Article Text
PDF
Abstract
Introduction For improving and optimising drug use in children, we previously developed a tool (including a series of criteria for identifying potentially inappropriate prescribing in children) by literature review and the two-round Delphi technique to prevent inappropriate medication prescriptions at the prescribing stage.
Objective To assess the prevalence of potentially inappropriate prescription (PIP) among hospitalised children and explore risk factors associated with PIP.
Design A retrospective cross-sectional study.
Setting A tertiary children’s hospital in China.
Participants Hospitalised children with complete medical records who received drug treatment and discharged from 1 January to 31 December 2021.
Outcome measures We evaluated the medication prescriptions by using a series of previously developed criteria for detecting the prevalence of PIP in hospitalised children and used logistic regression to explore the risk factors (including sex, age, number of drugs, number of comorbidities, days of hospitalisation and admission departments) for PIP in children.
Results A total of 87 555 medication prescriptions for 16 995 hospitalised children were analysed, and 19 722 PIPs were detected. The prevalence of PIP was 22.53%, and 36.92% of the children had at least one PIP during hospitalisation. The department with the highest prevalence of PIP was the surgical department (OR 9.413; 95% CI 5.521 to 16.046), followed by the paediatric intensive care unit (PICU; OR 8.206; 95% CI 6.643 to 10.137). ‘Inhaled corticosteroids for children with respiratory infections but without chronic respiratory diseases’ was the most frequent PIP. Logistic regression results showed that PIP was more likely to occur in male patients (OR 1.128, 95% CI 1.059 to 1.202) and younger patients (<2 years old; OR 1.974; 95% CI 1.739 to 2.241), and in those with more comorbidities (≥11 types; OR 4.181; 95% CI 3.671 to 4.761), concomitant drugs (≥11 types; OR 22.250; 95% CI 14.468 to 34.223) or longer hospital stay (≥30 days; OR 8.130; 95% CI 6.727 to 9.827).
Conclusions Medications for long-term hospitalised young children with multiple comorbidities should be minimised and optimised, to avoid PIP, reduce adverse drug reactions and ensure children’s medication safety. The surgery department and PICU had a high prevalence of PIP in the studied hospital and should be the focus of supervision and management in routine prescription review.
- PAEDIATRICS
- CLINICAL PHARMACOLOGY
- HEALTH SERVICES ADMINISTRATION & MANAGEMENT
- PUBLIC HEALTH
Data availability statement
Data are available upon reasonable request. The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
STRENGTHS AND LIMITATIONS OF THIS STUDY
This study was based on a hospital information system (real-world data), which allowed the study to have a large sample size and the findings to reflect real daily paediatric prescribing practices.
This is a single-centre study, so the generalisability of the findings is unknown.
Because this study is a retrospective database analysis, some information is not available; for example, we cannot know the temporal sequence of each diagnosis and each drug in children with many coexisting diseases or conditions, which could lead to some potentially inappropriate prescription (PIP) criteria for children with specific diseases or conditions be overestimated.
The prevalence of PIP in children with skin diseases has not been adequately assessed because the hospital where the study was conducted does not have a dermatology department.
Introduction
Inappropriate drug use is one of the major causes of preventable adverse drug events (ADEs), with significant implications for public health and healthcare costs.1–5 Despite tools such as the WHO Model List of Essential Medicines can help prescribers choose the most appropriate medicines for patients, the WHO estimates that 50% of medicines are still prescribed and used inappropriately.6 Timely detection and prevention of inappropriate medication prescriptions at the prescribing stage are expected to improve the rational use of drugs and reduce ADEs.
The latest definition divides potentially inappropriate prescription (PIP) into potentially inappropriate medication (PIM) and potential prescribing omission (PPO).7 PIM refers to the drug use when the potential risks of ADEs outweigh the potential clinical benefits (the risk of ADEs is high, or there is an unfavourable cost–effectiveness ratio or risk–benefit ratio); PPO is the omission of prescribing drugs that have been confirmed to be significantly beneficial to the patient’s longevity or quality of life in the absence of contraindications, that is, underprescribing of beneficial drugs.8
In the elderly, many tools have been developed to detect PIP, such as the Beers criteria,9 the Improved Prescribing in the Elderly Tool,10 the Screening Tool of Older Person’s Prescriptions and Screening Tool to Alert doctors to Right Treatment (the STOPP/START criteria),11 and the Medication Appropriateness Index.12 With the help of these tools, researchers have conducted many studies that detected prevalence rates of PIM in the elderly ranging from 32.4% to 66.8%; and found that the STOPP/START tool had significant beneficial effects on negative outcomes such as adverse drug reactions (ADRs), length of hospital stay, mortality and medication costs.5
Due to the particularity of the paediatric population—pharmacokinetic differences from adults and the relative lack of children’s drugs and medication information—the problem of off-label drug use13 14 and irrational drug use in children is serious.14–17 Children are a high-risk group for ADRs. In the paediatric population, the incidence of ADRs in inpatient children is 9.53% (6.81% to 12.26%) and in outpatient children is 1.46% (0.7% to 3.03%).18 The incidence of ADRs leading to hospital admission in children is 2.09% (1.02% to 3.77%).18 Prescribing appropriate medication prescription is the basis for optimising children’s medical and health services and reducing ADRs. Unfortunately, the development and research of PIP detection tools for children are insufficient and limited.19
To improve and optimise drug use in children, we previously developed a tool (including a series of criteria for identifying potentially inappropriate prescribing in children) by literature review and the two-round Delphi technique.20 This study aimed to use our developed tool to assess the prevalence of PIP (including PIM and PPO) in hospitalised children, explore risk factors associated with PIP in children and prepare for further conducting the prospective interventional study.
Methods
Study design and population
This study is a retrospective database analysis, in which we retrospectively analysed the medication prescriptions of hospitalised children discharged from the West China Second University Hospital (a tertiary hospital) from January to December 2021. Inclusion criterion: paediatric inpatients aged ≤18 years old. Exclusion criteria: patients with incomplete medical records and patients who did not use medications.
Data collection
The related data of paediatric inpatients were extracted from the Hospital Information System of the West China Second University Hospital, including admission department, hospitalisation number, sex, age, height, weight, body temperature, disease diagnosis, drug prescription (drug code, name, dosage form, dose, administration route, frequency, course) and length of hospital stay.
PIP assessment
Two pharmacists used our previously developed tool20 to evaluate the drug prescriptions of paediatric inpatients independently, and then they cross-checked the prescription review results. When the result was inconsistent, the two pharmacists reached an agreement through discussion or it was decided by a third clinician or pharmacist with a senior title who participated in the development of the tool.
Outcomes
The study outcomes were the prevalence rates of PIP, PIM and PPO in hospitalised children. Once a prescription met any PIM or PPO criteria, the prescription was considered a PIM or PPO. The number of PIPs was equal to the sum of the number of PPOs and PIMs.
Statistical analysis
Data cleansing, processing and analysis were performed by IBM SPSS V.23.0 and Excel V.2019. Continuous variable data were presented as median and IQR (25th–75th percentiles, Q1–Q3) or mean and SD depending on the normal distribution and the independent samples t-test or Mann-Whitney U test was used for comparison of differences between groups. Categorical variable data were presented as number (n) and frequency (%) and the Χ2 test was used for comparison of differences between groups. The significance level α was equal to 0.05, so p<0.05 indicated a statistically significant difference between groups. In addition, logistic regression was applied to identify risk factors associated with PIP (yes/no) in hospitalised children. Potential risk factors associated with PIP included sex, age, number of drugs, number of comorbidities, days of hospitalisation and admission departments. Univariate analysis was performed first, and if there were statistically significant factors (p<0.1) in the univariate analysis, multivariate logistic regression analysis would be performed.
Patient and public involvement
None.
Results
PIP in hospitalised children
Two pharmacists analysed 87 555 drug prescriptions of 16 995 hospitalised children discharged from 1 January to 31 December 2021. The demographic and clinical characteristics of these children were shown in table 1. There were significant differences between the inpatients with and without PIP in sex, age, length of hospital stay, number of comorbidities, number of drugs and admission departments (p<0.001).
Demographic and clinical characteristics of hospitalised children
A total of 19 722 PIPs (18 381 for PIM, 1341 for PPO) were detected and the prevalence rate of PIP was 22.53% (20.99% for PIM, 1.53% for PPO). A total of 6274 children (36.92%, 6274 of 16 995) had at least one PIP and the median number of PIP per child was 2 (figure 1).
Distribution of the number of potentially inappropriate prescriptions (PIPs) per hospitalised child.
The prevalence rates of PIP were comparatively higher in hospitalised children aged ≤2 years (OR 1.974; 95% CI 1.739 to 2.241), or with ≥11 diseases (4.181; 3.671 to 4.761), ≥11 drugs (22.250; 14.468 to 34.223) or hospital stay ≥30 days (8.130; 6.727 to 9.827). The department of surgery had the highest prevalence of PIP followed by the paediatric intensive care unit (PICU), department of respiratory and department of neurology (table 2 and online supplemental file 1).
Supplemental material
Prevalence of PIP in hospitalised children with different characteristics
The most frequent PIM in general hospitalised children was ‘antiepileptic drugs (AEDs) with liver enzyme induction (eg, phenytoin, carbamazepine, oxcarbazepine) in children’ (16.9% of all PIPs), and the most frequent PIM in hospitalised children with specific diseases or conditions was ‘inhaled corticosteroids (ICS) for children with respiratory infections without chronic respiratory diseases’ (21.0%) (online supplemental file 2).
Supplemental material
The common PPO in hospitalised children included ‘omission to prescribe therapeutic drug monitoring (TDM) to children with AEDs’ (4.4%), ‘omission to prescribe intestinal microecological preparations (Brucella, Bifidobacterium, Lactobacillus, etc) to non-immunocompromised children with acute or antibiotic-related diarrhoea’ (0.8%), ‘omission to prescribe ICS to children ≥6 years old with asthma’ (0.8%; online supplemental file 2).
Risk factors for PIP in children
The results of logistic regression showed that sex, age, length of hospital stay, number of comorbidities, number of drugs and admission departments were risk factors for PIP in hospitalised children. In males or the younger, or those with more comorbidities or drugs, or longer hospital stay, the PIP would more likely occur. Moreover, children in the surgery department were more likely to receive PIP (table 3).
Results of multivariate logistic regression analysis of risk factors associated with PIP in hospitalised children
Discussion
Our study found that the prevalence of PIP in hospitalised children was 22.53% (PIM 20.99%, PPO 1.53%), and 36.92% of children had at least one PIP during hospitalisation. Logistic regression results showed that the risk factors of children’s PIP included sex, age, length of hospital stay, number of comorbidities, number of drugs and admission departments. The younger children with more comorbidities, more concomitant drugs and longer hospital stays were at a higher risk of PIP. It is suggested that medications for long-term hospitalised young children with multiple comorbidities should be minimised and optimised, to avoid PIP, reduce ADRs and ensure children’s medication safety.
The POPI (Pediatrics: Omission of Prescriptions and Inappropriate prescriptions) is the first tool developed by French clinicians and pharmacists to detect PIP in children.21 It was used by Berthe-Aucejo et al to identify PIM and PPO in children aged <18 years in a hospital emergency department and a community pharmacy.22 After analysing 18 562 prescriptions of 15 973 children in the hospital emergency department and 4780 prescriptions of 2225 children in the community pharmacy, they found that the prevalence rates of children’s PIM and PPO in the hospital emergency department were 2.9% and 2.3%, 3.3% and 2.7% of children who experienced at least one PIM and PPO; in the community pharmacy, PIM and PPO accounted for 12.3% and 6.1% of all prescriptions, affecting 26.4% and 11.3% of children. Moreover, the study also found that age and number of concomitant drugs were risk factors for PIM and PPO. In 2016, 15 general practitioners, pharmacists and paediatricians from Ireland and the UK jointly developed a tool named PIPc criteria23 including 12 propositions on children’s PIP to detect the prevalence of PIP in children in primary care settings where a patient’s clinical information is limited. Then, researchers conducted a retrospective prescription review in children <16 years with this tool24 and found that the prevalence of PIM in children in primary care settings was 3.5% and the prevalence of PPO was 2.5%. Males had a higher risk of PPO compared with females (rate ratio 1.3; 95% CI 1.0 to 1.7, p<0.05). Consistent with the above studies, our study found that sex, age and number of concomitant drugs were risk factors for children’s PIP. Moreover, our study results showed that the length of hospital stay, number of comorbidities and admission departments also had an impact on PIP in hospitalised children. The higher prevalence of PIP in our study may be related to the fact that our subjects are hospitalised children who tend to have more complex conditions and take more drugs, and there are more PIP criteria in our tool (the number of PIP criteria in our tool vs POPI vs PIPc: 136 vs 82 vs 12).
‘ICS for children with respiratory infections without chronic respiratory diseases’ was the most frequent PIP, accounting for 21.0% of all PIPs. Bronchial asthma and chronic asthmatic bronchitis are the approved indications for the inhaled budesonide suspension on the label, but our study results found that it was frequently used for pneumonia, wheezing bronchitis, constrictive (obliterative) bronchitis, acute upper airway obstruction and other off-label indications. Yang et al retrospectively investigated the use of inhaled budesonide suspension in the outpatient and emergency department of a children’s hospital. Their study results showed that the inhaled budesonide suspension was used in 17 diseases including pneumonia, bronchitis, acute upper respiratory tract infection, etc. The proportions of inhaled budesonide suspension in pneumonia, bronchitis, acute wheezing bronchitis and acute laryngitis/acute laryngotracheitis were 38.70%, 22.90%, 19.20% and 12.70%, respectively, while in asthma, only accounted for 0.30%.25 ICS is widely used in paediatrics in China, especially in young children, but there is much off-label use, and clinicians and children have to bear risks of it. Relevant departments should formulate and improve relevant laws and regulations to ensure rational and safe medication for children and avoid medical risks (for example, the relevant departments should urge pharmaceutical companies to timely revise their drug labels based on clinical study evidence). Moreover, based on the off-label use of ICS in medical institutions, clinical pharmacists should form guidance on the reasonable off-label use of ICS by reviewing relevant evidence to regulate ICS use.
We detected 1303 prescriptions (6.6% of all PIPs) of oral or intravenous azithromycin or erythromycin in neonates ≤14 days, mainly in neonates infected with Ureaplasma urealyticum. U. urealyticum colonising the respiratory tract of premature infants is an important risk factor for bronchopulmonary dysplasia (BPD).26 27 Macrolide antibiotics (such as erythromycin, azithromycin and clarithromycin) have been shown to have antibacterial activity against U. urealyticum in vitro,28–30 and are expected to clear the Ureaplasma infection, inhibit the inflammatory response in the lungs and reduce the risk of BPD. However, due to insufficient relevant clinical studies, limitations of methodology in the existing studies (such as lack of appropriate controls, no blinding, small sample size, lower-than-expected colonisation rate of U. urealyticum in children, etc), inconsistent study results and insufficient pharmacokinetic data to support drug dosage selection, the macrolide antibiotic use in children infected with U. urealyticum is controversial.31–34 In most studies, treatment began on the seventh day after birth,31 but available clinical evidence suggested that systemic azithromycin or erythromycin would increase the risk of infantile hypertrophic pyloric stenosis in young children. The association was strongest when the exposure occurred in the first 2 weeks after birth, and persisted in children 2–6 weeks of age, but was weaker.35–37 Currently, given insufficient evidence on the benefit of macrolides for children infected with U. urealyticum and their risk of potentially serious adverse effects in neonates, routine use of macrolides for prevention and treatment of BPD caused by U. urealyticum infection in neonates is not recommended and paediatricians should fully weigh risks and benefits before using.
Sedative antihistamines such as diphenhydramine have potentially life-threatening side effects (such as respiratory depression), and the young may be more susceptible.38 In 2004, the Food and Drug Administration (FDA) reviewed 125 cases which reported adverse events caused by promethazine in children ≤16 years of age and found 38 cases of respiratory depression, apnoea or cardiac arrest. At the end of 2004, a ‘black box’ warning, including ‘a contraindication for use in children less than two years of age and a strengthened warning with regard to use in children two years of age or older’, was added to the label of promethazine hydrochloride (Phenergan) by the FDA.39 In December 2014, the Medicines and Healthcare products Regulatory Agency recommended antihistamines (brompheniramine, chlorpheniramine, diphenhydramine, doxylamine, promethazine and triprolidine) should no longer be used in children under 6 years.40 Our study detected 741 (3.8%) prescriptions of sedative antihistamines (diphenhydramine, promethazine, chlorpheniramine, etc) in children aged <2 years, which are potentially risky for children.
In addition, there are potential risks associated with antibiotic use in children. A total of 1115 (5.7%) and 796 (4.0%) antibiotic prescriptions were detected in children with acute upper respiratory tract infection with symptoms less than 4 days and in children with diarrhoea that was not confirmed to be caused by bacterial infection, respectively. In most cases, acute upper respiratory tract infection and diarrhoea in children are self-limiting, usually caused by viral infections and do not need antibiotics. The misuse of antibiotics will increase the risk of bacterial resistance.41–44
Although this study detected the largest number of children with PIP (1053 of 4129, 25.50%) in the neonatal department, the departments with the top two highest prevalence of PIP were the surgery department (58 of 76, 76.32%) and PICU (368 of 499, 73.75%), which may be related to the inexperience of surgeons in medication and children in the PICU with more complex conditions, more comorbidities and concomitant drugs, and longer hospital stays.45 46
These results show that our previously developed tool has good clinical applicability and can be used as an auxiliary tool for prescribers and prescription reviewers to identify PIP and regulate paediatric drug prescribing.
Strengths and limitations
Our study has several limitations. First, this is a single-centre study, so the generalisability of the findings is unknown. Second, because our study is a retrospective database analysis, some information is not available, such as we cannot know the temporal sequence of each diagnosis and each drug in children with many coexisting diseases or conditions, which could lead to some PIP criteria for children being overestimated (eg, 898 corticosteroid prescriptions were detected in children with fever without sepsis and toxic shock, and we speculated that they were used as antipyretic agents. But, in children with multiple comorbidities, prescribers prescribing glucocorticoids might not be for relieving fever but for other pharmacological purposes, which may lead to the prevalence of such PIP being exaggerated.). Our study found that PIP criteria for children with specific diseases or conditions are more appropriate for real-time prescription review and monitoring, while the PIP criteria for general children are easy to use in both retrospective prescription analysis and prospective prescription review and can detect PIP almost without access to the patient’s clinical information (these criteria may be helpful for community pharmacies). In addition, the applicability of PIP criteria for children with skin diseases could not be adequately assessed because the hospital where the study was conducted does not have a dermatology department. Finally, the prescriptions screened by our tool are only PIPs, and the screening results cannot be completely equal to the final rationality of the prescriptions. In special clinical situations, some children with complex conditions may have to use drugs in the PIP criteria after a thorough assessment of the patient’s condition by medical staff (prescriptions are reasonable in such cases). However, this is the first study to assess the prevalence of PIP and to explore risk factors for PIP in paediatric Chinese patients. This study validated the clinical applicability of our previously developed tool for identifying PIP in children, which is conducive to the use of this tool in clinical practice, promoting clinically rational drug use in children and reducing the risk of drug use and avoidable ADRs.
Directions for further research
This study has confirmed that our previously developed tool can detect potentially inappropriate prescribing behaviour in paediatric patients, so we will promote this tool application among medical staff and conduct a study to evaluate its reliability. If it is confirmed that the prescription screening results from this tool are stable and reliable, we will further conduct a multicentre prospective study to explore whether this tool can help clinicians and pharmacists to prescribe reasonable drugs and whether it can reduce the paediatric PIM, PPO and ADRs and improve children’s quality of life. Moreover, like the STOPP/START criteria47 and PIM-Check criteria48 49 for the elderly, we hope to integrate this tool into clinical decision support systems or prescription review software through computer coding algorithm to realise timely automatic identification and quantification of children’s PIP, which will contribute to the continuous improvement of medical quality.
Conclusion
Medications for long-term hospitalised young children with multiple comorbidities should be minimised and optimised, to avoid PIP, reduce ADRs and ensure children’s medication safety. Besides, the surgery department and PICU with a high prevalence of PIP should be the focus of supervision and management in routine prescription review.
Data availability statement
Data are available upon reasonable request. The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
This study was registered in the Ethics Committee of the West China Second University Hospital (record number 2022158).
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors Conceptualisation—SL and LH. Data curation—SL. Formal analysis—SL. Funding acquisition—LZeng. Methodology—SL and LH. Project administration—SL, LH and LZhang. Resources—SL, LH and LZhang. Software—SL. Supervision—SL, LH, LZeng, DY, Z-JJ, GC, HW and LZhang. Validation—SL and LH. Writing (original draft)—SL. Writing (review and editing)—SL, LH, LZeng, DY, Z-JJ, GC, HW and LZhang. LZhang is responsible for the overall content as the guarantor.
Funding This work was supported by Sichuan Province Science and Technology Plan Project (no. 2020YFS0035).
Disclaimer The funder did not participate in the work.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.