Assessing the impact of automated dispensing

In recent years, healthcare has seen a surge in technological advancement and automation. Once confined to science fiction, robotics is becoming increasingly commonplace in medicine, from the da Vinci surgical robot to microbot atherosclerotic plaque busters. Pharmacy is no exception to this robotic revolution, with the introduction of automated dispensing systems (ADS) designed to improve efficiency and minimise dispensing errors. Currently, there are three main types of ADS: pharmacy-based original pack-dispensing systems; repackaging systems; and ward-based ADS. This article will describe the impact of the different dispensing systems on workload and errors.

These systems are widely used in UK hospital pharmacies.(1) They automate medication storage, stock selection and, occasionally, product labelling. Medication is stored and retrieved from the ADS, based on recognition of the European Article Number barcode on the product by the ADS, pharmacy labelling and stock control software. During label generation, a signal is transmitted from the labelling software to the ADS initiating stock selection.

The ADS selects the requested medication from the shelf within the device and transfers the product to the delivery station via a conveyor belt or chute.(1) Some ADS have an integrated labelling device, which affixes the appropriate dispensing label to the product before transferring the medication to the delivery station.

The majority of research on original pack ADS originates from the UK. A detailed evaluation of the impact of automation on dispensary workload and dispensing near-misses was conducted by James and colleagues.(2) The researchers conducted a longitudinal case study at the pharmacy department of a 600-bed teaching hospital. A non-participant observer recorded details of dispensary workload using the validated event recording technique each working day for six weeks before (May-June 2007), and after (May-July 2009), the installation of the ADS (Rowa Speedcase).

During this data collection period, pharmacy staff also self-reported details of dispensing near-misses, defined as dispensing errors detected during dispensing before the medication had been issued from the pharmacy, on standardised data collection forms. The study findings revealed that automation significantly improved dispensing efficiency and reduced the rate of near-misses. Automation significantly increased the median dispensary workload by 43% from 9.(20) items/person/hour pre-automation to 13.(17) items/person/hour post-automation (p<0.001). Furthermore, the rate of dispensing near-misses decreased by 56% from 0.64% pre-automation to 0.28% post-automation (p<0.0001). However, there was no significant difference between the types of near-misses reported pre- and post-automation.(2)

Similar findings were reported by Fitzpatrick and colleagues, who evaluated the impact of installing a Consis Baxter pharmacy-based original pack dispenser on dispensing near-misses.(3) Data on near-misses were reported by pharmacy staff on a standardised form for five months pre-automation and four months post-automation.

The numbers of medication items dispensed were also recorded. It was reported that installation of the ADS had reduced the overall incidence of near-misses by 16% and that the number of items dispensed post-automation had increased by 19%. Supply of the wrong strength (-46%), expired medication (-37%), wrong drug (-22%), wrong quantity (-14%), wrong formulation (-4%), incorrect information on label (-19%) and failure to supply medication (-68%) all decreased post-automation. However, the incidence of labelling medication with the incorrect directions was increased by 35% post-automation.(3)

Franklin and colleagues also found that automation significantly reduced the incidence of dispensing near-misses at two London hospitals.(4) Hospital 1 had a Swisslog Packpicker ADS installed and hospital 2 employed a Rowa Speedcase to automate stock selection. Data on near-misses were self-reported by staff at the two hospitals over a period of two weeks before and after the installation of the ADS. The incidence of near-misses was reduced significantly, from 2.7% to 1% (95% CI: -1.4% to -2.1%) at hospital 1 and from 1.2% to 0.6% (95% CI: -0.3% to -0.9%) at hospital 2.

At both hospitals, the incidence of content/drug errors and labelling errors decreased. The time taken to label, select and assemble dispensed medication was determined by direct observation five days before and after the installation of the ADS at both hospital sites. There was no significant difference in the time taken to label and assemble medication pre- and post-automation at the two hospitals. However, automation of the stock selection stage had significantly reduced product selection time from 49 seconds to 32 seconds at hospital 1 (p=0.001) and from 21 to 0 seconds at hospital 2 (p<0.001).(4)

These systems are widely used by US and European pharmacies that dispense unit doses to hospital inpatients.(5) The systems operate by firstly removing medication from the manufacturers’ original pack, then repackaging the medication into a unit dose pack or blister card. Some repackaging units produce compliance packs or monitored dosage systems that contain each dose of medication to be administered at a particular time of day.(5) These compliance pack systems are used within some specialist UK pharmacies (for example, mental health hospitals) to produce monitored dosage systems for outpatients.(1)

Use of the Baxter ATC-212 automated repackaging system has been evaluated by a number of researchers. Klein and colleagues compared the accuracy and time spent by pharmacy staff filling unit dose carts manually and using the Baxter ATC-212 system.(6) Independent checking of the carts over a three-week period before and after the implementation of the ADS revealed that automation was associated with fewer filling errors (0.65%) compared with manual cart filling (0.84%).

This is consistent with the findings of previous research that reported manual cart filling was 92.62% accurate compared with the Baxter ATC-212 system, which was 99.98% accurate.(7) Klein and colleagues also reported that use of the ADS significantly reduced the time spent by technicians filling medication carts (mean time pre-automation = 456.6 minutes; mean time post-automation = 332.8 minutes).(6) However, the time spent checking the carts was not reduced significantly because the pharmacists were required to cut the strip of packaged medication produced by the ADS into individual doses during the checking process (mean time pre-automation = 210.8 minutes; mean time post-automation = 185.8 minutes).(6)

The use of the term ADS to describe these systems is misleading. These devices do not perform any of the characteristic stages of the dispensing process, namely, label generation, stock selection, medication assembly and product labelling. Indeed, these systems would be more accurately described as electronic storage cabinets and have been likened to vending machines and ATMs.(5,8)

These ward-based electronic storage devices consist of drug cabinets and/or drug trolleys. The drug cabinet and trolley consist of computer-controlled drawers. Dispensed medication supplied by the pharmacy as unit doses or manufacturers’ original packs are stored in patient-specific or product-specific drawers within the drug cabinet or drug trolley. When a patient’s details or specific product details are entered into the device’s computer system, the appropriate drawer opens, enabling the required medication to be selected and administered to the patient.(5,9)

In some systems, the electronic drug trolley can be docked with the drug cabinet. This enables a nurse to prepare the medication doses required for the next drug round, by transferring the medication from the drug cabinet to patient-specific drawers in the mobile drug trolley.(9)

Several research studies have been conducted in the US evaluating the impact of installing ward-based ADS on medication administration errors (MAEs). Borel and Rascati conducted a disguised observational study on three wards at a 600-bed hospital.(10) Nurses were observed administering medication at least two days per week before, and approximately two months after, the implementation of a Medstation Rx. The study found that MAEs decreased significantly following the installation of the Medstation Rx, from 16.9% (148/873) to 10.4% (97/929; p=0.001).(10) This finding was supported by the research of Barker and colleagues, who evaluated the impact of implementing a McLaughlin Dispensing System using a crossover study design.(11)

 Nurses were randomly assigned to use either the traditional unit-dose medication cart (control) or the ward-based ADS for seven days and were then switched to the alternative system for the remaining seven days. A trained pharmacist collected data on MAEs by directly observing nursing staff. During the 14 days of study, which equated to 10 hours observation per day, the rate of MAEs was significantly lower for the ADS (10.6%; 96/902) compared with the control system (15.9%; 139/873). Furthermore, the researchers found that the installation of the ADS had no significant effect on nursing workload, defined as opportunities for error per shift).(11)

Another study investigated errors in the stocking or filling of unit-dose cassettes compared with the filling of a Medstation Rx system for a 26-bed adult medicine ward.(12) A pharmacist checked the medicines within the cassettes and the Medstation Rx each day for six weeks and recorded the number of errors identified. The study identified that there were significantly fewer errors made in filling the Medstation Rx (0.61%) compared to filling the cassettes (0.89%; p=0.04). The study also reported that the mean waiting time for a new medication had decreased from 45 minutes for the traditional unit-dose system to one minute with the Medstation Rx. Furthermore, the time spent by pharmacists resolving drug distribution problems had decreased from 25% pre-automation to less than 5% post-automation.(12)

Research undertaken in the UK also demonstrates the positive impact of ward-based ADS on time to acquire a dose of medication. Ardern-Jones and colleagues evaluated the impact of installing a Medi-365 system in the Emergency Department of a district general hospital in England.(8) Prior to installing the Medi-365, medicines were stored in locked cupboards on the wards. Before and after the installation of the Medi-365, a researcher observed and timed the process of retrieving 50 medication doses from storage. It was found that the median time to acquire a dose of medication (from initial identification of the need for medication to the patient receiving the dose) decreased from 139.5 seconds pre-automation to 44 seconds post-automation (p<0.0001). Based on the number of doses of medication retrieved from the unit in a 24 hour period and time taken to retrieve doses, it was calculated that 7.1 hours/day of nursing time was saved by installing the Medi-365.(8)

Franklin and colleagues conducted a detailed evaluation of a closed loop electronic prescribing, automated dispensing, barcode patient identification and electronic medication administration record (ServeRx V.1:13 system).(13) This system was installed in a 28-bed general surgical ward of a London teaching hospital. Prior to the intervention, doctors prescribed medication by hand on medication charts and nurses retrieved medication for administration from drug trolleys and stock cupboards on the ward. The ServeRx system allowed doctors to prescribe medication electronically and medication was stored in an electronic drug cabinet. Before the drug round, nurses retrieved medication from the electronic drug cabinet and transferred the medication to patient specific drawers within an electronic drug trolley.

During administration, the patient’s identity was verified by barcode, resulting in the opening of the patient specific drawer in the electronic drug trolley. The evaluation was undertaken over a period of three-to-six months before and six-to-twelve months after the installation of the ServeRx system. Outcomes measured included prescribing errors; MAEs; time taken to prescribe; time taken by pharmacists to deliver a clinical service to the ward; time taken by nurses to conduct a drug round and nursing time spent on medicines related activities outside of the drug round. The closed loop system significantly reduced the incidence of prescribing errors (pre-intervention: 3.8%, 93/2450; post-intervention: 2%, 48/2353, p<0.001), MAEs (pre-intervention: 8.6%, 141/1644; post-intervention: 4.4%, 53/1178; p=0.0003) and time spent by nurses conducting the drug round (pre-intervention: mean = 50 minutes; post-intervention: mean = 40 minutes, p=0.006).

However, the time taken to prescribe a drug (pre-intervention: mean = 15 seconds; post-intervention: mean = 39 second, p=0.001) and the time taken by a pharmacist to provide a clinical service to the wards (pre-intervention: mean = 68 minutes; post-intervention: mean = 98 minutes, p=0.001) had increased significantly.

The increase in pharmacist workload was attributed to the availability of all electronic drug charts post-automation, whereas pre-intervention 22% of handwritten charts were unavailable.(13) In contrast to the findings of Ardern-Jones and colleagues,(8) the proportion of nursing time spent on medication-related activities (pre-intervention: 21.1%; post-intervention: 28.7%, p=0.006) increased significantly. This can be attributed to the requirement for nurses to transfer medication from the electronic storage cabinet to electronic drug trolley prior to administration.(13)

Automated dispensing systems, regardless of type, reduce the incidence of medication error and improve efficiency. Pharmacy-based original pack ADS improve the capacity of pharmacy departments to dispense accurately increasing numbers of medication by significantly decreasing the time spent selecting products.

Repackaging systems decrease dispensing errors and time spent by technicians filling the carts. Ward-based ADS are associated with fewer filling errors and significantly decrease MAEs. Where the ward-based ADS is linked to an electronic prescribing system, prescribing errors can also be reduced significantly. The use of ward-based ADS can improve accessibility to medication by decreasing the time taken to acquire new doses.

These systems also decrease the time taken to conduct a drug round. Nevertheless, the impact of ward-based ADS on pharmacy staff and nursing workload is system dependent. Ward-based ADS that have an integrated electronic prescribing system and mobile electronic drug trolley increase clinical pharmacy services to wards and increase nursing time spent on medicines-related activities.

Healthcare organisations are under increasing pressure to deliver efficient services while ensuring the quality and safety of patient care. Embracing technological advancement in automated dispensing systems has the potential to optimise medicines management, improve efficiency, maximise resources and safeguard patient safety.

• Pharmacy-based original pack automated dispensing systems (ADS), repackaging systems and ward-based ADS decrease the incidence of medication errors.
• Pharmacy-based original pack ADS reduce dispensing errors and improve the capacity to dispense increasing numbers of medicines.
• Repackaging systems decrease dispensing errors and time spent by technicians filling carts.
• Ward-based ADS decrease filling errors and administration errors; improve accessibility to medicines and decrease nurse time spent conducting drug rounds.
• ADS can improve efficiency while minimising medication errors.

1. Swanson D. Automated dispensing. An overview of the types of system available. H 2004;11:66-8.
2. James KL et al. The impact of automation on workload and dispensing errors in a hospital pharmacy. Int J Pharm Pract 2013;21:92-104.
3. Fitzpatrick R et al. Evaluation of an automated dispensing system in a hospital pharmacy dispensary. Pharm J 2005;274:763-5.
4. Franklin BD et al. An evaluation of two automated dispensing machines in UK hospital pharmacy. Int J Pharm Pract 2008;16:47-53.
5. Neuenschwander M. Limiting or increasing opportunities for errors with dispensing automation. Hosp Pharm 1996;31:1102-6.
6. Klein EG et al. Medication cart-filling time, accuracy and cost with an automated dispensing system. Am Hosp Pharm 1994;51:1193-6.
7. Kratz K, Thygesen C. A comparison of the accuracy of unit-dose cart fill with Baxter ATC-212 computerized system and manual filling. Hosp Pharm 1992;27:19-22.
8. Ardern-Jones J et al. The impact of the introduction of a ward-based automated medicines vending unit on nursing tasks and time in the Emergency Department. Int J Pharm Pract 2009;17:345-9.
9. Barber N et al. Safer, Faster, Better? Evaluating Electronic Prescribing. Report to the Patient Safety Research Programme. London: The School of Pharmacy;2006.
10. Borel JM, Rascati KL. Effect of an automated, nursing unit-based drug-dispensing device on medication errors. Am J Health Syst Pharm 1995;52:1875-9.
11. Barker KN et al. Effect of an automated bedside dispensing machine on medication errors. Am J Hosp Pharm 1984;41:1352-8.
12. Ray MD, Aldrich LT, Lew PJ. Experience with an automated point-of use unit-dose drug distribution system. Hosp Pharm 1995;30:18,20-23;27-30.
13. Franklin BD et al. The impact of a closed-loop electronic prescribing and administration system on prescribing errors, administration errors and staff time: a before-and-after study. QSHC 2007;16:279-84.