Article Text

Download PDFPDF

Retrospective study of irrational prescribing in French paediatric hospital: prevalence of inappropriate prescription detected by Pediatrics: Omission of Prescription and Inappropriate prescription (POPI) in the emergency unit and in the ambulatory setting
  1. Aurore Berthe-Aucejo1,2,
  2. Phuong Khanh Hoang Nguyen1,
  3. François Angoulvant2,3,
  4. Xavier Bellettre4,
  5. Patrick Albaret5,6,
  6. Thomas Weil1,
  7. Rym Boulkedid2,7,8,
  8. Olivier Bourdon1,6,9,10,
  9. Sonia Prot-Labarthe1,2,10
  1. 1Department of Pharmacy, AP-HP, Robert-Debré Hospital, Paris, France
  2. 2UMR-S1123, ECEVE; Inserm U1123, INSERM, Paris, Île-de-France, France
  3. 3Emergency Unit, AP-HP, Necker Hospital, Paris, Île-de-France, France
  4. 4Emergency unit, AP-HP, Robert-Debré Hospital, Paris, Île-de-France, France
  5. 5Pharmacy, Albaret Pharmacy, Cesson, France
  6. 6Clinical Pharmacy, Paris Descartes University, Paris, Île-de-France, France
  7. 7Clinical Epidemiology Unit, Robert-Debré Hospital, Paris, Île-de-France, France
  8. 8CIC-EC 1426, AP-HP, Robert-Debré Hospital, Paris, Île-de-France, France
  9. 9Laboratoire Educations et Pratiques de Santé, Paris XIII University, Bobigny, France
  10. 10Groupe Pédiatrie, Société Française de Pharmacie Clinique, Paris, France
  1. Correspondence to Dr Aurore Berthe-Aucejo; aurore.berthe84{at}


Background and objective Pediatrics: Omission of Prescription and Inappropriate prescription (POPI) is the first detection tool for potentially inappropriate medicines (PIMs) and potentially prescribing omissions (PPOs) in paediatrics. The aim of this study was to evaluate the prevalence of PIM and PPO detected by POPI regarding prescriptions in hospital and for outpatients. The second objective is to determine the risk factors related to PIM and PPO.

Design A retrospective, descriptive study was conducted in the emergency department (ED) and community pharmacy (CP) during 6 months. POPI was used to identify PIM and PPO.

Setting Robert-Debré Hospital (France) and Albaret community pharmacy (Seine and Marne).

Participants Patients who were under 18 years old and who had one or more drugs prescribed were included. Exclusion criteria consisted of inaccessible medical records for patients consulted in ED and prescription without drugs for outpatients.

Primary and secondary outcome measures PIM and PPO rate and risk factors.

Results At the ED, 18 562 prescriptions of 15 973 patients and 4780 prescriptions of 2225 patients at the CP were analysed. The PIM rate and PPO rate were, respectively, 2.9% and 2.3% at the ED and 12.3% and 6.1% at the CP. Respiratory and digestive diseases had the highest rate of PIM.

Conclusion This is the first study to assess the prevalence of PIM and PPO detected by POPI in a paediatric population. This study assessed PIMs or PPOs within a hospital and a community pharmacy. POPI could be used to improve drug use and patient care and to limit hospitalisation and adverse drug reaction. A prospective multicentric study should be conducted to evaluate the impact and benefit of implementing POPI in clinical practice.

  • inappropriate prescription
  • omission
  • tool
  • detection

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:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Strengths and limitations of this study

  • This study is the first to observe the prevalence of potentially inappropriate medicine (PIM) and potentially prescribing omission (PPO) in a paediatric population.

  • It is a retrospective and monocentric study. The prevalence of PIM and PPO may be underestimated (large number of prescriptions and absence of specific pathology). Some criteria could only be analysed in a prospective study. The lack of clinical information is the main limit to detection in a community setting.

  • Many omissions and inappropriate prescriptions can be easily detected with POPI despite limited clinical information.


Inappropriate prescribing is a known preventable cause of adverse drug events (ADEs) and has an important impact on public health and cost of care.1 2 In the literature, ADE is defined by ‘an injury resulting from medical intervention related to a drug’ (dose error, adverse drug reaction (ADR) and misuse of medication such as antibiotics).3–5 In the paediatric population, ADR during hospitalisation was estimated between 0.6% and 33.7% and between 1% and 1.5% for outpatients.6–9 Incidence of ADR leading to admission was evaluated between 1.8% and 17.7%.6 7 10 Many drugs were concerned in commonly used medication.11–13

The WHO estimated that 50% of medications are prescribed and used inappropriately.14 The most recent definition of inappropriate prescription (IP) encompasses potentially inappropriate medicines (PIMs) and potentially prescribing omissions (PPOs).15 In a report from the French National Authority for Health, PIMs are defined as ‘drugs being used in a situation in which the risks involved in treatment potentially outweigh the benefits, lack of demonstrated indication, high risk of ADE, or an unfavorable cost-effect or risk-benefit ratio exists’. PPO or underuse of appropriate medication is defined as the absence of initiation of an effective treatment in subjects with a condition for which one or several drug classes have demonstrated their efficacy. In an elderly population, which presents with age-related physiological changes and high prevalence of polypharmacy, various measures have been developed to detect PIM such as: Beers’ criteria, the Inappropriate Prescribing in the Elderly Tool, The Medication Appropriate Index and Screening Tool of Older Person’s prescriptions/Screening Tool to Alert doctor to Right Treatment (STOPP/START).16–21

Only the STOPP/START enables us to detect underprescription.15 Using these tools, many studies have been carried out which have detected that IPs range from 35% to 51% in the above population.22–26

Omission of prescriptions in geriatric population detected by the START tool concerned 58%–61% of patients.27 28 Negative outcomes related to an IP such as side effects, hospitalisation, mortality and utilisation of resources were also highlighted.2 21 29 30

Prescribing in a paediatric population is always challenging for physicians. It is often empirical and primarily based on safety and pharmacology information obtained in adults.31 This is a worry in a hospital or general practitioner setting and for the community pharmacists. With many off-label uses, they may be obligated to find alternative information sources and might even dispense infrequently for this vulnerable population.32 ADRs are three time higher in paediatric populations. This frequency is explained by the vulnerability of children, pharmacokinetic changes during childhood and paediatric off-label drug used.4 33 Large differences relating to treatment were seen within and between countries.6 34 Questions about the rationale of prescriptions could be asked.35 Optimising children’s care is based on rational prescribing and aims for a decrease in side effects.34 35 In order to improve the correct drug use and optimise practice, the first tool of detection for PIM and PPO was created by Prot-Labarthe et al in 2013. The tool was named Pediatrics: Omission of Prescriptions and Inappropriate prescriptions (POPI) (table 1).36 37 Presently, the complete tool has yet to be tested in clinical practice, and the prevalence of PIM and PPO is not known.

Table 1

Pediatrics: Omission of Prescriptions and Inappropriate prescriptions (POPI)

Our first aim is to assess the prevalence of PIM and PPO detected using POPI in hospital and outpatient care. This is its first application, focusing on prescriptions extracted from the emergency department (ED) and the community pharmacy (CP). Our second objective is to determine the risk factors related to PIM and PPO.



A retrospective and descriptive study was conducted in the ED of AP-HP Robert-Debré hospital (Paris)—the largest French paediatric hospital—and the Albaret community pharmacy (Seine and Marne). Inclusion criteria included patients who were under 18 years old and who had one or more drug prescriptions between 1 October 2014 and 31 March 2015. Prescription was defined as one or more lines of drugs prescribed by a physician. Exclusion criteria consisted of inaccessible medical records for ED patients and prescription without drugs for outpatients. POPI contains 101 criteria (76 PIMs, 25 PPOs). A literature review was done to obtain criteria. Criteria were categorised according to physiological systems (gastroenterology, respiratory infections, pain, neurology, dermatology and miscellaneous). Criteria were validated by a two-round Delphi consensus technique.37

Data collection

The prescriptions given on leaving hospital ED were extracted from the Urqual software V5 (McKesson Corp, Paris, France). Urqual is an emergency prescription software that is used in many French hospitals. Patient information including age, sex, weight, medical prescription and current diagnosis was collected. Medical histories and clinical examinations were consulted individually when necessary. Due to the significant amount of data, clinical files of the ED were analysed based on primary diagnosis. Prescriptions for secondary diagnosis were not evaluated. For this study, 82/101 criteria were analysed (table 1). Some criteria could not be used for a hospital setting.

The data extracted from Urqual software give only the first drug per prescription for each diagnosis (impossibility to extract all drugs for all prescriptions). To have every medications concerning the primary diagnosis, the prescription was then manually analysed for each diagnosis to evaluate presence of PIM/PPO. Consequently, the number of medications per prescription was not included. However, all prescriptions have been manually reviewed directly from medical files by two authors. For each targeted disorder, the prescription was analysed to detect PIMs or PPOs.

Data from the CP were obtained from the pharmacy management software OPUS (Computer PG, France). Patient’s age and drugs prescribed were collected. Current diagnosis and sex are not available in the OPUS software, so the number of patients per pathology and the number of prescriptions per pathology were lacking. Only drugs that did not require an assessment of diagnosis (eg, domperidone, metoclopramide…) were included (table 1) (28 criteria/101).

Among the five criteria including analgesics and antipyretics, only three were evaluated due to an overwhelming number of prescriptions and their association with many diseases. Pathologies analysed by POPI were the same in ED and in community. Summary of data and inclusion criteria are detailed in online supplementary appendix 1.

We consulted the software used by the ED by searching either: (1) per drug and by therapeutic class extension and (2) by main diagnosis for which a POPI item could matched. In each case,data were collected regardless of whether there was a PIM/PPO.

Statistical analysis

Data were presented as continuous variables (age, number of prescriptions by patient and number of medications per prescription) and were presented as median and IQR (25th–75th percentiles) or mean (SD), minimum and maximum depending on normal distribution.

Mixed effects logistic regression modelling for repeated measurements was applied to identify factors associated with PIM and PPO (yes/no) in the hospital and community settings. Unit of analysis was ‘the prescription’.

Univariate models were performed using different candidate factors as:

  • For model performed with hospital data: sex and age (0 days–2 years, 2–6 years, 6–12 years and 12–18 years).

  • For model performed with community data: age (0 days–2 years, 2–6 years, 6–12 years and 12–18 years) and number of medications (drugs) per prescription.

The model was constructed using the parameters of the univariate analysis, which showed at least a trend towards significance, with a cut-off of p=0.2. ORs with 95% CIs were estimated. Statistical significance was established at p<0.05. SPSS V.22 software and SAS V.9.4 were used for analysis.

Patient and public involvement

No patient and public involvement.


In the ED, 18 562 prescriptions for 15 973 patients were analysed. Around 11 500 prescriptions were reviewed manually concerning 9500 patients. Among the patients, 29% had at least two visits in 6 months. In the CP, 4780 prescriptions for 2225 patients were evaluated (figure 1). In ED and CP, 53% of patients had been issued one prescription, 21% with two and 26% with three or more prescriptions. The population’s characteristics and the frequency of pathologies were presented in table 2. Distribution of number of prescriptions by age category was described in the figure 2.

Figure 1

Flow chart indicating the course of the study. *Prescriptions with only one medical device, dietary supplement or hygiene product. ED, emergency department.

Table 2

Characteristics of the study population

Figure 2

Distribution of number of prescriptions according to age category in hospital and community settings.

In the hospital, POPI identified 541 PIMs in 2.9% of the prescriptions analysed. They were detected in 3.3% of the patients (n=530). PPOs were detected in 2.3% of prescriptions for 2.7% of patients. In the community, PIMs and PPOs represented 12.3% and 6.1% of all prescriptions, affecting 26.4% and 11.3% patients, respectively (table 3).

Table 3

Potentially inappropriate medications (PIMs) and potential prescription omissions (PPOs) identified by POPI

Table 4 presents the prevalence of PIMs (or PPOs) in the ED in patients with the targeted disorders. Patients with the targeted disorders represent the individuals who were at risk of each PIM/PPO. Table 5, however, presents the PIMs (or PPOs) as a proportion of the total number of PIMs (or PPOs) in the CP. Respiratory and digestive diseases had the highest rate of PIM in hospital and CP. For various illnesses, we removed one criterion involving medicines containing codeine because of their new contraindication in children under 12 years old.38 However, the prescription of codeine was observed in 18 cases. According to our comparison of PIMs detectable in both settings, out-of-hospital medication always presents with higher prevalence of PIMs (figure 3).

Table 4

Prevalence of PIMs and PPOs identified by POPI in hospital

Table 5

Most frequently occurring PIMs and PPOs identified by POPI in community setting

Figure 3

Comparison of PIMs detected in hospital and in outpatient care. NSAID, Non-steroidal anti-inflammatoy drug; PIMs, potentially inappropriate medicines. 

The analysis of criterion regarding the prescription of amoxicillin in milligram (IO-1) was not possible due to the fact that this drug is prescribed in great quantity. Among 100 prescriptions randomly assessed in hospital extractions, 97 prescriptions were inappropriate. Nonetheless, one analysis on acute otitis media alone identified a rate of 99.5% (807/811) of prescriptions issued without specification of the doses in milligram for oral amoxicillin. In community care, this was observed in 97% of prescriptions, in 13.2% of patients (table 5).

PIMs classed by age are presented in the figure 4. Potential factors associated with PIM or PPO are presented in online supplementary appendix 2a, b. On univariate analysis, only age was associated with risk of PIM or PPO in hospital setting. In a community setting, the number of drugs per prescription and different age categories were found to be significantly associated with a higher risk of PIM or PPO on univariate analysis. With a multivariable logistic regression model, the same results were obtained.

Figure 4

Total prescription and PIMs in both hospital and outpatient care: percentage distribution by age group. PIMs, potentially inappropriate medicines.


This is the first study to observe the prevalence of PIMs and PPOs in a paediatric population. In the literature, such tools focused on detecting PIMs/PPOs in a geriatric population.22 23 39 40 The two populations are not comparable. Respiratory and digestive pathologies are typical in children and are not so in geriatric populations, which are more concerned by cardiovascular and nervous central system diseases.22 23 41

Domperidone was frequently prescribed in a community setting, yet this drug is responsible for cardiac adverse effects such as QT prolongation. This side effect is described in the literature in adult populations and paediatric populations. The detection of this prescription will enable us to avoid cardiac risks.42–47

Prevalence of beta2 agonists or corticosteroids in an infant’s first case of bronchiolitis is 6.4% (25/386 cases), lower than that observed in a study of another French area in 2012 (41%).48–50 The use of beta2 agonists in a first case of bronchiolitis has no impact on oxygen saturation, length of hospitalisation or length of illness. They concurrently cause side effects as tachycardia, oxygen saturation and tremors.51 Implementation of guidelines has permitted to decrease beta2 agonist and corticosteroid use in a French hospital without increase morbidity.52

Unnecessary exposure to cough suppressants, pholcodine, nasal or oral decongestants was also observed frequently in this sector.53 In Norway, all drugs containing pholcodine were refused marketing authorisation in March 2007. As of this date, a decrease in sensitisation to suxamethonium used in anaesthesia and a decrease of 30%–40% cases of anaphylaxis was identified.54

Our tool enabled us to detect rare PIMs that carry heavy consequences, such as opioid use for migraines. The use of opioids for this disease induces a transition from episodic to chronic headaches and an increase of sensitivity to pain.55–57

Overuse of medication, in particular opioids, could contribute to the chronicity of headaches in 20%–30% of children and adolescents with chronic daily headaches.57

In the management of diarrhoea caused by gastroenteritis in hospitals, our study found that it was common to omit to prescribe oral rehydration solution (ORS): 14% (237/1643 cases). Even so, this rate is lower than that found in another national study in 2007 (29%).58 However, ORSs prevent hospitalisation in cases of acute gastroenteritis. In the UK, the use of ORSs has enabled a decrease from 300 deaths/year in 1970s to 25 deaths/year in 1980s.59 60 The need for ORS prescriptions was confirmed by the recommendation of the European Society for Paediatric Gastroenterology, Hepatology, and Nutrition in 2014.61

As estimated, children aged between 0 years and 12 years have the highest risk of presenting with a PIM, according to a multivariate analysis. No PIMs or PPOs were detected for patients aged less than 28 days. As we know, they are also affected by off-label drug prescriptions, which is consistent with reports from other sources.62 63 As with geriatrics, an increase in the number of medications used can be associated with PIM.23 Prescriptions issued from hospitals elicit fewer PIMs than those issued by the community. The main reason for this is that many drugs are not available in our hospital, such as cough suppressants, Rhinotrophyl (tenoate ethanolamine), domperidone and so on. This shows that many PIMs are preventable in a hospital settings. An efficient method for the prevention of PIM could be to focus on the prescribing habits of physicians and thus have an impact on the selection of drugs, thereby reducing the rate of PIM.

The data were extracted from a CP and the ED of a mother–child hospital during the winter months. The data focus on winter epidemics. An analysis of the year in its entirety would have found other PMIs/PPOs concerning different pathologies or events related to travel. While the Robert-Debré hospital offers subspecialised hospitalisation services (cardiology, nephrology, haematology and so on), the ED drains the more general activity. Likewise, the data coming from the CP provide a representative image of the paediatric prescriptions that could be found in other French pharmacies. Concerning a generalisation of our data to other countries, a study is in progress to specify which POPI items could be applicable internationally.

Our study has several limitations. First, it is a retrospective and monocentric study; the result in hospital could be underestimated. In addition, several criteria could not be analysed due to the large number of prescriptions (eg, those for fever or pain that are associated with many diseases) or absence of certain pathologies (mosquitos, lice, hyperactivity and so on). All drugs were not evaluated. Antibiotic prophylaxis, vitamin supplements, propositions for vaccination and so on can only be analysed in prospective studies. The lack of clinical information is the main limitation in detection in a community setting. This also constitutes a challenge for pharmaceutical care review in elderly patients.64 However, a certain amount of PIM was identified using POPI. Our study showed that there are many criteria that are easily identifiable and that could be detected without accessing clinical information. Moreover, community pharmacists, in their practice, can extrapolate diagnoses from their experience, from common indications or by interviewing their patients. The study presents a limitation regarding the URQUAL software, from which the number of medications per prescription could not be extracted.

This is the first study that permits an evaluation of the prevalence of PIM and PPO in paediatrics prescription. The detection of PIMs/PPOs would improve patient care and prevent hospitalisation and ADRs. A stepped wedge randomised cluster multicentre study will be conducted to prove if POPI decreases number of PIM and PPO. It is also necessary to evaluate the impact of this tool on reducing adverse drugs events, both in consultation or hospitalisation. The impact of pharmacists in providing appropriate prescriptions should be also evaluated. Subsequently, this tool may be offered to several professional societies such as the French Society for Pediatricians and the French Society of Clinical Pharmacy to make its use more widespread. The tool should be regularly updated to reflect recent events and to specify certain criteria.

To facilitate its use, this tool can be presented as a mobile app, a small handbook or installed into prescription software. In summary, we hope that POPI could be a practical option used to reduce medication errors and to improve the suitability of prescriptions. It provides rapid detection of PIM and PPO and can also open up discussion on the relationship between doctor and pharmacist to remedy the issues at hand.65


Our study was carried out in in two sectors, hospital and community, and provides a global view of PIM and PPO in paediatric patients. POPI has a clinical impact and plays a role in improving prescription quality in various sectors and patient care. POPI should be applied in different services to deepen and reinforce its utilisation. A prospective and multicentre study should be conducted to evaluate its impact and benefit in clinical practice.


We would like to thank the Emergency Department of Robert-Debré Hospital and the Albaret Pharmacy community for their data support. We would also like to thank the French Society of Clinical Pharmacy (paediatric group) for their support.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.


  • Contributors SP-L and AB-A conceptualised and designed the study, drafted the initial manuscript and approved the final manuscript as submitted. RB and PKHN carried out analyses, reviewed and revised the manuscript and approved the final manuscript as submitted. XB, TW and OB reviewed and revised the manuscript and approved the final manuscript as submitted. FA and PA supplied data from hospital and community pharmacy and reviewed and revised the manuscript and approved the final manuscript as submitted.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent Not required.

  • Ethics approval This project was approved by the local research ethics committee (n°2015/218).

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement We have no additional unpublished data.