered, not all had documented administrations, including 1 order each for desipramine, azithromycin, and promethazine.
Table 2.
Inappropriate Versus Total Prescribing
| Drug | Total Orders | Inappropriate Orders | Percent Inappropriate |
|---|---|---|---|
| Azithromycin | 3396 | 5 | 0.15 |
| Chlorpromazine | 499 | 0 | 0 |
| Codeine | 50 | 0 | 0 |
| Darunavir | 5 | 0 | 0 |
| Desipramine | 1 | 1 | 100 |
| Dicyclomine | 622 | 0 | 0 |
| Diphenoxylate atropine | 168 | 0 | 0 |
| Divalproex | 1373 | 0 | 0 |
| Erythromycin | 165 | 0 | 0 |
| Fluphenazine | 10 | 0 | 0 |
| Gentamicin ophthalmic ointment | 2 | 0 | 0 |
| Haloperidol | 3667 | 2 | 0.05 |
| Hexachlorophene | 0 | 0 | 0 |
| Ivermectin | 3 | 0 | 0 |
| Linaclotide | 29 | 0 | 0 |
| Meperidine | 1903 | 0 | 0 |
| Metoclopramide | 1679 | 0 | 0 |
| Midazolam | 31,778 | 27 | 0.08 |
| Mineral oil oral | 226 | 0 | 0 |
| Nitrofurantoin | 464 | 0 | 0 |
| Opium tincture | 15 | 0 | 0 |
| Perphenazine | 2 | 0 | 0 |
| Pimozide | 5 | 0 | 0 |
| Prochlorperazine | 3685 | 0 | 0 |
| Promethazine | 5994 | 2 | 0.03 |
| Propofol | 39,810 | 0 | 0 |
| Silver sulfadiazine | 48 | 0 | 0 |
| Sodium phosphate rectal | 0 | 0 | 0 |
| Sodium polystyrene sulfonate | 790 | 0 | 0 |
| Sulfamethoxazole | 1639 | 0 | 0 |
| Trifluoperazine | 0 | 0 | 0 |
| Valproic acid | 599 | 2 | 0.33 |
| Verapamil | 1711 | 0 | 0 |
Similarly, medications that were administered during the study period infrequently resulted in documented ADRs. No ADRs were identified through patient safety reports. Four possible KIDs List described ADRs were identified via chart review, and all were very low birth weight (<1500 g) neonates (<1 month of age) who had received midazolam and subsequently developed bilateral intraventricular hemorrhage of at least grade 2.
Although use of medications on the KIDs List was low, implementation of safety measures was similarly low. No identified medication had any clinical decision support (CDS) alert in place. The only medication with active dose-range checking was midazolam (1 of 38; 2.6%). Of the 33 medications evaluated, 8 (24%) had age or weight filtering in place for at least 1 order sentence for at least 1 of the hospitals. These medications were codeine, haloperidol, midazolam, naloxone, promethazine, propofol, sulfamethoxazole, and verapamil. The only medication with age or weight filtering at 1 hospital was promethazine. The exception to the low use of safety measures was availability on ADC override. Per policy, 33 of 38 medications (87%) are unavailable on override. The medications available on override are chlorpromazine, haloperidol, midazolam, naloxone, and propofol (Table 3). Of these, midazolam, naloxone, and propofol are to be on override in an ADC intended to primarily service the population in which they are recommended to be avoided.
Table 3.
Use of Safety Measures *
| Medication | Filtering | CDS Alerts | Dose-Range Checking | ADC Override Availability |
|---|---|---|---|---|
| Azithromycin | No | No | No | No |
| Benzocaine | No | No | No | No |
| Chlorpromazine | No | No | No | Yes |
| Codeine | Yes | No | No | No |
| Corticosteroid topical medium to very high potency | — | No | No | No |
| Darunavir | — | No | No | No |
| Desipramine | — | No | No | No |
| Dicyclomine | No | No | No | No |
| Diphenoxylate atropine | No | No | No | No |
| Divalproex | No | No | No | No |
| Erythromycin | No | No | No | No |
| Fluphenazine | No | No | No | No |
| Gentamicin ophthalmic ointment | No | No | No | No |
| Haloperidol | Yes | No | No | Yes |
| Hexachlorophene | No | No | No | No |
| Ivermectin | — | No | No | No |
| Lidocaine 2% viscous | No | No | No | No |
| Linaclotide | — | No | No | No |
| Loperamide | No | No | No | No |
| Meperidine | No | No | No | No |
| Metoclopramide | No | No | No | No |
| Midazolam | Yes | No | Yes | Yes |
| Mineral oil oral | No | No | No | No |
| Naloxone | Yes | No | No | Yes |
| Nitrofurantoin | No | No | No | No |
| Opium tincture | No | No | No | No |
| Perphenazine | No | No | No | No |
| Pimozide | No | No | No | No |
| Prochlorperazine | No | No | No | No |
| Promethazine | Yes | No | No | No |
ADC, automated dispensing cabinet; CDS, clinical decision support
* Dashes indicate not applicable
Discussion
Overall, use of KIDs List medications in targeted populations was low, as were identified adverse events related to use of KIDs List medications, although it is unknown how the rate of inappropriate orders compares with the percentage of patients in these categories. The only medication for which a possible adverse event was identified was intraventricular hemorrhage after administration of midazolam. Although there are many factors that can affect development of intraventricular hemorrhage and very low birth weight neonates are at risk even prior to administration of precipitating medications, it cannot be ruled out that the midazolam played a role in these patients developing the condition. Although only 4 potential adverse events were identified via chart review and none on review of adverse event reporting, it is hypothesized that adverse events may occur more frequently than documentation reveals, which is a limitation of the retrospective design of this study.
Unfortunately, implementation of safety measures was similarly low. Safety measure technologies are not a substitute for clinical judgment, but they can improve patient care when used as a tool rather than a rule. This is particularly true because a study of errors classified as malpractice found that 30% of errors were at least in part due to unclear guidelines and protocols or lack of access to guidelines and protocols,4 suggesting that error rates may be mitigated by providing clear guidance about when medications should and should not be used. The average clinician comfortably uses approximately 80 to 100 medications in daily practice, but 20 to 30 new medications are approved each year, and many of these medications are not initially labelled for pediatric use.7 This requires the clinician to consider and potentially use medications with which they are less familiar and for which fewer pediatric data is available, increasing risk of errors.7 Addition of clinical decision support systems have been found to decrease both intensive care unit and overall hospital length of stay, and in combination with computerized provider order entry enhancements decrease rates of serious errors by greater than 50%.8 Definition of error and ability to detect errors can significantly affect estimated rates of errors. A study in pediatric anesthesia that relied on chart review and self-report of errors reported an error rate of 0.03%,9 whereas a study of errors in newborns which relied on a pharmacist observing the ordering, preparation, and administration of medication reported an error rate of 2.3 errors per patient-day, with 84.8% of patients affected by an error, and 58.5% affected by multiple errors.10 Reporting of errors is limited by lack of feedback and time, failure to recognize an error has occurred, fear of discipline, and lack of understanding of what to report.11 Errors are also less likely to be reported if it takes further investigation to identify whether an error has occurred.7 A study comparing direct observation to reported errors found 383 errors when measured via direct observation, whereas only 2 errors were identified through error reports.10 Recognition of errors is important in mitigating patient harm because 73.5% of errors reach the patient,11 and 70% of errors require intervention or lead to harm.6 The biggest effect could likely be made by adding clinical decision support alerts when these medications are ordered in patients in whom they are recommended to be avoided. This could provide just-in-time education to providers and pharmacists to enhance their assessment of appropriate medication use. These clinical decision support alerts were prioritized and are now nearly fully operationalized in the electronic medical record.
On the other hand, although age and weight filtering are not fully implemented for KIDs List medications and could be optimized, this project will take significantly more information technology time and effort but likely provides less significant gains than addition of CDS alerts. Although hiding order sentences for patients in whom a medication is not recommended may be a deterrent to some providers, it does not prevent ordering of the medication, nor does it provide guidance on why a pre-built order sentence is unavailable. This may prompt providers less familiar with the medication to seek assistance with dosing, at which point they may receive information regarding potentially safer alternatives, but more experienced providers are likely to fill in the dosing information with which they are familiar without recognition of why the order sentence is unavailable. Requests for addition of filtering have been submitted for most medications not currently being filtered, but this enhancement is not yet implemented.
Similarly, dose-range checking is currently only used for midazolam. Because the evaluated KIDs List medications are recommended to be avoided, dose-range checking could be used to provide clinical decision support–like alerts, but dosing information itself is likely lower yield in terms of prescribing an appropriate medication. It also will take significantly more time to build dose-range checking compared with other interventions, so this, again, will be a lower priority to reduce use of medications on the KIDs List.
Disallowing removal of medication from an ADC prior to an order being placed does protect patients by ensuring a provider has considered the use of that medication prior to administration; however, of the 3 medications available on ADC override in areas serving the patient population in which they are intended to be avoided—midazolam, naloxone, and propofol—it will not be recommended to remove permission to override these medications because emergent use in other patients in these areas may still be warranted. Alerts on ADC override will be implemented for these medications.
Although there was no use of gentamicin ophthalmic ointment or propofol meeting KIDs List criteria for avoidance, it is unknown if this was influenced in part by shortages of these products which led to gentamicin ointment being unavailable at Women's and Children's Hospital during the study period, and propofol to be in short supply at both hospitals for most of the study period.
Finally, this evaluation only included medication use. As the KIDs List also includes recommendations regarding excipients, expanding the scope in the future to include excipients may further enhance appropriate medication use.
Conclusion
Overall use of and adverse events related to KIDs List medication use in populations in whom avoidance is recommended is appropriately low, but safety measure implementation is similarly low which leaves pediatric patients vulnerable.
Acknowledgments
Portions of this manuscript have been presented at Mid-Missouri Society of Health-System Pharmacists CE event in Columbia, MO, on May 6, 2021. Portions of this manuscript have been virtually presented at Midwest Pharmacy Residents Conference May 12–14, 2021.
ABBREVIATIONS
- ADC
automated dispensing cabinet
- ADR
adverse drug reaction
- CDS
clinical decision support
- KIDs
key potentially inappropriate drugs in pediatrics
Funding Statement
Disclosures. The authors declare no financial interest in any product or service mentioned in the manuscript, including grants, equipment, medications, employment, gifts, and honoraria. The authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Footnotes
Disclosures. The authors declare no conflicts.
Ethical Approval and Informed Consent. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national guidelines on human experimentation and have been approved by the appropriate committees at our institution. Given the nature of this study, institutional review board/ethics committee review and informed consent were not required.
References
- 1.The 2019 American Geriatrics Society Beers Criteria Update Expert Panel American Geriatrics Society 2019 updated AGS Beers Criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc . 2019;67(4):674–694. doi: 10.1111/jgs.15767. [DOI] [PubMed] [Google Scholar]
- 2.O'Mahony D, O'Sullivan D, Byrne S et al. STOPP/START criteria for potentially inappropriate prescribing in older people: version 2. Age Aging . 2015;44(2):213–218. doi: 10.1093/ageing/afu145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Meyers RS, Thackray J, Matson KL et al. Key potentially inappropriate drugs in pediatrics: the KIDs List. J Pediatr Pharmacol Ther . 2020;25(3):175–191. doi: 10.5863/1551-6776-25.3.175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Björkstén KS, Bergqvist M, Andersén-Karlsson E et al. Medication errors as malpractice--a qualitative content analysis of 585 medication errors by nurses in Sweden. BMC Health Serv Res . 2016;16(1):431. doi: 10.1186/s12913-016-1695-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hirata KM, Kang AH, Ramirez GV et al. Pediatric weight errors and resultant medication dosing errors in the emergency department. Pediatr Emerg Care . 2019;35(9):637–642. doi: 10.1097/PEC.0000000000001277. [DOI] [PubMed] [Google Scholar]
- 6.Glanzmann C, Frey B, Meier CR, Vonbach P. Analysis of medication prescribing errors in critically ill children. Eur J Pediatr . 2015;174(10):1347–1355. doi: 10.1007/s00431-015-2542-4. [DOI] [PubMed] [Google Scholar]
- 7.Rieder M. Adverse drug reacitons in children: pediatric pharmacy and drug safety. J Pediatr Pharmacol Ther . 2019;24(1):4–9. doi: 10.5863/1551-6776-24.1.4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Prgomet M, Li L, Niazkhani Z et al. Impact of commercial computerized provider order entry (CPOE) and clinical decision support systems (CDSSs) on medication errors, length of stay, and mortality in intensive care units: a systematic review and meta-analysis. J Am Med Inform Assoc . 2017;24(2):413–422. doi: 10.1093/jamia/ocw145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Leahy IC, Lavoie M, Zurakowski D et al. Medication errors in a pediatric anesthesia setting: incidence, etiologies, and error reduction strategies. J Clin Anesth . 2018;49:107–111. doi: 10.1016/j.jclinane.2018.05.011. [DOI] [PubMed] [Google Scholar]
- 10.Palmero D, Di Paolo ER, Stadelmann C et al. Incident reports versus direct observation to identify medication errors and risk factors in hospitalised newborns. Eur J Pediatr . 2019;178(2):259–266. doi: 10.1007/s00431-018-3294-8. [DOI] [PubMed] [Google Scholar]
- 11.Rishoej RM, Almarsdóttir AB, Christesen HT et al. Medication errors in pediatric inpatients: a study based on a national mandatory reporting system. Eur J Pediatr . 2017;176(12):1697–1705. doi: 10.1007/s00431-017-3023-8. [DOI] [PubMed] [Google Scholar]