Medication Administration Errors in Hospitals — Challenges and Recommendations for Their Measurement

Medication errors are a threat to patient safety. Those that result in patient harm occur in an estimated 1% to 2% of hospital inpatients (1,2) and contribute to an increased hospital stay of 4.6 to 10.3 days for each affected patient. (3–5) While errors may arise at any stage of the medication use process (prescribing, dispensing, administering and monitoring), research suggests prescribing and administration errors account for the largest percentage of all (39% and 38%, respectively). (6) However, medication administration errors (MAEs) are least likely to be intercepted before they reach the patient. (1,6) This is partly due to the narrow window of opportunity for detecting a MAE, which makes studying MAEs and developing suitable intervention strategies particularly problematic.

Since the publication of key reports worldwide, (7–9) several attempts have been made to adapt strategies from high-risk industries, such as aviation, to analyse and reduce risk in healthcare. (10,11) However, there are a number of key differences between such industries and healthcare. (7) First, front-line staff in high-risk industries are usually directly affected when an accident happens, while in the healthcare setting it is typically someone else, i.e., the patient, who is affected. Second, preventable harm in healthcare generally occurs to one patient at a time, rather than groups of patients, making incidents less visible at the organisational level unless a patient suffers severe harm. For MAEs, the problem of visibility is made more challenging by the difficulties of measuring and reporting MAEs in practice. (12,13) In this commentary, we highlight some of the main challenges associated with measuring MAE rates and make suggestions for the development of more practical proxy measures of MAE rates for use in everyday practice by healthcare professionals.

Challenges in Measuring MAE Rates

The first challenge is determining what to measure. In published research studies, MAEs are frequently defined as "a deviation from the physician's medication order as written on the patient's chart" (13,14); however, a number of important methodological differences exist among quantitative studies of MAEs. (13,14) Our recent systematic literature review explored these and their effect on reported MAE rates which revealed no single standard for determining MAE rates, even within one country. (14) Overall, we identified three MAE definitions, 44 MAE subcategories, and four denominators from 16 direct observation studies. The use of different denominators and MAE rate calculations not only influenced reported MAE rates, but made it difficult to interpret the literature surrounding the problem of MAEs. Furthermore, studies of MAE rates do not always publish sufficient information to interpret the findings. (14) For example, in the UK, MAEs occur in an estimated 5.6% of non-intravenous doses and 35% of intravenous doses, with intravenous doses about five times more likely to have an MAE. (14) If studies do not report the inclusion of intravenous doses, the conclusions drawn from the presented results can be severely limited. Other key information that should be reported includes the MAE definition used, inclusion and exclusion criteria, whether timing errors were included, and how many MAEs were possible for each dose.

Based on our recent literature review, we made a number of recommendations to guide future reporting of quantitative studies of MAEs. (14) For example, we suggest reporting the number of intravenous doses if intravenous doses were included, and whether or not 'when required' and/or 'once-only' medication orders were included. We based these recommendations on evaluating UK studies but believe the principles can be applied to other countries to help increase the interpretability of study findings and facilitate comparisons between studies.

The next challenge is how to measure. Within healthcare practice, incident report data are often used to assess medication error rates within an organisation. However, numerous reasons exist why MAEs may not ultimately be reported. Incident data requires someone to be aware that an error has occurred, know how to and be willing to report it, and then actually do so. For these reasons, it is estimated that only a very small proportion of MAEs are collected in organisational incident reporting systems. (15,16) We would therefore view incident report data as examples of errors that have occurred, rather than making any quantitative inferences about the underlying error rates.

Within the research field, measurement of MAE rates mainly involves direct observation of medication administration, usually by a pharmacist or a nurse. This practice allows for the detection of significantly more MAEs than incident reports or chart review (16, 17) and is generally considered to be the gold standard method. (13,17,18) However, like any method, it has a number of limitations. Firstly, it is possible that observation of an individual may induce behaviour change that could affect the occurrence of MAEs. Yet, previous research suggests that using a discreet and non-judgemental approach did not significantly affect MAE rates. (12,17) Second, observation is resource-intensive and likely to be impractical for routine regular monitoring. While in some clinical areas the majority of doses will be provided at four or five routinely scheduled times, in practice many are administered throughout the day. Efforts to maximise the observation of medication administration over long periods of time or its inclusion at all times of the day may increase the risk of observer-fatigue and possibly influence detected MAEs. Is there a more practical approach to measuring MAEs, one that is more quantitative than incident report data, but less resource intensive than extended periods of observation?

Proxy Measures of Patient Safety Associated with Medication Administration

In patient safety terms, MAEs are a proxy measure for actual patient harm; depending on the definitions used, research suggests that 0.6% to 21% of MAEs may lead to patient harm that could either result in prolonged hospitalisation or be considered potentially life-threatening. (19) However, MAEs are just one type of proxy measure. It has been suggested that approaches for identifying measures for improvement need to be both reactive and proactive. (20) Based on the concept of institutional resilience in healthcare, reactive measures, such as MAEs (proxy measure of harm) and adverse drug events (actual measure of harm), provide important information about incidents that have occurred in the past so that lessons may be learnt. By contrast, proactive proxy measures act as an early warning indicator of potential problems and underlying conditions that may contribute to future incidents. While there is clearly a place for using direct observation to study MAEs, more practical methods to regularly monitor the quality and safety of the medication administration process are required.

Proactive measures involve regular monitoring of the essential processes and defences (20) which relate to the prevention or amelioration of risk from potential and actual patient harm. This approach therefore requires an understanding of the conditions and specific factors associated with future incidents that occur locally. Acquiring and understanding the context in which incidents occur is vital, and thus identifying appropriate proxy measures requires local knowledge and expertise. Based on our clinical and research experience, we suggest exploring measures such as duration of scheduled drug administration rounds and number of pedometer steps as potential proactive measures of patient safety associated with medication administration. To determine whether or not such measures and other potential proxy measures of patient safety for medication administration might be useful in practice, we recommend the consideration of the following factors:

1. Evidence of a relationship between the proxy measure and the quality or safety of medication administration
2. Time required to collect data
3. Training required to collect data
4. Time to report the data
5. Reports that are intuitive and facilitate interpretation and analysis

The data could be analysed using approaches such as statistical process control (SPC), which has been used with demonstrable benefits in healthcare. (21,22) Benneyan et al (21) provides a useful overview of SPC and its application in healthcare. Briefly, the SPC approach would facilitate interpretation of what would otherwise be potentially crude measures, such as duration of drug administration rounds, by factoring in the inherent day-to-day variations associated with drug administrations. Consequently, this method enables the identification of potential problems when the parameter being measured falls over or under the expected limits of variation.

The increasing use of computerised prescriber order entry and electronic medication administration systems provides an opportunity to incorporate regular monitoring (using SPC or other approaches) of other patient safety measures that are less practical to assess with paper systems; for example, some specialized health information systems automatically flag and report timeliness of time-critical dose administrations and dose omissions. Given the challenges of measuring MAEs in healthcare, a medication administration 'dashboard,' which shows the results from routine proxy measures monitored in a way that is easily accessible and viewable by relevant staff, may offer healthcare professionals a better way to 'see' medication administration problems in a more timely manner, thus enabling action to be taken to reduce the risk of MAEs.

Conclusions

Patient safety is a global priority, and MAEs remain an important proxy measure of patient harm. However, a number of challenges associated with MAE measurement exist. In this commentary, we summarise what we believe are the main challenges in understanding and interpreting published MAE rates. Additionally, by applying concepts from institutional resilience and SPC to the assessment of medication administration, we make suggestions towards future research and the development of more practical proxy measures for routine monitoring of the safety of medication administration. More collaborative work between researchers and healthcare service providers is now required to translate such theory into the practice of increasing patient safety.

Authors

Monsey McLeod, MPharm, MSc, PhD
Centre for Medication Safety and Service Quality, Imperial College Healthcare NHS Trust and The UCL School of Pharmacy, London, UK

Nick Barber, BPharm, PhD
Centre for Medication Safety and Service Quality, Imperial College Healthcare NHS Trust and The UCL School of Pharmacy, London, UK

Bryony Dean Franklin, BPharm, PhD
Centre for Medication Safety and Service Quality, Imperial College Healthcare NHS Trust and The UCL School of Pharmacy, London, UK

Disclaimer

The views and opinions expressed are those of the author and do not necessarily state or reflect those of the National Quality Measures Clearinghouse™ (NQMC), the Agency for Healthcare Research and Quality (AHRQ), or its contractor ECRI Institute.

Potential Conflicts of Interest

Dr. McLeod and Professor Barber state no personal or family financial, business or professional conflicts of interest with respect to this expert commentary.

Professor Franklin states no personal financial or family financial/other conflict of interest with respect to this expert commentary. Professor Franklin reports the following business/professional interest: expert advisor for the Royal Pharmaceutical Society of Great Britain.

References

1. Bates DW, Cullen DJ, Laird N, Petersen LA, Small SD, Servi D, et al. Incidence of adverse drug events and potential adverse drug events. Implications for prevention. JAMA. 1995 Jul 5;274(1):29–34.
2. Neale G, Woloshynowych M, Vincent C. Exploring the causes of adverse events in NHS hospital practice. J R Soc Med. 2001;322–30.
3. Bates D, Spell N, Cullen D, Burdick E, Laird N, Petersen L, et al. The costs of adverse drug events in hospitalized patients. JAMA. 1997;277(4):307–11.
4. Vincent C, Neale G, Woloshynowych M. Adverse events in British hospitals: preliminary retrospective record review. BMJ. 2001 Mar 3;322:517–9.
5. Pinilla J, Murillo C, Carrasco G, Humet C. Case-control analysis of the financial cost of medication errors in hospitalized patients. Eur J Health Econ. 2006 Mar;7(1):66–71.
6. Leape L, Bates DW, Cullen DJ, Cooper J, Demonaco HJ, Gallivan T, et al. Systems analysis of adverse drug events. JAMA. 1995 Jul 5;274(1):35–43.
7. Kohn L, Corrigan J, Donaldson M. To err is human: building a safer health system. Washington (DC): National Academy Press; 1999.
8. Department of Health. An organisation with a memory. London: The Stationary Office; 2000.
9. Australian Council for Safety and Quality in Health Care. Safety and quality council second national report on patient safety. Canberra; 2002.
10. Reason J. Understanding adverse events: human factors. Qual Health Care. 1995 Jun 1;4(2):80–9.
11. Vincent C, Taylor-Adams S, Stanhope N. Framework for analysing risk and safety in clinical medicine. BMJ. 1998;316(April):1154–7.
12. Barker K, McConnell WE. The problems of detecting errors in hospitals. Am J Hosp Pharm. 1962;19:361–9.
13. Allan EL, Barker KN. Fundamentals of medication error research. Am J Hosp Pharm. 1990 Mar;47(3):555–71.
14. McLeod MC, Barber N, Franklin B. Methodological variations and their effects on reported medication administration error rates. BMJ Qual Saf. 2013 Jan 15;22(4):278–89.
15. Barker K, Flynn E, Pepper G, Bates D, Mikeal RL. Medication errors observed in 36 health care facilities. Arch Intern Med. 2002 Sep 9;162(16):1897–903.
16. Ghaleb MA, Barber N, Franklin BD, Wong IC. The incidence and nature of prescribing and medication administration errors in paediatric inpatients. Arch Dis Child. 2010 Feb;95(2):113–8.
17. Dean B, Barber N. Validity and reliability of observational methods for studying medication administration errors. Am J Health Syst Pharm. 2001;58:54–9.
18. Ferner RE. The epidemiology of medication errors: the methodological difficulties. Br J Clin Pharmacol. 2009;67(6):614–20.
19. McLeod M. Medication administration processes and systems - exploring the effects of systems-based variation on the safety of medication administration in the UK National Health Service. University College London School of Pharmacy; 2013.
20. Carthey J, de Leval MR, Reason J. Institutional resilience in healthcare systems. Qual Health Care. 2001 Mar 1;10(1):29–32.
21. Benneyan JC. Statistical process control as a tool for research and healthcare improvement. Qual Saf Health Care. 2003 Dec 1;12(6):458–64.
22. Thor J, Lundberg J, Ask J, Olsson J, Carli C, Härenstam KP, et al. Application of statistical process control in healthcare improvement: systematic review. Qual Saf Health Care. 2007 Oct;16(5):387–99.