Eye-tracking devices measure “where, how, and when people see” by exposing the cornea to near-infrared radiation and analyzing images of eye movements based on a measurement method called the corneal reflection method [12] . This study used the Tobii Pro Glasses 2 (Tobii Corp.) ( Figure 1 ) eye-tracking system to examine the optimal distance and angle between a pharmacist and a simulated patient (SP) during patient-compliance instructions. This system defines a saccade as movements of the line of sight at a speed of 100 deg/sec within 20 msec, and fixation is defined as a period between saccades. The eye-tracking system consists of the Tobii Pro Glasses 2, a head unit, a scene camera, and a recording unit [13] .

This study was conducted with the cooperation of seven female SPs. To avoid the influence of gender difference, only female SPs were conducted for this study. Each SP was asked to act out scenarios of diseases specific to women, such as “endometriosis” and “vaginal candida.” The corresponding pharmacists were a man in his 40 s (10 years’ experience as a pharmacist) and a woman in her 20 s (5 years’ experience); regarding the number of times they had met with the SPs, this was mostly their first time. In addition, to avoid the influence of change in tone of voice, the same pharmacist provided all medication instructions.

In previous research on interpersonal distance, Kikuzawa showed in a survey that “the optimal distance during interaction within a residence is 0.8 to 1.2 m” and “the minimum distance during dialog by gender is: female to female at 0.55 m, and this distance is shorter when the direction of approach is from behind” [14] . In addition, Hashimoto et al. report that “for people who are not talking to each other, the limit for approaching one another is approximately 1.5 m” [15] . Thus, people have a limit to the distance that others can approach them. Therefore, if the pharmacist gets too close to the patient, it may interfere with medication instruction. Building on these previous studies, in this study, a pharmacy and a hospital room were assumed as the settings where patient interactions would occur. In addition, male and female pharmacists were alternated to provide patient-compliance instructions to each SP. In the sessions between pharmacists and patients that assumed a pharmacy setting, the participants were positioned to face each other across a desk, and measurement positions were set between the pharmacist and the SP at ① 50 cm and ② 70 cm, at ③ 90-degree angle with a 70 cm desk in between for medication instruction in three patterns ( Figure 2 ).

In the investigation between the pharmacist and patient in the hospital room setting, the pharmacist provided medication instructions to a patient sitting in the center of the bed. To determine the pharmacist’s optimal position and height, the pharmacist-patient position was measured at three positions: ① 0-degree (just beside), ② 45-degree, and ③ 90-degree (in front). Additionally, male and female pharmacists performed patient-compliance instructions for SPs—while sitting in the 0- and 45-degree conditions, and while standing in the 90-degree condition, for a total of six patterns ( Figure 3 ).

SP eye movements were measured in all 12 patterns of pharmacy and hospital settings as described above, and all SPs were video recorded during the medication instruction.

Hashimoto et al.’s “conversation scale” was used to measure SPs’ subjective evaluations at each measurement point to assess whether the selected location was appropriate for conversation [15] . Conversation was measured on a 5-point scale, with the left end being “too far away,” the center “optimal,” and the right end “too close.” ( Figure 4 )

The analysis was divided into four scenes: introducing the medication instructions (from confirmation of patient name to patient reply); confirming the reason for the medical examination (today you were seen for [medical condition name] ~); medication guidance (today’s medicine ~); and the closing scene (this concludes the explanation ~). The comparative study was then conducted on the two scenes where patients were thought to be more aware of the interpersonal situation: the introduction and confirmation scenes.

The data obtained were linked to video footage and analyzed using the analysis software, Tobii Pro Lab (Tobii Corp.). IBM SPSS ver.26 (IBM Corp.) was used for data analysis.

To account for the effects of individual differences, pupil diameters were compared using the average pupil diameter for the introduction and confirmation scenes for each SP.

Regarding blink count, the number of blinks was visually measured using a counter based on front-facing video of the SP during the medication instructions, and the blink rate per unit time {blink count/time (min)} was calculated. Furthermore, the effects of individual differences were taken into account and compared using the rate of change in the blink rate of each SP {(blink rate at each point − average blink rate at all points (introduction/confirmation scenes only))/average blink rate at all points × 100}.

Regarding saccades, considering the variation in measurement time, the ratio of saccades per total time of each SP introduction/confirmation scene {total saccade time (msec)/total time (msec) × 100} was calculated.

As for fixation, the rate of fixation per total time during the introduction/confirmation scene of each SP {total fixation time (msec)/total time (msec) × 100} was calculated.

The Friedman test and Wilcoxon’s signed-rank test were performed in the analysis. After the Friedman test, multiple comparisons were performed on those for which significant differences were observed.

This study was conducted with the approval of Kinjo Gakuin University Ethics Review Committee (No. H17015).

The conversation scale survey results suggest that, in the pharmacy setting, the pharmacist-patient distance was too close to the pharmacist, which interfered with conversation in situations where the pharmacist-patient was at the 50 cm and 90-degree ( Table 1 ). On the other hand, in the hospital room setting, when the pharmacist was standing in front of the patient, it was suggested that the pharmacist was too far away, which interfered with the conversation ( Table 2 ). In addition, when the pharmacist was standing at the 0-degree (just beside), it was suggested that the patient was too close to the pharmacist, which interfered with the conversation ( Table 2 ).

Pupil diameter is understood to increase due to emotional discomfort and/or mental stress [16] . The results for the pharmacy setting showed that the pupils were significantly more mydriatic at the 50 cm point than at the other points (p < 0.01). The results for the hospital room setting showed the most mydriasis at the 0-degree point. Mydriasis was also significantly greater at the 0-degree point than at the 90-degree point (p < 0.01).

In addition, a comparison of pupil diameter by gender showed that male pharmacists were significantly more mydriatic when they gave medication instructions (p < 0.01) ( Figure 5 ).

Blinking, the so-called “blink” of the eye, is thought to reflect a person’s psychological state and personality traits. It is said to increase when anxiety or anger is felt and be suppressed when looking at an object or staring without looking away [8] [17] .

The results for the pharmacy setting showed that blinking occurred more frequently at the 50 cm point than at the other points ( Figure 6 ). Comparisons were made for the hospital room setting by gender, but there were no significant differences.

Saccades are voluntary movements that can serve as indicators of spatial attention being paid to emotional stimuli [5] . Accordingly, as the gaze becomes unstable when the mind is agitated, we hypothesized that the frequency of saccades would increase in scenes where subjects were distracted and restless. The pharmacy setting results showed no significant difference in the incidence of saccades. On the other hand, When the likelihood of saccades occurrence in the hospital room setting was compared, saccades were most likely to occur when patient medication instructions were given at the 0-degree point and least likely to occur at the 90-degree point. Because the Friedman test showed significant differences, a multiple comparisons test was performed, yielding the following results: significant differences were found between the 0-degree point and 45-degree point, the 0-degree point and 90-degree point (n = 7, p < 0.05). A comparison by gender of pharmacists giving patient medication instructions showed a significant increase in the occurrence of saccades when the pharmacist giving patient medication instructions was male rather than female (n = 7, p < 0.05) ( Figure 7 ).

Used to send a signal opposite to “averting one’s fixation,” a classic sign of rejection, fixation is a physiological reaction, the frequency of which we assumed would increase as subjects’ attention is directed toward the other person. A comparison of fixation rates at each point in the pharmacy and hospital room settings showed no differences between points. Comparisons of fixation rates were made for the pharmacy and hospital room setting by gender, but there were no significant differences ( Figure 8 ).