Medication administration errors studied through the mixed-methods lens

Background: In acute-care settings interruptions and workload increase medication administration errors (MAEs). However, MAEs are studied less within mental-health settings, where nursing staff specialise in psychiatric therapeutics and the mode of administration is different; e.g. in psychiatry, patients are not necessarily resting in hospital beds making bedside administration redundant. Existing research in psychiatry suggests MAEs are linked to morning rounds associated with more interruptions, medicines to administer, activities and staff shortages. An audit at the current Trust in 2014 found 13−52% of patients had at least one ‘blank box’ on their drug chart in the preceding seven days on seven mental-health wards necessitating further research. In addition, the Trust spans mental-health (psychiatric) and community wards (some with patient lockers for medication administration), providing an opportunity to examine differences across therapeutic areas and modes of administration. Aim: The study aims to document the rate of MAEs across all inpatient wards at Berkshire Healthcare NHS Foundation Trust (BHFT) psychiatric and community hospital wards, and to investigate the interrelationship between error rates and a range of possible contributing factors as well as the causes behind the MAEs. The aim eventually is to produce recommendations for improving medication administration practices in this type of setting. Methodology: This research used a mixed-method approach by adopting direct observation and semi-structured interviews to examine in depth the rate and causes of MAEs within mental-health and community hospital wards. The researcher visited all 19 wards (9 community hospital wards, 10 psychiatric wards) between July and October 2015 making 65 separate observations in total. Different modes of administration were noted alongside any interruptions and other potential contributing factors. The different modes of administration observed included administering at the patient ‘bedside’ prepared in clinic room (applied to psychiatric wards only), using patient lockers (applied to community hospital wards only) or a drug trolley (applied to community hospital wards only), or via a ‘queue’ where by patients were called to wait outside the clinic room (applied to psychiatric wards only) or a ‘mixed’ mode whereby medicines were given at the bedside or via a queue with the medication prepared in clinic room (applied to psychiatric wards only). In this research, data were entered into SPSS (v21) and analysed descriptively and using the Poisson Regression Model. These findings were discussed in structured interviews with nurses, pharmacists and compared with the researcher’s own notes to describe the causes of errors using the organisational accident causation model as a theoretical framework, which were then triangulated with the quantitative findings. Findings: In total 2237 opportunities for error were observed with 367 MAEs, resulting in a total error rate of 16.4%, 2.4% of which were wrong time errors. When the number of MAEs was separated to procedural errors versus clinical errors, the clinical error rate was 7.7%. The most frequent type of MAE was expiry errors (32% - a type of procedural error) followed by omissions (23% - a type of clinical error). Two modes of administration (‘bedside-prepared in clinic room’ and using ‘bedside-patient locker’) and the non-psychiatric drug group increased the risk of procedural errors. Two administration times (08:00 and 12:00 only), nurse grade (band-5 nurses) and two modes of administration (using ‘bedside-patient locker’ and ‘mixed’) increased the risk of clinical errors. Twelve interviews were completed with eight nurses and four pharmacists. The main active failure category for clinical errors was understood to be ‘lapses’, the main error-producing condition ‘staff workload’ and the main latent condition ‘safety culture and priorities’. The main active failure category for procedural errors was judged to be ‘situational violations’ by the researcher but this was not the view of nurses and pharmacists who painted a more blameless picture of the workplace. Conclusion: The findings show that expiry errors, a major component of procedural errors, take place across both psychiatric and community hospital ward types specifically when medication is given at the patient bedside and is prepared either in the clinic room or given via patient lockers, and that this is likely because of staff workload and/or lack of staff knowledge. The findings also suggest that omissions, a major component of clinical errors, were associated with the mixed mode of administration on psychiatric wards and the 12:00 administration time, and occurred because of a range of reasons that included workload, miscommunication and staff-related factors. Another type of clinical error was the wrong time error which was associated mainly with the patient lockers and the 08:00 administration time, mainly because of the high staff workload. The results suggest that patient lockers are not as safe as perceived because they are implicated in both clinical and procedural errors. In addition, the distinct contributing factors identified in this study can provide a means through which the occurrence of MAEs can be addressed. This mixed-method study makes a novel contribution to knowledge as the first study to compare mental-health and community hospital wards using the direct observation method, researching different modes of administration, and then comparing and contrasting the perceptions of the researcher with nurse and pharmacist opinions for causes of MAEs