Takase et al.

Journal of Pharmaceutical Health Care and Sciences (2022) 8:24

https://doi.org/10.1186/s40780-022-00255-w

RESEARCH ARTICLE Open Access

Evaluating the safety and efficiency

of robotic dispensing systems
Tomoki Takase1* , Norio Masumoto1, Naoki Shibatani1, Yusaku Matsuoka1, Fumiaki Tanaka1,
Masaki Hirabatake1, Hiroko Kashiwagi1, Itaru Nishioka2, Hiroaki Ikesue1, Tohru Hashida1, Naoshi Koide1,3,4 and
Nobuyuki Muroi1

Abstract
Background: Although automated dispensing robots have been implemented for medication dispensing in Japan,
their effect is yet to be fully investigated. In this study, we evaluated the effect of automated dispensing robots and
collaborative work with pharmacy support staff on medication dispensing.

dispensing robot (Drug Station®), which is operated by pharmacy support staff, (2) automated dispensing robot for
Methods: A robotic dispensing system integrating the following three components was established: (1) automated

powdered medicine (Mini DimeRo®), and (3) bar-coded medication dispensing support system with personal digital
assistance (Hp-PORIMS®). Subsequently, we evaluated the incidences of dispensing errors and dispensing times
before and after introducing the robotic dispensing system. Dispensing errors were classified into two categories,
namely prevented dispensing errors and unprevented dispensing errors. The incidence of dispensing errors was
calculated as follows: incidence of dispensing errors = total number of dispensing errors/total number of medication
orders in each prescription.
Results: After introducing the robotic dispensing system, the total incidence of prevented dispensing errors was
significantly reduced (0.204% [324/158,548] to 0.044% [50/114,111], p < 0.001). The total incidence of unprevented dis-
pensing errors was significantly reduced (0.015% [24/158,548] to 0.002% [2/114,111], p < 0.001). The number of cases
of wrong strength and wrong drug, which can seriously impact a patient’s health, reduced to almost zero. The median
dispensing time of pharmacists per prescription was significantly reduced (from 60 to 23 s, p < 0.001).
Conclusions: The robotic dispensing system enabled the process of medication dispensing by pharmacist to be
partially and safely shared with automated dispensing robots and pharmacy support staff. Therefore, clinical care for
patients by pharmacists could be enhanced by ensuring quality and safety of medication.
Keywords: Pharmacist, Robot, Dispensing device, Dispensing error, Dispensing time

Background contribute toward maximizing patient safety and effi-

Healthcare systems are rapidly shifting from a single cacy of pharmacotherapy, from hospital to commu-
hospital-based care module to a community-based nity care. In April 2019, Japanese Ministry of Health,
regional collaborative care system. Pharmacists can Labour, and Welfare released a notification titled “the
conception of medication dispensing” to pharmacists
to ensure sufficient time for clinical care of patients
*Correspondence: t-takase@kcho.jp [1]. This notification describes that preparing medi-
1
Department of Pharmacy, Kobe City Medical Center General Hospital, 2‑1‑1, cines is one of delegable works from pharmacists to
Minatojima Minamimachi, Chuo‑ku, Kobe, Hyogo 650‑0047, Japan pharmacy support staff. In most hospitals, dispensing
Full list of author information is available at the end of the article

© The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​
mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Takase et al. Journal of Pharmaceutical Health Care and Sciences (2022) 8:24 Page 2 of 9

system/process includes verifying the appropriateness Robotic dispensing system

of the prescription, such as the dose, in individual dis- Newly implemented dispensing devices
orders and drug–drug interactions as well as manually We introduced the robotic dispensing system integrating
selecting medicines from shelves. Because dispens- the following three components (Fig. 1) in February 24,
ing and verification processes are complex, preventing 2021.
human errors is difficult. Thus, dispensing errors can
inevitably occur at a certain rate [2, 3]. However, inci- a) Automated dispensing robot
dents because of dispensing errors can cause iatrogenic
harm in patients. Therefore, minimizing human errors
(Drug Station®, Yuyama Co., Ltd., Osaka, Japan)
We implemented an automated dispensing robot
in drug dispensing is essential. Because highly skilled
pharmacists are required to prevent human errors in which stores a maximum of 1,200 single unit pack-
manual medication dispensing, delegating even part ages of oral medicines such as tablets, capsules,
of the dispensing process (preparing prescribed medi- powders, liquids, and topical medications. This
cines) from pharmacists to pharmacy support staff is robot is linked to our hospital computerized physi-
difficult.
GX®, Fujitsu, Ltd., Tokyo, Japan). Pharmacists or
cian order entry (CPOE) system (HOPE/EGMAIN-
Recently, automated dispensing robots have been imple-
mented in Japan. They have achieved remarkable reduc- pharmacy support staff pick up the ordered quan-
tion in dispensing errors and improved the efficiency tity of medicines according to the instructions on
of dispensing processes [2, 4, 5]. The ability of robots to
cally moves to the handling slots of Drug Station®
the screen from the storage bins, which automati-
provide fast and accurate dispensing allows pharmacists
to spend more time on clinical care for patients, thereby by ordered prescription data in the CPOE system.
adding value to their clinical role [2, 4, 5]. Despite several After the medicines pick up by a pharmacist or a
potential advantages of integrating automated dispensing pharmacy support staff, the number of medicines is
robots and collaboration with pharmacy support staff, the graphically confirmed by using the built-in camera,
safety and the efficiency of such systems is yet to be fully and/or their weight are confirmed by the built-in
evaluated in Japan. electronic scale. Because of these reliable functions,
We established the “robotic dispensing system” with this robot can support preparing medications accu-
the following three components: (1) automated dispens- rately for both pharmacists and pharmacy support
ing robot operated by pharmacy support staff, (2) auto- staff.
mated dispensing robot for powdered medicine, and (3)
bar-coded medication dispensing support system with In our hospital formulary, a total of 749 oral or topi-
personal digital assistance (PDA). Notably, pharmacy cal medicines were approved by the Pharmacy and
support staff engaged in preparing prescribed medicines Therapeutic Committee (P&T Committee). Among
using the automated dispensing robot in the robotic dis- them, 623 (83.2%) were stored in the automated dis-
pensing system. There is no independent organization pensing robot. The remaining 126 medicines could
of pharmacy technicians in Japan. Therefore, we trained not be stored in the robot because they require low-
pharmacy support staff to collaborate with pharmacists temperature storage; strict legal controls, such as opi-
for medication dispensing. oid analgesics; or are packed in large packaging that
We investigated reduction in dispensing errors and could not fit in the storage bin.
dispensing time before and after introducing the robotic b) Automated dispensing robot for powdered medicine
dispensing system comprising collaborative working
for powdered medicine (Mini DimeRo®, Yuyama Co.,
We also implemented an automated dispensing robot
model with pharmacists and pharmacy support staff.
Ltd., Osaka, Japan), which was linked to our hospi-
tal CPOE system. After physicians ordered pow-
Methods dered medicines, this robot automatically weighed
Study site and packed them. This system can prepare powdered
Kobe City Medical Center General Hospital is a 768-bed medicines precisely, and the time required to pre-
acute phase hospital in Japan. An average of 500 prescrip- pare powdered medicines was considerably shorter
tions including single or multiple medication orders per than that in using conventional automatic packaging
prescription are handled each day in the hospital phar- machines [4]. In our hospital, among the 71 pow-
macy. We analyzed proportions of prescriptions using dered medicines approved by the P&T Committee,
each dispensing device in the study periods, as described 42 (59.2%) were stocked into the cassettes of the
later. robot.
Takase et al. Journal of Pharmaceutical Health Care and Sciences (2022) 8:24 Page 3 of 9

Fig. 1 Newly implemented dispensing devices. a Automated dispensing robot (Drug Station®), (1) outside view, (2) storage bins and robotic arms,
(3) slot. b Automated dispensing robot for powdered medicine (Mini DimeRo®). c Bar-coded medication dispensing support system with using PDA
(Hp-PORIMS.®)

c) Bar-coded medication dispensing support system 17:30 on weekdays. Other than those time, pharmacists
executed all flows of medication dispensing steps.

support system with PDA (Hp-PORIMS®, Yuyama Co.,

We implemented a bar-coded medication dispensing
Operation flow before and after introducing the robotic
Ltd., Osaka, Japan), which was connected to our hos- dispensing system
pital CPOE system. This system control prescriptions Before introducing the robotic dispensing system (until
and medicine packages by bar coding and collates using February 23, 2021), pharmacists verified each prescrip-
PDA whether the drug is picked correctly according tion and manually prepared and dispensed medicines.
to the prescription. This system considerably reduced Dispensed medicines and the prescriptions were sub-
dispensing errors [6]. This system was used for 126 sequently verified by another pharmacist. After intro-
oral or topical medicines, which could not be stored in ducing the robotic dispensing system (since February
the automated dispensing robot. Additionally, 60 self- 24, 2021), pharmacists verified each prescription, and
injectable drugs, approved by the P&T Committee, pharmacists or pharmacy support staff prepared medi-
were also dispensed using this system. cines using the automated dispensing robot. With the
exception of medicines stored in the automated dis-
pensing robot, medicines were collated using bar-coded
Role of pharmacy support staff medication dispensing support system with PDA. Sub-
Pharmacy support staff engaged in preparing pre- sequently, the prepared and dispensed medicines, and
scribed medicines using the automated dispensing their prescriptions were verified by another pharma-
robot. Pharmacy support staff worked from 8:45 to cist. Pharmacists trained pharmacy support staff on the
Takase et al. Journal of Pharmaceutical Health Care and Sciences (2022) 8:24 Page 4 of 9

operation of the robotic dispensing system for 2 weeks. before and after introducing the robotic dispensing
Pharmacy support staff then started operating indepen- system: periods 1, 2, and 3 (Fig. 2). The participating
dently since June 1, 2021. pharmacists numbered 59, 60, and 60, respectively, in
We defined the study periods as follows (Fig. 2): period this study, on dispensing errors for periods 1, 2 and 3,
1 (before introduction: between March 2020 and August respectively.
2020), period 2 (early phase after introduction: between The incidence of dispensing errors was calculated as
March 2021 and May 2021), and period 3 (collaborative follows:
phase after introduction: between June 2021 and August incidence of dispensing errors = total number of dis-
2021) to evaluate the after-mentioned incidences of dis- pensing errors/total number of medication orders in each
pensing errors and dispensing time. prescription.

10]: wrong drug (e.g., caused by similar name: “Norvasc®

Types of dispensing errors were defined as follows [8–

tablet” and “Nolvadex® tablet”), wrong quantity (caused

Incidences of dispensing errors
We classified dispensing errors into two categories,
namely prevented dispensing errors and unprevented by miscount), wrong strength (caused by selection error:
dispensing errors [7]. The prevented dispensing errors e.g., “bisoprolol tablet 0.625 mg” and “bisoprolol tablet
denoted errors detected by pharmacists before the medi- 2.5 mg”), wrong dosage form (e.g., “diclofenac supposi-
cines provided from the pharmacy to clinical wards or tory” and “diclofenac tablet”), and others.
outpatients. By contrast, unprevented dispensing errors
denoted errors that were detected by other medical staff Dispensing time
or patients after the medicines provided from the phar- We randomly selected the prescriptions (i.e., each phar-
macy to clinical wards or to outpatients. Each incident macist picked up the prescriptions without looking at the
was recorded when pharmacists, other medical staff, or contents) in period 1, 2, and 3 (Fig. 2), and compared the
patients detected the errors. The incidences of prevented time spent on dispensing per prescription in each period.
and unprevented dispensing errors were compared Additionally, the dispensing time was classified into three

Fig. 2 Dispensing process flowchart before and after introducing the robotic dispensing system in the study period. Before introducing the robotic
dispensing system (period 1), pharmacists manually prepared and dispensed medicines. After introducing the robotic dispensing system (period 2,
3), pharmacists or pharmacy support staff prepared medicines using automated dispensing robot. EMR: electronic medical records
Takase et al. Journal of Pharmaceutical Health Care and Sciences (2022) 8:24 Page 5 of 9

categories as follows: (a) work time of pharmacists, (b) USA). To compare the incidences of dispensing errors
work time of robot, and (c) work time of pharmacy sup- and dispensing time among the study periods, the Bon-

regardless of the use of Drug Station®. A total of 10 phar-

port staff. This study involved all types of prescriptions ferroni corrections were applied to determine the level of
significance for each group (p < 0.0167).
macists with 1–42 years of experience (one year, n = 6;
four years, n = 1; eleven years, n = 1; twenty-six years,
Results
n = 1; forty-two years, n = 1) and 5 pharmacy support
Proportions of prescriptions using the robotic dispensing
staff participated in the study. Pharmacy support staff
system among the study periods
took turns in operating the automated dispensing robot.
Proportions of prescriptions using the robotic dis-
of Drug Station® for each prescription. We defined the
The work time of robot was defined as the operating time
pensing system in each study period are displayed in
Table 1. The number of prescriptions dispensed were
of Drug Station®, because Mini DimeRo® automatically
work time of robot, considering only the operating time
77,199, 51,482, and 54,822, the total number of medi-
cation orders per prescription dispensed were 158,548,
intervention unlike Drug Station®. In periods 2 and 3, the
weighs and packs powdered medicines without human
106,611 and 114,111 in periods 1, 2, and 3, respectively.
In period 2, the number of drugs using the automated
PORIMS® were included into the work time of pharma-
dispensing times of powdered medicines and using Hp-
dispensing robot, automated dispensing robot for pow-
dered medicine, and bar-coded medication dispensing
cists. In period 1, the work time of pharmacists (defined
support system were 81,073 (76.0%), 4,380 (4.1%), and
as the time required for manually preparing and dispens-
9,252 (8.7%), respectively. In period 3, proportions were
ing medicines by pharmacist) was measured. In period
87,742 (76.9%), 4,091 (3.6%), and 9,975 (8.7%), respec-
2, the work time of robot and pharmacists (including the
tively. The dispensing device use rates were similar in
Station® by one pharmacist and dispensing medicines
total time required for preparing medicines using Drug
periods 2 and 3.
by another pharmacist) were measured. In period 3, the
work time of robot, and the work time of pharmacy sup- Effects of the robotic dispensing system reducing

medicines using Drug Station® by pharmacy support

port staff (defined as the time required for preparing dispensing errors
The dispensing errors related to oral or topical medi-
staff ), and the work time of pharmacists (defined as the cines, and self-injectable drugs were detected in the study
time required for dispensing medicines by pharmacist) periods. The incidences of prevented dispensing errors
were measured. by error type were wrong quantity (0.107% vs 0.026%
vs 0.028%), wrong strength (0.052% vs 0.003% vs 0%),
Statistical analysis wrong drug (0.025% vs 0.005% vs 0.001%), wrong dos-
Categorical data were presented as numbers (percentage) age form (0.010% vs 0% vs 0%), others (0.010% vs 0.021%
and were compared between groups using Fisher’s exact vs 0.015%), and total (0.204% vs 0.054% vs 0.044%) in
test. Continuous data are presented as medians (inter- periods 1, 2, and 3, respectively (Table 2). Among them,
quartile ranges), and the Mann–Whitney U test was used wrong quantity, wrong strength, wrong drug, wrong dos-
to compare the groups. All statistical analyses were per- age form and total were significantly reduced in periods
formed using JMP 14.2.0 (SAS Institute Inc., Cary, NC, 2 and 3 compared with those in period 1 (all p < 0.001).

Table 1 Proportions of prescriptions using the robotic dispensing system

Period 1 Period 2 Period 3

Number of prescriptions dispensed, n 77,199 51,482 54,822

Total number of medication orders per prescription dis- 158,548 (100%) 106,611 (100%) 114,111 (100%)
pensed, n (%)
Number of medication orders using dispensing devices, n (%)
Drug Station® 0 (0%) 81,073 (76.0%) 87,742 (76.9%)
Mini DimeRo® 0 (0%) 4,380 (4.1%) 4,091 (3.6%)
®
Hp-PORIMS 0 (0%) 9,252 (8.7%) 9,975 (8.7%)
Not used above 3 dispensing devices 158,548 (100%) 11,907 (11.2%) 12,303 (10.8%)
Proportions of prescriptions before and after introducing the robotic dispensing system among three periods of trial are displayed. Period 1 (between March 2020 and
August 2020) is the time before introducing the robotic dispensing system, period 2 (between March 2021 and May 2021) indicates early phase after introducing the
robotic dispensing system, and period 3 (between June 2021 and August 2021) indicates the collaborative phase after introducing the robotic dispensing system
Takase et al. Journal of Pharmaceutical Health Care and Sciences (2022) 8:24 Page 6 of 9

Table 2 Incidence of dispensing errors before and after introducing the robotic dispensing system
Period 1 Period 2 Period 3
Number (%) Number (%) p-values vs Period 1 Number (%) p-values vs Period 1 p-values
vs Period
2

Prescribed medications 158,548 (100%) 106,611 (100%) 114,111 (100%)

Prevented dispensing errors
Wrong quantity 170 (0.107%) 28 (0.026%) < 0.001* 32 (0.028%) < 0.001* 0.897
Wrong strength 83 (0.052%) 3 (0.003%) < 0.001* 0 (0%) < 0.001* 0.113
Wrong drug 39 (0.025%) 5 (0.005%) < 0.001* 1 (0.001%) < 0.001* 0.113
Wrong dosage form 16 (0.010%) 0 (0%) < 0.001* 0 (0%) < 0.001* NA
Others 16 (0.010%) 22 (0.021%) 0.031 17 (0.015%) 0.292 0.339
TOTAL 324 (0.204%) 58 (0.054%) < 0.001* 50 (0.044%) < 0.001* 0.290
Unprevented dispensing errors
Wrong quantity 12 (0.008%) 3 (0.003%) 0.123 0 (0%) 0.002* 0.113
Wrong strength 1 (0.001%) 0 (0%) 1.000 0 (0%) 1.000 NA
Wrong drug 4 (0.003%) 0 (0%) 0.154 0 (0%) 0.145 NA
Wrong dosage form 2 (0.001%) 0 (0%) 0.519 0 (0%) 0.513 NA
Others 5 (0.003%) 2 (0.002%) 0.709 2 (0.002%) 0.707 1.000
TOTAL 24 (0.015%) 5 (0.005%) 0.013* 2 (0.002%) < 0.001* 0.274
The prevented dispensing errors were detected by pharmacists before the medicines were provided from the pharmacy to clinical wards or to outpatients. By
contrast, unprevented dispensing errors indicates errors detected by other medical staff or patients after the medicines were provided from the pharmacy to clinical
wards or to outpatients. The incidences of prevented and unprevented dispensing errors among three periods of trial are displayed. Period 1 (between March 2020
and August 2020) is the period before introducing the robotic dispensing system, period 2 (between March 2021 and May 2021) indicates the early phase after
introducing the robotic dispensing system, and period 3 (between June 2021 and August 2021) indicates the collaborative phase after introducing the robotic
dispensing system. The incidences were presented as the number of errors divided by the numbers of total medication orders in each prescription
Abbreviation: NA Not applicable
*
Statistically significant after adjustment using the Bonferroni correction (p < 0.0167 for Fisher’s exact test)

No significant difference was observed in all types of the and 3, respectively. The work time of pharmacists, robot,
incidences of prevented dispensing errors between peri- and pharmacy support staff in each period are presented
ods 2 and 3. in Table 3. The work time of pharmacists in period 3
The incidences of unprevented dispensing errors by (median of 23 s) was significantly lower than that in
error type were wrong quantity (0.008% vs 0.003% vs periods 1 (median of 60 s) or 2 (median of 69 s) (both
0%), wrong strength (0.001% vs 0% vs 0%), wrong drug p < 0.001). Although the total dispensing time signifi-
(0.003% vs 0% vs 0%), wrong dosage form (0.001% vs 0% cantly increased from periods 1 (median of 60 s) to 2
vs 0%), others (0.003% vs 0.002% vs 0.002%), and total (median of 87 s) (p < 0.001), it recovered to the original
(0.015% vs 0.005% vs 0.002%) in periods 1, 2, and 3, level in period 3 (median of 61 s).
respectively (Table 2). Among them, wrong quantity was
significantly reduced in period 3 compared with that in Discussion
period 1 (p = 0.002). Total unprevented dispensing errors Replacing manual dispensing, which requires human
were significantly reduced in periods 2 and 3 compared resources, with automated dispensing robots is critical
with those in period 1 (p = 0.013 and p < 0.001, respec- for enhancing clinical care for patients by pharmacist.
tively). No significant difference was observed in all Additionally, collaboration with pharmacy support staff
types of the incidences of unprevented dispensing errors is critical. In this study, we established the robotic dis-
between periods 2 and 3. pensing system by using automated dispensing robots
and collaborating with pharmacy support staff; subse-
Effects of the robotic dispensing system reducing quently, we evaluated the safety and the efficiency of
the dispensing time per prescription those systems. The results of this study clearly revealed
The characteristics of prescription used for evaluat- that the incidences of dispensing errors were significantly
ing the dispensing time are presented in Supplemen- reduced immediately after introducing the robotic dis-
tary Table 1. The number of prescriptions for evaluating pensing system (period 2), and these reduced incidences
dispensing time were 223, 184, and 310 in periods 1, 2, were maintained after collaboration with pharmacists
Takase et al. Journal of Pharmaceutical Health Care and Sciences (2022) 8:24 Page 7 of 9

Table 3 Dispensing time per prescription (second) before and after introducing the robotic dispensing system
Period 1 (n = 221) Period 2 (n = 181) Period 3 (n = 310)
Median (IQR), Median (IQR), p–values vs Median (IQR), p–values vs p–values vs Period
seconds seconds Period 1 seconds Period 1 2

Work time
Pharmacists 60 (26–176) 69 (40–130) 0.364 23 (12–48) p < 0.001* p < 0.001*
Robot – 19 (12–38) – 17 (10–32) – –
Pharmacy sup- – – – 21 (10–45) – –
port staff
TOTAL 60 (26–176) 87 (54–167) p < 0.001* 61 (35–140) 0.556 p < 0.001*
®
The work time of robot is defined as the operating time of Drug Station for each prescription. In period 1, the work time of pharmacists (defined as the time required

total time required for preparing medicines using Drug Station® by one pharmacist and dispensing medicines by another pharmacist) were measured. In period 3,
for manually preparing and dispensing medicines by pharmacist) was measured. In period 2, the work time of robot and the work time of pharmacists (including the

the work time of robot and the work time of pharmacy support staff (defined as the time required for preparing medicines using Drug Station® by pharmacy support
staff ) and the work time of pharmacists (defined as the time required for dispensing medicines by pharmacist) were measured
Abbreviation: IQR Interquartile range
*
Statistically significant after adjustment using the Bonferroni correction (p < 0.0167 for Mann–Whitney U test)

and pharmacy support staff (period 3). Additionally, the the results of this study reflect the characteristics of each
dispensing time of pharmacists was significantly reduced system. The error type of others was not changed after
after introducing the system (period 3). To the best of our introducing the robotic dispensing system (periods 2
knowledge, this is the first study that evaluates the safety and 3). These cases included human errors, such as put-
and efficiency of implementing automated dispensing ting medicines into wrong labeled paper bag, and these
robots and a collaborative working model with pharma- cases should be focused on even when we use robotic
cists and pharmacy support staff in Japan. dispensing systems.
The incidence of unprevented dispensing errors in the Similar to prevented dispensing errors, the inci-
previous studies was 0.003–0.047% [11, 12]. The inci- dences of total unprevented dispensing errors were sig-
dence of unprevented dispensing errors in this study at nificantly reduced from 0.015% in period 1 to 0.005% (a
baseline (period 1) was 0.015%, which is consistent with reduction rate of 66.7%) and 0.002% (a reduction rate
previous reports. Thus, the accuracy of dispensing pro- of 86.7%) in periods 2 and 3, respectively. After intro-
cess in our hospital appeared to be within the general ducing the robotic dispensing system (periods 2 and 3),
level in Japan. Therefore, the results in this study can be the absence of unprevented dispensing errors related
generalized to other institutions nationwide. to wrong strength, wrong drug, or wrong dosage form
The incidences of total prevented dispensing errors revealed a remarkable safety benefit.
was significantly reduced from 0.204% in period 1 to The dispensing process is extremely complicated
0.054% (a reduction rate of 73.5%) and 0.044% (a reduc- because pharmacists must select accurate medicines,
tion rate of 78.4%) in periods 2 and 3, respectively. These quantities, strengths, etc., from among more than 2,000
results were consistent among wrong quantity, wrong medicines. These are written on each patient’s pre-
strength, wrong drug, and wrong dosage form. Notably, scription. In addition, dispensing errors can cause seri-
the number of cases of wrong strength, wrong drug, and ous iatrogenic harm to patients. Therefore, the partial
wrong dosage form were nearly zero in period 3. The involvement of pharmacy support staff in the dispens-
error type of wrong quantity, which was the most fre- ing procedure is not accelerated nationwide. Although
quent type of errors in period 1, was reduced to approxi- no original study has been published on collaborative
mately 25% in period 2, and subsequently the incidence work with pharmacists and pharmacy support staff in
of this type of error was similar in period 3. The most Japan, some studies have been published in the world

we used Hp-PORIMS® (data not shown). These results

case of wrong quantity in periods 2 and 3 occurred when [13–17]. Numerous hospitals face significant short-
age of hospital pharmacists, who are necessary for

contrast with Hp-PORIMS®, which requires manual

were consistent with those of a previous study [6]. In enhancing clinical care for patients in Japan. This issue

checking of the number of medicines, Drug Station®

has been highlighted herein for the first time, with the
intention of resolving it. We demonstrated that the
can confirm the number of medicines using the built-in robotic dispensing system enabled the medication dis-
camera and/or built-in electric scale. We consider that pensing duties of the pharmacist to be partially and
Takase et al. Journal of Pharmaceutical Health Care and Sciences (2022) 8:24 Page 8 of 9

safely shared with automated dispensing robots and dispensing robots and pharmacy support staff. Addition-
pharmacy support staff in Japan. ally, after introducing the robotic dispensing system, the
To date, numerous automated dispensing devices have incidences of total prevented or unprevented dispens-
been implemented, and safety and efficiency in the dis- ing errors were reduced by approximately 80% than that
pensing process has been achieved by implementing before introducing the system (from periods 1 to 3).

mated dispensing robot (Drug Station®) and the auto-

these devices [2, 4, 5, 13–19]. We introduced the auto- Among them, the number of cases of wrong strength and
wrong drug, which can cause serious iatrogenic harm to

DimeRo®) [4], the bar-corded dispensing support system

mated dispensing robot for powdered medicines (Mini patients, were reduced to almost zero, and these results

(Hp-PORIMS®) [6]. Although the effect of the latter two

exhibit clinical implications for safe dispensing. The

devices have been reported [4, 6], that of Drug Station®

results of the study suggest that the robotic dispensing
system by using automated dispensing robots and col-
have not been reported. In Japan, press-through package laborating with pharmacy support staff is one of the ways
(PTP, also known as blister pack) sheet, bottle, sachet etc. to enhance clinical care for patients ensuring quality and
are available as pharmaceutical packaging types of oral or safety of medication by pharmacists.
topical medicines. Therefore, a system for preventing dis- This study had some limitations. First, the pre-
pensing errors and improving the efficiency of dispens- scribed medicines and pharmacists were not exactly
ing processes for several types of medicines was required. the same in each period, because of the patient and
Previously, the effect of implementing the automated dis- staff turnover. Secondly, we performed an uncontrolled

tablets or capsules (robo-pick®, Yuyama Co., Ltd., Osaka,

pensing robot, which was only available for PTP sheets of before-after study, because dispensing could not be
randomized after introducing the automated dispens-

tion® can not only store PTP sheets but also other dosage
Japan), was reported in Japan [2]. However, Drug Sta- ing robots. Although this method is often used in other
studies [13–17, 19], it is inherently susceptible to bias
forms (topical medications, etc.) and their various types owing to the lack of a control group. Therefore, a quasi-
of pharmaceutical packaging. Additionally, because the experimental method such as an interrupted time
automated dispensing robot is equipped with the visual series analysis is needed for incorporating long-term
and gravimetric verification of each medicine, we could outcomes in future studies.
delegate part of the dispensing process from pharma-
cists to pharmacy support staff. The results of this study
revealed that the incidence of dispensing errors after Conclusions
introducing the robotic dispensing system (period 2) The robotic dispensing system enabled the process of
were reduced. Additionally, the reduced rate was main- medication dispensing by pharmacist to be partially
tained after starting collaboration with the pharmacists and safely shared with automated dispensing robots
and pharmacy support staff (period 3). Notably, the num- and pharmacy support staff. Therefore, clinical care for
ber of cases of wrong strength and wrong drug, which patients by pharmacist could be enhanced by ensuring
can seriously impact a patient’s health [20], reduced to quality and safety of medication.
almost zero immediately after introducing the robotic
dispensing system. Abbreviations

cation dispensing was shared with Drug Station® and

In this study, the work time of pharmacists in medi- PDA: Personal digital assistance; CPOE: Computerized physician order entry;
P&T Committee: Pharmacy and Therapeutic Committee; PTP: Press-through
package.
pharmacy support staff after introducing the robotic
dispensing system in period 3. Consequently, the work
time of pharmacists in medication dispensing was signifi- Supplementary Information
The online version contains supplementary material available at https://​doi.​
cantly reduced after introducing the robotic dispensing org/​10.​1186/​s40780-​022-​00255-w.
system. The total dispensing time significantly increased
from periods 1 to 2 and recovered to the original level in Additional file 1: Supplementary Table 1. Characteristics of prescription
period 3. Possible causes of these differences included the used for evaluating dispensing time.
unexpected behaviors or inexperienced operators of the
automated dispensing robot in period 2. Thus, we consid- Acknowledgements
ered that the effects of the robotic dispensing system for Not applicable.
the dispensing time were reflected in period 3. Authors’ contributions
To summarize, introducing the robotic dispensing TT, Norio M, NS, HI, NK, and Nobuyuki M conceived and designed this study.
system enabled the process of medication dispensing TT and Norio M collected data. TT analyzed dispensing errors. TT, NS, IN, and
NK analyzed dispensing time. YM, FT, MH, HK, and TH supervised the conduct
by pharmacist to be partially shared with automated of this study. TT, HI, and Nobuyuki M drafted the manuscript, and all authors
Takase et al. Journal of Pharmaceutical Health Care and Sciences (2022) 8:24 Page 9 of 9

contributed substantially to its revision. All authors read and approved the codes and total numbers. Iryo Yakugaku (Jpn J Pharm Health Care Sci).
final manuscript. 2008;34:997–1003.
12. Muramatsu H, lketani O, Kaneko K, Tsuda S, Hayakawa T, Yamayoshi Y, et al.
Funding Development of a bar-code dispensing system for drugs associated with
Not applicable. a high risk for adverse events. J Jpn Soc Hosp Pharm. 2012;48:337–40.
13. Temple J, Ludwig B. Implementation and evaluation of carousel dispens-
Availability of data and materials ing technology in a university medical center pharmacy. Am J Health Syst
All data generated or analyzed during this study are included in this published Pharm. 2010;67:821–9.
article. 14. Lin AC, Huang YC, Punches G, Chen Y. Effect of a robotic prescription-
filling system on pharmacy staff activities and prescription-filling time.
Am J Health Syst Pharm. 2007;64:1832–9.
Declarations 15. Oswald S, Caldwell R. Dispensing error rate after implementation of
an automated pharmacy carousel system. Am J Health Syst Pharm.
Ethics approval and consent to participate 2007;64:1427–31.
Not applicable. 16. Rodriguez-Gonzalez CG, Herranz-Alonso A, Escudero-Vilaplana V, Ais-
Larisgoitia MA, Iglesias-Peinado I, Sanjurjo-Saez M. Robotic dispensing
Consent for publication improves patient safety, inventory management, and staff satisfaction in
Not applicable. an outpatient hospital pharmacy. J Eval Clin Pract. 2019;25:28–35.
17. Berdot S, Korb-Savoldelli V, Jaccoulet E, Zaugg V, Prognon P, Le LMM, et al.
Competing interests A centralized automated-dispensing system in a French teaching hos-
The authors declare that they have no competing interests. pital: return on investment and quality improvement. Int J Qual Health
Care. 2019;31:219–24.
Author details 18. Chapuis C, Roustit M, Bal G, Schwebel C, Pansu P, David-Tchouda S, et al.
1
Department of Pharmacy, Kobe City Medical Center General Hospital, 2‑1‑1, Automated drug dispensing system reduces medication errors in an
Minatojima Minamimachi, Chuo‑ku, Kobe, Hyogo 650‑0047, Japan. 2 Deloitte intensive care setting. Crit Care Med. 2010;38:2275–81.
Analytics, Deloitte Touche Tohmatsu LLC, 3‑2‑3, Marunouchi, Chiyoda‑ku, 19. James KL, Barlow D, Bithell A, Hiom S, Lord S, Pollard M, et al. The impact
Tokyo 100‑8360, Japan. 3 Social Solution Initiative, Osaka University, 2‑8, of automation on workload and dispensing errors in a hospital pharmacy.
Yamadaoka, Suita, Osaka 565‑0871, Japan. 4 Research Center On Ethical, Legal Int J Pharm Pract. 2013;21:92–104.
and Social Issues, Osaka University, 2‑8, Yamadaoka, Suita, Osaka 565‑0871, 20. Tsuji T, Irisa T, Tagawa S, Kawashiri T, Ikesue H, Kokubu C, et al. Differences
Japan. in recognition of similar medication names between pharmacists and
nurses: a retrospective study. J Pharm Health Care Sci. 2015;1:19.
Received: 20 April 2022 Accepted: 7 August 2022

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in pub-
lished maps and institutional affiliations.
References
1. Ministry of Health, Labour and Welfare. Official notification. 2019. https://​
www.​mhlw.​go.​jp/​conte​nt/​00049​8352.​pdf. Accessed 21 Aug 2021.
2. Hamada A, Sakaguchi N, Okamoto K, Senoo N, Hosoda T. Reduction
of dispensing errors by introducing optical character readers and
automated tablet dispensing machines to community pharmacies. Jpn J
Pharm Health Care Sci. 2014;40:174–9.
3. Tsuji T, Irisa T, Ohata S, Kokubu C, Kanaya A, Sueyasu M, et al. Relationship
between incident types and impact on patients in drug name errors: a
correlational study. J Pharm Health Care Sci. 2015;1:11.
4. Kobayashi K, Uenoyama K, Ito T, Takahashi T, Kondo A, Tikatani H, et al.
Efforts in facilitating work through the introduction of a powdered medi-
cine dispensing robot. J Jpn Soc Hosp Pharm. 2019;55:402–8.
5. Tabata H, Kaya N, Inagaki Y, Shibanami A. Survey on use of a dispensing
medicine robot (DimeRo) and considerations for efficient operation. J Jpn
Soc Hosp Pharm. 2018;54:175–9.
6. Kanzaki H, Tanaka Y, Konuma T, Nishihara S, Manabe Y, Inoue T, et al.
Prevention of drug dispensing errors by using personal digital assistance
and recording the number of agents. Jpn J Pharm Health Care Sci.
2017;43:430–7.
7. Aldhwaihi K, Schifano F, Pezzolesi C, Umaru N. A systematic review of
the nature of dispensing errors in hospital pharmacies. Integr Pharm Res Ready to submit your research ? Choose BMC and benefit from:
Pract. 2016;5:1–10.
8. Flynn EA, Barker KN, Gibson JT, Pearson RE, Berger BA, Smith LA. Impact of • fast, convenient online submission
interruptions and distractions on dispensing errors in an ambulatory care • thorough peer review by experienced researchers in your field
pharmacy. Am J Health Syst Pharm. 1999;56:1319–25. • rapid publication on acceptance
9. Nickman NA, Drews FA, Tyler LS, Kelly MP, Ragsdale RJ, Rim M. Use of
multiple methods to measure impact of a centralized call center on aca- • support for research data, including large and complex data types
demic health system community pharmacies. Am J Health Syst Pharm. • gold Open Access which fosters wider collaboration and increased citations
2019;76:353–9. • maximum visibility for your research: over 100M website views per year
10. Flynn EA, Barker KN, Carnahan BJ. National observational study of pre-
scription dispensing accuracy and safety in 50 pharmacies. J Am Pharm At BMC, research is always in progress.
Assoc (Wash). 2003;43:191–200.
11. Katakura M, Toda N, Kunimoto Y, Takahashi K, Nakamura K, Masuko H, Learn more biomedcentral.com/submissions
et al. Error prevention through written check of medicine identification