Eye-tracking-based analysis to improve the efficiency and safety of prescription audit

Eye-tracking-based analysis to improve the efficiency and safety of prescription audit | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Eye-tracking-based analysis to improve the efficiency and safety of prescription audit Toshikazu Tsuji, Masayuki Tanaka, Shiori Iwane, Naoko Kizu, Shigeru Hasebe, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8139848/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Pharmacists need to explore more efficient method of prescription audits to ensure safe and reliable medical therapy for patients. In this study, pharmacists’ gaze movements during the prescription audit process were analyzed using an eye-tracking system. Methods First, two methods for displaying drug’s dose and usage during prescription audits were developed: “set phrase display” and “numerical table display.” Second, three key items were defined: (a) drug name, (b) drug dose and usage, and (c) dose period. The number of gaze points for these items were designated as Gaze 1, Gaze 2, and Gaze 3, and the time required for verification was recorded as Time. Third, three pairs of models (A 1 -A 2 , B 1 -B 2 , and C 1 -C 2 ) with different levels of difficulty for prescription auditing were prepared, along with one pair of models (Q 1 -Q 2 ) that included a drug–drug interaction. This study aimed to demonstrate the differences in pharmacists' information processing between the two expression methods (set-phrase display and numerical table display) using four pairs of prescription auditing models. Results A total of 22 pharmacists participated in the study. During the prescription audit process, gaze movements followed the pattern “set phrase display > numerical table display” in all three models (A 1 -A 2 , B 1 -B 2 , and C 1 -C 2 ). Significant differences between model pairs were observed in Gaze 2 and Time, both favoring “set phrase display > numerical table display”. Furthermore, significant differences in gaze movements between models Q 1 -Q 2 were consistent with those observed in the three model pairs. Recognition rates for models Q 1 -Q 2 were 72.7% (8/11) and 100% (11/11), respectively. Recognition times were 20.5 ± 6.2 and 16.0 ± 3.8, respectively. No significant differences were observed between them. Conclusions The results demonstrated that the “numerical table display” method not only enables prescription audit to be performed without unnecessary procedures but also improves the accuracy of drug–drug interaction judgments. In other words, introducing the “numerical table display” method for drug dose and usage can enhance both efficiency and safety in prescription audit. Eye-tracking method prescription audit thought process efficiency and safety set phrase display numerical table display Figures Figure 1 Figure 2 Figure 3 Figure 4 Background Many medical institutions, including Kyushu University Hospital, have implemented initiatives to prevent dispensing errors caused by pharmacists, and several successful outcomes have been reported to date [ 1 – 15 ]. Pre-checking prescription content (hereinafter referred to as prescription audits) is also essential for ensuring safe and reliable medical therapy for patients. Pharmacists conduct prescription audits accurately and efficiently within a predetermined timeframe to prevent duplicate medications and unsafe dosages. For this reason, pharmacists need to explore more efficient methods of prescription auditing to ensure safe and reliable medical therapy for patients. However, few studies have examined improvements in prescription auditing efficiency and patient safety [ 16 – 19 ]. Considering these circumstances, improving efficiency and safety in prescription auditing first requires that pharmacists understand their own thought processes during auditing. In our previous studies, we used an eye-tracking module that followed individual’s gaze movements in real time from camera-captured video and clarified the thought processes of 12 or 22 pharmacists in simulated dispensing environments [ 20 – 23 ]. Our recent findings revealed that introducing visual information using colors or symbols (➁| ▮--|) into drug location information in prescription content not only increased dispensing efficiency but also reduced the dispensing errors. In this study, the gaze movements of pharmacists during prescription audits were analyzed using an eye-tracking system. First, two methods for indicating drug’s dose and usage in prescription audit were prepared: the widely used “set phrase display (1 tablet at a time, once a day, after breakfast)” and the unprecedented “numerical table display (- | 1 - - | -)”. Second, three pairs of prescription audit models with different levels of difficulty were prepared, along with one pair of models that included a drug–drug interaction. This approach enabled analysis of differences in both efficiency and safety between “set phrase display” and “numerical table display,” in relation to drug dose and usage display methods. Color recognition and processing are closely linked to human brain function [ 24 , 25 ], but differences in information processing between letters and numbers are not fully understood. Therefore, our goal was to demonstrate the difference in pharmacists' information processing between “set phrase display” and “numerical table display” using four pairs of prescription audit models. Methods Analysis of gaze movements using a visual line tracing system Eye-tracking, which uses corneal reflection of infrared rays to analyze gaze points and eye movements, is applied in many fields, including medicine, psychology, and pedagogy [26–29]. In this study, pharmacists’ gaze movements during prescription audits were examined using a wearable eye tracker (Tobii Pro Glasses 3; Tobii Technology K.K.). Gaze movements were classified into two categories: fixation (stationary for ≥100 ms) and saccade (rapid eye movements). Fixations and saccades were analyzed using motion videos recorded with dedicated software (Tobii Pro Lab Analyzer, Tobii Technology K.K.). Target pharmacists to participate in this study The inclusion criteria for pharmacists were as follows: First, to ensure accurate eye-movement measurement, pharmacists needed to read prescription information clearly on a large monitor while wearing soft contact lenses or with the naked eye. Second, pharmacists were required to have at least 12 months of prescription audit experience at Kyushu University Hospital to ensure high verification quality. Finally, pharmacists had to provide informed consent to participate in the study. Target drugs used in the prescription audit Nineteen drugs dispensed at Kyushu University Hospital were used in this study: Amlodipine OD 5 mg, Pravastatin Sodium 10 mg, Eliquis ® 5 mg, Feburic ® 20 mg, Brotizolam OD 0.25 mg, Famotidine OD 20 mg, Medrol ® 4 mg, Aspara-CA 200 mg, Pitavastatin Calcium OD 2 mg, Geninax ® 200 mg, Methotrexate 2 mg, Foliamin ® 5 mg, Predonine ® 5 mg, Rosuvastatin OD 2.5 mg, Magnesium Oxide 250 mg, MS Contin ® 10 mg, Calonal ® 200 mg, Tegretol ® 100 mg, and Levofloxacin 250 mg. Verification slides The slides verifying prescription audits were prepared using Microsoft PowerPoint ® 2016. Each slide contained basic patient information at the top, including name, age (sex), body weight, height, and creatinine clearance. Prescription information was displayed in the center of each slide and consisted of three items: (a) drug name, (b) drug dose and usage, and (c) dose period. In this study, four pairs of models (A 1 -A 2 , B 1 -B 2 , C 1 -C 2 , and Q 1 -Q 2 ) were prepared, and five target drugs were included in each verification. Importantly, the method of indicating (b) drug dose and usage on prescription slides was classified into two types: “set phrase display” and “numerical table display.” In the “set phrase display,” prescription contents were presented as words and phrases, such as “1 tablet at a time, once a day, after breakfast.” In the “numerical table display,” prescription contents were shown using only numerical figures, such as “ - | 1 - - | -,” and explanatory comments about drug administration after meals were omitted. The specifications of the prescription audit information are presented in Table 1, and the order of verification was randomized. Definition of the four pairs of models First, we set up three pairs of models (A 1 -A 2 , B 1 -B 2 , and C 1 -C 2 ) to compare the difference in efficiency according to the difficulty level of prescription audit between “set phrase display” and “numerical table display.” Second, we set up another pair of models (Q 1 -Q 2 ) to compare the difference in safety related to prescription content that included a drug interaction between “set phrase display” and “numerical table display.” A summary of the four pairs (A 1 -A 2 , B 1 -B 2 , C 1 -C 2 , and Q 1 -Q 2 ) is provided below: Low-difficulty models (A 1 -A 2 ): The dosage and administration were appropriate, and there were no drug interactions among the five drugs. Moderate-difficulty models (B 1 -B 2 ): The dosage and administration were appropriate, but prescription content was a little complicated as it contained a drug of unequal quantities. High-difficulty models (C 1 -C 2 ): The dosage and administration were appropriate, but prescription content was complicated as it contained a drug of unequal quantities and two drugs for the designated days of the week. Prescription question models (Q 1 -Q 2 ): There was an obvious drug interaction between the two drugs (magnesium oxide 250 mg and levofloxacin 250 mg). Verification procedure An outline of the verification task using the eye-tracking method is shown in Figure 1. Two notebook computers connected to 27-inch monitors were used to display the verification slides. The prescription audit area (34 × 60 cm) was shown on the main monitor (prescription audit monitor), while the side monitor for prescription question (prescription question monitor) was positioned to the left of the prescription audit monitor. A pharmacist wearing an eye tracker was positioned 100 cm from the prescription audit monitor, and their gaze movements were investigated during the prescription audit process. The Tobii Pro Lab Analyzer, which records motion videos, was used to assess parameters such as gaze point (circle center), gaze duration (circle size), and movement of sight lines (line between circle centers). The eye tracker was calibrated to ensure accuracy before a series of verification tasks was conducted. Pharmacists practiced with several training slides in advance to familiarize themselves with the verification process. The primary aim of the verification task was to maintain a smooth prescription audit. If the pharmacists identified an issue in the prescription audit, they point to the prescription question monitor. The six major steps for dispensing verification are as follows. 1) The pharmacist maintained gaze at a specified position. 2) When the pharmacist indicated the “Next” signal, the assistant switched to the next prescription audit slide. 3) The pharmacist verified the dosage and administration of each drug and checked for interactions among the five drugs. 4) When a prescription question was deemed necessary, the pharmacist pointed to the prescription question monitor. 5) After the pharmacist signaled “Next”, the assistant switched to a rest slide. 6) A sequence of verifications using 10 or more prescription audit slides was repeated, with sufficient breaks taken as needed. Verification items and classifications Verification of prescription audits using the four pairs of models (A 1 -A 2 , B 1 -B 2 , C 1 -C 2 , and Q 1 -Q 2 ) requires pharmacists to visually confirm three key items: (a) drug name, (b) drug dose and usage, and (c) dose period. Furthermore, pharmacists had to shift their gaze to assess the interactions among the five drugs accurately. In this study, verification items were defined to compare gaze points and verification time, and differences in gaze movements of pharmacists were analyzed between two display methods for (b) drug dose and usage: “set phrase display” and the “numerical table display.” A summary of the four classifications (Gaze 1, Gaze 2, Gaze 3, and Time) is provided below. Gaze 1: Total number of gaze points in the area of (a) drug name Gaze 2: total number of gaze points in the area of (b) drug dose and usage Gaze 3: total number of gaze points in the area of (c) dose period Time: total time required to verify the prescription audit for the five target drugs. Measurement of prescription question In this study, a drug interaction between magnesium oxide 250 mg and levofloxacin 250 mg was established using the prescription question models Q 1 –Q 2 . The content of the drug interaction was the reduced absorption of levofloxacin caused by the simultaneous administration of magnesium oxide. First, the presence or absence of a prescription question was determined based on each pharmacist’s performance in indicating the prescription question monitor. Second, the rate of recognizing the drug interaction between magnesium oxide and levofloxacin was calculated as “Recognition rate”. This calculation was performed by dividing the number of pharmacists who appropriately judged the drug interaction by the total number of target pharmacists. Furthermore, the time required to recognize the drug interaction was measured as “Recognition time.” A summary of the two classifications showing the recognition states (Recognition rate and Recognition time) is provided below. Recognition rate: the proportion of pharmacists who recognized a drug interaction between magnesium oxide and levofloxacin Recognition time: the time required to recognize a drug interaction between magnesium oxide and levofloxacin Data analysis Using the gaze category data (fixation and saccade) from the recorded motion video, we analyzed six classifications: Gaze 1, Gaze 2, Gaze 3, Time, Recognition rate, and Recognition time. Data are presented as the mean ± standard deviation. Significant differences in the four classifications (Gaze 1, Gaze 2, Gaze 3, and Time) between the model pairs A 1 -A 2 , B 1 -B 2 , and C 1 -C 2 were analyzed between the same pharmacists using the paired t -test. Significant differences in the five classifications (Gaze 1, Gaze 2, Gaze 3, Time, and Recognition time) between the Q 1 -Q 2 models were analyzed across different pharmacists using Student’s t-test, while difference in another classification (Recognition rate) was analyzed using Fisher's exact test. A P- value <0.05 was considered statistically significant, whereas P- values <0.01 and <0.001 were considered highly significant. Statistical analyses were performed using JMP Pro 16 software. Results Basic information of pharmacists A total of 22 pharmacists (9 men and 13 women; mean age, 30.1±5.9 years) participated in this study. Data from these 22 pharmacists were analyzed for each model pair (A 1 -A 2 , B 1 -B 2 , and C 1 -C 2 ) using the paired t-test. Also, pharmacists participating in Q 1 and Q 2 verifications were 11 (4 men and 7 women) and 11 (5 men and 6 women) with an average age of 28.9±4.7 and 31.3±6.9 years, respectively. Data from the 11 pharmacists participating in Q 1 and Q 2 verifications were analyzed using Student’s t-test or Fisher's exact test. Comparison of four classifications between two display methods using three pairs of models The difference in gaze movements of 22 pharmacists during the prescription audit process between “set phrase display” and “numerical table display” was analyzed using three model pairs (A 1 -A 2 , B 1 -B 2 , and C 1 -C 2 ). The data for the four classifications (Gaze 1, Gaze 2, Gaze 3, and Time) are presented below, and the relationships between each pair of models are shown in Figure 2. Significant differences between models A 1 -A 2 were observed in Gaze 2 and Time using the paired t -test ( P 0.72, P 0.79, P <0.001), whereas no strong positive correlation was observed among the other classifications. Similarly, significant differences between models B 1 -B 2 were observed in Gaze 2 and Time ( P 0.75, P 0.70, P <0.001). Furthermore, significant differences between models C 1 and C 2 were observed in gaze 2 and time ( P <0.0001 and P 0.86, P 0.83, P <0.001). Model A 1 : Gaze 1, 10.5±3.3; Gaze 2, 18.2±5.1; Gaze 3, 5.5±1.4; Time, 14.9±4.2 Model A 2 : Gaze 1, 12.2±5.4; Gaze 2, 9.3±3.7; Gaze 3, 5.1±2.2; Time, 12.5±13.9 Model B 1 : Gaze 1, 16.9±5.9; Gaze 2, 26.3±11.0; Gaze 3, 6.4±3.2; Time, 20.5±5.6 Model B 2 : Gaze 1, 19.1±8.3; Gaze 2, 14.0±6.0; Gaze 3, 6.7±2.8; Time, 17.7±6.6 Model C 1 : Gaze 1, 15.5±7.4; Gaze 2, 35.5±15.1; Gaze 3, 8.0±4.0; Time, 24.8±8.0 Model C 2 : Gaze 1, 17.4±6.4; Gaze 2, 21.4±9.9; Gaze 3, 7.6±4.6; Time, 19.7±6.3 Comparison of gaze averages per necessary spot in area (b) between two display methods using three pairs of models The difference in gaze points focusing on area (b) drug dose and usage was analyzed by using low-, moderate-, and high-difficulty models (A 1 -A 2 , B 1 -B 2 , and C 1 -C 2 ), since this area was the source of differences in gaze movements between “set phrase display” and “numerical table display.” In the area of (b) drug dose and usage, the number of necessary gaze spots that must be visually recognized to perform prescription audit (necessary spot) and the average of gaze points per necessary spot (gaze average) for each model pair are as follows. The relationship between each pair of models is shown in Figure 3. Significant differences in gaze averages per necessary spot were observed between models A 1 -A 2 using the paired t -test ( P =0.0110), but not between the other two models. Low-difficulty models A 1 -A 2 : Necessary spots (15, 6); Gaze averages (1.22±0.34, 1.55±0.61) Moderate-difficulty models B 1 -B 2 : Necessary spots (15, 7); Gaze averages (1.75±0.73, 2.00±0.85) High-difficulty models C 1 -C 2 : Necessary spots (17, 10); Gaze averages (2.09±0.89, 2.14±0.99) Comparison of four classifications and two recognition states between two display methods using prescription question models The difference in safety during the prescription audit process between “set phrase display” and “numerical table display” was analyzed using a pair of prescription question models (Q 1 -Q 2 ). The data for the four classifications (Gaze 1, Gaze 2, Gaze 3, and Time) are shown below, and the relationships between models Q 1 -Q 2 are shown in Figure 4a. Significant differences between the 11 pharmacists were observed in Gaze 2 and Time using Student’s t-test ( P <0.0001 and P = 0.0336, respectively), and no strong positive correlations were found among the four classifications in either model Q 1 or Q 2 . Model Q 1 : Gaze 1, 15.5±6.9; Gaze 2, 32.8±11.0; Gaze 3, 5.1±3.1; Time, 23.8±6.1 Model Q 2 : Gaze 1, 17.9±4.9; Gaze 2, 15.3±4.8; Gaze 3, 5.8±3.4; Time, 18.7±4.0 Furthermore, the number of pharmacists who recognized a drug interaction between magnesium oxide and levofloxacin was 8 and 11 in models Q 1 and Q 2 , respectively. Recognition rates were 72.7% (8/11) and 100% (11/11), respectively; no significant difference was observed using Fisher’s exact test ( P = 0.2143). Recognition times in models Q 1 and Q 2 were 20.5±6.2 (n=8) and 16.0±3.8 (n=11), respectively; and no significant difference between them was found using Student’s t-test ( P = 0.0651). The data for the two recognition states (Recognition rate and Recognition time) are presented below, and the relationships between models Q 1 -Q 2 are shown in Figure 4b. Model Q 1 : Recognition rate, 72.7% (8/11); Recognition time, 20.5±6.2 (n=8) Model Q 2 : Recognition rate, 100% (11/11); Recognition time, 16.0±3.8 (n=11) Discussion In this study, we aimed to elucidate the thought processes of pharmacists during prescription audits using an eye-tracking system. The gaze movements of 22 pharmacists were compared between two methods of indicating drug dose and usage—set-phrase display and numerical table display—using four pairs of prescription audit models (A 1 -A 2 , B 1 -B 2 , C 1 -C 2 , and Q 1 -Q 2 ). The findings demonstrated that the “numerical table display” method allowed prescription audits to be performed more efficiently and improved the accuracy of judging drug interactions. In summary, these results suggest that pharmacists can conduct prescription audit work more effectively and safely by utilizing the “numerical table display” method. When using the “set phrase display” method (e.g., 1 tablet at a time, once a day, after breakfast), pharmacists are forced to grasp the overall drug’s dose and usage while memorizing multiple words and phrases separately, since they must read and understood them one by one. Furthermore, if there is a drug interaction in the prescription, the pharmacist must check the timing of administration by moving the visual lines not only vertically but also diagonally. In summary, prescription audits using the “set phrase display” method require more complex gaze movements, which may increase the risk of losing stored memory. In contrast, when using the “numerical table display” method (e.g., - | 1 - - | -), pharmacists can grasp the drug’s dose and usage by processing them as a numerical table and can check for drug interactions by just moving their gaze vertically. According to research on human memory capacity, short-term memory typically holds 7±2 items, but this can be reduced to 4±1 items when information is complex or when interference occurs [30–32]. In other words, the key point of this study was that pharmacists’ capacity to process prescription information varied depending on whether drug dose and usage were expressed as a combination of words and phrases (set phrase display) or as numerical figures (numerical table display). First, significant differences in the four classifications (Gaze 1, Gaze 2, Gaze 3, and Time) when using the low-difficulty models A 1 -A 2 were observed in Gaze 2 and Time, with A 1 >A 2 ( P B 2 , P C 2 , P <0.0001, P <0.0001). These results indicate that prescription audits using the “numerical table display” method was more efficient than those using “set phrase display,” regardless of the model’s difficulty levels (Figure 2). Furthermore, strong positive correlations were observed between Gaze 2 and Time in models A 1 , A 2 , B 1 , B 2 , C 1 , and C 2 (r>0.70, P <0.001, respectively), suggesting that considerable time was spent processing information on drug dosage and usage during prescription audits. Therefore, it is assumed that the decrease of Gaze 2 (gaze points in the area of drug dose and usage) caused by using the “numerical table display” method had a great impact on the reduction of Time (time required to perform prescription audits). However, it is unclear why Gaze 1 (gaze points in the area of drug name) was slightly higher with the “numerical table display” than with the “set phrase display” in each model, suggesting that the reduced workload in area (b) may have led to the polite check in area (a). Second, when focusing on the gaze averages per necessary spot in area (b) drug dose and usage, there was a significant difference between models A 1 -A 2 ( P =0.0110) but not between the other two pairs of models. These results indicate that the confirmation frequency per necessary spot using the “numerical table display” method was equal to or higher than that of the “set phrase display” method, regardless of the model’s difficulty levels (Figure 3). Although the details of these relationships according to the model’s difficulty levels are unclear, the necessary spots in the “numerical table display” consisted only of numerical figures, and there were also fewer spots to check, suggesting that pharmacists could select the necessary items that need to be cross-checked and confirm them more prudently. Another factor may be that numerous figures indicating drug’s dose and usage are arranged in one vertical row, making it easier for pharmacists to judge whether there are any drug interactions. For these reasons, the use of the “numerical table display” method appears to enable pharmacists to confirm target points more carefully and in less time. Third, when using the prescription question models Q 1 -Q 2 , significant differences in the four classifications (Gaze 1, Gaze 2, Gaze 3, and Time) were observed for Gaze 2 and Time, with Q 1 >Q 2 ( P <0.0001, P =0.0336). These results demonstrate that prescription audits using the “numerical table display” method was performed more efficiently than with the “set phrase display,” even when drug interactions were present in the prescription (Figure 4a). Conversely, for the two recognition states (Recognition rate, Recognition time) in models Q 1 and Q 2 , the formers were 72.7% (8/11) and 100% (11/11), while the latter were 20.5±6.2 and 16.0±3.8, with no significant differences observed (Figure 4b). However, it is important to note that all 11 pharmacists were aware of the drug interaction in model Q 2 , moreover, the task was performed in less time ( P =0.0651). In other words, these results suggest that pharmacists could conduct prescription audits more effectively and safely by using the “numerical table display” method. This study has some limitations. First, the “numerical table display” method used in this study may not apply to electronic medical charts across all medical environments. Second, when using the “numerical table display” method, explanatory comments for drugs taken before, immediately before, or just after meals need to be written in the remarks column. However, to the best of our knowledge, this is the first study to evaluate the thought processes of more than 20 pharmacists during prescription audit tasks and to clarify both efficiency and safety using an eye-tracking system. Conclusions We analyzed the differences in the gaze movements of 22 pharmacists between two methods–set phrase display and numerical table display–in prescription audits using an eye-tracking system. Employing four model pairs (A 1 -A 2 , B 1 -B 2 , C 1 -C 2 , and Q 1 -Q 2 ) allowed us to elucidate pharmacists’ thought processes across multiple prescription audit patterns. The results demonstrated that the use of the “numerical table display” method not only enables prescription audits to be performed without the complicated procedures but also improves the accurate judgement of drug interactions. In other words, by introducing the “numerical table display” method into the area of drug dose and usage, it is possible to enhance both efficiency and safety in prescription audits. Declarations Ethics approval and consent to participate This study was approved by the Clinical Trials Ethics Committees of Kyushu University Hospital (approval number: 20222011) and Setsunan University (approval number: 2023-018). After explaining the research content, the pharmacists provided written informed consent. This study complied with the Ethical Guidelines for Medical and Health Research Involving Human Participants. Consent for publication Not applicable. Availability of data and materials All data generated or analyzed during this study are included in this published article. Competing interests The authors declare that they have no competing interests. Funding Not applicable. Authors’ contributions T.T.,M.T., and S.Iwa. conducted analyses and drafted the manuscript. S.H., N.K., M.K., and K.N. performed statistical analyses. S.Ishi., T.H., H.W., and M.U. revised the manuscript. All the authors discussed the results and approved the final version of the manuscript. Acknowledgments Not applicable. Author information 1 Laboratory of Clinical Pharmacy, Faculty of Pharmaceutical Sciences, Setsunan University, Osaka, Japan 2 Department of Social Pharmacy, Faculty of Pharmaceutical Sciences, Setsunan University, Hirakata, Osaka, Japan 3 Department of Pharmacy, Kyushu University Hospital, Fukuoka, Japan 4 Department of Pharmacy, Fukuoka Tokushukai Hospital, Fukuoka, Japan References Anto B, Barlow D, Oborne CA, Whittlesea C. Incorrect drug selection at the point of dispensing: a study of potential predisposing factors. Int J Pharm Pract. 2011;19:51–60. https://doi.org/10.1111/j.2042-7174.2010.00072.x Beso A, Franklin BD, Barber N. The frequency and potential causes of dispensing errors in a hospital pharmacy. Pharm World Sci. 2005;27:182–90. https://doi.org/10.1007/s11096-004-2270-8 Ashcroft DM, Quinlan P, Blenkinsopp A. Prospective study of the incidence, nature and causes of dispensing errors in community pharmacies. Pharmacoepidemiol Drug Saf. 2005;14:327–32. https://doi.org/10.1002/pds.1012 Cheung KC, Bouvy ML, De Smet PA. Medication errors: the importance of safe dispensing. Br J Clin Pharmacol. 2009;67:676–80. https://doi.org/10.1111/j.1365-2125.2009.03428.x Lambert BL. Predicting look-alike and sound-alike medication errors. Am J Health Syst Pharm. 1997;54:1161–71. https://doi.org/10.1093/ajhp/54.10.1161 Lambert BL, Lin SJ, Chang KY, Gandhi SK. Similarity as a risk factor in drug-name confusion errors: the look-alike (orthographic) and sound-alike (phonetic) model. Med Care. 1999;37:1214–25. https://doi.org/10.1097/00005650-199912000-00005 Lambert BL, Chang KY, Lin SJ. Descriptive analysis of the drug name lexicon. Drug Inf J. 2001;35:163–72. https://doi.org/10.1177/009286150103500119 Lambert BL, Chang KY, Lin SJ. Effect of orthographic and phonological similarity on false recognition of drug names. Soc Sci Med. 2001;52:1843–57. https://doi.org/10.1016/S0277-9536(00)00301-4 Lambert BL, Donderi D, Senders JW. Similarity of drug names: comparison of objective and subjective measures. Psychol Mark. 2002;19:641–61. https://doi.org/10.1002/mar.10029 Katakura M, Toda N, Kunimoto Y, Takahashi K, Nakamura K, Masuko H, et al. Error prevention through written check of medicine identification codes and total numbers. Jpn J Pharm Health Care Sci. 2008;34:997–1003. https://doi.org/10.5649/jjphcs.34.997 Niwa T, Komori Y, Ishihara M, Takahashi S, Matsuura K, Goto C. Evaluation of the countermeasures for the prevention of dispensing errors with risk analysis. J Jpn Soc Hosp Pharm. 2008;44:1487–90 Ikeda K, Kominami S, Morita S, Kita H, Uno M. Effectiveness of dedicated dispensing paper in preventing dispensing errors. Jpn J Pharm Health Care Sci. 2011;37:431–6. https://doi.org/10.5649/jjphcs.37.431 Watanabe H, Yoshida M, Nakahara A, Futagami S, Onoue R, Tsuji T, et al. Measures for prevention of dispensing errors based on ISO 9001 certified management system and their evaluation. Jpn J Pharm Health Care Sci. 2006;32:824–34. https://doi.org/10.5649/jjphcs.32.824 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. https://doi.org/10.1186/s40780-015-0011-x Tsuji T, Irisa T, Tagawa S, Kawashiri T, Ikesue H, Kokubu C, et al. Differences in recognition of similar medication names between pharmacists and nurses: a retrospective study. J Pharm Health Care Sci. 2015;1:19. https://doi.org/10.1186/s40780-015-0017-4 Morrison C, MacRae Y. Promoting safer use of high-risk pharmacotherapy: impact of pharmacist-led targeted medication reviews. Drugs Real World Outcomes. 2015;2:261–71. https://doi.org/10.1007/s40801-015-0031-8 Hammar T, Hellström L, Ericson L. The use of a decision support system in Swedish pharmacies to identify potential drug-related problems—effects of a national intervention focused on reviewing elderly patients’ prescriptions. Pharmacy (Basel). 2020;8:118. https://doi.org/10.3390/pharmacy8030118 Shermock SB, Shermock KM, Schepel LL. Closed-loop medication management with an electronic health record system in U.S. and Finnish hospitals. Int J Environ Res Public Health. 2023;20:6680. https://doi.org/10.3390/ijerph20176680 Nagata K, Tsuji T, Suetsugu K, Muraoka K, Watanabe H, Kanaya A, et al. Detection of overdose and underdose prescriptions—an unsupervised machine learning approach. PLOS One. 2021;16:e0260315. https://doi.org/10.1371/journal.pone.0260315 Tsuji T, Nagata K, Sasaki K, Suetsugu K, Watanabe H, Kanaya A, et al. Elucidation of occurrence mechanism of dispensing errors based on gaze movement of pharmacists using eye tracking method. Jpn J Pharm Health Care Sci. 2019;45:493–503. https://doi.org/10.5649/jjphcs.45.493 Tsuji T, Nagata K, Sasaki K, Matsukane R, Ishida S, Kawashiri T, et al. Analysis of the thinking process of pharmacists in response to changes in the dispensing environment using the eye-tracking method. J Pharm Health Care Sci. 2022;8:23. https://doi.org/10.1186/s40780-022-00254-x Tsuji T, Nagata K, Tanaka M, Hasebe S, Yukita T, Uchida M, et al. Eye-tracking-based analysis of pharmacists’ thought processes in the dispensing work: research related to the efficiency in dispensing based on right-brain thinking. J Pharm Health Care Sci. 2024;10:21. https://doi.org/10.1186/s40780-024-00341-1 Tsuji T, Nagata K, Tanaka M, Shiori I, Hasebe S, Yuto N, et al. Right-brain utilization in pharmacists’ dispensing processes: an eye-tracking analysis of efficiency and safety using error-induction models. J Pharm Health Care Sci. 2025;11:13. https://doi.org/10.1186/s40780-025-00443-4 Barnett KJ. Colour knowledge: the role of the right hemisphere in colour processing and object colour knowledge. Laterality. 2008;13:456–67. https://doi.org/10.1080/13576500802146387 Pinna B, Deiana K. New conditions on the role of color in perceptual organization and an extension to how color influences reading. Psihologija. 2014;47:319–51. https://doi.org/10.2298/PSI1403319P Kato Y, Kobayashi T, Hasebe D, Kano H, Saito C. A study in subjective evaluation and gaze point analysis of facial symmetry: analysis using eye-tracking. Jpn J Jaw Deform. 2009;19:184–92. https://doi.org/10.5927/jjjd.19.184 Vervoort T, Trost Z, Prkachin KM, Mueller SC. Attentional processing of other’s facial display of pain: an eye tracking study. Pain. 2013;154:836–44. https://doi.org/10.1016/j.pain.2013.02.017 Lim JZ, Mountstephens J, Teo J. Emotion recognition using eye-tracking: taxonomy, review and current challenges. Sensors (Basel). 2020;20:2384. https://doi.org/10.3390/s20082384 Wolf A, Ueda K. Contribution of eye-tracking to study cognitive impairments among clinical populations. Front Psychol. 2021;12:590986. https://doi.org/10.3389/fpsyg.2021.590986 Ishibashi A. Human factors and error prevention. J Natl Inst Public Health. 2002;51:232–44 Miller GA. The magical number seven plus or minus two: some limits on our capacity for processing information. Psychol Rev. 1956;63:81–97. https://doi.org/10.1037/h0043158 Cowan N. The magical number 4 in short-term memory: a reconsideration of mental storage capacity. Behav Brain Sci. 2001;24:87–114; discussion 114. https://doi.org/10.1017/S0140525X01003922 Table Table 1 is available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files Table1.pdf Table 1. List of verification information Object model and prescription information for (a) drug name, (b) drug dose and usage, and (c) dose period are shown. Regarding the expression method of (b) drug dose and usage, the set phrase display (one tablet at a time, once a day, after breakfast) and numerical table display (- | 1 - - | -) represent the same contents. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8139848","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":554028602,"identity":"b6076c5b-4c2b-4cfd-b4e1-6a2001e2d1af","order_by":0,"name":"Toshikazu Tsuji","email":"data:image/png;base64,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","orcid":"","institution":"Setsunan University","correspondingAuthor":true,"prefix":"","firstName":"Toshikazu","middleName":"","lastName":"Tsuji","suffix":""},{"id":554028603,"identity":"4978ca66-785a-4832-a3c0-6895560a400c","order_by":1,"name":"Masayuki Tanaka","email":"","orcid":"","institution":"Setsunan University","correspondingAuthor":false,"prefix":"","firstName":"Masayuki","middleName":"","lastName":"Tanaka","suffix":""},{"id":554028604,"identity":"69c3ef0a-9566-4351-80e4-be2265eef696","order_by":2,"name":"Shiori Iwane","email":"","orcid":"","institution":"Setsunan University","correspondingAuthor":false,"prefix":"","firstName":"Shiori","middleName":"","lastName":"Iwane","suffix":""},{"id":554028605,"identity":"f89481f7-34bb-460c-81d9-24c502662789","order_by":3,"name":"Naoko Kizu","email":"","orcid":"","institution":"Setsunan University","correspondingAuthor":false,"prefix":"","firstName":"Naoko","middleName":"","lastName":"Kizu","suffix":""},{"id":554028606,"identity":"569c9d41-3e9d-4beb-aab2-1d34ba06230c","order_by":4,"name":"Shigeru Hasebe","email":"","orcid":"","institution":"Setsunan University","correspondingAuthor":false,"prefix":"","firstName":"Shigeru","middleName":"","lastName":"Hasebe","suffix":""},{"id":554028607,"identity":"e10bc9b2-138c-4d36-89ce-bd554f714332","order_by":5,"name":"Maho Kikuta","email":"","orcid":"","institution":"Setsunan University","correspondingAuthor":false,"prefix":"","firstName":"Maho","middleName":"","lastName":"Kikuta","suffix":""},{"id":554028608,"identity":"7a760f74-e341-4d61-b0aa-183aff1f4a78","order_by":6,"name":"Kenichiro Nagata","email":"","orcid":"","institution":"Kyushu University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Kenichiro","middleName":"","lastName":"Nagata","suffix":""},{"id":554028609,"identity":"59a8b790-f442-4d00-8e72-d9947ba3f5cc","order_by":7,"name":"Shigeru Ishida","email":"","orcid":"","institution":"Kyushu University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Shigeru","middleName":"","lastName":"Ishida","suffix":""},{"id":554028610,"identity":"b035e8aa-4692-4ef7-b881-d05e2660e06e","order_by":8,"name":"Takeshi Hirota","email":"","orcid":"","institution":"Kyushu University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Takeshi","middleName":"","lastName":"Hirota","suffix":""},{"id":554028611,"identity":"9a9a0a14-f678-47f2-a362-4ec9e35f9529","order_by":9,"name":"Hiroyuki Watanabe","email":"","orcid":"","institution":"Fukuoka Tokushukai Hospital","correspondingAuthor":false,"prefix":"","firstName":"Hiroyuki","middleName":"","lastName":"Watanabe","suffix":""},{"id":554028613,"identity":"cb1f79cd-af1e-4f26-91bb-efe0330d2850","order_by":10,"name":"Mayako Uchida","email":"","orcid":"","institution":"Kyushu University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Mayako","middleName":"","lastName":"Uchida","suffix":""}],"badges":[],"createdAt":"2025-11-18 01:53:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8139848/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8139848/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97370926,"identity":"673941ee-6779-4094-9a79-5b1bbd896080","added_by":"auto","created_at":"2025-12-03 16:28:09","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":44387,"visible":true,"origin":"","legend":"","description":"","filename":"RevisedEyetrackingbasedanalysistoimprovetheefficiencyandsafetyofprescriptionauditToshikazuTsuji.docx","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/74658ab7b52e4f1033d7c255.docx"},{"id":97345385,"identity":"6753bf3f-959b-49df-be1b-785022d38371","added_by":"auto","created_at":"2025-12-03 11:44:49","extension":"json","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11778,"visible":true,"origin":"","legend":"","description":"","filename":"6147b48c63d246448fbfa3a741e5f3b7.json","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/d81d1a4ff6a4353fa60e8709.json"},{"id":97345392,"identity":"730f5bdd-8f65-42b5-bc45-601ac1e58f1f","added_by":"auto","created_at":"2025-12-03 11:44:49","extension":"xml","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":97823,"visible":true,"origin":"","legend":"","description":"","filename":"6147b48c63d246448fbfa3a741e5f3b71enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/a82fac617d23721dc617af74.xml"},{"id":97345394,"identity":"bb26bd2c-ce66-4024-806f-1c6b159c1252","added_by":"auto","created_at":"2025-12-03 11:44:49","extension":"pdf","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":269082,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1Revised.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/cc890eecad03bf46cfb2bc54.pdf"},{"id":97370639,"identity":"25e40c45-77d2-4d44-b008-47968fe2d90c","added_by":"auto","created_at":"2025-12-03 16:27:44","extension":"pdf","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":116797,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/61ea5883f40e8f0370d9221d.pdf"},{"id":97345389,"identity":"df342991-2664-48a6-b586-5ba0c8996ee8","added_by":"auto","created_at":"2025-12-03 11:44:49","extension":"pdf","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":104657,"visible":true,"origin":"","legend":"","description":"","filename":"Figure3.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/0a0c2a0e8d833c6e51d2febf.pdf"},{"id":97345396,"identity":"dab2fb6b-cbf3-4b34-ba4b-d7a762754ed1","added_by":"auto","created_at":"2025-12-03 11:44:49","extension":"pdf","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":120093,"visible":true,"origin":"","legend":"","description":"","filename":"Figure4ab.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/b23689bc8df79f4201c83af6.pdf"},{"id":97370899,"identity":"f68729e3-ec33-4ff2-bef4-8418cd85270d","added_by":"auto","created_at":"2025-12-03 16:28:07","extension":"xml","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":99024,"visible":true,"origin":"","legend":"","description":"","filename":"6147b48c63d246448fbfa3a741e5f3b71structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/e4c4f32e866471cfcb606bdf.xml"},{"id":97345391,"identity":"5bc2db49-2dfc-4778-85e0-6258e6c0c17d","added_by":"auto","created_at":"2025-12-03 11:44:49","extension":"html","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":107346,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/dee6ff2d4a5c8baf677bf0ab.html"},{"id":97345383,"identity":"2d6a2d1a-3fd6-4371-a634-55e9ba7e094c","added_by":"auto","created_at":"2025-12-03 11:44:49","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":203536,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eOutline of verification process using the eye-tracking method\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGaze movement data acquired using the Tobii Pro Lab Analyzer were classified into two main categories: fixation (stagnation for a certain time) and saccades (quick movements of the eyeballs). We analyzed the gaze movements of pharmacists during the verification process by showing the two monitors of prescription audits and prescription questions (34 cm × 60 cm). Wearing an eye tracker, the pharmacist was positioned on a chair 100 cm from the main monitor and verified four pairs of models (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e \u003c/sub\u003eC\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e \u003c/sub\u003eand Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e) in random order. In the enlarged diagram on the right, “O.W., M, A, N, B.S.” stand for “on waking, morning, afternoon, night, before sleeping,” respectively. \u003c/p\u003e","description":"","filename":"Binder11.png","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/00140ae2294c22e08b0906dd.png"},{"id":97345382,"identity":"883b61f5-064c-48c3-a56d-ab01fc9e749b","added_by":"auto","created_at":"2025-12-03 11:44:49","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":122749,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of four classifications between two display methods using three pairs of models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis figure depicts the differences in the four classifications (Gaze 1, Gaze 2, Gaze 3, and Time) between the two display methods (set-phrase display and numerical table display) according to three pairs of models A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2 \u003c/sub\u003e(Low-difficulty), B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e (Moderate-difficulty),\u003csub\u003e \u003c/sub\u003eand C\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e (High-difficulty). **\u003cem\u003eP\u003c/em\u003e \u0026lt;0.01, ***\u003cem\u003eP\u003c/em\u003e\u0026lt;0.001 using the paired t-test\u003c/p\u003e","description":"","filename":"Binder12.png","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/3f17bd1ab39cca02e9c85b49.png"},{"id":97345384,"identity":"df46038a-e333-4145-a70a-888477b504b9","added_by":"auto","created_at":"2025-12-03 11:44:49","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":98336,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of gaze averages per necessary spot between two display methods using three pairs of models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis figure depicts the difference in gaze averages per necessary spot between the two display methods according to the difficulty level of the prescription audit using three pairs of models (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e \u003c/sub\u003eand C\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e). *\u003cem\u003eP\u003c/em\u003e \u0026lt;0.05 using the paired t-test\u003c/p\u003e","description":"","filename":"Binder13.png","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/ab55490e86987b3260ecea82.png"},{"id":97370584,"identity":"34854e2a-51b2-47e9-b410-233ee8c993d7","added_by":"auto","created_at":"2025-12-03 16:27:39","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":83997,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003e(a) Comparison of four classifications between two display methods using prescription question models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis figure depicts the differences in the four classifications (Gaze 1, Gaze 2, Gaze 3, and Time) between the two display methods using prescription question models (Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e). *\u003cem\u003eP\u003c/em\u003e \u0026lt;0.05, ***\u003cem\u003eP\u003c/em\u003e\u0026lt;0.001 using Student’s t-test\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(b) Comparison of two recognition states between two display methods using prescription question models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis figure depicts the difference in the two recognition states (Recognition rate and Recognition time) between the two display methods using prescription question models (Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e).\u0026nbsp;\u003c/p\u003e","description":"","filename":"Binder14.png","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/1f6146100589e24109fa1757.png"},{"id":98797923,"identity":"fc8bae22-ce59-41f9-9a3a-9f65a6d3b900","added_by":"auto","created_at":"2025-12-22 14:04:10","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1599954,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/71b7506e-5f5c-498a-9b53-87ea8a91e5ee.pdf"},{"id":97370873,"identity":"22a5c60c-25d1-4ca0-8bdc-8809e3c7d8c5","added_by":"auto","created_at":"2025-12-03 16:28:05","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":289871,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTable 1. List of verification information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eObject model and prescription information for (a) drug name, (b) drug dose and usage, and (c) dose period are shown. Regarding the expression method of (b) drug dose and usage, the set phrase display (one tablet at a time, once a day, after breakfast) and numerical table display (- | 1 - - | -) represent the same contents.\u003c/p\u003e","description":"","filename":"Table1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8139848/v1/a6e8065baa3de0bd5c97894d.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Eye-tracking-based analysis to improve the efficiency and safety of prescription audit","fulltext":[{"header":"Background","content":"\u003cp\u003eMany medical institutions, including Kyushu University Hospital, have implemented initiatives to prevent dispensing errors caused by pharmacists, and several successful outcomes have been reported to date [\u003cspan additionalcitationids=\"CR2 CR3 CR4 CR5 CR6 CR7 CR8 CR9 CR10 CR11 CR12 CR13 CR14\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Pre-checking prescription content (hereinafter referred to as prescription audits) is also essential for ensuring safe and reliable medical therapy for patients. Pharmacists conduct prescription audits accurately and efficiently within a predetermined timeframe to prevent duplicate medications and unsafe dosages. For this reason, pharmacists need to explore more efficient methods of prescription auditing to ensure safe and reliable medical therapy for patients. However, few studies have examined improvements in prescription auditing efficiency and patient safety [\u003cspan additionalcitationids=\"CR17 CR18\" citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Considering these circumstances, improving efficiency and safety in prescription auditing first requires that pharmacists understand their own thought processes during auditing.\u003c/p\u003e\u003cp\u003eIn our previous studies, we used an eye-tracking module that followed individual\u0026rsquo;s gaze movements in real time from camera-captured video and clarified the thought processes of 12 or 22 pharmacists in simulated dispensing environments [\u003cspan additionalcitationids=\"CR21 CR22\" citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Our recent findings revealed that introducing visual information using colors or symbols (➁| ▮--|) into drug location information in prescription content not only increased dispensing efficiency but also reduced the dispensing errors. In this study, the gaze movements of pharmacists during prescription audits were analyzed using an eye-tracking system. First, two methods for indicating drug\u0026rsquo;s dose and usage in prescription audit were prepared: the widely used \u0026ldquo;set phrase display (1 tablet at a time, once a day, after breakfast)\u0026rdquo; and the unprecedented \u0026ldquo;numerical table display (- | 1 - - | -)\u0026rdquo;. Second, three pairs of prescription audit models with different levels of difficulty were prepared, along with one pair of models that included a drug\u0026ndash;drug interaction. This approach enabled analysis of differences in both efficiency and safety between \u0026ldquo;set phrase display\u0026rdquo; and \u0026ldquo;numerical table display,\u0026rdquo; in relation to drug dose and usage display methods.\u003c/p\u003e\u003cp\u003eColor recognition and processing are closely linked to human brain function [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e], but differences in information processing between letters and numbers are not fully understood. Therefore, our goal was to demonstrate the difference in pharmacists' information processing between \u0026ldquo;set phrase display\u0026rdquo; and \u0026ldquo;numerical table display\u0026rdquo; using four pairs of prescription audit models.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eAnalysis of gaze movements using\u003c/strong\u003e \u003cstrong\u003ea visual line tracing system\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEye-tracking, which uses corneal reflection of infrared rays to analyze gaze points and eye movements, is applied in many fields, including medicine, psychology, and pedagogy [26\u0026ndash;29]. In this study, pharmacists\u0026rsquo; gaze movements during prescription audits were examined using a wearable eye tracker (Tobii Pro Glasses 3; Tobii Technology K.K.). Gaze movements were classified into two categories: fixation (stationary for \u0026ge;100 ms) and saccade (rapid eye movements). Fixations and saccades were analyzed using motion videos recorded with dedicated software (Tobii Pro Lab Analyzer, Tobii Technology K.K.).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTarget pharmacists to participate in this study\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe inclusion criteria for pharmacists were as follows: First, to ensure accurate eye-movement measurement, pharmacists needed to read prescription information clearly on a large monitor while wearing soft contact lenses or with the naked eye. Second, pharmacists were required to have at least 12 months of prescription audit experience at Kyushu University Hospital to ensure high verification quality. Finally, pharmacists had to provide informed consent to participate in the study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTarget drugs used in the prescription audit\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNineteen drugs dispensed at Kyushu University Hospital were used in this study: Amlodipine OD 5 mg, Pravastatin Sodium 10 mg, Eliquis\u003csup\u003e\u0026reg;\u003c/sup\u003e 5 mg, Feburic\u003csup\u003e\u0026reg;\u0026nbsp;\u003c/sup\u003e20 mg, Brotizolam OD 0.25 mg, Famotidine OD 20 mg, Medrol\u003csup\u003e\u0026reg;\u0026nbsp;\u003c/sup\u003e4 mg, Aspara-CA 200 mg, Pitavastatin Calcium OD 2 mg, Geninax\u003csup\u003e\u0026reg;\u0026nbsp;\u003c/sup\u003e200 mg, Methotrexate 2 mg, Foliamin\u003csup\u003e\u0026reg;\u0026nbsp;\u003c/sup\u003e5 mg, Predonine\u003csup\u003e\u0026reg;\u0026nbsp;\u003c/sup\u003e5 mg, Rosuvastatin OD 2.5 mg, Magnesium Oxide 250 mg, MS Contin\u003csup\u003e\u0026reg;\u0026nbsp;\u003c/sup\u003e10 mg, Calonal\u003csup\u003e\u0026reg;\u0026nbsp;\u003c/sup\u003e200 mg, Tegretol\u003csup\u003e\u0026reg;\u0026nbsp;\u003c/sup\u003e100 mg, and Levofloxacin 250 mg.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVerification slides\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe slides verifying prescription audits were prepared using Microsoft PowerPoint\u003csup\u003e\u0026reg;\u003c/sup\u003e 2016. Each slide contained basic patient information at the top, including name, age (sex), body weight, height, and creatinine clearance. Prescription information was displayed in the center of each slide and consisted of three items: (a) drug name, (b) drug dose and usage, and (c) dose period. In this study, four pairs of models (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eC\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eand Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e) were prepared, and five target drugs were included in each verification.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eImportantly, the method of indicating (b) drug dose and usage on prescription slides was classified into two types: \u0026ldquo;set phrase display\u0026rdquo; and \u0026ldquo;numerical table display.\u0026rdquo; In the \u0026ldquo;set phrase display,\u0026rdquo; prescription contents were presented as words and phrases, such as \u0026ldquo;1 tablet at a time, once a day, after breakfast.\u0026rdquo; In the \u0026ldquo;numerical table display,\u0026rdquo; prescription contents were shown using only numerical figures, such as \u0026ldquo; - | 1 - - | -,\u0026rdquo; and explanatory comments about drug administration after meals were omitted. The specifications of the prescription audit information are presented in Table 1, and the order of verification was randomized.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDefinition of the four pairs of models\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFirst, we set up three pairs of models (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eand C\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e) to compare the difference in efficiency according to the difficulty level of prescription audit between \u0026ldquo;set phrase display\u0026rdquo; and \u0026ldquo;numerical table display.\u0026rdquo; Second, we set up another pair of models (Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e) to compare the difference in safety related to prescription content that included a drug interaction between \u0026ldquo;set phrase display\u0026rdquo; and \u0026ldquo;numerical table display.\u0026rdquo; A summary of the four pairs (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eC\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eand Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e) is provided below:\u003c/p\u003e\n\u003cp\u003eLow-difficulty models (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e): The dosage and administration were appropriate, and there were no drug interactions among the five drugs.\u003c/p\u003e\n\u003cp\u003eModerate-difficulty models (B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e): The dosage and administration were appropriate, but prescription content was a little complicated as it contained a drug of unequal quantities.\u003c/p\u003e\n\u003cp\u003eHigh-difficulty models (C\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e): The dosage and administration were appropriate, but prescription content was complicated as it contained a drug of unequal quantities and two drugs for the designated days of the week.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ePrescription question models (Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e): There was an obvious drug interaction between the two drugs (magnesium oxide 250 mg and levofloxacin 250 mg).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVerification procedure\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAn outline of the verification task using the eye-tracking method is shown in Figure 1. Two notebook computers connected to 27-inch monitors were used to display the verification slides. The prescription audit area (34 \u0026times; 60 cm) was shown on the main monitor (prescription audit monitor), while the side monitor for prescription question (prescription question monitor) was positioned to the left of the prescription audit monitor. A pharmacist wearing an eye tracker was positioned 100 cm from the prescription audit monitor, and their gaze movements were investigated during the prescription audit process.\u0026nbsp;The Tobii Pro Lab Analyzer, which records motion videos, was used to assess parameters such as gaze point (circle center), gaze duration (circle size), and movement of sight lines (line between circle centers).\u003c/p\u003e\n\u003cp\u003eThe eye tracker was calibrated to ensure accuracy before a series of verification tasks was conducted. Pharmacists practiced with several training slides in advance to familiarize themselves with the verification process. The primary aim of the verification task was to maintain a smooth prescription audit. If the pharmacists identified an issue in the prescription audit, they point to the prescription question monitor. The six major steps for dispensing verification are as follows.\u003c/p\u003e\n\u003cp\u003e1) The pharmacist maintained gaze at a specified position.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2) When the pharmacist indicated the \u0026ldquo;Next\u0026rdquo; signal, the assistant switched to the next prescription audit slide. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e3) The pharmacist verified the dosage and administration of each drug and checked for interactions among the five drugs.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e4) When a prescription question was deemed necessary, the pharmacist pointed to the prescription question monitor.\u003c/p\u003e\n\u003cp\u003e5) After the pharmacist signaled \u0026ldquo;Next\u0026rdquo;, the assistant switched to a rest slide.\u003c/p\u003e\n\u003cp\u003e6) A sequence of verifications using 10 or more prescription audit slides was repeated, with sufficient breaks taken as needed.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVerification items and classifications\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eVerification of prescription audits using the four pairs of models (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eC\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eand Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e) requires pharmacists to visually confirm three key items: (a) drug name, (b) drug dose and usage, and (c) dose period. Furthermore, pharmacists had to shift their gaze to assess the interactions among the five drugs accurately. In this study, verification items were defined to compare gaze points and verification time, and differences in gaze movements of pharmacists were analyzed between two display methods for (b) drug dose and usage: \u0026ldquo;set phrase display\u0026rdquo; and the \u0026ldquo;numerical table display.\u0026rdquo; A summary of the four classifications (Gaze 1, Gaze 2, Gaze 3, and Time) is provided below.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eGaze 1: Total number of gaze points in the area of (a) drug name\u003c/p\u003e\n\u003cp\u003eGaze 2: total number of gaze points in the area of (b) drug dose and usage\u003c/p\u003e\n\u003cp\u003eGaze 3: total number of gaze points in the area of (c) dose period\u003c/p\u003e\n\u003cp\u003eTime: total time required to verify the prescription audit for the five target drugs.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMeasurement of prescription question\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn this study, a drug interaction between magnesium oxide 250 mg and levofloxacin 250 mg was established using the prescription question models Q\u003csub\u003e1\u003c/sub\u003e\u0026ndash;Q\u003csub\u003e2\u003c/sub\u003e. The content of the drug interaction was the reduced absorption of levofloxacin caused by the simultaneous administration of magnesium oxide. First, the presence or absence of a prescription question was determined based on each pharmacist\u0026rsquo;s performance in indicating the prescription question monitor. Second, the rate of recognizing the drug interaction between magnesium oxide and levofloxacin was calculated as \u0026ldquo;Recognition rate\u0026rdquo;. This calculation was performed by dividing the number of pharmacists who appropriately judged the drug interaction by the total number of target pharmacists. Furthermore, the time required to recognize the drug interaction was measured as \u0026ldquo;Recognition time.\u0026rdquo; A summary of the two classifications showing the recognition states (Recognition rate and Recognition time) is provided below.\u003c/p\u003e\n\u003cp\u003eRecognition rate: the proportion of pharmacists who recognized a drug interaction between magnesium oxide and levofloxacin\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eRecognition time: the time required to recognize a drug interaction between magnesium oxide and levofloxacin\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData analysis\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUsing the gaze category data (fixation and saccade) from the recorded motion video, we analyzed six classifications: Gaze 1, Gaze 2, Gaze 3, Time, Recognition rate, and Recognition time. Data are presented as the mean \u0026plusmn; standard deviation. Significant differences in the four classifications (Gaze 1, Gaze 2, Gaze 3, and Time) between the model pairs A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eand C\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e were analyzed between the same pharmacists using the paired \u003cem\u003et\u003c/em\u003e-test. Significant differences in the five classifications (Gaze 1, Gaze 2, Gaze 3, Time, and Recognition time) between the Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e models were analyzed across different pharmacists using Student\u0026rsquo;s t-test, while difference in another classification (Recognition rate) was analyzed using Fisher\u0026apos;s exact test. A \u003cem\u003eP-\u003c/em\u003evalue \u0026lt;0.05 was considered statistically significant, whereas \u003cem\u003eP-\u003c/em\u003evalues \u0026lt;0.01 and \u0026lt;0.001 were considered highly significant. Statistical analyses were performed using JMP Pro 16 software.\u0026nbsp;\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eBasic information of pharmacists\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA total of 22 pharmacists (9 men and 13 women; mean age, 30.1\u0026plusmn;5.9 years) participated in this study. Data from these 22 pharmacists were analyzed for each model pair (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eand C\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e) using the paired t-test. Also, pharmacists participating in Q\u003csub\u003e1\u003c/sub\u003e and Q\u003csub\u003e2\u003c/sub\u003e verifications were 11 (4 men and 7 women) and 11 (5 men and 6 women) with an average age of 28.9\u0026plusmn;4.7 and 31.3\u0026plusmn;6.9 years, respectively. Data from the 11 pharmacists participating in Q\u003csub\u003e1\u003c/sub\u003e and Q\u003csub\u003e2\u003c/sub\u003e verifications were analyzed using Student\u0026rsquo;s t-test or Fisher\u0026apos;s exact test.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eComparison of four classifications between two display methods using three pairs of models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe difference in gaze movements of 22 pharmacists during the prescription audit process between \u0026ldquo;set phrase display\u0026rdquo; and \u0026ldquo;numerical table display\u0026rdquo; was analyzed using three model pairs (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eand C\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e). The data for the four classifications (Gaze 1, Gaze 2, Gaze 3, and Time) are presented below, and the relationships between each pair of models are shown in Figure 2. Significant differences between models A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e were observed in Gaze 2 and Time using the paired \u003cem\u003et\u003c/em\u003e-test (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001, \u003cem\u003eP\u0026nbsp;\u003c/em\u003e= 0.0019, respectively).\u0026nbsp;In addition, Gaze 2 and Time showed a strong positive correlation in models A\u003csub\u003e1\u0026nbsp;\u003c/sub\u003e(r\u0026gt;0.72, \u003cem\u003eP\u003c/em\u003e\u0026lt;0.001) and A\u003csub\u003e2\u003c/sub\u003e (r\u0026gt;0.79, \u003cem\u003eP\u003c/em\u003e\u0026lt;0.001), whereas no strong positive correlation was observed among the other classifications. Similarly, significant differences between models B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e were observed in Gaze 2 and Time (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001, \u003cem\u003eP\u0026nbsp;\u003c/em\u003e= 0.0068), and a strong positive correlation was noted in models B\u003csub\u003e1\u0026nbsp;\u003c/sub\u003e(r\u0026gt;0.75, \u003cem\u003eP\u003c/em\u003e\u0026lt;0.001) and B\u003csub\u003e2\u003c/sub\u003e (r\u0026gt;0.70, \u003cem\u003eP\u003c/em\u003e\u0026lt;0.001). Furthermore, significant differences between models C\u003csub\u003e1\u003c/sub\u003e and C\u003csub\u003e2\u003c/sub\u003e were observed in gaze 2 and time (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001 and \u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001, respectively), and a strong positive correlation was found in models C\u003csub\u003e1\u0026nbsp;\u003c/sub\u003e(r\u0026gt;0.86, \u003cem\u003eP\u003c/em\u003e\u0026lt;0.001) and C\u003csub\u003e2\u003c/sub\u003e (r\u0026gt;0.83, \u003cem\u003eP\u003c/em\u003e\u0026lt;0.001).\u003c/p\u003e\n\u003cp\u003eModel A\u003csub\u003e1\u003c/sub\u003e: Gaze 1, 10.5\u0026plusmn;3.3; Gaze 2, 18.2\u0026plusmn;5.1; Gaze 3, 5.5\u0026plusmn;1.4; Time, 14.9\u0026plusmn;4.2\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eModel A\u003csub\u003e2\u003c/sub\u003e: Gaze 1, 12.2\u0026plusmn;5.4; Gaze 2, 9.3\u0026plusmn;3.7; Gaze 3, 5.1\u0026plusmn;2.2; Time, 12.5\u0026plusmn;13.9 \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eModel B\u003csub\u003e1\u003c/sub\u003e: Gaze 1, 16.9\u0026plusmn;5.9; Gaze 2, 26.3\u0026plusmn;11.0; Gaze 3, 6.4\u0026plusmn;3.2; Time, 20.5\u0026plusmn;5.6\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eModel B\u003csub\u003e2\u003c/sub\u003e: Gaze 1, 19.1\u0026plusmn;8.3; Gaze 2, 14.0\u0026plusmn;6.0; Gaze 3, 6.7\u0026plusmn;2.8; Time, 17.7\u0026plusmn;6.6\u003c/p\u003e\n\u003cp\u003eModel C\u003csub\u003e1\u003c/sub\u003e: Gaze 1, 15.5\u0026plusmn;7.4; Gaze 2, 35.5\u0026plusmn;15.1; Gaze 3, 8.0\u0026plusmn;4.0; Time, 24.8\u0026plusmn;8.0\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eModel C\u003csub\u003e2\u003c/sub\u003e: Gaze 1, 17.4\u0026plusmn;6.4; Gaze 2, 21.4\u0026plusmn;9.9; Gaze 3, 7.6\u0026plusmn;4.6; Time, 19.7\u0026plusmn;6.3\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eComparison of gaze averages per necessary spot in area (b) between two display methods using three pairs of models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe difference in gaze points focusing on area (b) drug dose and usage was analyzed by using low-, moderate-, and high-difficulty models (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eand C\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e), since this area was the source of differences in gaze movements between \u0026ldquo;set phrase display\u0026rdquo; and \u0026ldquo;numerical table display.\u0026rdquo; In the area of (b) drug dose and usage, the number of necessary gaze spots that must be visually recognized to perform prescription audit (necessary spot) and the average of gaze points per necessary spot (gaze average) for each model pair are as follows. The relationship between each pair of models is shown in Figure 3.\u0026nbsp;Significant differences in gaze averages per necessary spot were observed between models A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e using the paired \u003cem\u003et\u003c/em\u003e-test (\u003cem\u003eP\u003c/em\u003e=0.0110), but not between the other two models.\u003c/p\u003e\n\u003cp\u003eLow-difficulty models A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u0026nbsp;\u003c/sub\u003e: Necessary spots (15, 6); Gaze averages (1.22\u0026plusmn;0.34, 1.55\u0026plusmn;0.61)\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eModerate-difficulty models B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u0026nbsp;\u003c/sub\u003e: Necessary spots (15, 7); Gaze averages (1.75\u0026plusmn;0.73, 2.00\u0026plusmn;0.85) \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eHigh-difficulty models C\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u0026nbsp;\u003c/sub\u003e: Necessary spots (17, 10); Gaze averages (2.09\u0026plusmn;0.89, 2.14\u0026plusmn;0.99) \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eComparison of four classifications and two recognition states\u003c/strong\u003e \u003cstrong\u003ebetween two display methods using prescription question models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe difference in safety during the prescription audit process between \u0026ldquo;set phrase display\u0026rdquo; and \u0026ldquo;numerical table display\u0026rdquo; was analyzed using a pair of prescription question models (Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e). The data for the four classifications (Gaze 1, Gaze 2, Gaze 3, and Time) are shown below, and the relationships between models Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e are shown in Figure 4a. Significant differences between the 11 pharmacists were observed in Gaze 2 and Time using Student\u0026rsquo;s t-test (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001 and \u003cem\u003eP\u0026nbsp;\u003c/em\u003e= 0.0336, respectively), and no strong positive correlations were found among the four classifications in either model Q\u003csub\u003e1\u0026nbsp;\u003c/sub\u003eor Q\u003csub\u003e2\u003c/sub\u003e.\u003c/p\u003e\n\u003cp\u003eModel Q\u003csub\u003e1\u003c/sub\u003e: Gaze 1, 15.5\u0026plusmn;6.9; Gaze 2, 32.8\u0026plusmn;11.0; Gaze 3, 5.1\u0026plusmn;3.1; Time, 23.8\u0026plusmn;6.1\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eModel Q\u003csub\u003e2\u003c/sub\u003e: Gaze 1, 17.9\u0026plusmn;4.9; Gaze 2, 15.3\u0026plusmn;4.8; Gaze 3, 5.8\u0026plusmn;3.4; Time, 18.7\u0026plusmn;4.0 \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFurthermore, the number of pharmacists who recognized a drug interaction between magnesium oxide and levofloxacin was 8 and 11 in models Q\u003csub\u003e1\u003c/sub\u003e and Q\u003csub\u003e2\u003c/sub\u003e, respectively. Recognition rates were 72.7% (8/11) and 100% (11/11), respectively;\u0026nbsp;no significant difference was observed using Fisher\u0026rsquo;s exact test (\u003cem\u003eP\u0026nbsp;\u003c/em\u003e= 0.2143). Recognition times in models Q\u003csub\u003e1\u003c/sub\u003e and Q\u003csub\u003e2\u003c/sub\u003e were 20.5\u0026plusmn;6.2 (n=8) and 16.0\u0026plusmn;3.8 (n=11), respectively; and no significant difference between them was found using Student\u0026rsquo;s t-test (\u003cem\u003eP\u0026nbsp;\u003c/em\u003e= 0.0651). The data for the two recognition states (Recognition rate and Recognition time) are presented below, and the relationships between models Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e are shown in Figure 4b.\u003c/p\u003e\n\u003cp\u003eModel Q\u003csub\u003e1\u003c/sub\u003e: Recognition rate, 72.7% (8/11); Recognition time, 20.5\u0026plusmn;6.2 (n=8)\u003c/p\u003e\n\u003cp\u003eModel Q\u003csub\u003e2\u003c/sub\u003e: Recognition rate, 100% (11/11); Recognition time, 16.0\u0026plusmn;3.8 (n=11) \u0026nbsp;\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we aimed to elucidate the thought processes of pharmacists during prescription audits using an eye-tracking system. The gaze movements of 22 pharmacists were compared between two methods of indicating drug dose and usage\u0026mdash;set-phrase display and numerical table display\u0026mdash;using four pairs of prescription audit models (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eC\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eand Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e). The findings demonstrated that the \u0026ldquo;numerical table display\u0026rdquo; method allowed prescription audits to be performed more efficiently and improved the accuracy of judging drug interactions. In summary, these results suggest that pharmacists can conduct prescription audit work more effectively and safely by utilizing the \u0026ldquo;numerical table display\u0026rdquo; method.\u003c/p\u003e\n\u003cp\u003eWhen using the \u0026ldquo;set phrase display\u0026rdquo; method (e.g., 1 tablet at a time, once a day, after breakfast), pharmacists are forced to grasp the overall drug\u0026rsquo;s dose and usage while memorizing multiple words and phrases separately, since they must read and understood them one by one. Furthermore, if there is a drug interaction in the prescription, the pharmacist must check the timing of administration by moving the visual lines not only vertically but also diagonally. In summary, prescription audits using the \u0026ldquo;set phrase display\u0026rdquo; method require more complex gaze movements, which may increase the risk of losing stored memory. In contrast, when using the \u0026ldquo;numerical table display\u0026rdquo; method (e.g., - | 1 - - | -), pharmacists can grasp the drug\u0026rsquo;s dose and usage by processing them as a numerical table and can check for drug interactions by just moving their gaze vertically. According to research on human memory capacity, short-term memory typically holds 7\u0026plusmn;2 items, but this can be reduced to 4\u0026plusmn;1 items when information is complex or when interference occurs [30\u0026ndash;32]. In other words, the key point of this study was that pharmacists\u0026rsquo; capacity to process prescription information varied depending on whether drug dose and usage were expressed as a combination of words and phrases (set phrase display) or as numerical figures (numerical table display).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFirst, significant differences in the four classifications\u0026nbsp;(Gaze 1, Gaze 2, Gaze 3, and Time)\u0026nbsp;when using the low-difficulty models A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e were observed in Gaze 2 and Time, with A\u003csub\u003e1\u003c/sub\u003e\u0026gt;A\u003csub\u003e2\u003c/sub\u003e (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001, \u003cem\u003eP\u003c/em\u003e=0.0019). Furthermore, similar differences were observed between the moderate-difficulty models B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e for Gaze 2 and time (B\u003csub\u003e1\u003c/sub\u003e\u0026gt;B\u003csub\u003e2\u003c/sub\u003e, \u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001, \u003cem\u003eP\u0026nbsp;\u003c/em\u003e= 0.0068) and between the high-difficulty models C\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e for Gaze 2 and Time (C\u003csub\u003e1\u003c/sub\u003e\u0026gt;C\u003csub\u003e2\u003c/sub\u003e, \u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001, \u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001). These results indicate that prescription audits using the \u0026ldquo;numerical table display\u0026rdquo; method was more efficient than those using \u0026ldquo;set phrase display,\u0026rdquo; regardless of the model\u0026rsquo;s difficulty levels (Figure 2). Furthermore, strong positive correlations were observed between Gaze 2 and Time in models A\u003csub\u003e1\u003c/sub\u003e, A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e, B\u003csub\u003e2\u003c/sub\u003e, C\u003csub\u003e1\u003c/sub\u003e, and C\u003csub\u003e2\u003c/sub\u003e (r\u0026gt;0.70, \u003cem\u003eP\u003c/em\u003e \u0026lt;0.001, respectively), suggesting that considerable time was spent processing information on drug dosage and usage during prescription audits. Therefore, it is assumed that the decrease of Gaze 2 (gaze points in the area of drug dose and usage) caused by using the \u0026ldquo;numerical table display\u0026rdquo; method had a great impact on the reduction of Time (time required to perform prescription audits). However, it is unclear why Gaze 1 (gaze points in the area of drug name) was slightly higher with the \u0026ldquo;numerical table display\u0026rdquo; than with the \u0026ldquo;set phrase display\u0026rdquo; in each model, suggesting that the reduced workload in area (b) may have led to the polite check in area (a).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSecond, when focusing on the gaze averages per necessary spot in area (b) drug dose and usage, there was a significant difference between models A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e (\u003cem\u003eP\u003c/em\u003e=0.0110) but not between the other two pairs of models. These results indicate that the confirmation frequency per necessary spot using the \u0026ldquo;numerical table display\u0026rdquo; method was equal to or higher than that of the \u0026ldquo;set phrase display\u0026rdquo; method, regardless of the model\u0026rsquo;s difficulty levels (Figure 3). Although the details of these relationships according to the model\u0026rsquo;s difficulty levels are unclear, the necessary spots in the \u0026ldquo;numerical table display\u0026rdquo; consisted only of numerical figures, and there were also fewer spots to check, suggesting that pharmacists could select the necessary items that need to be cross-checked and confirm them more prudently.\u0026nbsp;Another factor may be that numerous figures indicating drug\u0026rsquo;s dose and usage are arranged in one vertical row, making it easier for pharmacists to judge whether there are any drug interactions. For these reasons, the use of the \u0026ldquo;numerical table display\u0026rdquo; method appears to enable pharmacists to confirm target points more carefully and in less time.\u003c/p\u003e\n\u003cp\u003eThird, when using the prescription question models Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e, significant differences in the four classifications\u0026nbsp;(Gaze 1, Gaze 2, Gaze 3, and Time)\u0026nbsp;were observed for Gaze 2 and Time, with Q\u003csub\u003e1\u003c/sub\u003e\u0026gt;Q\u003csub\u003e2\u003c/sub\u003e (\u003cem\u003eP\u003c/em\u003e\u0026lt;0.0001, \u003cem\u003eP\u003c/em\u003e=0.0336). These results demonstrate that prescription audits using the \u0026ldquo;numerical table display\u0026rdquo; method was performed more efficiently than with the \u0026ldquo;set phrase display,\u0026rdquo; even when drug interactions were present in the prescription (Figure 4a). Conversely, for the two recognition states (Recognition rate, Recognition time) in models Q\u003csub\u003e1\u003c/sub\u003e and Q\u003csub\u003e2\u003c/sub\u003e, the formers were 72.7% (8/11) and 100% (11/11), while the latter were 20.5\u0026plusmn;6.2 and 16.0\u0026plusmn;3.8, with no significant differences observed (Figure 4b). However, it is important to note that all 11 pharmacists were aware of the drug interaction in model Q\u003csub\u003e2\u003c/sub\u003e, moreover, the task was performed in less time (\u003cem\u003eP\u003c/em\u003e=0.0651). In other words, these results suggest that pharmacists could conduct prescription audits more effectively and safely by using the \u0026ldquo;numerical table display\u0026rdquo; method.\u003c/p\u003e\n\u003cp\u003eThis study has some limitations. First, the \u0026ldquo;numerical table display\u0026rdquo; method used in this study may not apply to electronic medical charts across all medical environments. Second, when using the \u0026ldquo;numerical table display\u0026rdquo; method, explanatory comments for drugs taken before, immediately before, or just after meals need to be written in the remarks column. However, to the best of our knowledge, this is the first study to evaluate the thought processes of more than 20 pharmacists during prescription audit tasks and to clarify both efficiency and safety using an eye-tracking system.\u0026nbsp;\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eWe analyzed the differences in the gaze movements of 22 pharmacists between two methods\u0026ndash;set phrase display and numerical table display\u0026ndash;in prescription audits using an eye-tracking system. Employing four model pairs (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eC\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e,\u003csub\u003e\u0026nbsp;\u003c/sub\u003eand Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e) allowed us to elucidate pharmacists\u0026rsquo; thought processes across multiple prescription audit patterns. The results demonstrated that the use of the \u0026ldquo;numerical table display\u0026rdquo; method not only enables prescription audits to be performed without the complicated procedures but also improves the accurate judgement of drug interactions. In other words, by introducing the \u0026ldquo;numerical table display\u0026rdquo; method into the area of drug dose and usage, it is possible to enhance both efficiency and safety in prescription audits.\u0026nbsp;\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was approved by the Clinical Trials Ethics Committees of Kyushu University Hospital (approval number: 20222011) and Setsunan University (approval number: 2023-018). After explaining the research content, the pharmacists provided written informed consent. This study complied with the Ethical Guidelines for Medical and Health Research Involving Human Participants.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data generated or analyzed during this study are included in this published article.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eT.T.,M.T., and S.Iwa. conducted analyses and drafted the manuscript. S.H., N.K., M.K., and K.N. performed statistical analyses. S.Ishi., T.H., H.W., and M.U. revised the manuscript. All the authors discussed the results and approved the final version of the manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003csup\u003e1\u003c/sup\u003e Laboratory of Clinical Pharmacy, Faculty of Pharmaceutical Sciences, Setsunan University, Osaka, Japan\u003c/p\u003e\n\u003cp\u003e\u003csup\u003e2\u003c/sup\u003e Department of Social Pharmacy, Faculty of Pharmaceutical Sciences, Setsunan University, Hirakata, Osaka, Japan\u003c/p\u003e\n\u003cp\u003e\u003csup\u003e3\u003c/sup\u003e Department of Pharmacy, Kyushu University Hospital, Fukuoka, Japan\u003c/p\u003e\n\u003cp\u003e\u003csup\u003e4\u003c/sup\u003e Department of Pharmacy, Fukuoka Tokushukai Hospital, Fukuoka, Japan\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAnto B, Barlow D, Oborne CA, Whittlesea C. Incorrect drug selection at the point of dispensing: a study of potential predisposing factors. Int J Pharm Pract. 2011;19:51\u0026ndash;60. https://doi.org/10.1111/j.2042-7174.2010.00072.x\u003c/li\u003e\n\u003cli\u003eBeso A, Franklin BD, Barber N. The frequency and potential causes of dispensing errors in a hospital pharmacy. Pharm World Sci. 2005;27:182\u0026ndash;90. https://doi.org/10.1007/s11096-004-2270-8\u003c/li\u003e\n\u003cli\u003eAshcroft DM, Quinlan P, Blenkinsopp A. Prospective study of the incidence, nature and causes of dispensing errors in community pharmacies. Pharmacoepidemiol Drug Saf. 2005;14:327\u0026ndash;32. https://doi.org/10.1002/pds.1012\u003c/li\u003e\n\u003cli\u003eCheung KC, Bouvy ML, De Smet PA. Medication errors: the importance of safe dispensing. Br J Clin Pharmacol. 2009;67:676\u0026ndash;80. https://doi.org/10.1111/j.1365-2125.2009.03428.x\u003c/li\u003e\n\u003cli\u003eLambert BL. Predicting look-alike and sound-alike medication errors. Am J Health Syst Pharm. 1997;54:1161\u0026ndash;71. https://doi.org/10.1093/ajhp/54.10.1161\u003c/li\u003e\n\u003cli\u003eLambert BL, Lin SJ, Chang KY, Gandhi SK. Similarity as a risk factor in drug-name confusion errors: the look-alike (orthographic) and sound-alike (phonetic) model. Med Care. 1999;37:1214\u0026ndash;25. https://doi.org/10.1097/00005650-199912000-00005\u003c/li\u003e\n\u003cli\u003eLambert BL, Chang KY, Lin SJ. Descriptive analysis of the drug name lexicon. Drug Inf J. 2001;35:163\u0026ndash;72. https://doi.org/10.1177/009286150103500119\u003c/li\u003e\n\u003cli\u003eLambert BL, Chang KY, Lin SJ. Effect of orthographic and phonological similarity on false recognition of drug names. Soc Sci Med. 2001;52:1843\u0026ndash;57. https://doi.org/10.1016/S0277-9536(00)00301-4\u003c/li\u003e\n\u003cli\u003eLambert BL, Donderi D, Senders JW. Similarity of drug names: comparison of objective and subjective measures. Psychol Mark. 2002;19:641\u0026ndash;61. https://doi.org/10.1002/mar.10029\u003c/li\u003e\n\u003cli\u003eKatakura M, Toda N, Kunimoto Y, Takahashi K, Nakamura K, Masuko H, et al. Error prevention through written check of medicine identification codes and total numbers. Jpn J Pharm Health Care Sci. 2008;34:997\u0026ndash;1003. https://doi.org/10.5649/jjphcs.34.997\u003c/li\u003e\n\u003cli\u003eNiwa T, Komori Y, Ishihara M, Takahashi S, Matsuura K, Goto C. Evaluation of the countermeasures for the prevention of dispensing errors with risk analysis. J Jpn Soc Hosp Pharm. 2008;44:1487\u0026ndash;90\u003c/li\u003e\n\u003cli\u003eIkeda K, Kominami S, Morita S, Kita H, Uno M. Effectiveness of dedicated dispensing paper in preventing dispensing errors. Jpn J Pharm Health Care Sci. 2011;37:431\u0026ndash;6. https://doi.org/10.5649/jjphcs.37.431\u003c/li\u003e\n\u003cli\u003eWatanabe H, Yoshida M, Nakahara A, Futagami S, Onoue R, Tsuji T, et al. Measures for prevention of dispensing errors based on ISO 9001 certified management system and their evaluation. Jpn J Pharm Health Care Sci. 2006;32:824\u0026ndash;34. https://doi.org/10.5649/jjphcs.32.824\u003c/li\u003e\n\u003cli\u003eTsuji 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. https://doi.org/10.1186/s40780-015-0011-x\u003c/li\u003e\n\u003cli\u003eTsuji T, Irisa T, Tagawa S, Kawashiri T, Ikesue H, Kokubu C, et al. Differences in recognition of similar medication names between pharmacists and nurses: a retrospective study. J Pharm Health Care Sci. 2015;1:19. https://doi.org/10.1186/s40780-015-0017-4\u003c/li\u003e\n\u003cli\u003eMorrison C, MacRae Y. Promoting safer use of high-risk pharmacotherapy: impact of pharmacist-led targeted medication reviews. Drugs Real World Outcomes. 2015;2:261\u0026ndash;71. https://doi.org/10.1007/s40801-015-0031-8\u003c/li\u003e\n\u003cli\u003eHammar T, Hellstr\u0026ouml;m L, Ericson L. The use of a decision support system in Swedish pharmacies to identify potential drug-related problems\u0026mdash;effects of a national intervention focused on reviewing elderly patients\u0026rsquo; prescriptions. Pharmacy (Basel). 2020;8:118. https://doi.org/10.3390/pharmacy8030118\u003c/li\u003e\n\u003cli\u003eShermock SB, Shermock KM, Schepel LL. Closed-loop medication management with an electronic health record system in U.S. and Finnish hospitals. Int J Environ Res Public Health. 2023;20:6680. https://doi.org/10.3390/ijerph20176680\u003c/li\u003e\n\u003cli\u003eNagata K, Tsuji T, Suetsugu K, Muraoka K, Watanabe H, Kanaya A, et al. Detection of overdose and underdose prescriptions\u0026mdash;an unsupervised machine learning approach. PLOS One. 2021;16:e0260315. https://doi.org/10.1371/journal.pone.0260315\u003c/li\u003e\n\u003cli\u003eTsuji T, Nagata K, Sasaki K, Suetsugu K, Watanabe H, Kanaya A, et al. Elucidation of occurrence mechanism of dispensing errors based on gaze movement of pharmacists using eye tracking method. Jpn J Pharm Health Care Sci. 2019;45:493\u0026ndash;503. https://doi.org/10.5649/jjphcs.45.493\u003c/li\u003e\n\u003cli\u003eTsuji T, Nagata K, Sasaki K, Matsukane R, Ishida S, Kawashiri T, et al. Analysis of the thinking process of pharmacists in response to changes in the dispensing environment using the eye-tracking method. J Pharm Health Care Sci. 2022;8:23. https://doi.org/10.1186/s40780-022-00254-x\u003c/li\u003e\n\u003cli\u003eTsuji T, Nagata K, Tanaka M, Hasebe S, Yukita T, Uchida M, et al. Eye-tracking-based analysis of pharmacists\u0026rsquo; thought processes in the dispensing work: research related to the efficiency in dispensing based on right-brain thinking. J Pharm Health Care Sci. 2024;10:21. https://doi.org/10.1186/s40780-024-00341-1\u003c/li\u003e\n\u003cli\u003eTsuji T, Nagata K, Tanaka M, Shiori I, Hasebe S, Yuto N, et al. Right-brain utilization in pharmacists\u0026rsquo; dispensing processes: an eye-tracking analysis of efficiency and safety using error-induction models. J Pharm Health Care Sci. 2025;11:13. https://doi.org/10.1186/s40780-025-00443-4\u003c/li\u003e\n\u003cli\u003eBarnett KJ. Colour knowledge: the role of the right hemisphere in colour processing and object colour knowledge. Laterality. 2008;13:456\u0026ndash;67. https://doi.org/10.1080/13576500802146387\u003c/li\u003e\n\u003cli\u003ePinna B, Deiana K. New conditions on the role of color in perceptual organization and an extension to how color influences reading. Psihologija. 2014;47:319\u0026ndash;51. https://doi.org/10.2298/PSI1403319P\u003c/li\u003e\n\u003cli\u003eKato Y, Kobayashi T, Hasebe D, Kano H, Saito C. A study in subjective evaluation and gaze point analysis of facial symmetry: analysis using eye-tracking. Jpn J Jaw Deform. 2009;19:184\u0026ndash;92. https://doi.org/10.5927/jjjd.19.184\u003c/li\u003e\n\u003cli\u003eVervoort T, Trost Z, Prkachin KM, Mueller SC. Attentional processing of other\u0026rsquo;s facial display of pain: an eye tracking study. Pain. 2013;154:836\u0026ndash;44. https://doi.org/10.1016/j.pain.2013.02.017\u003c/li\u003e\n\u003cli\u003eLim JZ, Mountstephens J, Teo J. Emotion recognition using eye-tracking: taxonomy, review and current challenges. Sensors (Basel). 2020;20:2384. https://doi.org/10.3390/s20082384\u003c/li\u003e\n\u003cli\u003eWolf A, Ueda K. Contribution of eye-tracking to study cognitive impairments among clinical populations. Front Psychol. 2021;12:590986. https://doi.org/10.3389/fpsyg.2021.590986\u003c/li\u003e\n\u003cli\u003eIshibashi A. Human factors and error prevention. J Natl Inst Public Health. 2002;51:232\u0026ndash;44\u003c/li\u003e\n\u003cli\u003eMiller GA. The magical number seven plus or minus two: some limits on our capacity for processing information. Psychol Rev. 1956;63:81\u0026ndash;97. https://doi.org/10.1037/h0043158\u003c/li\u003e\n\u003cli\u003eCowan N. The magical number 4 in short-term memory: a reconsideration of mental storage capacity. Behav Brain Sci. 2001;24:87\u0026ndash;114; discussion 114. https://doi.org/10.1017/S0140525X01003922\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Table","content":"\u003cp\u003eTable 1 is available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Eye-tracking method, prescription audit, thought process, efficiency and safety, set phrase display, numerical table display","lastPublishedDoi":"10.21203/rs.3.rs-8139848/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8139848/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e\u003cp\u003ePharmacists need to explore more efficient method of prescription audits to ensure safe and reliable medical therapy for patients. In this study, pharmacists\u0026rsquo; gaze movements during the prescription audit process were analyzed using an eye-tracking system.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eFirst, two methods for displaying drug\u0026rsquo;s dose and usage during prescription audits were developed: \u0026ldquo;set phrase display\u0026rdquo; and \u0026ldquo;numerical table display.\u0026rdquo; Second, three key items were defined: (a) drug name, (b) drug dose and usage, and (c) dose period. The number of gaze points for these items were designated as Gaze 1, Gaze 2, and Gaze 3, and the time required for verification was recorded as Time. Third, three pairs of models (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e, and C\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e) with different levels of difficulty for prescription auditing were prepared, along with one pair of models (Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e) that included a drug\u0026ndash;drug interaction. This study aimed to demonstrate the differences in pharmacists' information processing between the two expression methods (set-phrase display and numerical table display) using four pairs of prescription auditing models.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eA total of 22 pharmacists participated in the study. During the prescription audit process, gaze movements followed the pattern \u0026ldquo;set phrase display\u0026thinsp;\u0026gt;\u0026thinsp;numerical table display\u0026rdquo; in all three models (A\u003csub\u003e1\u003c/sub\u003e-A\u003csub\u003e2\u003c/sub\u003e, B\u003csub\u003e1\u003c/sub\u003e-B\u003csub\u003e2\u003c/sub\u003e, and C\u003csub\u003e1\u003c/sub\u003e-C\u003csub\u003e2\u003c/sub\u003e). Significant differences between model pairs were observed in Gaze 2 and Time, both favoring \u0026ldquo;set phrase display\u0026thinsp;\u0026gt;\u0026thinsp;numerical table display\u0026rdquo;. Furthermore, significant differences in gaze movements between models Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e were consistent with those observed in the three model pairs. Recognition rates for models Q\u003csub\u003e1\u003c/sub\u003e-Q\u003csub\u003e2\u003c/sub\u003e were 72.7% (8/11) and 100% (11/11), respectively. Recognition times were 20.5\u0026thinsp;\u0026plusmn;\u0026thinsp;6.2 and 16.0\u0026thinsp;\u0026plusmn;\u0026thinsp;3.8, respectively. No significant differences were observed between them.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e\u003cp\u003eThe results demonstrated that the \u0026ldquo;numerical table display\u0026rdquo; method not only enables prescription audit to be performed without unnecessary procedures but also improves the accuracy of drug\u0026ndash;drug interaction judgments. In other words, introducing the \u0026ldquo;numerical table display\u0026rdquo; method for drug dose and usage can enhance both efficiency and safety in prescription audit.\u003c/p\u003e","manuscriptTitle":"Eye-tracking-based analysis to improve the efficiency and safety of prescription audit","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-03 11:44:44","doi":"10.21203/rs.3.rs-8139848/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"79b6fdcb-e5fc-4737-b622-e97314d6c23c","owner":[],"postedDate":"December 3rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-12-22T02:54:10+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-03 11:44:44","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8139848","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8139848","identity":"rs-8139848","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}