Tens of millions of mistakes throughout the process of providing medication in England result in avoidable costs to the NHS of upwards of £98 million and at least 1700 lost lives, every year, according to a new peer-reviewed study, published in the British Medical Journal.
Rachel Elliott, Professor of Health Economics at the University of Manchester, who co-authored the report Economic analysis of the prevalence and clinical and economic burden of medication error in England, explained to Sputnik why the system of prescribing and delivering medication is what needs to be addressed rather than simply blaming individual healthcare workers.
Sputnik: What are the increasingly complex healthcare needs that appear to be fuelling medication error within the England?
Professor Rachel Elliott: Medication error is a global problem, not confined to England or Europe. High quality data recording systems in the NHS in England mean that we have a good idea of the level of some types of medication errors.
Increasingly complex medical needs, and the introduction of many new medications, have resulted in Adverse Drug Events being recognised as a key global issue. This has led to the World Health Organization’s (WHO) Third Global Patient Safety Challenge: Medication Without Harm. It aims to reduce the global level of severe, avoidable harm related to medications by 50% between 2017 and 2022.
Increasingly complex needs are mainly due to the increased prevalence of chronic illness managed by medicines (hypertension, ischaemic heart disease, asthma and COPD, type 2 diabetes, different type of arthritis, mental health conditions, chronic pain, dementia) in an ageing population, with increased prevalence of multimorbidity, such that many people are taking multiple medicines for multiple conditions.
Sputnik: How exactly did you determine whether a medical error was “minor”, “moderate” or had the potential to cause “serious harm”?
Professor Rachel Elliott: Linking errors to harm requires information about which errors persist through the medication use process, and the impact on patients and healthcare utilisation. Studies identified through our rapid systematic review of recent UK-based studies found very little good quality data that linked harm to errors. This is partly due to challenges in following up patients from error to harm, and attributing harm to errors.
To deal with this evidence gap, studies have ranked errors by subjective judgement of potential of errors to cause harm, some using expert panel-derived criteria to divide errors into “minor”, “moderate”, or “severe”. This approach does not allow estimation of harm that has occurred but can help to understand what errors could lead to the most severe harm. Five of the studies we used to estimate error prevalence assessed the proportion of errors with potential to cause minor, moderate or severe harm.
We used the data from multiple studies, based on their judgement of whether an error could cause “minor”, “moderate”, or “severe” harm. See below for an illustration:
| Study | Method for determining the severity of errors | Examples of mild errors | Examples of moderate errors | Examples of severe errors |
| PRACTICE | A panel of 5 judges assessed each error using a visual analogue scale from zero to 10, then classified errors with scores 0-2 as minor, 3-7 as moderate, 8-10 as severe | 1-year-old girl prescribed amoxicillin 123mg/ml suspension twice during the same consultation. One was for 2.5ml TDS for one week, and the other for 5ml for one week | 64-year-old patient was prescribed ibuprofen 400mg to be taken three times daily after a road accident. No concomitant medication was prescribed for gastric protection. Patient also on aspirin for peripheral vascular disease. | 62-year-old patient with documented allergy to penicillin; prescribed a course of oral flucloxacillin. |
Sputnik: Why do you think care homes have the highest prevalence of medication error compared to the other health care sectors such as hospitals?
Professor Rachel Elliott: The prescribing error rate is similar to that in hospitals. People in care homes are generally an older, frail population, with a much increased prevalence of multimorbidity, such that many people are taking multiple medicines for multiple chronic and acute conditions. Therefore, there is more opportunity for error than in community dwelling primary care patients.
Sputnik: Your study says that “Error rates are lowest in primary care, but because of the sector's size, these account for nearly 4 out of every 10”. Can you explain this further?
Professor Rachel Elliott: The error rates are lowest, but the vast majority of prescribing activity takes place in primary care. In our supplementary appendix, we reported that 1,104 million items were dispensed in primary care in 2016.7 Therefore, this means that even a low error rate translates into larger total numbers of errors.
Sputnik: What are the limits in terms of how you have conducted your study?
Professor Rachel Elliott: As we note in our report there are four key limitations on the estimates of burden of errors:
(1) the assumption that avoidable Adverse Drug Reaction/Adverse Drug Event correspond to medication errors,
(2) generalisability of the source studies to the NHS,
(3) lack of primary data to inform estimates, and
(4) assumptions about the valuation of healthcare resource use associated with errors.
Sputnik: Is there a stage in the process, ie point of administration, prescription or dispensing, which is easiest to tackle medication error, and why?
Professor Rachel Elliott: I cannot state that there is an easiest point in the process to tackle error. It is necessary to know where the errors occur and why they occur if we want to tackle error. There are many causes of error, some individual, and local, some organisational.
Sputnik: What needs to be done in order to reduce medication error?
Professor Rachel Elliott: Three key things:
(1) We need to facilitate routine data collection of clinically important errors and link them to outcome data to allow identification of priority areas for targeting interventions. The UK is a world leader in research in this area, and also has extensive high-quality primary care and secondary care data sources, that if harnessed and integrated, could already be used to record errors in some parts of the medication use process, particularly prescribing and monitoring, and link them to outcomes.
(2) Interventions to identify and reduce medication errors need to be designed and implemented thoughtfully given that the medicines use process is such a significant part of day-to-day work. One approach that is popular is Clinical decision support systems which consist of point-of-care alerts to clinicians that relates to prescribing that may be hazardous or inappropriate.
More complex interventions such as the PINCER intervention, an education and outreach intervention combined with practice-level error report generation, which consisted of pharmacists working with prescribers, have been shown to be effective and cost-effective, but can be costly and time consuming so need to be applied thoughtfully.
(3) Another important development that is needed is to change the culture around reporting errors. We know that current self-reporting systems are thought to detect only 7-15% of all incidents including medication errors.
It is clear from the work carried out in this report that medication errors are a system failure, so the third action should focus on changing cultures to remove personal blame, which will improve self-reporting figures, and allow systems to be improved.
