Objective To determine if the introduction of patient bedside medicine lockers leads to a safer and faster medicine administration round.
Methods The undisguised observer technique was utilised to observe the medicine administration round on four wards, two medical wards in Antrim Area Hospital (AAH) and two surgical wards in Causeway Hospital (CH), both before and after the introduction of patient bedside medicine lockers. All non-intravenous medicine administrations during the morning medicines administration round were observed and timed before and after the introduction of the lockers. Medicine administration errors (MAEs), time taken and reasons for delays were recorded and analysed.
Results The MAE rate and the time spent on the medicine administration round both decreased after the introduction of patient bedside medicine lockers. The MAE rate dropped from 8.3% to 1.3% (p < 0.001) in the AAH site and from 9.9% to 3.2% (p = 0.029) in the CH site; the time spent per patient on medicine administration decreased from 6.80 ± 5.44 min pre-intervention to 3.03 ± 1.87 min post-intervention and from 7.35 ± 6.24 min pre-intervention to 6.95 ± 5.39 min post-intervention in AAH (p < 0.001) and CH (p > 0.05), respectively.
Conclusions The introduction of patient bedside medicine lockers resulted in safer and faster medicine administration rounds.
Statistics from Altmetric.com
The Integrated Medicines Management model introduced in the Northern Health and Social Care Trust (NHSCT) in Northern Ireland has led to improved patient outcomes.1–3 The evolving model has been extended to incorporate a custom designed patient bedside medicine locker to reduce risks associated with medicines administration.
The administration of medicine is one of the most error prone steps in the medicines management process and accounts for 44% of all detected errors.4 Between September 2006 and June 2009, the National Patient Safety Agency (NPSA) received reports of 27 deaths, 68 severe harms and 21 383 other patient safety incidents relating to omitted or delayed medicines. Wider evidence suggests that the true rate of harm may be much higher as medication administration events are often not reported. The NPSA has called for action to reduce harm from omitted and delayed medicines.5
The goal of any medicine administration system is to ensure that, once validated, each medicine dose is administered in accordance with the prescriber's instructions. In the hospital setting, nurses are ultimately responsible for accurate and appropriate medicine administration as part of a disparate and demanding set of nursing duties.6 Medicine administration occupies 30–50% of nurses' time in hospitals and is one of the highest risk activities that a nurse performs.7,8 The worldwide incidence of medicine administration errors (MAEs) varies between 6.6% and 44.6% for all doses administered by nurses.6,9–12
An MAE is a dose administered, or omitted, that deviates from that prescribed.13 MAEs can be due to system and/or human failure.6 System factors include the type of drug, its route of administration, the physical appearance and availability of the drug, its packaging, warnings, checking protocols and documentation. Human factors are related to interruptions, distractions, fatigue, and the health professional's knowledge and experience. In 2000, the Department of Health stated that “human error is a consequence rather than a cause”, implying that systems are at fault rather than individuals.14 Review of the literature demonstrates that the effectiveness and quality of a medicine administration system can be successfully measured by determining the incidence of MAEs.15
Decentralising the medicine administration system so that medicines are held close to the patient may enhance safety and efficiency and helps minimise the number of interruptions and distractions that occur during medicine administration to a particular patient.16
The main objective of the present study was to determine whether the introduction of patient bedside medicine lockers results in safer and faster medicine administration rounds when compared with the traditional medicine trolley-based medicine administration system.
The study involved pre- and post-intervention assessment, conducted in two study site teaching hospitals in Northern Ireland, that is, two medical wards in Antrim Area Hospital (AAH) and two surgical wards in Causeway Hospital (CH). MAEs were detected using the undisguised observation technique.17 This technique has been shown to be valid and reliable, and does not influence the likelihood of nurses making or not making an error.6,17,18
Prior to the installation of the patient bedside medicine lockers (referred to throughout the article as ‘the intervention’), medicines were stored on the ward in two medicine trolleys and a ward stock cupboard. A pharmacy technician was responsible for ward medicine stock and replenished the medicine trolleys each day and the stock cupboard each week. Medicines administration was carried out by two nurses simultaneously, each nurse carrying out medicines administration on one half of the ward using one medicine trolley. After the intervention, a 28-day supply of each medicine (where applicable) was dispensed and stored in custom designed patient bedside medicine lockers, labelled for the patient; ward stock was held in a separate cupboard as before. A pharmacy technician was responsible for stocking the lockers (as required) on a daily basis and the stock cupboard each week.
Medicines were prescribed on a medicines prescription list (Kardex) and according to hospital policy, the nurse confirmed the patient's identity and allergy status and signed the Kardex each time a medicine was administered. If the nurse was unable to administer a medicine, for example, because the medicine was not available, the reason for the omission was recorded on the Kardex.
The morning medicine administration round was observed over a period of 8 consecutive days in AAH excluding weekends, before and after the introduction of patient bedside medicine lockers. The equivalent observation periods in CH were 9 and 5 days, respectively. The administration of all non-intravenous, scheduled doses was observed. Pharmacy staff were deployed as observers following a period of training which covered the observation technique, completion of the data collection form, confidentiality and when to intervene (eg, to avoid patient harm), that is, it was agreed that the observer would intervene prior to administration if an error was about to be made (including imminent omissions when the medicine was available on the ward). Data collection was piloted under the supervision of a pharmacist experienced in MAE observation, to ensure observer competence.
Approval was obtained from the trust director of nursing and the study explained in full to the ward managers prior to commencing data collection. Nursing staff were assured of confidentiality and protection of identity.
An MAE was defined as any dose of medicine that deviated from the prescriber's valid prescription and was categorised according to previously published criteria adapted for this purpose (Box 1).19 The observer recorded the medicines administration round start time, end time and the time taken per patient during the ward round. The medicines administration round was deemed to have started when the nurse opened the first patient Kardex and to have ended when the last patient was administered their medicine. The observer also recorded whether the patient's identity and allergy status were confirmed prior to medicine administration.
The observer directly observed the nurse throughout the medicines administration round and recorded details of all possible prescribed medicines and their doses on a custom-designed data collection form. In addition to the recording of MAEs, reasons for delays were also documented (Box 2). The MAE rate was expressed as a percentage of the total opportunities for error, that is, the sum of all doses observed as being administered plus all doses omitted, in accordance with the literature.13
Data were entered into SPSS for Windows (V.18) for statistical analysis. Standard statistical methodologies were employed to assess the impact of patient bedside medicine lockers on the MAE rate and the time taken to complete the medicines administration round.
Number and timing of patient medicine administrations
In AAH, data were available for 355 patients (164 pre- and 191 post-intervention). A total of 1014 medicine doses were observed out of 1045 (97%) potential doses prior to the intervention. Post-intervention, 1256 doses were observed out a total of 1266 (99%) potential doses. In CH, data were available for 447 patients (379 pre- and 68 post-intervention). A total of 1688 medicine doses were observed out of 1823 (93%) potential doses prior to the intervention. Post-intervention, 253 doses were observed out a total of 259 (98%) potential doses.
In AAH, the mean (±SD) time per patient for the medicine administration round was 6.80 ± 5.44 min pre-intervention and 3.03 ± 1.87 min post-intervention. This difference of 3.77 min per patient was statistically significant (p < 0.001). Extrapolating these findings across the 27-bed wards could potentially contribute an additional 101.8 min to nursing staff time per ward. In CH, the mean time per patient for the medicine administration round was 7.35 ± 6.24 min pre-intervention and 6.95 ± 5.39 min post-intervention. This difference, however, was not statistically significant (p > 0.05).
The MAE rates are detailed pre- and post-intervention for both study sites in table 1. The total MAE rate was 8.3% before and 1.3% after the introduction of the patient medication lockers at the AAH site. The figures for CH were 9.9% before and 3.2% after locker introduction. During the study period, the observer intervened in all potential MAE occasions. Omitted doses of medicines available on the ward which were corrected included a number of high risk drugs such as methotrexate.
Delays on medicine administration rounds
The necessity for the nurse to visit the ward stock cupboard or fridge for medicines was a common cause of delays both pre- and post-intervention (54% and 60% at the AAH site and 39% and 55% at the CH site, respectively). Delays caused by an interruption to the medicine round by factors outside the control of the nursing staff also occurred frequently (28% pre- and 39% post-intervention in AAH and 44% pre- and 43% post-intervention in CH, respectively) (table 2).
The allergy status and the identity of the patient were checked 88% of the time pre- and 99% of the time post-intervention at the AAH site and 97.9% of the time pre- and 98.3% of the time post-intervention at the CH site.
The requirement to minimise risk throughout the medicines management process, together with the increasing demands on nursing staff time, drives the need to improve the efficiency and accuracy of the medicine administration process.
The introduction of individual patient bedside medicine lockers ensured that medicines were selected from each patient's individual supply rather than from an often crowded medicine trolley containing multiple medicines for all patients on the ward. It was expected that the patient bedside medicine lockers would also make medicine administration more patient-centred. A further safety benefit is the reduction in the opportunity for a member of the public on the ward to illicitly remove drugs during the medicine administration round.20
The mean time taken, per patient, to complete the morning medicines administration round on the study wards was 6.80 ± 5.44 min pre-intervention and 3.03 ± 1.87 min post-intervention at the AAH site (p < 0.001). It was anticipated that with the introduction of the patient bedside medicine lockers, nurses would be able to locate the medicines required for each patient more quickly, thereby speeding up the medicine administration round and this indeed was the case. A similar study in Canada found that a decentralised medicine distribution system also resulted in significant time savings; on average each nurse saved 23 min per 12 h shift on medicine distribution.16 However, in CH, the reduction in the mean time per patient for the medicine administration round was not statistically significant (7.35 ± 6.24 min pre-intervention vs 6.95 ± 5.39 min post-intervention). It should be noted, however, that one patient (post-intervention) required PEG (percutaneous endoscopic gastrostomy) delivery of his medicines which added significantly to the time taken overall. Due to the nature of the surgical speciality, a larger sample size may demonstrate a significant time-saving effect as seen in the medical wards. Therefore, further research is needed.
The most common medicine-related delay was caused by nurses going to locate medicines in the ward stock cupboard or ward fridge when the required medicine could not be located or the supply of the medicine was depleted. This occurred for 15.7% of all observed medicine doses pre-intervention and 6.6% of observed doses post-intervention at the AAH site and 19.4% of observed doses pre-intervention and 20.2% of observed doses post-intervention at the CH site.
The efficiency and quality of a medicine administration system can be reflected by the MAE rate. At the AAH site, the total MAE rate decreased from 8.3% pre-intervention to 1.3% after the introduction of patient bedside medicine lockers. The most frequent types of MAE identified prior to the introduction of the patient bedside medicine locker were omitted doses (combined omitted dose-available and omitted dose-not available) and the wrong formulation of the medicine being administered to the patient. This supports the assumption that the correct medicines are more easily identified and located in the patient bedside medicine locker compared with the traditional medicine trolley. At the CH site, the total MAE rate decreased from 9.9% before to 5.9% after the introduction of patient bedside medicine lockers. Omitted dose was also the most frequent type of MAE identified prior to the introduction of the patient bedside medicine locker (both available and not available on the ward).
A study examining the accuracy of medication administration by nurses at a large tertiary hospital found that 4.89% of doses had errors and the most frequent errors were wrong technique (34%), wrong time (34%) and omission (19.5%), while on average each nurse saved 12 min per 12 h shift through not searching for medicines.21 Observational studies of medication administration errors in UK hospitals suggest that the MAE rate is 3.0–8.0%, and approximately half of these errors are due to the drug concerned not being available on the ward.16 Omission errors have been reported as the most common type of administration error in observational studies conducted in general hospitals.22 Reasons for dose omissions have been reported to involve a non-stock drug being required, inadequate storage space, failure to notice the dose due on the Kardex, patient transfer, difficulty in locating the drug prescribed, and confusion about where drugs are stored (eg, other than the trolley, for example, in the stock cupboard).23,24
Similar results were also demonstrated in a study that compared a ward bay drug distribution system and a bedside system (patient bedside locker), which revealed an error rate of 18% with the ward bay system compared to 7% with the bedside system.25 Barker and associates observed medication administration in 36 acute and long-term care facilities and concluded that errors occurred in nearly one in every five doses administered.26 Complete eradication of MAEs due to non-availability of medicines at ward level cannot be achieved since late admissions make some missed doses unavoidable. While some errors can be prevented, human error cannot be completely eliminated.27
The introduction of patient bedside medicine lockers was shown to result in a safer and faster medicine administration round in medical wards. Significant improvements in safety were also noted in surgical wards, however, further research is required to determine why the same improvement in the administration time was not achieved. These findings support the implementation of this service within Northern Ireland. The Integrated Medicines Management model is continuing to evolve and automated medicines administration systems are being considered as part of ongoing quality improvement within the NHSCT.
Contributors All authors contributed to the writing of this article.
Funding This study was carried out as a service development activity within the study site hospitals and as such no external funding was required or used.
Competing interests None.
Provenance Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.