Article Text
PDF
Abstract
Background and aims Unlicensed and off label drug use in children is common and leads to well-recognised problems. This study aimed to determine whether a relationship exists between medication errors and licence status.
Methods Reports of errors in a UK children's hospital from 2004 to 2006 were analysed in terms of licence status and degree of harm and compared to the incidence of unlicensed and off label drug use in the hospital.
Results 20 of 158 (13%) errors were considered to have caused moderate harm and 12 of these involved unlicensed/off label drugs. 138 (87%) caused no or low harm. None caused severe harm. Unlicensed drug usage was significantly more likely to be associated with errors than licensed use in both children and neonates.
Conclusion Unlicensed drug use appears to be associated with medication errors in neonates and children. Medication errors causing moderate harm were significantly more likely to be associated with both unlicensed and off label than licensed drugs.
Statistics from Altmetric.com
Background
Use of unlicensed and off label drugs in children is common and leads to well-recognised problems, for example lack of suitable formulations and adequate prescribing information and adverse drug reactions.1,–,3 The British National Formulary for Children has brought improvements and is intended to provide adequate prescribing information for neonatal and paediatric drugs regardless of licence status, however, errors continue to happen. This study aimed to determine whether there is a relationship between medication errors and drug licence status.
Methods
Prescribing in the hospital studied involves handwriting prescriptions which also form the drug administration record and are the vehicle for dispensing/ordering of drugs. Staff detecting medication errors are required to report them. Errors are reviewed by a multidisciplinary team aiming to reduce the risk of recurrence. All reports from January 2004 to December 2006 were reviewed and classified in terms of the licence status of drugs involved using methodology previously described.1 Drugs prepared extemporaneously or purchased as ‘specials’ were unlicensed. Drugs used outside the licence in terms of patient age, dose, route or indication were off label. Each incident was classified according to the degree of harm caused using UK National Patient Safety Agency (NPSA) definitions.4 Neonatal and paediatric errors were analysed separately and compared to the incidence of unlicensed and off label drug use in the hospital, which was known from previous work.2 3 The paediatric data were collected from a ward treating general paediatric medical and acute surgical patients and the neonatal data from a unit offering all levels of care except surgery, extracorporeal membrane oxygenation and nitric oxide therapy. Between-group differences were analysed by χ2 tests (proportions) and the standard errors of the ORs were obtained using Cornfield approximation.
Results
One hundred and fifty-eight reports were analysed. One hundred and thirty-one reports had been excluded as they were duplicates, contained insufficient information or mentioned supply problems only.
Table 1 compares the licence status of drugs involved in medication errors and those used generally in the hospital. It is estimated that 7% of all prescriptions written on paediatric wards are unlicensed,3 whereas medication errors were associated with 17% of unlicensed prescriptions. Unlicensed drugs were significantly more likely to be associated with errors than licensed drugs (OR 2.24 (95% CI 1.31 to 3.87), p=0.003).
Medication errors and licence status
In the neonatal unit the unlicensed drug error rate was 38%, whereas unlicensed drug use in neonates is estimated to be 10%.2 Unlicensed drug usage in neonates appeared significantly more likely to be associated with errors compared to licensed use (OR 5.81 (95% CI 2.32 to 14.55), p=0.0001).
There were 55 prescribing errors, including eight unlicensed (15%) and four (7%) off label prescriptions. Incorrect doses were most common. Five 10-fold errors occurred involving four licensed medicines (intravenous fluids, intravenous hydrocortisone, oral furosemide and subcutaneous insulin) and one unlicensed preparation (oral morphine). Twenty-two dispensing errors involved nine (41%) unlicensed and five (23%) off label medicines. Six of the unlicensed drug errors were labelling errors in extemporaneously prepared medicines. Eighty-one administration errors included 17 unlicensed (21%) and 17 (21%) off label prescriptions. Thirty-seven administration errors involved drugs not being given/given late (19 licensed, 9 unlicensed, 9 off label).
Most errors (138, 87%) caused no or low harm. Twenty (13%), however, caused moderate harm and 12 (60%) of these involved unlicensed and off label drugs (table 2). None caused severe harm or death. Medication errors causing moderate harm were significantly more likely to be associated with unlicensed and off label drugs than licensed drugs (OR 2.81 (95% CI 1.1 to 7.17), p=0.03).
Degree of harm caused by errors (definitions adapted from the NPSA4)
Discussion
This study found that unlicensed drug use in both children and neonates appears to be significantly associated with medication errors. Both unlicensed and off label drug use were significantly associated with medication errors causing moderate harm. With 30–65% of prescriptions for children and neonates in hospital being for unlicensed or off label drugs, this is of great concern.2 3
Dispensing errors seemed most likely to be associated with unlicensed drugs, mainly involving labelling/packaging errors. This may be expected since for these products labels often have to be custom produced to include details of individual product contents/strength, etc, whereas for licensed products these are available in a computerised list.
Administration errors also commonly involved unlicensed and off label drugs, many of which were drugs being given late or not given. The reasons for this are unclear. Possibilities include special storage conditions required by some products, for example refrigeration causing delays in nurses finding them, busy staff or short staffed shifts. Nine neonatal administration errors involved off label gentamicin doses prescribed 24, 36 or 48 hourly and given early, late, missed, to the wrong patient or at an incorrect frequency. This has been recognised as a national problem and is the subject of a NPSA alert advising steps to reduce this risk.5
Prescribing errors in the main involved licensed drugs with a smaller number of unlicensed and off label errors, most of which involved unlicensed oral liquid preparations prepared extemporaneously or purchased as ‘specials’. One error resulted in a patient being readmitted to hospital with increased seizures after a general practitioner prescribed a more dilute strength of phenobarbital liquid, an unlicensed product available in many strengths. Unlicensed liquid preparations are made in different strengths using different formulations. This may result in problems when communication processes are not effective between tertiary, secondary and primary care sectors. Problems at the interfaces of care have been highlighted by the NPSA despite the overall lack of reporting of errors from primary care.4
While off label drug use did not appear directly associated with an increased risk of errors and indeed appeared less implicated than licensed medicines in some cases, it was involved in 35% of errors causing moderate harm. The potential risks of off label use therefore should not be dismissed. It must also be remembered that 98 (62%) medication errors including four out of five 10-fold errors involved licensed drugs.
It is acknowledged that the comparator data were collected earlier than the study data, however, the proportions of licensed and unlicensed/off label drugs suggest that drug usage was very similar during the two time periods. The main limiting factor was that the study relied on staff reporting of errors, resulting in low numbers of reported errors. Information on exact numbers of drugs prescribed, dispensed and administered over the relevant time periods would have put the figures into context, however systems for such data collection are unavailable in most UK hospitals. Prescription numbers are however estimated to be of the order of 110 000 during the period studied making the reported error rate around 0.1% of prescriptions. The data analysed were obtained from a general paediatric ward and a neonatal intensive care unit typical of many in the UK and therefore the findings described are likely to be applicable to many UK hospitals caring for children.
In conclusion, unlicensed drug use appears to be associated with an increased risk of medication errors in children and particularly in neonates. It appears to be particularly associated with dispensing errors but is also prevalent in prescribing and administration errors. Off label drug use does not appear to increase overall risk, although it is prominent in errors causing moderate harm. Medication errors causing moderate harm were significantly more likely to be associated with unlicensed and off label than licensed drugs.
Footnotes
-
Funding The University of Nottingham funded this study.
-
Competing interests None.
-
Provenance and peer review Not commissioned; externally peer reviewed.