Background
The care home sector is an increasingly important provider of long-term care for older people. A review of the international literature has recently identified that research in the area of quality and safety is lacking, especially for residential homes which have no on-site nursing staff [
1]. A number of authors have identified prescribing and management of medication more generally as key areas for improved patient safety in care homes [
2‐
9]. In England, over 18,000 homes currently provide beds for more than 453,000 people, compared to 167,000 beds in hospitals. The majority of residents are older people with complex health needs. Six out of ten are cared for in a residential home with no on-site nurses. In such homes the management of prescribed medication is undertaken by non-nursing, social care staff who may have had no formal training in safe practice [
10]. In nursing homes, which must have a registered nurse (RN) on site 24 hours per day to meet regulation requirements, medicine administration is one of the many tasks carried out by busy RNs. In both settings, prescribing decisions are the responsibility of the general practitioner (GP) or the hospital physician.
It is known that in England 45% of all care homes in 2005 failed to meet the minimum standard for medication management [
4], and that this figure remains high at 28% in 2010 [
11]. A cross-sectional study of a sample of 256 residents in 55 UK care homes found that 69.5% had been exposed to one or more medication errors [
12]; these included mistakes made by GPs in prescribing, dispensing errors by pharmacies, and administration errors made by care home staff.
To guard against drug administration errors in hospital care settings, electronic medication administration recording (eMAR) has been widely implemented to replace paper-based systems [
13]. eMAR systems have now been developed for use in long-term residential care environments. It is reported that safety is now being further improved in hospitals by the use of barcode technology integrated with eMAR systems [
14]. Similar systems are being developed for use in long-term residential care.
We report on a study of the first barcode medication management system specifically developed for use in UK residential and nursing homes, with external pharmacy-led data capture, processing and record management. The main aim of the research was to examine the incidence of potential medication administration errors (MAEs) in nursing and residential homes using the barcode medication administration (BCMA) system. A further objective was to compare observed error rates and response to system-alerts for residential and nursing homes. Other aspects of the system such as bar-coded dispensing, clinical readings, and stock management were not considered in the present study.
Discussion
Medication management covers the whole process from prescribing, through to dispensing and finally administration of medicines. Errors in any one of these steps can have serious consequences for the patient. Although such errors are acknowledged to be preventable [
10], currently they still result in considerable morbidity, mortality and healthcare utilisation by older people [
2‐
6]. According to the United States (US) Food and Drug Administration, over 770,000 patients are injured annually because of medication errors [
18]. Administration errors account for 38% of these events. In the US, it is reported that up to 35% of older people in the community may experience some form of adverse medication event each year [
19]. The incidence is thought to be even higher among nursing home residents [
20]. In Italy, up to 30% of hospital admissions in older people are related to such events [
21]. In the UK, 9% of hospital admissions for people aged 60 and over are as a result of 'poisonings by drugs, medicaments and bio substances' [
22]. In 2005, 76,692 admissions to English hospitals were associated with an adverse drug reaction; this number increased by 45% over the period 1998 to 2005 and 59% of all cases involved patients aged over 60 years [
23].
Older people are at increased risk of medication-related adverse events due to a combination of factors including multiple medication (polypharmacy) and age-related changes in the body's response to medicines [
24]. Polypharmacy is extremely common in care homes, with residents generally reported to receive seven or more items each [
25]. Residents in the current study received an average of nine different drugs. In such a situation, the risk of incorrect administration of a prescribed medication is high, with the potential to result in a large number of adverse events [
26]. To date, most studies of improving medication safety in care homes have focused on prescribing [
27]. Relatively little research has examined administration of prescribed medicines and how the safety of this might be improved. A comprehensive literature review has drawn attention to a general lack of evidence on this aspect of safety and quality improvement in care homes, in particular residential homes [
1]. The current research provides the first incidence figures for medication administration errors in UK residential, as well as nursing, homes.
With the number of people aged 75 and over in the UK projected to nearly double by 2033, increasing from 4.8 to 8.7 million [
28], the quality of clinical care provided to older people will increasingly affect national patient safety. The care sector in the UK relies heavily on residential homes which have no on-site nursing and, as older people's clinical needs increase, innovative ways of providing clinical care and increasing the expertise of non-nursing, social care staff will be required [
29]. New technology may have a part to play in this, especially for aspects such as medication. Historically, quality improvement interventions for preventing medication errors have included labour intensive manual medication reviews, inspection of prescription requests and authorised prescriptions, stock checks, inspection of dispensed items and audit of medication administration charts [
30]. More recently, systematic reviews of the literature have provided evidence that computerised support systems can improve prescribing and dispensing practices, but there is limited evidence of their impact on administration of medication for older people [
31,
32].
The present study is the first to assess the introduction of a barcode medication administration system in UK long-term residential care. In terms of the level and pattern of errors we observed, research from the US has found a similarly high level of medication administration errors, with 'wrong time' the most frequently observed error in hospitals, skilled nursing facilities and assisted living environments [
33‐
35]. Our findings indicate that, over a three month observation period, 90% of residents were exposed to at least one potential administration error. This figure mirrors that reported in an earlier study of care homes in England although, because a smaller number of administrations was examined in this study and sampling methods differed (See Additional file
3), rates cannot be directly compared [
12]. In the present study, overall error rates were higher in nursing homes, where RNs undertake medication rounds.
The pre-study survey suggests that errors are linked to system and behaviour factors rather than a lack of education or training. Staff in both settings identified interruptions to medicine rounds as the major cause of errors, supported by several other authors [
36‐
43]. Neither staff group associated errors in medication administration with lack of training. The fact that recorded levels of confidence were not linked to qualification levels would seem to support this. Instead, it appears that the concentration necessary for safe administration of medicines is interrupted by competing demands upon staff time. Our pre-study survey also indicated that RNs in nursing homes appeared to be generally less aware of the potential for errors to occur in administration than their residential home social care staff counterparts.
Some explanation for the higher incidence of medication administration errors observed in nursing homes may lie firstly in the greater complexity of decision-making underpinning the process of administering medication for RNs as opposed to social care staff [
44‐
46]. Since nursing home patients are generally more seriously ill and therefore may have
more complex medication regimes, this will inevitably raise pressure on staff and increase the risk of administration errors by nurses [
47]. In other studies, it has been reported that higher grade nurses are generally more prone to making medication errors than those of a lower grade [
48]. Secondly, RNs with their higher level responsibilities are also more likely to have to
multitask when undertaking medication rounds, further increasing the risk of error [
49]. Researchers have recently reported 4.8 ± 6.6 interruptions per medication round for nurses in long-term care facilities [
50]. Studies on hospital wards have also shown that the more frequent the interruptions the greater the number of errors [
51]. Thirdly, RNs may
employ critical thinking and clinical judgement, using their knowledge of the patient to make decisions regarding the timing of medication which counter what is prescribed by the physician [
45].
The higher incidence of errors recorded for RNs in the study contrasts with their lower pre-study recall of previous near misses. This may be linked to the fact that RNs have a tendency to focus on
'reportable' errors that have actually occurred more so than ones that have been averted [
52], and therefore may not acknowledge the latter. Other research evidence also suggests that conduct of routine, time-consuming tasks such as repeat medication rounds can lead nurses into
complacency and a diminished sensitivity towards the
potential for harm resulting from medication errors [
53,
54]. Medication rounds occupy approximately one-third of nursing time in long-term residential care [
50]. In contrast, for social care staff who do not have a robust professional and educational framework or clinical training to support them [
55], the sense of being 'stressed' when administering medication could reduce any complacency and increase recall of near misses.
Although there is some research to demonstrate that nursing staff adopt unsafe work-around practices with electronic medication administration systems [
56], interestingly there was no evidence of this in the present study. At the same time, there was a high level of acceptability for the system among nursing and social care staff. The final level of non-compliance with medication administration as prescribed was very low in both settings (0.075% of administrations). Presumably, this is because the PCS system by its very nature is extremely difficult to circumvent, with all data management undertaken outside the care home setting and feedback provided on every alert to the care home manager. Compliance may also be linked to the facts that the system was implemented in an institutional setting and that it provided automatic system-initiated alerts. A recent review of the evidence on computerised prescribing decision-support systems concludes that these perform better in institutional rather than ambulatory settings, and when decision support is initiated automatically by the system as opposed to user initiation [
57]. In a context in which communication between shifts is imperfect, or there is a high level of agency use, a system with built-in safeguards may also be expected to be more effective.
The main limitations associated with the present study include the relatively small number of care homes studied and the disparity in numbers of nursing and social care staff, and the absence of agreed criteria for valuing the different types of medication administration error observed. In terms of the latter, although a number of approaches have been attempted to categorising medication errors for older people in hospital, community and general practice [
58‐
60], it is only very recently that this debate has extended to care homes [
61]. Furthermore, existing criteria concentrate almost exclusively on identifying errors in prescribing, and only rarely errors in administration [
62]. There is therefore no consensus currently on the relative importance of different types of administration errors in care homes. Even so, certain errors (such as attempting to give medication to the wrong resident or attempting to give medication that had been discontinued) can clearly be considered as more serious. A further limitation is the assumption that the introduction of the system did not alter behaviour and make staff more careful. However, any Hawthorne effect was minimised by the disguised observational technique used.
Finally, our findings from residential care homes suggest that social care staff in nursing homes might also be trained to administer basic medication using the barcode medication administration system. This would leave registered nurses free to focus on more complex medication regimes and free up valuable nurse time for other tasks in the care home. More research is required into the decision-making of nurses during medication rounds before delegation to care staff in a nursing home setting can be recommended. The potential also exists for data from the system to be used to assess other aspects of medication management, such as prescribing, and provide a low cost decision-support system. A preliminary analysis of PCS prescribing data on antipsychotics and comparison with national guidelines has already demonstrated various short-comings [
63]. Research is currently underway to examine prescribing patterns for a range of further medications and to bench-mark these across care homes and GP practices.
Competing interests
The authors declare that they have no competing interests. DW and SN received support through an educational grant given to the University of the West of England by Pharmacy Plus Ltd for the study. The study design, methods and materials were prepared by an independent academic team (the authors) with IPR retained by the lead University.
Authors' contributions
DW, AS and SN conceived the idea for the study. SN obtained the data, undertook the surveys and completed the statistical analysis. AS and DW wrote the manuscript which was reviewed by all authors. All authors had full access to all data in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. All authors read and approved the final manuscript.