2025 Volume 47 Issue 2 Pages 55-67
This study aimed to clarify the association between readiness to return to work, cognitive function, and work continuation after returning to work in workers with major depressive disorder. We assessed whether cognitive assessments in THINC-it could be used as intermediate variables to explain the association between The Psychiatric rework readiness scale (PRRS) total score and work continuation one year after returning to work in workers with major depressive disorder. Twenty-four individuals participated. The psychiatric rework readiness scale total score was significantly associated with work continuation one year after return to work (β = 0.197, 95% CI [0.002, 0.391]). PRRS total score and the five-item Perceived Deficits Questionnaire-Depression (PDQ-D5) were also significantly associated (β = 0.072, 95% CI [0.014, 0.130]). However, the indirect effects were not significant for PDQ-D5, Spotter, Symbol Check, Codebreaker, or Trails. Better readiness for return to work in workers with major depressive disorder was associated with a better assessment of the PDQ-D5. Five cognitive assessments as mediating variables did not explain the association between the evaluation of readiness for return to work and work continuation one year later.
Workers with major depressive disorder are at high risk of taking more sick leave even after returning to work. Sick leave caused by major depressive disorder is common, accounting for more than half of all leave, because of mental health problems in Japan [1]. Even after returning to work, 40% of workers with major depressive disorder take more sick leave within three years [2]. Repeated sick leave due to mental health problems has been suggested as a factor leading to further and longer duration of sick leave [2, 3]. Extending the duration of work continuation after returning to work and avoiding further sick leave are critical.
Readiness to return to work means whether a worker on sick leave is able to return to work without a recurrence of their illness [6]. Better readiness to return to work is associated with a lower likelihood that workers with major depressive disorder will take further sick leave later [4, 5]. Approaches to enhancing readiness to return to work in workers with major depressive disorder include pharmacotherapy, psychotherapy, participation in a Re-Work program, and discussion with workplace managers and occupational health professionals [4, 7, 8]. A Re-Work program is designed to assist employees on sick leave with their mental health problems and includes group psychotherapy and occupational therapy [8].
The Psychiatric rework readiness scale (PRRS) is a measure of readiness to return to work in workers with mental health problems [9]. It contains eight subscales: basic life rhythm, symptoms, social aptitude, support resources, relationship with the workplace, work ability, preparation status, and health management. A higher PRRS total score is associated with longer work continuation after return to work for workers with major depressive disorder [4, 5].
Lower cognitive dysfunction on return to work has been associated with a lower likelihood that workers with major depressive disorder will take further sick leave [10, 11]. Cognitive dysfunction is one of the symptoms of major depressive disorder [12], and can remain even after depressive episodes have stopped [13–16]. Ongoing cognitive dysfunction in people with major depressive disorder makes them more prone to recurrent depressive episodes [17, 18] and decreases their work performance [19–22]. Previous studies have found that better performance on cognitive assessments, such as the 3-back task and the Trail Making Test Part B (TMT-B) on return to work were associated with work continuation after return to work for workers with major depressive disorder [10, 11].
Readiness to return to work may be associated with reduced cognitive dysfunction in workers with major depressive disorder. Better readiness to return to work reflects the results of effective approaches, such as pharmacotherapy, psychotherapy, and Re-Work programs, which may contribute to reduced cognitive dysfunction. All these approaches have been suggested to be effective in reducing cognitive dysfunction in people with major depressive disorder [14, 23–25]. We suggest that better readiness to return to work reflects the overall effect of these approaches and is associated with reduced cognitive dysfunction. However, no previous studies have investigated the association between readiness to return to work and cognitive function. One related study reported improved PRRS, a measure of readiness to return to work, and reduced cognitive dysfunction after participation in Re-Work programs [25]. Although this previous study did not specifically examine the association between readiness to return to work and cognitive function, the findings suggest a potential relationship between the two.
It is possible that cognitive function on return to work might mediate the association between evaluation of readiness to return to work and work continuation for workers with major depressive disorder (Figure 1). Higher readiness to return to work may be associated with lower cognitive dysfunction, potentially prolonging the duration of work continuation. No studies have explicitly examined the association between these three factors. Clarifying these associations could lead to more effective guidance by occupational health professionals for workers on sick leave with major depressive disorder.
We formulated two hypotheses to clarify the association between readiness to return to work, cognitive function, and work continuation after return to work of workers with major depressive disorder:
Hypothesis 1: Greater readiness to return to work in workers with major depressive disorder is associated with higher cognitive function.
Hypothesis 2: Cognitive function on return to work in workers with major depressive disorder serves as an intermediate variable, explaining the association between readiness to return to work and work continuation one year later.
Study Design
This was a cohort study of workers with major depressive disorder. We evaluated the association between readiness to return to work, cognitive function on return to work, and work continuation one year later. All participants gave written informed consent before participating in this study. This study was conducted under the approval of the Ethics Committee of Medical Research, University of Occupational and Environmental Health, Japan (R3-021).
Setting and Participants
Participants were enrolled from September 2021 to March 2023 and were followed until February 2024. Participants were recruited at seven medical institutions in Fukuoka Prefecture. All of the medical institutions were registered with the Japanese Association of Rework for Depression. A baseline survey was conducted with participants on sick leave, including THINC-it, PRRS, and other data. Follow-up was conducted every three months after return to work to assess work continuation, and this follow-up continued for up to one year or until work continuation was interrupted.
The study included participants who met the following criteria: (1) between 20 and under 60 years of age, (2) diagnosed with major depressive disorder by a psychiatrist, (3) in remission from their condition, (4) participating in a Re-Work program, and (5) having a confirmed return to work. Exclusions were those with central nervous system disorders, severe psychiatric symptoms, alcohol or other substance addiction, or intellectual disabilities. The reason for limiting the study to Re-Work program participants was that they were less encumbered by the demands of participation than those who were not participating in a Re-Work program. Re-work participants attend medical facilities relatively often, but other patients typically only visit once every few weeks. It would therefore have been necessary to ask them to attend a medical institution specifically for the study at a time just before their return to work.
Variables
Work continuation was evaluated by whether participants continued to work consecutively from the date of their return to work to one year later. We defined the categorical variable as 1 for continued and 0 for discontinued. The interval between the date of return to work and the date of discontinuation was defined as the duration of work continuation. The date of return to work was defined as the date announced for this, even if the worker worked reduced hours or on a trial basis. Participants were considered not to have continued working if they resigned or were absent after that date for more than 30 days. Follow-up was conducted every three months by e-mail or telephone.
The cognitive function of the participants was assessed using the five cognitive assessments included in the THINC-it. The THINC-it is a computerized cognitive assessment tool. It has been validated for the assessment of cognitive function in patients with major depressive disorder and healthy controls [26, 27]. It consists of the subjective cognitive assessment five-item Perceived Deficits Questionnaire-Depression (PDQ-D5) and objective cognitive assessments, including Spotter, Symbol Check, Codebreaker, and Trails. Spotter, Symbol Check, Codebreaker, and Trails are based on the Choice Reaction Time, 1-back task, Digit Symbol Substitution Test, and TMT-B. The TMT-B has been reported to be associated with work continuation after return to work [11]. Each test result was converted to a z-score using the results of Japanese workers in employment from Shibaoka et al [28]. All z-scores were adjusted so that a higher score would indicate higher cognitive function. In this study, the THINC-it for Android was administered in Japanese on an 8-inch tablet.
Readiness to return to work was evaluated using the PRRS total score and subscales. The PRRS is an interview assessment scale designed to evaluate the overall readiness to return to work of workers with mental health problems [9]. The scale has been validated for inter-rater reliability and internal consistency [4]. The PRRS contains eight subscales: basic life rhythm, symptoms, social aptitude, support resources, relationship with the workplace, work ability, preparation status, and health management. The total PRRS score calculated from these subscales has been reported to be significantly predictive of work continuation after return to work [4, 5]. The PRRS is a 23-item instrument that uses a 4-point Likert scale, with higher scores indicating better readiness to return to work. One indicator of readiness to return to work is an average score of at least three points for each question, or a total PRRS score of at least 69 points [25]. All of the surveys were conducted by the same interviewer.
As part of the baseline survey, we obtained information about age, gender, years of education, marital status, industry, occupation, existence of occupational health professional, number of previous sick leave, and K6 score. The industry was categorized using the Japan Standard Industrial Classification [29]. Occupation was classified into managerial positions and non-managerial positions. The K6 is a scale developed by Kessler et al [30], and its Japanese version has been verified for reliability and validity [31]. In this study, we used the cutoff value of 4/5 proposed by Sakurai et al as the optimal cutoff for screening equivalent psychological stress [32].
Statistical Analysis
The data at baseline were analyzed using descriptive statistics, and the duration of work continuation since the return to work was analyzed using the Kaplan-Meier method.
Models were constructed using the five cognitive assessments, readiness for return to work, and work continuation (Figure 2). On the basis of a previous study, the five cognitive functions from THINC-it were used simultaneously [28]. Following Baron & Kenny’s approach [33], we conducted four steps. First (Path c), we performed a logistic regression analysis with work continuation as the dependent variable and the PRRS total score as the independent variable. Second (Path a1–5), we conducted separate simple regression analyses with each cognitive assessment as the dependent variable and the PRRS total score as the independent variable. Third (Path b1–5), we conducted a logistic regression analysis with work continuation as the dependent variable and the PRRS total score and five cognitive assessments as independent variables. Fourth, if the first to third steps were significant, we compared the absolute values of the coefficients of the PRRS total score in Path c and the absolute values of the coefficients of the PRRS total score in Paths c’.
Furthermore, separate single regression analyses were conducted with each cognitive assessment as the dependent variable and the eight subscales of the PRRS as independent variables.
All analyses used R version 4.3.2. We defined P < 0.05 as statistically significant and P < 0.1 as marginally significant. No covariates were entered into the regression analysis because of the small number of study participants.
Participants
Of the 52 individuals who met the inclusion criteria, 24 (46.2%) consented to participate in this study, and all were included in the analysis (Figure 3). All of the participants were followed until they stopped work or for one year after return to work. Three of the participants were unavailable for interview, thus the reason for their inability to continue their work remains unknown. However, no data were missing for the variables used in the statistical analysis.
Descriptive statistics
The characteristics of the participants are shown in Table 1. The mean age was 44.79 (SD = 8.31), with just one participant in their 20s (4.2%). Most of the participants were non-managers (n = 22, 91.7%). All of the participants worked in different workplaces, and the industries in which they worked varied widely. However, occupational health professionals were present in almost all of the workplaces (n = 23, 95.8%). Approximately half (n = 14, 58.3%) of the participants were returning from their first period of sickness absence. The K6 score was higher than 5 in approximately 40% (n = 10, 41.7%).
| n = 24 | |
|---|---|
| Age | |
| Mean (SD) | 44.79 (8.31) |
| 20 - 29, n (%) | 1 (4.2) |
| 30 - 39, n (%) | 5 (20.8) |
| 40 - 49, n (%) | 9 (37.5) |
| 50 - 60, n (%) | 9 (37.5) |
| Gender | |
| Female, n (%) | 9 (37.5) |
| Male, n (%) | 15 (62.5) |
| Education | |
| Mean (SD), years | 14.67 (2.39) |
| Marital status | |
| Married, n (%) | 14 (58.3) |
| Single, n (%) | 10 (41.7) |
| Industry | |
| Construction, n (%) | 1 (4.2) |
| Manufacturing, n (%) | 3 (12.5) |
| Electricity, gas, heat supply and water, n (%) | 3 (12.5) |
| Transport and postal activities, n (%) | 1 (4.2) |
| Wholesale and Retail trade, n (%) | 3 (12.5) |
| Finance and insurance, n (%) | 3 (12.5) |
| Real estate and goods rental and leasing, n (%) | 1 (4.2) |
| Scientific research, professional and technical services, n (%) | 1 (4.2) |
| Education, learning support, n (%) | 1 (4.2) |
| Medical, health care and welfare, n (%) | 1 (4.2) |
| Compound services, n (%) | 1 (4.2) |
| Government, except elsewhere classified, n (%) | 5 (20.8) |
| Occupation | |
| Managerial position, n (%) | 2 (8.3) |
| Non-managerial position, n (%) | 22 (91.7) |
| Occupational health professionals | |
| Presence, n (%) | 23 (95.8) |
| Absence, n (%) | 1 (4.2) |
| Sickness absence episodes | |
| 1, n (%) | 14 (58.3) |
| 2, n (%) | 6 (25.0) |
| ≥3, n (%) | 4 (16.7) |
| K6 | |
| ≤4, n (%) | 14 (58.3) |
| ≥5, n (%) | 10 (41.7) |
n : Number of participants, SD : Standard Deviation.
The data on work continuation, the PRRS total score and subscales, and cognitive assessment on return to work are shown in Table 2, and the work continuation curve is illustrated in Figure 4. Overall, 16 participants (66.7%) continued to work. The mean PRRS total score was 77.88 (SD = 5.51). The mean total score for the Work ability subscale, which consists of three items, was 8.88 (SD = 1.51). When converted to a mean score per item, this equates to approximately 2.96, which is below the recommended threshold of three points for readiness to return to work. In contrast, all of the other subscales exhibited mean per-item scores above three points. The mean z-scores for the PDQ-D5, Symbol Check, Codebreaker, and Trails were 0.01 (SD = 0.87), –0.07 (SD = 0.69), –0.08 (SD = 0.74), and –0.09 (SD = 0.74), which were all near 0. The mean z-score for the Spotter was –0.42 (SD = 0.86), which was low compared to other cognitive assessments.
| n = 24 | |
|---|---|
| Work continuation | |
| Continued, n (%) | 16 (66.7) |
| Discontinued, n (%) | 8 (33.3) |
| Readiness for RTW Mean (SD) | |
| PRRS total score | 77.88 (5.51) |
| Basic life rhythm | 10.58 (1.47) |
| Symptoms | 19.58 (2.75) |
| Social aptitude | 7.00 (0.66) |
| Support resources | 7.17 (1.01) |
| Relationship with the workplace | 7.42 (0.88) |
| Work ability | 8.88 (1.51) |
| Preparation status | 6.21 (1.72) |
| Health management | 10.96 (1.23) |
| Cognitive function Mean (SD) | |
| PDQ-D5 | 0.01 (0.87) |
| Spotter | -0.42 (0.86) |
| Symbol Check | -0.07 (0.69) |
| Codebreaker | -0.08 (0.74) |
| Trails | -0.09 (0.74) |
n : number of participants, SD : standard deviation, RTW: return to work, PDQ-D5: five-item Perceived Deficits Questionnaire–Depression, PRRS: Psychiatric Rework Readiness Scale. PRRS total score range is 23-96, with subscale scores as follows: Basic life rhythm (3-12), Symptoms (6-24), Social aptitude (2-8), Support resources (2-8), Relationship with the workplace (2-8), Work ability (3-12), Preparation status (2-8), and Health management (3-12).
The association between the PRRS total score, cognitive assessments, and work continuation
The results of the analyses for the model are shown in Figure 5. The total effect (path c), the PRRS total score, was significantly associated with work continuation (β = 0.197, 95% CI [0.002, 0.391]). Indirect effects were not significant for any of the cognitive assessments. The PRRS total score was significantly associated with PDQ-D5 (path a1, β = 0.072, 95% CI [0.014, 0.130]). However, PDQ-D5 was not significantly associated with work continuation (path b1, β = 0.043, 95% CI [–1.366, 1.452]). Spotter, Symbol Check, Codebreaker, and Trails were not significant for either path a2–5 or path b2–5. None of the indirect effects were significant, and we therefore did not compare the absolute values of the coefficients of the PRRS total score for path c with the absolute values of the coefficients of the PRRS total score for path c’.
The associations between the PRRS subscales and cognitive assessments
The associations between the PRRS subscales and each cognitive assessment are shown in Table 3. The association between Symptoms and Trails was significant (β = –0.119, 95% CI (–0.226, –0.013)), and the association with Codebreaker was marginally significant (β = –0.121, 95% CI (–0.212, 0.008)). The association between Support resources and PDQ-D5 was marginally significant (β = 0.330, 95% CI (–0.022, 0.681)). The association between Relationship with the workplace and PDQ-D5 was significant (β = 0.536, 95% CI (0.172, 0.901)). The association between health management and Symbol Check was significant (β = 0.260, 95% CI (0.041, 0.479)).
| PDQ-D5 | Spotter | Symbol Check | Codebreaker | Trails | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| β | 95%CI | β | 95%CI | β | 95%CI | β | 95%CI | β | 95%CI | |
| PRRS total score | 0.072* |
(0.011, 0.133) |
-0.011 |
(-0.080, 0.057) |
0.000 | (-0.055, 0.055) | -0.014 | (-0.073, 0.044) | -0.020 | (-0.078, 0.038) |
|
Basic life rhythm |
0.201 | (-0.044, 0.446) | -0.173 | (-0.420, 0.074) | -0.141 | (-0.339, 0.056) | -0.141 | (-0.355, 0.073) | -0.032 | (-0.253, 0.189) |
| Symptoms | 0.091 | (-0.042, 0.225) | -0.042 | (-0.179, 0.095) | -0.048 | (-0.157, 0.061) | -0.121† | (-0.212, 0.008) | -0.119* | (-0.226, -0.013) |
|
Social aptitude |
0.242 | (-0.329, 0.813) | -0.089 | (-0.665, 0.487) | 0.200 | (-0.254, 0.654) | 0.238 | (-0.248, 0.724) | -0.108 | (-0.600, 0.384) |
|
Support resources |
0.330† | (-0.022, 0.681) | -0.297 | (-0.651, 0.058) | -0.106 | (-0.406, 0.193) | -0.086 | (-0.409, 0.238) | -0.085 | (-0.406, 0.237) |
|
Relationship with the workplace |
0.536* | (0.172, 0.901) | -0.069 | (-0.500, 0.363) | 0.047 | (-0.299, 0.393) | 0.280 | (-0.071, 0.631) | 0.126 | (-0.240, 0.492) |
| Work ability | 0.143 | (-0.103, 0.388) | 0.150 | (-0.093, 0.393) | 0.074 | (-0.125, 0.273) | 0.002 | (-0.215, 0.219) | -0.004 | (-0.219, 0.212) |
|
Preparation status |
-0.109 | (-0.327, 0.108) | 0.012 | (-0.210, 0.233) | 0.028 | (-0.149, 0.205) | 0.095 | (-0.091, 0.281) | 0.118 | (-0.064, 0.301) |
|
Health management |
0.157 | (-0.146, 0.460) | 0.232 | (-0.059, 0.523) | 0.260* | (0.041, 0.479) | 0.074 | (-0.190, 0.338) | 0.038 | (-0.226, 0.302) |
β : Beta coefficient, 95% CI: 95% Confidence Interval, PRRS: Psychiatric Rework Readiness Scale, PDQ-D5: Perceived Deficits Questionnaire-Depression 5 item. * P<0.05 †P<0.1
We tested two hypotheses to clarify the association between the evaluation of readiness to return to work, cognitive function, and work continuation after return to work in workers with major depressive disorder. Hypothesis 1 was supported by one of the five cognitive assessments. Workers with major depressive disorder demonstrated an association between their PRRS total scores and PDQ-D5, but there was no association with Spotter, Symbol Check, Codebreaker, or Trails. Hypothesis 2 was rejected for all five cognitive assessments. The intermediate variables PDQ-D5, Spotter, Symbol Check, Codebreaker, and Trails could not explain the association between PRRS total score and work continuation in workers with major depressive disorder. They were not associated with work continuation in particular.
The discrepancies in outcomes between the PDQ-D5 and other cognitive assessments observed in Hypothesis 1 may have been due to differences in assessment methods. Previous studies have found that PRRS total score is associated with severity of major depressive disorder, and social functioning [34], and that subjective cognitive function was associated with severity of major depressive disorder [13, 16, 19, 35–37] and social functioning [19, 21, 22, 35]. However, other studies reported that objective cognitive function was not associated with severity of major depressive disorder or social functioning [36, 38]. The PDQ-D5 assessed subjective cognitive function, whereas the Spotter, Symbol Check, Codebreaker, and Trails assessed objective cognitive function. In other words, the discrepancy in the results may have depended on whether the assessment of cognitive function was subjective or objective.
Based on the PRRS subscales and the results of each cognitive assessment, we considered it appropriate to use the PRRS total score to assess its association with cognitive function. Some of the results of the PRRS subscales and each cognitive assessment in this study were difficult to explain, as better readiness to return to work was associated with poorer cognitive recovery. This may have been due to limitations of the sample size and statistical instability caused by analyzing the PRRS subscales separately. Indeed, in two previous studies on the PRRS, the association with duration of work continuation was significant for the PRRS total score, but the results for the PRRS subscales were inconsistent [4, 5]. Horii et al reviewed their own study and that of Sakai et al and suggested a cautious interpretation of the validity of the PRRS subscales due to the lack of association of Preparation status, the subscale most expected to be associated with the duration of work continuation, and issues related to small sample sizes [5]. In light of the aforementioned results and the findings of previous studies, it seems reasonable to conclude that total scores, rather than subscale scores from the PRRS, should be used when assessing the association between the PRRS and cognitive function.
The rejection of Hypothesis 2 can be explained in two ways. The first is that the cognitive assessments may not have adequately assessed cognitive function related to work continuation. The Symbol Check was based on a 1-back task [39]. A study by Hori et al found that the 3-back task was significantly associated with work continuation, but the less difficult 0-back and 2-back tasks were not [10]. We consider that the Symbol Check was less difficult than the 2-back task and therefore did not adequately assess cognitive function related to work continuation. Similar problems may exist with the other assessments used in this study. The second issue is that there may be other factors that are more important to work continuation than cognitive function. A study by Yamashita et al found that the TMT-B, on which the Trails is based, was significantly associated with work continuation [11]. However, the participants in our study differed from Yamashita et al’s participants in the percentage who worked in each type of industry. It is possible that the factors that matter for work continuation are work content and work environment after return to work, rather than cognitive function.
Furthermore, the results of the evaluation of readiness to return to work and cognitive function indicated that the work continuation after return to work in this study may be associated with work content and work environment. The participants exhibited better readiness to return to work. A PRRS total score of 69 or higher, which equates to an average of at least three points on each PRRS item, is indicative of a return to work [25]. In addition, Horii et al reported that a cutoff score of 64/65 points indicates moderate predictability of work continuation six months after returning to work [5]. The PRRS total score for all of the participants was 65 points or above, with only one participant not achieving a score of 69 or above. Additionally, the cognitive function of the participants was comparable to that of working Japanese workers, contrary to the anticipated outcome that it would be lower than that of the average worker. This may be attributed to the fact that all of the participants were engaged in the Re-Work program, which may have contributed to reduced cognitive dysfunction [25]. These findings suggest that the participants’ condition at the time of return to work was less problematic. In contrast, the rate of continued work after one year was 66.7% in this study, which was lower than that observed in previous studies of rework program participants, which were 83.2% [40] and 76.2% [41], respectively. However, the manner in which the work continuation curve declined one year later was similar to that observed in previous studies, with the rate of work continuation decreasing throughout the year without being biased toward a specific time period [4, 40]. Therefore, we considered the possibility that work continuation was not related to factors to which workers were exposed at a specific time after their return to work, but to factors to which they were continuously exposed after their return to work, such as work content and work environment.
Our results suggest two avenues for future research. First, Hypothesis 2 needs to be tested using a different cognitive assessment. This study used only a subset of the cognitive assessments used in studies of major depressive disorder. Second, Hypothesis 2 needs to be tested with factors other than cognitive function. In particular, it is important to assess the hypothesis against factors known to be associated with return to work, such as work content and work environment.
This study had several limitations. The first is that the participants were limited to those also participating in the Re-Work program. Re-Work programs typically span several months, and participants must therefore be employed in an environment that permits them to take extended sick leave. They must also be able to afford to participate in the Re-Work program because the cost of participation is in addition to usual medical expenses. To enable the results to be generalized, future studies should include workers who were not participating in the Re-Work program. The second is that many of the eligible candidates did not agree to participate. Some may have been either less anxious or less confident about return to work. The third is that the analysis assumed a causal relationship between readiness to return to work and cognitive function. However, it is also possible that the reverse causal relationship exists. In this study, data on readiness to return to work and cognitive function were collected concurrently. If cognitive function had been assessed initially, it might have been possible to establish a model in which cognitive function influences readiness to return to work. The fourth is that the sample size was small, which may have affected the statistical power of this study. Future studies with larger sample sizes are warranted.
In this study, greater readiness to return to work in workers with major depressive disorder was associated with better subjective cognitive function, as assessed by the PDQ-D5. However, the five cognitive assessments measured by the THINC-it as intermediate variables did not explain the association between readiness to return to work and work continuation one year later.
We would like to express our gratitude to all of the participants who contributed to the study. We also thank Kaname Clinic, Kawano Clinic, Kurosaki Central Clinic, Shiranui Clinic, Studio Rika Clinic, Tateiwa Hospital, and Yahata Kohsei Hospital for their assistance in recruiting participants, and Melissa Leffler, MBA, from Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
