Patient preferences for advanced therapies in ulcerative colitis using conjoint analysis
Article information
Abstract
Background/Aims
Selecting an optimal advanced therapy for ulcerative colitis (UC) is difficult because of the increasing number of available therapies. This study assessed UC patients’ preferences for drug profiles in decision-making regarding advanced therapies using conjoint analysis.
Methods
A web-based survey was conducted from October to November 2023 in patients with UC aged ≥ 18 years with prior oral 5-aminosalicylic acid treatment (UMIN000052327). We quantified the importance of drug attributes (location of administration, route/frequency of administration, speed of onset-of-action, maintenancesustainability, risk of serious adverse events within 1 year, and novelty of the drug) and the part-worth utility of attribute levels in mild and severe symptom scenarios, including among employed versus unemployed patients.
Results
Of 372 patients who completed the survey, 365 were evaluated. Patient preferences were generally highly individualized. The route/frequency of administration was the most important attribute in both the mild and severe symptom scenarios. Oral administration was preferred in the mild symptom scenario, whereas no specific preference was observed in the severe symptom scenario. The route/ frequency of administration was more valued in the mild symptom scenario than in the severe one, whereas speed of onset of action was more valued in the severe symptom scenario. No significant difference was found in the preference for drug profiles between employed and unemployed patients.
Conclusions
Patient preferences for the route/frequency of administration, as well as other drug profiles, change with disease severity but demonstrate substantial interindividual variability. Therefore, shared decision-making is important to incorporate patients’ perspectives into the selection of advanced therapies.
INTRODUCTION
Ulcerative colitis (UC) is a chronic inflammatory bowel disease (IBD) of unknown etiology, presenting with repeated relapses and remission of symptoms [1,2]. The disease course of UC requires lifelong treatment, as it is a relatively early onset disease that peaks in patients aged 30–40 years, with a lack of curative treatment options [1,2]. Major symptoms of UC include increased blood in stool, stool frequency with diarrhea, abdominal pain, bowel urgency, and fatigue [1,2], which are reported to decrease patients’ quality of life, affect work disability, and influence psychological and social well-being [1,3-5].
In recent years, several advanced therapies (biologics, Janus kinase inhibitors, and calcineurin inhibitors) have emerged for patients with steroid-resistant and steroid-dependent UC, shifting the treatment target from symptomatic remission to endoscopic healing [1]. Nevertheless, studies comparing the efficacy and safety of these advanced therapies are limited, and the observed remission rates remain modest at 20%–30%, suggesting a therapeutic ceiling [6]. In addition, guidelines for UC do not mention any specific sequence of choice for advanced therapy in UC [7,8]. Furthermore, differences in the mechanism of action and route/frequency of administration of advanced therapies complicate treatment selection. Under these uncertain circumstances, it is challenging to select an optimal therapeutic agent, considering the various values and wishes of the patients. To support this challenge, conjoint analysis is widely used to clarify the decision-making process while trading off competing factors when studying the profiles of drugs preferred by patients with IBD [9-17]. However, to the best of our knowledge, no studies have examined how the severity of UC or patient employment status influences the choice of drug profile.
Our study objective was to investigate the drug profiles of patients with UC that are important in decision-making when selecting advanced therapies under 2 different severities of UC using conjoint analysis. In addition, the association between the preferred drug profiles and patient characteristics, including employment status, was evaluated.
METHODS
1. Study Design
This cross-sectional web-based survey was conducted among patients with UC between October 2023 and November 2023. This study was registered with the University Hospital Medical Information Network (UMIN000052327). The protocol was approved by the Ethical Review Committee for Human Tissue Research of the Mitsubishi Chemical Group Corporation on September 20, 2023 (IRB No. H-23-017-00). This study was conducted in compliance with the Declaration of Helsinki of the World Medical Association (modified October 2013) and the Ethical Guidelines for Medical and Health Research Involving Human Subjects (Notification No. 1 of the Ministry of Education, Culture, Sports, Science and Technology; Ministry of Health, Labour and Welfare; and Ministry of Economic Industry in 2021). All the patients provided informed consent to participate in the survey via the study website.
Patients were recruited from the IBD patient panel “IBD Plus” offered by QLife, Inc. (Tokyo, Japan). Patients were initially screened for age, UC diagnosis, and experience with drugs for UC treatment. Thereafter, patients with UC aged ≥ 18 years who were previously treated with oral 5-aminosalicylic acid (5-ASA) and willing to provide informed consent to participate in the online survey were enrolled.
2. Conjoint Analysis Survey
Patient preference for advanced therapies was evaluated based on the relative importance scores of drug attributes (location of administration, route/frequency of administration, time to symptom improvement [speed of onset of action], proportion of patients with sustained improvement in symptoms at 1 year [maintenance-sustainability], risk of serious adverse events within 1 year, and novelty of the drug [time since launch and amount of long-term use data]) and part-worth utility of the attributes’ levels in both the mild and severe symptom scenarios. Patient characteristics associated with the relative importance scores of drug attributes were also assessed. The 6 drug attributes and their assigned levels for the analysis are listed in Table 1. The levels included for each attribute were selected based on a review of clinical trial data for biologics and Janus kinase inhibitors in UC [18-26] and other relevant studies using conjoint analysis of IBD [9-17].

Drug Attributes and Assigned Attribute Levels for Advanced Therapies for Ulcerative Colitis Included in the Conjoint Survey
An orthogonal design was used to construct 24 conjoint cards that displayed a single hypothetical drug profile, each containing a unique combination of 6 attributes and attribute levels. The attribute level “intravenous infusion” was consistently combined with the location of administration being “administered at a hospital or clinic,” and the attribute level “daily oral administration” was combined with “can be administered at home.” Because patient preference may be influenced by UC symptoms, the card sorting task was assessed under uniform mild and severe symptom scenarios across patients (Fig. 1).

Example of sample choice cards in stages 2–4. Patients selected 1 of 3 cards in stage 1 and 1 of 2 cards in stages 2–4.
For the conjoint analysis, patients completed the card sorting task in 4 stages until each patient had chosen the single most preferred card. In stage 1, patients were presented with 8 sets of 3 cards each, randomly selected from a pool of 24 cards, and they selected the most preferred card from each of the 8 sets (a total of 8 cards were selected). In stage 2, the 8 cards selected in stage 1 were presented with 4 sets of 2 cards each, and the patients selected the most preferred card from each of the 4 sets (a total of 4 cards were selected). In stage 3, the 4 cards selected in stage 2 were presented with 2 sets of 2 cards each, and the patients selected the preferred card from each of the 2 sets (a total of 2 cards were selected). In stage 4, the patients chose the most preferred card from the 2 cards selected in stage 3.
Patient characteristics, including age, sex, age at diagnosis, disease duration, extent of disease, and disease status (most severe symptoms in the past and current symptoms [within the past 3 days] assessed as stool frequency score, rectal bleeding score, abdominal pain, and bowel urgency), were recorded. In addition, prior treatment, the route of administration of prior biologics, type of medical institutions visited, travel time to the hospital (one way), burden of hospital visits, and details of employment status were captured.
Regarding employment status, permanent employees, temporary and atypical employees, self-employment, freelancers, and part-time workers were defined as employed patients, whereas homemakers, students, and leave of absence/unemployment/retirement due to or unrelated to UC were defined as unemployed patients. For employed patients, information regarding working hours, working overtime, option to telework, option to flexible work hours, business trips with stay, ease of taking leave for medical treatment, and whether superiors and colleagues know UC was also collected.
3. Statistical Analysis
Assuming an allowable error margin of 5% and a response ratio (proportion of respondents for a given response) of 50%, the estimated sample size was approximately 380 patients [27]. Low-quality responses, including a response time of < 5 minutes, missing responses, and inconsistent responses, were excluded.
Each card was scored according to the number of times it was selected by each patient (minimum: 0; maximum: 4). Using these scores, a conjoint analysis was performed individually for each patient, and the part-worth utility of patient preferences was calculated for each patient. Part-worth utilities are centered at zero, with higher values representing a higher degree of preference. The importance score of drug attributes for each patient was calculated as the difference between the part-worth utility for the most preferred and least preferred levels for each attribute and displayed relative to the importance of other attributes, adding up to 100% [9-17]. Medication attributes with large differences in part-worth utilities between levels indicated higher relative importance [9-17]. Data on part-worth utility and importance scores were summarized for both mild and severe symptom scenarios.
Unpaired t-tests were performed to assess differences in part-worth utilities and relative importance scores between employed and unemployed patients. The following tests were performed as post hoc analyses. Differences in part-worth utilities and relative importance scores between the mild and severe symptom scenarios were assessed using paired t-tests. The paired t-test was used for comparison between 2 levels of part-worth utilities, and Dunnett’s test against the reference standard was used for comparison between 3 or more levels of part-worth utilities. As employment status varied among employed patients, part-worth utility was confirmed as a post hoc analysis by limiting the patient population to employed patients with time constraints. Employed patients with time constraints were defined as those who worked full-time without flexible work hours or telework choice.
For the top 3 drug profiles that showed high relative importance (mild symptom scenario: route/frequency of administration, maintenance-sustainability, risk of serious adverse events within 1 year; severe symptom scenario: route/frequency of administration, speed of onset of action, and risk of serious adverse events within 1 year), multiple regression analyses were performed for the entire population to identify factors associated with individual relative importance scores of different attributes. The individual relative importance score was set as an objective variable, and demographics (age, sex, and employment status), disease duration, extent of disease, prior treatment, current medical institution, burden of hospital visits, and the most severe symptoms in the past (stool frequency, rectal bleeding, abdominal pain, and bowel urgency) were explanatory variables. The explanatory variables were selected based on their clinical importance after confirming the correlation between them to eliminate multicollinearity. The base-case multiple regression analysis for the entire population used the “most severe symptoms in the past,” and the “most severe symptoms in the past” was replaced with “current symptoms” for the sensitivity analysis. For the employed population, in addition to the variables used in the base-case multiple regression analysis (excluding the presence or absence of employment), employment status, working hours, overtime work, option to telework, option to work flexible hours, business trips with stay, ease of taking leave for medical treatment, and whether superiors and colleagues know your UC were added as explanatory variables. For the unemployed population, in addition to the variables used in the base-case multiple regression analysis (excluding the presence or absence of employment), similar employment status details were added as explanatory variables in the multiple regression analysis.
Continuous variables were presented as mean and standard deviation (SD), and categorical variables were presented as frequencies and percentages. Statistical significance was set at two-tailed P<0.05. Conjoint analysis, multiple regression analysis, and data tabulation were performed using Excel add-in software developed by ISTAT, Inc. (Tokyo, Japan).
RESULTS
1. Patient Disposition
Of the 372 patients with UC who completed the survey, those with medications that are not approved for concomitant use in standard UC therapy (n = 2), inadequate rectal bleeding score (n = 1), diagnosis of both Crohn’s disease and UC (n = 1), and survey completion in < 5 minutes (n = 3) were excluded, and data from 365 patients were evaluated.
2. Baseline Characteristics
The mean ± SD age of the patients was 42.9 ± 11.8 years, and 66.6% were women (Table 2). The mean ± SD disease duration was 9.2 ± 8.8 years, and 55.9% had pancolitis. For the “most severe symptoms in the past,” most patients reported the highest severity category for all assessed symptoms, including stool frequency score, rectal bleeding score, abdominal pain, and bowel urgency. Conversely, for the “current symptoms,” most patients reported the mildest severity category for all 4 symptoms. Prior medications included topical 5-ASA/steroid (77.8%), oral/injectable steroids (66.3%), thiopurines (37.0%), and advanced therapies (46.3%). In addition, 58.4% of the patients felt the burden of visiting the hospital for UC treatment or somewhat felt it.
A total of 267 (73.2%) patients were employed, of whom 58.8% were permanent employees, 67.8% worked full-time, 61.0% did not work overtime, 30.7% had the option to telework, and 25.5% had the option for flexible work hours. The majority of patients reported that it was easy or somewhat easy to take leave for medical treatment (85.0%) and had revealed their UC to supervisors and colleagues entirely or in part (88.0%).
3. Conjoint Analysis
The part-worth utilities for drug attribute levels were assessed for each patient in the mild and severe symptom scenarios (Fig. 2). Regarding the route/frequency of administration, in the mild symptom scenario, patients showed a preference for oral administration. In contrast, no specific preference for the route/frequency of administration was observed in the severe symptom scenario. Regarding the novelty of the drug, in the mild symptom scenario, patients tended to prefer drugs with extensive long-term use data. In the severe symptom scenario, patients tended to prefer newly launched drugs even with limited long-term use data. A higher speed of onset of action, higher sustained improvement in symptoms at 1 year (maintenance-sustainability), and lower risk of serious adverse events within 1 year were preferred for both the mild and severe symptom scenarios; patients also tended to prefer drugs that could be administered at home over those requiring administration in a hospital or clinic. The preference for speed of onset of action was particularly evident in the severe symptom scenario.

Part-worth utilities for each drug attribute in (A) the mild symptom scenario and (B) the severe symptom scenario. Attribute levels with higher part-worth utilities were preferred more than those with lower part-worth utilities. The larger the differences, the stronger the preference. Data are presented as mean±standard deviation for 365 patients. The average coefficients of determination for the conjoint analysis conducted on each patient to calculate individual part-worth utilities were 0.733 for the mild symptom scenario and 0.729 for the severe symptom scenario. Inter-level comparisons: aP<0.01 (paired t-test for 2 levels and paired Dunnett’s test against the Ref [reference] for 3 or more levels). Comparison between the severe and mild symptom scenarios: bP<0.05, cP<0.01 (paired t-test).
The relative importance of drug attributes assessed for each patient is shown in Fig. 3. The relative importance score of the route/frequency of administration was highest for both symptom severities; this relative importance score was higher in the mild (mean ± SD, 34.1% ± 11.4%) than in the severe symptom scenario (28.3% ± 9.4%). In the mild symptom scenario, the risk of serious adverse events within 1 year (18.2% ± 8.4%) and a sustained improvement in symptoms at 1 year (maintenance-sustainability, 16.8% ± 7.4%) were the next most important attributes. The next most important attributes in the severe symptom scenario were speed of onset of action (18.2% ± 8.7%) and the risk of serious adverse events within 1 year (17.8% ± 8.0%). The relative importance given to the speed of onset of action was higher in the severe symptom scenario (18.2% ± 8.7%) than in the mild symptom scenario (10.5% ± 5.9%), and the importance of maintenance-sustainability was higher in the mild symptom scenario (16.8% ± 7.4%) than in the severe symptom scenario (14.9% ± 7.4%).
4. Comparison between Employed and Unemployed Patients
The part-worth utility for each drug attribute in employed patients (n = 267) was similar to that in unemployed patients (n = 98) in both the mild and severe symptom scenarios (Fig. 4). In the mild symptom scenario, there was no significant difference in the part-worth utility of being able to administer the drug at home between employed and unemployed patients (mean ± SD: 0.13 ± 0.37 vs. 0.05 ± 0.35, P=0.066); however, a post hoc analysis focusing on employed patients with time constraints (defined as those working full-time without flexible work hours or telework choice, n = 116) showed that the part-worth utility of being able to administer the drug at home was higher than that in unemployed patients (mean ± SD: 0.15 ± 0.36 vs. 0.05 ± 0.35, P=0.034). The relative importance scores for all drug attributes did not differ between employed and unemployed patients, regardless of whether the symptom scenarios were mild or severe (Fig. 5).

Part-worth utilities for each drug attribute by employment status in (A) the mild symptom scenario and (B) the severe symptom scenario. Data are presented as mean±standard deviation. Comparison of employment status: aP<0.05 (unpaired t-test).
5. Multiple Regression Analysis
Multiple regression analyses for the entire population (n = 365) assessed the association between the top 3 preferred drug attributes with high relative importance (route/frequency of administration, maintenance-sustainability [only mild symptom scenario], speed of onset of action [only severe symptom scenario], and risk of serious adverse events within 1 year) and patient characteristics selected based on clinical importance. Several factors of individual patient characteristics associated with the relative importance scores of the different attributes were identified, all exhibiting small absolute values of the regression coefficients (Tables 3 and 4). The coefficients of determination for the multiple regression analyses ranged from 0.033 to 0.094 across the entire population.

Multivariable Linear Regression Models with Relative Importance Scores and Patient Characteristics for the Entire Population

Sensitivity Analysis of Multivariable Linear Regression Models with Relative Importance Scores and Patient Characteristics for the Entire Population
In the severe symptom scenario, patients aged 40–59 years were more likely to consider a choice in the route/frequency of administration than those aged < 40 years, and patients aged < 40 years gave consideration to speed of onset of action than those aged ≥ 40 years (Table 3). Patients aged ≥ 60 years were more likely to be concerned about the risk of serious adverse events within 1 year than those aged < 40 years in the mild symptom scenario. Patients attending the clinic were more likely to prefer maintenance-sustainability than those attending the hospital in the mild symptom scenario. Patients who had ≥ 5 stool frequencies more than before UC as the most severe symptom in the past were less concerned about the risk of serious adverse events within 1 year compared with those with ≤ 4 stool frequencies in the severe symptom scenario.
As a sensitivity analysis, multiple regression analysis was performed using data on current UC symptoms instead of data on the most severe symptoms in the past (Table 4). The results of the sensitivity analysis additionally demonstrated that patients with current abdominal pain were less concerned about maintenance-sustainability compared with those without abdominal pain in the mild symptom scenario. Patients whose current stool frequency increased by at least 1 more than before UC placed greater importance on the route/frequency of administration in the severe symptom scenario than those with the same stool frequency as before UC. Patients currently experiencing bowel urgency placed less emphasis on the route/frequency of administration in the severe symptom scenario than those without such urgency.
The results of multiple regression analyses among employed patients (n = 267) showed that in the severe symptom scenario, employed patients who worked short hours were less concerned about the route/frequency of administration than those who worked full-time (Table 5). Multiple regression analyses among the unemployed population (n = 98) showed that in the severe symptom scenario, patients with “leave of absence, unemployment, or retirement not related to UC” were more concerned about the risk of serious adverse events within 1 year than those who were “homemakers” (Table 6).

Multivariable Linear Regression Models with Relative Importance Scores and Patient Characteristics for the Employed Population
DISCUSSION
This is the first study to investigate the drug profiles of advanced therapies preferred by patients in mild and severe symptom scenarios and how patient characteristics, including employment status, affect advanced therapy choices using the conjoint analysis. Patient preferences were generally highly individualized. Regardless of the symptom severity, the most important drug profile preferred by patients was the route/frequency of administration. Although oral administration was preferred in the mild symptom scenario, there was no trend toward specific route/frequency of administration in the severe symptom scenario. Patients also valued the speed of onset of action in the severe symptom scenario. There was no significant difference in the preference for a specific drug profile between employed and unemployed patients. Multiple regression analyses identified several factors of patient characteristics that were associated with the relative importance scores of the route/frequency of administration, maintenance-sustainability, speed of onset of action, or risk of serious adverse events within 1 year; however, the absolute values of the coefficients of determination were small.
Most patients in this study had currently mild symptoms of UC; however, most had experienced severe symptoms in the past, suggesting that they were able to respond to the survey by imagining a severe symptom scenario based on their previous experience. In this study, patients placed the highest importance on the route/frequency of administration in both the mild and severe symptom scenarios. In contrast, previous studies have reported that patients placed the highest importance on effectiveness [10,13,16], safety [14], and both the route/frequency of administration and effectiveness [9,11]. Although direct comparisons cannot be made, one of the reasons for this difference in the drug attributes with high importance among reports may be differences in the drug attributes and levels set in each study.
The current finding that the route/frequency of administration had the highest relative importance in both the mild and severe symptom scenarios indicates a strong patient preference for the route/frequency of administration. In the mild symptom scenario, patients showed a preference for oral administration. However, in the severe symptom scenario, no clear preference was observed for any specific route/frequency of administration. The decrease in the part-worth utility of oral administration in the severe symptom scenario compared with the mild one implies that some patients who prefer oral administration when the symptoms are mild may prefer intravenous or subcutaneous injections when the symptoms are severe. Nevertheless, this does not mean that the route/ frequency of administration was not prioritized under the severe symptom scenario but suggests that the preference for the route/frequency of administration varied widely among individuals. Indeed, while the relative importance of the route/ frequency of administration slightly decreased in the severe symptom scenario compared with the mild one, it remained the most prioritized profile. Concurrently, the importance of speed of onset of action was higher in the severe symptom scenario than in the mild one, suggesting that patients also desired immediate effects in addition to the route/frequency of administration when experiencing severe symptoms. Our findings also suggest that drug profile priorities may change over time due to fluctuations in the symptom severity, even within the same patient.
There are few reports on the relationship between employment status and patients’ preferred drug profiles; however, previous studies have reported both the presence [11] and absence [9] of an impact of employment status on the importance of the route/frequency of administration. In this study, we did not observe any difference in the importance of drug profiles among employed and unemployed patients in both mild and severe symptom scenarios. This could be attributed not only to differences in the sample size and definition of employment in previous studies but also to the characteristics of the employed patients, including employment conditions and work environments; in particular, in this study, over 60% worked no overtime and over 90% had no business trips with stay, that is, those with little time constraints at work. When focusing on patients working full-time without flexible work hours or telework options, the part-worth utility value of being able to administer at home was higher than that of unemployed patients, suggesting that diverse working styles may affect treatment preferences. Therefore, it is important to make treatment decisions by considering not only employment status but also patients’ working styles. These findings suggest that preference for advanced therapy cannot be simply determined or predicted by the presence or absence of employment.
In this study, younger patients valued the speed of onset of action in the severe symptom scenario more than elderly patients, who placed higher value on the risk of serious adverse events within 1 year in the mild symptom scenario. These findings are consistent with the results of a previous study [11]. As described above, although this study identified several factors associated with the importance of each drug profile, the absolute values of regression coefficients were small. Furthermore, the coefficients of determination, which assess the extent to which the relative importance of drug profiles can be explained by a multiple regression model based on patient characteristics, were low, ranging from 0.033 to 0.094 for the entire population. These results align with research from outside Japan [9,11,15], which reported that treatment preferences among patients with UC are highly individualized and variable. This makes it difficult to predict treatment preferences solely based on patient characteristics.
In this study, the patients were required to make trade-offs between hypothetical drugs online considering drug attributes based on currently available advanced therapies for UC. This approach may cause patients to behave differently than if they were making real-world decisions with physicians. However, there are gaps in treatment satisfaction and preferences between physicians and patients with UC [28,29], and patients do not fully communicate their desire for treatment to their physicians [30]. Therefore, we believe that this study was able to capture the honest perspectives of patients, which can be difficult for physicians to ascertain in clinical practice. In recent years, shared decision-making (SDM) has been emphasized in the treatment of IBD to improve patient satisfaction [31,32]. Our findings suggest that, because patient preferences vary, there is a necessity for SDM, and in addition to disease stage and employment status, it is important to communicate individual treatment preferences to physicians. Moreover, because these factors may change with the patient’s life stage, regular SDM is required.
The present study had some limitations. First, this study included self-reported data, such as the extent of disease, treatment history, and UC symptoms. Second, selection bias (age distribution and sex differences) resulting from online surveys may have existed. The male-to-female ratio of patients with UC is similar in Japan and other countries [1,2,33]; however, as in some online survey studies [10,16], the percentage of women was higher than that of men in this survey. As sex was not identified as a factor influencing the importance of drug profiles in the multiple regression analyses, a higher proportion of women may not have had a major impact on the study results. In addition, the percentage of elderly patients was lower based on reports of age distribution of patients with UC in Japan and adult patients outside Japan [31,34,35]; however, we were able to capture more detailed data of the employed patients, which is the population of interest in this study. Finally, although our analysis incorporated nearly all factors presumed to affect patient decision-making, we cannot completely exclude the possibility that unrecognized factors might have also impacted patients’ treatment preferences.
In conclusion, patients placed the highest relative importance on the route/frequency of administration. Patient preferences not only change within the same patient depending on disease severity but also demonstrate substantial interindividual variability. Therefore, it is difficult to predict the best treatment based on patient characteristics only from physicians’ perspectives, and SDM is essential to incorporate patients’ perspectives into the decision-making for advanced therapies.
Notes
Funding Source
This study was supported by Mitsubishi Tanabe Pharma Corporation and Janssen Pharmaceutical K.K.
Conflict of Interest
Kobayashi T reports support for this work from Mitsubishi Tanabe Pharma Corporation; institutional grants or contracts from AbbVie GK, Activaid, Alfresa Pharma Corporation, JMDC Inc., Gilead Sciences Inc., Nippon Kayaku Co. Ltd., Eli Lilly Japan K.K., Mochida Pharmaceutical Co. Ltd., Janssen Pharmaceutical K.K., Pfizer Japan Inc., Takeda Pharmaceutical Co. Ltd., Bristol-Myers Squibb, Mitsubishi Tanabe Pharma Corporation, Zeria Pharmaceutical Co. Ltd., Otsuka Holdings, EA Pharma Co. Ltd., JIMRO Co. Ltd., Kyorin Pharmaceutical Co. Ltd., and Google Asia Pacific Pte Ltd.; payment or honoraria for lectures, presentations, speakers bureaus, manuscript writing, or educational events from Takeda Pharmaceutical Co. Ltd., Activaid, Alfresa Pharma Corporation, Zeria Pharmaceutical Co. Ltd., Kyorin Pharmaceutical Co. Ltd., Nippon Kayaku Co. Ltd., Mitsubishi Tanabe Pharma Corporation, AbbVie GK, Pfizer Japan Inc., Janssen Pharmaceutical K.K., Thermo Fisher Diagnostics K.K., JIMRO Co. Ltd., and Galapagos; and payment for expert testimony from Janssen Pharmaceutical K.K., EA Pharma Co. Ltd., Kissei Pharmaceutical Co. Ltd., Takeda Pharmaceutical Co. Ltd., Activaid, Pfizer Japan Inc., Nippon Kayaku Co. Ltd., Alfresa Pharma Corporation, Kyorin Pharmaceutical Co. Ltd., AbbVie GK, Mochida Pharmaceutical Co. Ltd., and Mitsubishi Tanabe Pharma Corporation. Mizuno N, Sato N, Kawaguchi Y, Ikawa Y, and Komorita N report support for this work from Mitsubishi Tanabe Pharma Corporation and are employees of Mitsubishi Tanabe Pharma Corporation. Ishikawa H reports support for this work from Mitsubishi Tanabe Pharma Corporation. Except for that, no potential conflict of interest relevant to this article was reported.
Data Availability Statement
The data underlying this article will be shared upon reasonable request by the corresponding author.
Author Contributions
Conceptualization: Kobayashi T, Mizuno N, Sato N, Kawaguchi Y, Ikawa Y, Ishikawa H. Investigation: Kobayashi T, Mizuno N, Sato N, Kawaguchi Y, Ikawa Y, Ishikawa H. Methodology: Kobayashi T, Mizuno N, Sato N, Kawaguchi Y, Ikawa Y, Ishikawa H. Project administration: Mizuno N. Visualization: Kobayashi T, Mizuno N, Sato N, Kawaguchi Y, Ikawa Y, Komorita N, Ishikawa H. Writing – original draft: Kobayashi T, Mizuno N. Writing – review and editing: Kobayashi T, Mizuno N, Sato N, Kawaguchi Y, Ikawa Y, Komorita N, Ishikawa H. Approval of the final manuscript: All authors.
Additional Contributions
The authors thank all respondents for completing the questionnaire and QLife Inc. who assisted with the administration of the questionnaire. The authors also thank Tamio Kan, Yasuo Shiga, and Naoko Himeno of iStat, Inc. for planning and analyzing the conjoint analysis and Annirudha Chillar, MD, PhD, of Cactus Life Sciences (part of Cactus Communications) for providing medical writing support, which was funded by Mitsubishi Tanabe Pharma Corporation and Janssen Pharmaceutical K.K.