Prescribing errors are the most common type of medical errors and can result in harm particularly in young children. Doctors were enrolled in a programme of written assessment in prescribing skills and individualized feedback. Pharmacists audited the impact. The setting was the paediatric wards and neonatal unit of a District General Hospital. 16 doctors were tested and received feedback. A total of 110 errors were identified in this test, out of a 51 were classified as major including wrong dose and frequency, and prescribing medication the patient had an allergy to. Audit of impact of this intervention revealed a reduction of errors from 47 to 21, and patients affected from 19 to 11 per 100 (
Medication errors are considered to be the most common type of medical error, and have been defined by the UK Department of Health and US National Coordinating Council for Medication Error and Prevention as “…any preventable event that may cause or lead to inappropriate medication use or patient harm, while the medication is in the control of the health care professional, patient, or consumer.” [
Objectives of our project were to reduce prescribing errors putting paediatric patients at risk of harm; to reduce the risk associated with lack of necessary information for safe transcribing; to reduce the risk associated with lack of acting on prescriptions with errors.
A baseline audit was undertaken on all prescribing errors noted on daily (week days) review of all drug charts by pharmacists on their routine visits of the paediatric wards and the neonatal unit for about a month in April and May 2008. All drug charts with incorrect prescribing including nonadherence to guidance given in the British National Formulary for children [
Prescribing skills were subsequently assessed by asking paediatric trainee doctors to complete 5 tasks relating to prescribing as part of the induction programme on commencement of their post at Luton and Dunstable Hospital NHS Foundation Trust, United Kingdom. The task comprised of transcribing a drug chart containing deliberate errors, two scenarios requiring prescription, and two tasks with instructions for prescription of intravenous drugs (see Appendix
Two months after this assessment and feedback, a further in-depth audit of prescribing errors was performed in November 2008 using the same approach as the baseline audit.
Statistical comparison between prescribing errors of audit and re-audit was by reference to the total number of emergency admissions and chi-square test (with Yates correction if sample size was <30) or Fisher’s exact test (if sample size was less than 5) with the assumption that major errors in the majority of cases did not occur more than once per patient and with regards to comparison of dose errors; when results of test and retest were compared, more than once per task. Median error rates were compared using the Mann-Whitney test. A
A total of 16 junior and middle grade doctors underwent assessment and feedback. These constituted the majority (about 80%) of prescribing doctors on paediatric and neonatal units, who change 3 to 6 monthly on training rotations.
Compared to middle grade doctors (more than 2 years experience in pediatrics), junior doctors had overall more errors in prescribing but the same amount of errors concerning prescribing medication the patient was allergic to (see Table
Result of assessment of prescribing of paediatric doctors.
Middle grade ( | Junior grade ( | |
---|---|---|
Total number of errors1 | 45 | 65 |
Major errors2 | 10 | 27 |
Potentially fatal errors relating to allergy | 7 | 7 |
1Errors affecting dose and frequency, transcribing errors regarding duration, wrong application, missing dose, unit, timing or signature, illegible writing, wrong spelling.
2Defined as prescribing a drug the patient is allergic to, prescription of the wrong dose or unit, guessing the dose a patient has been on, not prescribing a required drug, and choice of the wrong frequency for application of a drug.
For potentially fatal transcribing errors, it was noted that despite transferring the penicillin allergy information onto the new drug chart 11/16 candidates prescribed co-amoxiclav by reproduction from the template. Doctors who did not commit this type of prescribing error had significantly less other errors than doctors who did (median (range) of 3 (0–5) versus 6 (4–9),
Investigation into the effectiveness of feedback on prescribing errors regarding intravenous drugs used on NICU revealed that, in the two test items requiring calculation of doses for intravenous infusions, in 11 trainees who required retesting, the error rate did not decrease significantly from 10/22 to 8/22 tasks completed
There was an overall reduction of prescribing errors and number of patients affected but no significant reduction of major drug errors (see Table
Results of audits of all drug charts with errors in prescribing before and after assessment of prescribing skills with feedback in September 2008.
Audit 21.04.2008 to 16.05.2008 (number of emergency admissions | Reaudit 01.11.2008 to 30.11.2008 (number of emergency admissions | |
---|---|---|
Total number of errors (per 100 emergency admissions) | 188 (47) | 120 (21)1 |
Number of patients with drug errors on their charts (per 100 emergency admissions) | 79 (19) | 67 (11)2 |
Major errors3 (per 100 emergency admissions) | 36 (8) | 35(6)4 |
1Statistical comparison not possible as variable number of errors per admission.
2Chi-square test: Chi-square 10.77,
3Defined as prescribing a drug the patient is allergic to, prescription of the wrong dose or unit, guessing the dose a patient has been on, not prescribing a required drug, and choice of the wrong frequency for application of a drug.
4Chi-square test: chi-square 2.53,
In the general paediatric patients, comparison of major drug errors before (28, 7% of emergency admissions) and after (24, 4% of emergency admissions) the intervention revealed a nonsignificant (
In the category of major errors, there was no significant reduction on the neonatal intensive care unit, where there were 11 versus 8 before the intervention indicating a nonsignificant increase from 30 to 41 per hundred emergency admissions to NICU. Other error categories showed significant improvement and included inappropriately prescribing the dose as volume, which occurred in 7 charts of the first and 3 of the second audit. Omissions of dose, unit, frequency, time of administrations, start date, and allergy information occurred in 108 in April/May (26 per 100 emergency admissions) and in 52 in November (9 per 100 emergency admissions), a marked reduction not assessable by statistical analysis because of variable number of errors per patient. The allergy section on the drug chart was not completed in 8 patients of the baseline and 1 patient in the re-audit, which was a significant reduction related to number of emergency admissions (
Both objectives of a reduction of the number of errors and improved prescribing practice by improvement of information given in the prescription were achieved. We demonstrated an improvement of prescribing practice by less omissions and more attention to information on allergies by introduction of a module of assessment and feedback by a pharmacist. Our rate of 47 errors per 100 emergency admissions noted at the baseline audit was similar to the rate of 55 per hundred admissions reported previously [
A major concern is that the majority of junior and middle grade doctors transcribed a potentially fatal error of the prescription of co-amoxiclav in a patient they recognized (by putting penicillin in the allergy section) as allergic to penicillin. Not committing this transcription error was associated with a marked reduction in other major errors making this error a marker for decreased ability to reflect on the prescribing process. The significant risk associated with the tendency to reproduce errors during transcribing has been recognized previously, and an assessment which included tasks of transcribing drugs, which were dosed to high, revealed that 60.3% of inappropriately high doses were missed and the error reproduced with only 4/34 participants able to detect all inappropriate doses [
Future research needs to determine whether prolonged training in calculation of drug doses and critical transcribing can avoid prescribing errors. This should be integrated in training at the undergraduate and postgraduate level and form part of examinations to provide incentive for enhancement of specific skills in this area in doctors.
An intervention combining a comprehensive multifaceted assessment and detailed pharmacist-led feedback can lead to a reduction of prescribing errors in paediatric patients. A single intervention to improve prescribing is not sufficient to address calculation errors specific to preparation of intravenously administered drugs.
Instructions: Please complete drug charts with the tasks set and with the help of a BNF for children. If a task cannot be completed with the information given please A drug chart is given (attached) and the information needs to be transferred to a new drug chart. A 6-year-old child (weight 23 kg) is admitted with respiratory distress and generalized expiratory wheeze and chest pain not responding to paracetamol. Past history reveals that the child is asthmatic and responded previously with bronchospasm to application of nonsteroidal anti-inflammatory drugs. Please fill in a prescription chart with medication addressing the problems (prescribe steroid, bronchodilator as nebulizer, and pain relief). A 10-year-old child (weight 25 kg) is admitted with inability to swallow because of a peritonsillar abscess. Past medical history reveals that the child is HIV positive, on antiretroviral treatment and has a history of anaphylactic shock in response to ingestion of co-amoxiclav. Please issue a prescription chart with medication addressing the problems. Please prescribe a morphine infusion for a 25-week gestation baby weighing 600 g on the drug chart provided. Dose to be administered is 10 to 20 microgrammes per kg per hour. The nursing staff will use a syringe pump and the drug should be given in a 25 mL volume and 10 microgrammes per kg per hour should be given as 0.1 mL/hour. Please write a fluid prescription chart for a 1.5 kg, 2-day-old baby, who is on 120 mL/kg/day total daily fluids. The baby is on a morphine infusion running at 0.2 mL/hour and a dopamine infusion at 0.2 mL/hour. The baby needs 10% dextrose and 3 mmol/kg/day sodium and 2 mmol/kg/day potassium. The electrolytes are normal.
Instructions: Please complete drug charts with the tasks set and with the help of a BNF for children. If a task cannot be completed with the information given, please A parent gives you the following information about his 11-year-old 20 kg child’s medication: Baclofen one tablet three times daily, Tegretol 2 tablets three times a day. Information given by parents contains only drug names and not doses. What would you do to arrive at an accurate prescription (Give three examples)? Please prescribe a prostaglandin E2 infusion (dinoprostone) to maintain a duct patency in a 2-kg baby with congenital heart disease. Dose is 10 nanogrammes per kg per minute. Dilute in 25 mL normal saline and give at a rate of 2 mL per hour. Baby H is a 2-day-old 28-week gestation baby. He weighs 800 g. He is hypotensive and needs inotropes. Please prescribe a dopamine infusion dose 10 microgrammes per kg per minute at a rate of 0.2 mL/hour in a total volume of 25 mL of 5% dextrose. A parent gives you his/her 6-year-old child’s medication doses and you are supposed to prescribe the medication on a drug chart as the child has become an inpatient: iron supplement: 5 mL syrup once a day, paracetamol oral suspension 15 mL four times a day, piriton 50 mL four times a day.
The project did not require ethical approval or consent as it fulfilled the criteria for clinical audit set by the National Research Ethics Service of the National Patient Safety Agency of the United Kingdom for clinical audit including design and conduct to produce information to inform delivery of best care and evaluation of service delivery against a standard [