Unveiling Preferences in Closed Communities: Development of a Discrete Choice Experiment (DCE) Questionnaire to Elicit Ultra-Orthodox Women Preferences for Video Consultations in Primary Care

Abstract

Background

Video consultations in primary care settings demonstrate substantial benefits, including improved accessibility, reduced waiting times, and enhanced health management. These services could particularly benefit ultra-Orthodox women in Israel, who typically manage large families and face unique healthcare access challenges as primary caregivers. However, eliciting preferences within this closed religious community presents distinct methodological challenges because of cultural sensitivities and religious restrictions regarding technology use.

Objective

We aimed to develop and validate a culturally sensitive, discrete choice experiment questionnaire for eliciting ultra-Orthodox women’s preferences regarding video versus in-clinic consultations in primary care settings.

Methods

A three-stage mixed-methods approach was employed: (1) 33 semi-structured interviews with key stakeholders (women, men, rabbis, and healthcare providers) to identify attributes and levels; (2) an attribute-ranking exercise with 88 ultra-Orthodox women to refine attributes; and (3) cognitive interviews with 15 women to validate the discrete choice experiment questionnaire.

Results

Four key attributes emerged as most important for ultra-Orthodox women when choosing between video and in-clinic consultations: (1) consultation timing (regular hours/after 20:00); (2) travel time; (3) waiting time; and (4) familiarity with the healthcare provider. Importantly, the study revealed the necessity for a dedicated device exclusively for healthcare provider communication, closed to open Internet networks, as a fundamental prerequisite for implementing video consultations in this community. Additional unique findings emerged through this methodological process, contributing to the understanding of technological adoption in closed religious patients’ communities.

Conclusions

This study provides a comprehensive example of implementing pre-discrete choice experiment stages while addressing unique considerations of a special population. The findings provide a framework for developing inclusive telemedicine services for traditionally underserved populations.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40271-025-00734-w.

Key Points for Decision Makers

The process of developing the discrete choice experiment survey can serve as a valuable tool for initiating dialogue about healthcare preferences within closed communities. This tailored approach, considering unique cultural contexts, potentially paves the way for a broader acceptance of video consultations, telemedicine, and other healthcare innovations, while respecting the community’s values and traditions.

Engaging multiple stakeholders within the closed community (women, men, religious leaders, healthcare providers) throughout all stages of the discrete choice experiment development is crucial for identifying and selecting valid, culturally relevant attributes. This tailored approach ensures cultural sensitivity by incorporating community-specific values and norms into the discrete choice experiment design.

Employing a mixed-methods approach, including cognitive interviews and attribute ranking exercises, provides rich data to guide necessary adaptations of the discrete choice experiment. Methodological adjustments, such as using community-matched interviewers, obtaining religious leadership approval, and adapting language and data collection methods, ensure cultural sensitivity and suitability for use within closed communities.

Introduction

Telemedicine, particularly video consultations (VCs) in primary care settings, has demonstrated substantial benefits when used instead of traditional in-clinic consultations (ICCs). Video consultation services can increase accessibility for geographic and ethnic cultural peripheries, improve chronic disease management, decrease hospitalizations, and reduce costs [1–5]. Video consultations also reduce waiting times and travel costs for patients while creating positive patient experiences. Additionally, these services have shown potential in addressing healthcare disparities among various populations, offering convenient and efficient access to medical care [6, 7].

In Israel, telemedicine consultations in primary care settings have been successfully implemented through four major health maintenance organizations (HMOs), offered at no additional cost to patients [8–11]. These HMOs operate under Israel’s National Health Insurance Law, which mandates universal healthcare coverage. Telemedicine services are funded through a combination of government allocations and HMO resources, ensuring equal access across the population. Approximately 76% of the general Israeli population has used some form of telemedicine service, including VCs, phone consultations, and digital correspondence with healthcare providers [12]. While telemedicine adoption in the general population has increased, driven by widespread smartphone usage [8, 13], the ultra-Orthodox (UO) religious community remained outside these services because of rabbinic proscriptions and resistance to technology and especially smartphone use [14, 15].

UO Community and Telemedicine

The UO population in Israel represents a religious minority with a unique lifestyle based on strict Jewish Halacha religious laws [16]. This community currently numbers approximately 1.2 million people, with women comprising roughly 600,000 of this population. During recent decades, smartphone use was prohibited in UO communities because of concerns about exposure to secular content and potential threats to spiritual integrity [17]. In 2005, the Rabbinical Committee for Communication Affairs restricted technology use to “kosher phones” only, effectively preventing telemedicine implementation in this population [18].

The UO community, comprising 13.6% of Israel’s population, could significantly benefit from telemedicine services, especially women who often juggle multiple responsibilities as primary caregivers and breadwinners [16]. Ultra-Orthodox women, typically managing large families with an average of 6.6 children, face unique challenges in accessing healthcare because of their busy lifestyles and cultural norms [19]. They play a crucial role in their families’ health management, often being responsible for healthcare decisions and interactions with medical providers [20].

Recent years have shown a gradual shift in attitudes towards technology within the UO community. Employment needs, educational requirements, and daily necessities have led some community members to seek solutions that allow the use of advanced technologies while preserving community values. The COVID-19 pandemic accelerated this trend [17]. Current data indicate that approximately half of the UO population now owns a smartphone, and about 68% have access to the Internet through various means [13, 21]. However, technology adoption is critically dependent upon tailoring the technology to rabbinical guidance, reflecting a nuanced approach to modernization within traditional frameworks. These changes make VCs increasingly relevant for UO women [22].

The adaptation of telemedicine in public healthcare for special religious, ethnic, and unique populations holds significant importance [23]. Understanding UO women’s preferences regarding VCs is crucial for developing tailored telemedicine solutions that respect cultural sensitivities while improving healthcare access.

Preferences Research Within Closed Communities

Conducting preferences research within closed communities such as the UO community presents unique methodological challenges. These include access difficulties, language and cultural barriers, and the need for gender-matched methods [24]. The challenge intensifies when researching sensitive topics such as smartphone usage for telemedicine purposes [25]. Additionally, community members often provide inaccurate responses to external surveys, a phenomenon known as “social desirability bias” [26, 27].

A discrete choice experiment (DCE) can effectively assess patient preferences for specific medical services or products. However, because of the unique methodological challenges within the UO community outlined above, special care must be taken in designing the DCE questionnaire. This careful approach will ensure that the attributes and levels included in the experiment are both relevant and precise, leading to more accurate insights into patient preferences [28]. This research aims to develop a DCE survey to elicit preferences for VCs in primary care from UO women, including survey testing to establish its validity and feasibility in this unique population.

Methods

Discrete choice experiment questionnaire development was undertaken in three sequential stages following best-practice guidelines [29–31]. Qualitative and quantitative methods were incorporated into the stages of the DCE questionnaire development. The methods and three stages of DCE development are described in the following section and presented in Table 1. Stage 1 aimed to identify attributes and levels through a review of primary care service-related patients’ preferences literature and 33 interviews with key stakeholders in the UO community: women, men, primary care physicians treating the UO population, religious influencers, and rabbis. After analyzing the data collected through interviews regarding attributes and levels relevant to UO women usage of VCs, stage 2 took place via a ranking task. At Stage 2, the aim was to select the most important attributes and to refine their levels based on feedback from the target population. Another sample of 88 UO women was asked to rank identified attributes and provide detailed input about the proposed levels to ensure they were relevant and appropriate for their community context. The questionnaire with the ranking exercise is included in the Electronic Supplementary Material (ESM). After Stage 2, the DCE survey was developed, and Stage 3 included the pre-testing of the survey. Fifteen cognitive interviews with UO women were conducted to assess face and content validity and to provide additional information regarding questionnaire design issues.

Table 1.

Stages in the development of the DCE questionnaire

n	Stage	Research methods	Sample and tools
1	Attribute and level development	Qualitative	Literature review and personal semi-structured interviews with 32 stakeholders: UO women, men, community influencers
2	Attribute selection and level refinement	Quantitative	Ranking exercise and levels survey among 88 UO women
Developing the DCE survey: experimental design and task construction: fractional factorial using a hand catalog and excluding a dominant choice set and 12 labeled forced-choice tasks (in-clinic/video consultations). Questionary creation: task appearance, graphics, demographic characteristics
3	Pre-testing DCE questionnaire	Qualitative and qualitative	Cognitive interviews with 15 UO women
Analysis of respondent feedback, conduction of modifications and finalization of the DCE questionnaire

DCE discrete choice experiment, UO ultra-Orthodox

The study was approved by the Institutional Ethics Committee of the University of Haifa. All participants signed the consent form. The methods and results are reported according to the Consolidated Criteria for Reporting Qualitative Research (COREQ) guidelines [32] and the DIRECT checklist for DCE studies [33].

Stage One: Attribute and Level Identification

Literature Review

A literature review was conducted, examining preferences for VC use and primary healthcare service utilization both globally and in Israel, focusing on the general population and special populations such as the UO Jews and other minority groups. We searched major health and social sciences databases including PubMed and Google Scholar from 2014 to 2024. Search terms included “telemedicine,” “video consultation,” “primary care,” “preferences,” “discrete choice experiment,” “ultra-Orthodox,” “smartphones,” “religious communities,” and combinations of these terms. Additional sources were identified through reference lists of relevant articles. Articles were included if they addressed preferences for telemedicine in primary care settings, cultural considerations in healthcare delivery, or healthcare utilization patterns in religious communities. The review informed our attribute selection process by identifying commonly studied attributes (waiting times, travel considerations, appointment availability) and noting gaps regarding religious considerations, such as rabbinical approval and large family needs that are significant in UO communities. These findings were combined with stakeholder interviews to ensure both universal and culturally specific attributes were included in our DCE, ensuring our selection was evidence based and culturally appropriate.

The Interview Guides

The interview guides were created by a diverse team of researchers, including a public health physician and nurse (SH, BM) and researchers from social and behavioral sciences (ADZ, IC, LG), one of whom is a UO woman. The interview guides encompassed a wide range of questions, exploring women’s utilization of primary healthcare services and the key factors they consider when deciding to consult pediatricians, family doctors, or family health station public health nurses. The guides also related topics crucial to women’s decision-making process when contemplating various healthcare access options, including their prior experiences with VCs, telemedicine, smartphone usage patterns, and other relevant factors influencing their healthcare choices. We also employed the think-aloud protocol [34, 35] in interviews to gain deeper insights into the decision-making processes of UO women regarding their choice between VCs and ICCs. This method allowed women to verbalize their thoughts as they considered various healthcare scenarios, providing data on the cognitive processes underlying their preferences. The guides are provided in the ESM.

Sample

Data collection took place between March and June 2024. We conducted semi-structured personal interviews with 33 key stakeholders: UO women, men, community religious influencers, primary care practitioners that serve the communities, and UO rabbis. “Influencers” were defined as individuals who held formal or semi-formal leadership positions within the community, primarily ‘askanim’ (community activists). ‘Askanim’ are officially recognized community representatives who serve as intermediaries between the UO community and external institutions, managing community affairs and advocating for their needs with government bodies, healthcare organizations, and other secular institutions. Their formal role and established connections make them crucial gatekeepers for implementing new initiatives within the community. To ensure all relevant UO stakeholders’ perspectives are represented [36–39], the research included members from all three UO sub-communities: Lithuanians, Hasidic, and Sephardic. While Israel’s UO community includes various subgroups, including modern Haredi, nationalist Haredi, and groups such as Eidah Charedit and Neturei Karta, our focus on the three main streams (Lithuanian, Sephardic, and Hasidic) allowed us to represent approximately 90% of Israel’s UO population. This scope was chosen to ensure representation of the majority population while maintaining feasible study parameters. The characteristics of stakeholders interviewed are presented in Table 2. While the perspectives of women are the most important, as their preferences are intended to be mapped through a future DCE questionnaire, in populations such as UO communities, there are additional perspectives that influence women’s decisions, including rabbis, religious influencers, and men. Therefore, we included these stakeholders to gain a comprehensive view of the diverse preferences and influences within the community. This approach allows us to capture the complex interplay of cultural, religious, and social factors that shape women’s healthcare choices in this unique population. Snowball sampling was used to recruit UO stakeholders and later, the purposeful sampling was added to ensure that responders’ characteristics were varied not only in terms of the UO sub-community but in age, geographical areas within Israel, and urban and rural settings. We ensured representation of UO women affiliated with four HMOs in Israel (e.g., Clalit, Maccabi, Leumit, Meuhedet). As the study progressed, we used purposeful sampling to fill gaps in our participant profile, ensuring a balanced representation across these characteristics. The sample size was determined using the ‘data saturation’ principle, with data collection continuing until no new insights emerged, typically requiring about 30 participants for a qualitative sample. The final sample consisted of 33 UO respondents, out of them 22 are UO women.

Table 2.

Stage one: personal interviews’ respondents’ characteristics

Characteristics	UO women [n = 22]	Other stakeholders: men, primary care practitioners, religious influencers, and rabbis [n = 11] (3 UO men, 4 family physicians and nurses treating UO community; 4 rabbis and influencers)
Female sex	22 (100%)	5 (45%)
Mean age, years ± SD	32.4 (6.7)	43.2 (11.6)
Family status	Married (100%)	Married (100%)
Children, n	4.7	5.1
UO community	Sephardic (29%), Hasidic (40%), Lithvanian (31%)	Sephardic (22%), Hasidic (33%), Lithvanian (22%), Dati Leumi, modern Orthodox (23%)
Living area	North (24%), center (40%), Jerusalem area (18%), South (18%)	North (25%), center (42%), Jerusalem area (21%), South (12)
Health maintenance organization	C (42%), Mac (33%), Meu (24%), Leu (1%)	C (42%), Mac (27%), Meu (31%)

SD standard deviation, UO ultra-Orthodox

Special Methodological Considerations

Conducting research within the UO community presents unique challenges, particularly when addressing sensitive topics such as smartphone ownership and usage, which are often subject to strict rabbinic and societal regulations. To navigate these complexities, we implemented several methodological adaptations. The interview guide used formal language, avoiding potentially immodest expressions. Interviews were conducted by a female UO researcher to ensure cultural compatibility. Following guidelines for sensitive topic research [40], we carefully selected appropriate locations for data collection, employed indirect questioning techniques, and maintained personal research diaries. The interviewers participated in regular supervision meetings to address any arising issues. Additionally, we sought endorsement from community leaders and provided clear explanations about data confidentiality and usage. We also remained flexible in our approach, allowing participants to choose their preferred method of communication and timing for interviews, further accommodating the community’s unique needs and sensitivities. Some interviews, for example, were chosen to be performed when children were absent, to ensure privacy discussing smartphone usage. Use of smartphones for recording was negotiated with each participant. Some participants requested verification of kosher approval and content filter stickers on the interviewer’s smartphone. When interviewing rabbis, plural and respectful forms of address were used, reflecting hierarchical norms within the community. Cultural and religious practices, such as leaving the door open during interviews to avoid ‘yichud’ (seclusion with a member of the opposite sex), were strictly observed during interviews. Questions about health and faith were framed using religious terminology and concepts, facilitating more authentic responses. The interviewer’s familiarity with UO culture and terminology was crucial in understanding nuanced responses.

For interviews that were not recorded (n = 9), detailed field notes were taken during the session and transcribed the same day to ensure data accuracy. The research team included two senior researchers with extensive qualitative research experience (more than 10 years), and a doctoral candidate who completed advanced qualitative research coursework and participated in multiple qualitative studies. One researcher had specific expertise in DCE attribute and level mapping methodology, ensuring methodological rigor for this specialized research type.

Procedures

Two researchers conducted the interviews: one of them is a UO woman. The sampling strategy varied by stakeholder group. The UO female researcher recruited UO women by approaching them at children’s playgrounds near family health stations in several cities. Interested women received formal invitation letters explaining the study details, and interviews were scheduled at their convenience. The other researcher primarily conducted interviews with UO men, primary care practitioners, influencers, and rabbis. Rabbis were recruited through researchers’ connections with the Ministry of Health and community ‘askanim,’ ensuring representation across all three UO populations. Healthcare providers were purposively sampled to ensure representation across different HMOs, geographical areas, and settlement types: UO or mixed cities. The interviews were conducted in Hebrew, with durations that ranged from 40 to 120 min. All respondents signed the consent form. With the consent of 24 out of 33 participants, these interviews were audio-recorded and subsequently transcribed verbatim. To ensure data security, all information was stored on the researcher’s personal computer, protected by a dedicated password in addition to the computer’s general access password. Furthermore, all identifying information was removed from the transcripts and replaced with codes to maintain participant anonymity.

Interviews’ Data Analysis

The research team employed an inductive approach to analyze the data, utilizing a thematic content analysis. The process began with researchers immersing themselves in the interview data, repeatedly reading transcripts and listening to recordings to familiarize themselves with each participant’s narrative. Initial impressions were documented, noting potential attributes and levels related to VC versus ICC choices. Themes were then examined across narratives, identifying connections, similarities, and differences. Redundant themes were eliminated, and remaining themes were categorized and subcategorized. The frequency of each theme’s occurrence was noted using a, b, and c designations. The team then engaged in a comparative analysis of individual interpretations, discussing discrepancies and reaching a consensus on attributes and levels. To mitigate potential biases, researchers openly discussed their personal perspectives on telemedicine in primary care prior to the analysis. An interactive coding tree was constructed using a plastic board and paper scraps, allowing for flexible reorganization during the analysis. The themes were then analyzed across different stakeholder narratives, identifying connections, similarities, and differences. The analysis revealed four major domains: accessibility factors (including time constraints, travel considerations, waiting times), provider relationship elements (trust, familiarity, communication style), technical and religious considerations (device acceptability, rabbinical approval, privacy concerns), and family management aspects (childcare coordination, home responsibilities). The frequency of each domain’s occurrence was noted using a, b, and c designations. These four domains were further refined through a cross-stakeholder analysis to identify recurring themes that could serve as measurable attributes for the DCE. The final key themes represented identified attributes, each containing subcategories reflecting possible levels. For example, within the accessibility domain, the theme of “time management” emerged consistently across narratives, leading to three distinct attributes: consultation timing (regular hours/after 20:00), travel time, and waiting time. From the provider relationship domain, the theme of “continuity of care” emerged as the attribute “familiarity with healthcare provider”. This process resulted in the identification of 11 distinct attributes, each with multiple levels. Demographic data were entered and analyzed using Microsoft Excel 2021 (Microsoft Corporation, Redmond, WA, USA).

Stage Two: Attribute Ranking and Level Refinement

To refine the attributes identified in the qualitative phase, we conducted an attribute-ranking exercise. Following the initial phase where 11 relevant attributes were identified, the second stage involved a questionnaire in Hebrew designed for women to rank these attributes. A sample of 88 UO women was recruited from playgrounds near primary care clinics and family health stations. Once again, we combined snowball sampling with purposive sampling to ensure sufficient diversity among women from various sub-populations and geographic locations.

The questionnaire consisted of three parts. In the first part, women ranked the attributes and selected the “four most important factors when considering consultations with primary care providers for children or family doctors”. The second part included questions about different levels; for example, to refine the levels of the “travel time to clinic” variable, women were asked about their walking and driving times to their children’s clinic. The third part comprised sociodemographic questions. A complete questionnaire is attached in the ESM.

The frequency with which each attribute was selected was tabulated using Microsoft Excel software. This quantitative approach allowed us to prioritize the attributes based on participant preferences. The four attributes that received the highest rankings within each stakeholder group were then selected for inclusion in a future DCE. This process ensured that the attributes used in the subsequent DCE were not only grounded in the qualitative data but also validated and prioritized by a broader sample of potential users. This approach helps to enhance the relevance and validity of the forthcoming DCE by focusing on the attributes deemed most important by the target population.

Developing the DCE Survey

The development of the DCE design and survey adhered to established best practices and guidelines [31]. This approach ensured that our DCE methodology was robust and reliable, and aligned with current standards in the field of health economics research.

We chose a labeled-choice experiment to compare preferences between VCs and in-clinic consultations. This labeled design approach was chosen for its enhanced feasibility, realism, and potential to increase the validity of results. Given that a full factorial design would have resulted in an overwhelming 32 (2⁵) scenarios, we instead employed a fractional factorial design. This design was crafted using a manual catalog-based procedure, prioritizing optimized level balance (ensuring each attribute level appears with equal frequency) and orthogonality (maintaining uncorrelated occurrences of levels across different attributes). Through this method, we created 12 hypothetical choice tasks, each presenting two alternatives with forced binary responses. For each task, participants were asked to indicate their preference between the two consultation methods (VC vs ICC). This approach allowed us to efficiently capture participant preferences while maintaining statistical robustness and minimizing the respondent burden. Our DCE survey was structured into two sections: (1) the DCE choice’s tasks, including a dominant option set to assess internal consistency of responses and (2) sociodemographic questions. The dominant option featured one clearly superior alternative, serving as a quality check for participant engagement and understanding.

The DCE survey was developed based on the results of interviews and ranking exercise. The willingness-to-pay variable was not included in our questionnaire to maintain realistic scenarios that reflect the actual situation in Israel, where VCs are provided as a free service by all HMOs and family health centers. In addition, during the initial interview phase, it emerged clearly that a fundamental prerequisite for remote consultation with healthcare providers was the use of a dedicated device that would be closed to open Internet networks, designated exclusively for consultations with medical providers, and would receive rabbinical approval as “kosher device”. Therefore, the condition that appeared in the DCE questionnaire before the choice tasks explicitly stated that the consultation would be conducted through a dedicated device.

Stage Three: Pre-testing DCE Survey

Cognitive Interviews

Cognitive interviews [41, 42] to test the validity and acceptability of the DCE survey were conducted with 15 UO women, providing a total of 2700 observations. Ultra-Orthodox women were recruited in several cities through purposive sampling based on age, community, and geographical area. Participants were encouraged to think aloud as they navigated through the survey, vocalizing their thoughts and reactions. Upon completion of the survey, the researcher posed targeted questions to gauge the relevance and significance of the attributes in decision-making processes related to a consultation type VC or ICC. These inquiries also aimed to assess participants’ perceptions of the DCE task’s complexity, the time required to complete the survey, and the clarity of instructions provided. The researcher meticulously documented all verbal responses, capturing nuanced insights into the participants’ thought processes and experiences with the survey instrument. This approach allowed for a comprehensive evaluation of the survey’s effectiveness and accessibility within the target population, providing valuable feedback for refinement and optimization of the DCE design. Participants’ selections in the DCE tasks were gathered using paper-based questionnaires.

A qualitative content analysis was conducted using the same steps as the Stage 1 content analysis to provide insights into respondents’ thought processes and decision-making strategies. The analysis of “think-aloud” data also allowed for a trade-off analysis to determine if respondents were making full trade-offs with all attributes. To check if the direction of attribute choices in the questionnaire was correct, we conducted an attribute direction analysis. This involved comparing stated preferences in interviews with choices made in the DCE and using a simple count analysis of choices for each attribute level. In addition, we analyzed the content of all documented verbal responses of UO women about the experiences with the survey instrument—its clarity and accessibility.

Results

Stage One: Attribute and Level Identification

All three UO communities, Lithuanian, Hasidic, and Sephardic, were represented in the study. Participants varied in age and geographical area. The characteristics of the participants in personal interviews are displayed in Table 2. After analyzing interview data, 11 attributes with two levels each were identified in the patient group. Table 3 shows the identified attributes, their levels, and supporting quotations from the interviews. At this stage, the attribute levels were exploratory and required further refinement to establish precise levels. For example, timing of visits emerged as a significant factor in choosing the type of visit, but it was not clear what exact hour was considered by UO women as “after hours,” requiring additional research to establish more precise time thresholds that align with their daily religious and family routines.

Table 3.

Stage one: stakeholders’ quotations contributing to themes for attributes and attributes and levels identified

n	Attribute	Description	Levels	Stakeholders’ quotations
1	Waiting time before consultationa	Standard waiting time in-line before the appointment	5 min; 40 min	“There’s chaos in clinics. You arrive and have to wait a long time. Until they let you in. This way [video visit] I won't need to wait” “The biggest advantage is not having to wait in the clinic, where everyone is sick, with children, and it’s a mess”
2	Travel timea	Travel time to the healthcare facility	15 min; 40 min	“I prefer to minimize visits. I don’t have time to travel, anything that can be done from home, I prefer” “I’m paying so much already, I don’t see any reason to invest time and travel to Ashdod or Yavne”
3	Time until next appointment	Standard waiting period for non-urgent medical appointments	2 days; 1 week	“[With video consultations] perhaps there would be an earlier available appointment, because today I have to wait half a year for a pediatric orthopedist. For the pediatrician, I don't make appointments but simply show up at the clinic, that’s how it works there, I know the secretary”
4	Quality of visit	Uninterrupted provider attention to patient during the consultation	There are interruptions, the physician is not attentive enough to me; there are no interruptions, the physician is attentive to me	“When I’m at the clinic, there is a lot of ‘background noise’ ... there are other patients waiting to see the doctor... and when it’s finally my turn and I'm with the doctor, there’s always a patient who knocks, wanting to check if the doctor is available or if he can ask a quick question ... or the secretary might come in, or there’s a phone ringing in the background ... In a video consultation, however, I have a few minutes of quiet with no interruptions – the doctor is focused solely on me and my reason for contacting her.”
5	Familiarity with the practitionera	Known/familiar healthcare provider with an established relationship	I am familiar with the practitioner; I am not familiar with the practitioner	“Sometimes you want that regular nurse, the one you see regularly, who knows your child" "If it’s a nurse who already knows the child and it's not the first time, I feel I could do video ... ” “I’m a bit hesitant about having a video consultation if I don't know the doctor ... Sometimes you get various people you may or may not connect with, may or may not trust ... Sometimes it's also very brief like: ‘What do you want?’ and ‘bye’ ...”
6	Purpose of visit	Severity level of medical concern and perceived complexity of the medical issue	Small problem; big problem	“I don’t think a doctor can provide proper judgment or answer about a major problem through video. If there’s something serious, then not exactly ... they can't give an accurate diagnosis for everything” “Video can be suitable for referrals to tests (blood/ultrasound) or for 'specific' guidance regarding a minor problem” “When there’s something more serious, in a critical health condition, you need to go to primary care clinic ... a video consultation won’t help”
7	Service availabilitya	Timing of available service in the primary care facility	Regular services hours; after-hours (after 20:00)	“If it were possible to speak with doctors via video after working hours when the children are asleep, that would be a significant advantage” “After eight in the evening when the children are asleep, I have time to take care of myself and the children [health], and yes, I want to consult with the well-baby clinic nurse” “I know how to describe to her [the doctor] the type of cough ... I could have the video consultation when the child is already asleep. I would, for example, schedule the call in the afternoon, and she would get back to me in the evening ... that would be a significant advantage”
8	Weather	Weather conditions for traveling to the healthcare facility	Bad weather; good weather	“Especially in bad weather I wouldn’t invest so much time just to get to the clinic and sit in front of the doctor when I could have the same appointment over video, on my own terms ... Just think about all the preparations ... and what if it’s raining outside? Or too hot? It’s enough that I’m dealing with a sick child, and he’s dealing with some illness or pain, then also to contend with the weather?! Not for me ...”
9	Modesty considerations	Modesty-related challenges, extensive in clinic settings vs limited in the home environment	More modesty issues to be concerned about; less modesty issues to be concerned about	“Without a doubt, video has a modesty advantage – after all, the woman is not in the same room with the doctors”
10	Privacy considerations	Privacy considerations regarding potential encounters with community members at the clinic’ waiting room	More privacy issues to be concerned about; less privacy issues to be concerned about	“I don’t need to go to the clinic and meet people who might know me there. This way [through VC] my visit remains between me and the nurse”
11	Type of practitioner	Type of consulting healthcare professional	Physician; nurse	“There are consultations that I find more comfortable having with nurses rather than doctors. For example, it’s common for nurses to also be lactation or sleep consultants – I consult with our well-baby clinic nurse often” “There were even questions I asked the pediatrician, and he referred me to a nurse who specializes in infant massage. I prefer nurses for care like wound dressing and treatments, rather than doctors ... Sometimes, nurses are even more seasoned and experienced than doctors”

VC video consultation

^aAttributes selected for the final discrete choice experiment survey following later stages 2 and 3

Respondents noted that later hours, when household tasks are completed, allowed more time for addressing health issues. However, the exact definition of “regular” versus “non-regular” visit hours needed clarification. Similarly, the specific levels for travel time were not yet clearly defined. To address these uncertainties, we included questions about attribute levels in the next stage of the study, which involved a ranking exercise. This approach helped us gather more precise information about respondents’ preferences and better define the attributes’ levels.

Attribute Ranking and Level Refinement

Sample Characteristics

Eighty-eight UO women responded to the survey with ranking exercise. Their characteristics are included in Table 4. In addition to sociodemographic questions, women responded to questions related to attributes’ levels and to women’s health habits and technology use. Regarding accessibility to health services, the average time to reach a family doctor’s clinic was 15.6 min walking, while for a pediatrician’s clinic, it was 18.7 min. Eighty-five percent of the sample indicated that they—UO women—are responsible for health issues in the family, six women stated it was their husband’s responsibility, and eight women responded that both they and their husband are jointly responsible for family healthcare. Regarding technology use, 65 women reported using the Internet to search for health and medical information in the previous 3 months, and 51 out of 88 women indicated they used video capabilities on their mobile phones for various purposes.

Table 4.

Stage two: attribute-ranking exercise: UO women’s characteristics

Characteristics	Responders [n = 88]
Female sex	88 (100%)
Mean age, years ± SD	31.9 (5.9)
Married, family status	88 (100%)
Children, n	4.1
UO community	Sephardic (36.4%), Hasidic (18.2%), Lithvanian (38.6%), NA (6/8%)
Living area	North (25%), center (28%), Jerusalem area (22%), South (18%), NA (7%)
Part of revival movementa (baaley tshuva)	15 (17%)
Level of education achieved	Seminary or high school education (33.4%), post-secondary Seminary without academic credentials (37.9%), and academic degree holders (12.5%), NA (16.2%)
Employment status	Employed (67.8%), homemakers/unemployed (22.7%), no response (9.5%)
Husband employment status	Full-time Torah study (44.8%), full-time employment (25.1%), combined Torah study and work (21.8%), no response (8.3%)
Health maintenance organization	C (36%), Mac (33%), Meu (21%), Leu (4%), NA (6%)
Responsibility for family healthcare	Wife (85%), husband (6%), both partners (5%), NA (4%)
Average walking and driving time to the family physician clinic	15.6 min; 7.7 min
Average walking and driving time to the family health station	18.7 min; 8.6 min
Used the Internet for health information	57 (65%)
Used mobile video capabilities	45 (51%)

NA, SD standard deviation, UO ultra-Orthodox

^aRevival movement refers to individuals who have not be born into chosen to become Orthodox Jewish (baalei teshuva), representing a significant religious and social phenomenon in Israel [26]

Ranking Results

Key attributes that were ranked as the four most important attributes for each stakeholder group were selected for inclusion in the DCE quantitative questionnaires. Figure 1 presents the ranking-exercise results showing the number of times each attribute was selected as one of the four most relevant attributes. The final list of attributes selected for the DCE, based on the top four most highly ranked attributes, included: visiting hours (regular clinic hours/after-hours 20:00+), travel time (40 min, 15 min, 0 min), waiting time, and familiarity with the provider—physician or nurse, as described in Table 3.

Fig. 1.

Ultra-Orthodox women’s rankings of attributes. ‘‘Imagine you want to see a family physician or a pediatrician in an HMO or a nurse in a Family Station to receive advice on a non-urgent, medical matter. You can set an appointment in the clinic or via a video application on a computer or a special kosher and approved video-device (not a smartphone). Please choose the four most important factors when considering a video versus an in-clinic visit.” The bars represent the number of times each attribute was chosen as one of the four most relevant. HMO health maintenance organization, phys

Pre-testing Results

Cognitive interviews with 15 UO women regarding the DCE questionnaire provided valuable insights into the accessibility of the developed DCE survey for UO women and its cultural sensitivity, the validity of attributes, levels, and scenario pairs, and offered comments on phrasing and errors found in the questionnaire that served as a basis for correcting several technical issues, especially in sociodemographic questions in the survey. All feedback was implemented, and corrections were made accordingly. The complete list of modifications is provided in the ESM. Figure 2 illustrates an example of a choice task from the survey.

Fig. 2.

Example of a discrete choice experiment task in the survey. Please choose your preferred option between the two ways of visiting doctors and nurses in community clinics. min minute

Acceptability and Cultural Sensitivity of the DCE Survey

The DCE survey method in general was accessible and culturally sensitive. Several women experienced difficulties understanding the purpose and method of the choice tasks, largely owing to unfamiliarity with such research methods and the concept of trading off attributes. To address this, we implemented several adaptations. For instance, we incorporated religiously appropriate analogies, such as ‘etrog’ (citron) selection for ‘Sukkot,’ the Jewish harvest festival, which involves decision making based on several citron parameters, a process that resonated with the participants’ cultural context. Women were asked to describe their considerations behind each choice in each pair of scenarios, and we observed that most of them traded off all the attributes rather than using simplifying heuristics. Additionally, all women correctly answered the dominant set. We found that the ability to perform the DCE choice task was related to the respondents’ cognitive abilities and may be related to the education level as previously described in the literature [43]. To ensure cultural sensitivity, we consulted with community leaders and adjusted the language and presentation of the survey, using pictures that adhered to modesty standards and removing any content that could be perceived as immodest or contrary to religious norms. Our conclusion was that to obtain meaningful results in the future DCE survey, it is essential that each respondent has the opportunity to ask the interviewer questions. Therefore, budgetary and other resource allocations should be planned accordingly for the future DCE procedure. Eight comments regarding terminology, phrasing, and questionnaire construction were collected during the pre-testing phase, along with one technical error identification.

Validity of Attributes, Levels, and Scenarios

The validity of the DCE method was assessed by analyzing verbal feedback from cognitive interviews and review of choices made in the DCE tasks. Choice review showed that attributes and levels behaved as expected and aligned with the verbal feedback. For instance, attribute levels that were perceived as beneficial, such as shorter waiting times, showed a positive sign, indicating a gain in utility. Conversely, those perceived as detrimental, such as longer travel times, indicated a loss in utility. Importantly, the verbal feedback and preference estimates for the ‘time of day’ attribute showed an interesting pattern. The UO women’s choices in the survey associated with after-hours (after 20:00) consultations suggested that women gained utility from this option, which was supported by their verbal feedback indicating a preference for healthcare services that do not interfere with family and work responsibilities. The ‘familiarity with the healthcare provider’ attribute demonstrated expected behavior in both verbal feedback and preference estimates. Ultra-Orthodox women reported a strong preference for familiar providers. This consistency further validates the inclusion of this attribute in the DCE. A significant finding from the interviews was that clinic travel times are considerably longer than simple distance calculations might suggest. Respondents detailed how the total time investment encompasses multiple components: preparing multiple children for the visit, dressing appropriately for the public place, transportation time (either walking or driving), finding parking, and navigating to the clinic itself. Based on these insights, the travel time attributes in the final questionnaire were modified to better reflect these real-world circumstances.

Overall, pre-test test results suggested that the attributes and levels selected were meaningful and relevant to the UO women, allowing for valid preference elicitation. Levels that were found to be imprecise were modified and appropriately adjusted in the final version of the survey.

The final DCE questionnaire was developed, and data collection is currently beginning with a sample of UO women. The final survey is attached in the ESM.

Discussion

This study presents an approach to developing a DCE questionnaire to elicit preferences for VCs among UO women in primary care settings. The research addresses a critical gap in healthcare accessibility for a traditionally underserved population, highlighting the potential of VCs to overcome cultural and logistical barriers. The study’s three-stage methodology of attribute identification, refinement, and pre-testing demonstrates an approach to DCE development that aligns with best practices in health preference research [44, 45], while incorporating unique strategies to address the specific challenges of conducting research within this closed community. Our qualitative findings revealed important insights about healthcare decision making in UO communities, demonstrating how four broad domains of accessibility, provider relationships, technical-religious considerations, and family management were transformed into specific DCE attributes that reflect both practical needs and cultural sensitivities of the target population.

A key strength of this study is its recognition of the complex sociocultural factors influencing healthcare decisions within the UO community. By including diverse stakeholders (women, men, rabbis, and healthcare providers) in the initial stages of attribute identification, the researchers ensured a comprehensive understanding of factors influencing VC adoption. The analysis revealed broad agreement across stakeholder groups on most key factors, though notably, modesty concerns were emphasized more strongly by men and religious leaders than by women themselves. This approach is particularly important in closed communities where decision making often involves multiple influences [46, 47]. The identified attributes reflect both practical considerations (e.g., waiting time, travel time) and cultural factors (e.g., modesty, privacy, familiarity with providers), highlighting the importance of tailoring telemedicine solutions to the specific needs and values of the target population.

The approach to overcoming methodological challenges in this closed community, such as using culturally appropriate language and female UO interviewers, provides valuable lessons for conducting research in similar settings. These strategies align with recommendations for culturally sensitive research methods [48] and highlight the importance of researcher reflexivity and cultural competence in health preference studies.

Our findings confirm the significance of two well-established time-related factors in primary care preference research: travel time and waiting time at the clinic. Prior research has consistently shown these factors to be crucial determinants in healthcare utilization and patient satisfaction [49–51], particularly in primary care settings [52, 53]. Travel time preferences may also be influenced by modesty concerns around public transportation use among some community members, though the extent of this factor’s impact requires further investigation. Interestingly, our study revealed an unexpected finding regarding time to the next available appointment. While this factor has been identified as significant both globally and in the general Israeli population [54], it did not emerge as crucial among our UO women participants. This finding could be explained by the unique appointment-making culture within the UO community, where walk-in visits are prevalent rather than scheduled appointments [55]. Additionally, the timing of service availability emerged as significant, particularly the possibility of receiving healthcare services after 20:00, which is currently unavailable in the Israeli public healthcare system. This aligns with recent studies emphasizing the importance of extended service hours for working populations [56, 57]. Provider familiarity emerged as another crucial factor, aligning with previous research on continuity of care in culturally distinct communities [54]. This factor is particularly relevant in the telehealth contexts, where maintaining the patient–provider relationship presents unique challenges and opportunities [58, 59].

These findings have important implications for healthcare policy and practice, providing a foundation for developing tailored telemedicine services that are more likely to be accepted and utilized by this population. This approach could potentially lead to improved healthcare access and outcomes for UO women and their families.

Limitations

The study has several limitations. Despite using both snowball and purposive sampling methods, the inherent limitations of snowball sampling may have introduced selection bias, potentially over-representing certain sub-groups while under-representing others within the UO community [60, 61]. The participating women may represent a more moderate segment of the UO community, characterized by a greater openness to cooperation with external entities. Consequently, the preferences of more conservative women or those from more extreme UO factions remain unknown. Additionally, findings may not be directly transferable to UO communities outside Israel because of differences in healthcare systems and funding mechanisms.

Another limitation relates to our attribute ranking methodology. By asking participants to select the four most important attributes rather than ranking all 11 attributes, we were unable to capture the full preference ordering or understand the relative importance of non-selected attributes. While this approach reduced the cognitive burden and aligned with our goal of identifying the most crucial attributes for the DCE, it may have missed potentially valuable information about the relative importance of the remaining attributes. Future research might benefit from employing both selection and full ranking methods to provide a more comprehensive understanding of attribute preferences.

Practical Implications

While conducting culturally appropriate research requires significant resources, including extensive cultural adaptations, extended research time, and substantial budget allocation for conducting lengthy individual meetings, these investments are crucial for meaningful engagement with closed communities. The effort required for cultural adaptation and individual attention should not deter researchers from undertaking such studies, as they are essential for developing effective healthcare solutions for traditionally underserved populations.

Conclusions

This study contributes insights to the broader discourse on DCE survey development for preference elicitation in culturally sensitive and historically underrepresented closed communities. This research serves as a guide for researchers facing similar challenges in diverse cultural settings, ultimately advancing the field of patient preference research and improving the applicability of DCE methodology in complex sociocultural environments.

Supplementary Information

Below is the link to the electronic supplementary material.

Declarations

Funding

Open access funding provided by University of Haifa. This research was funded by the Israel Ministry of Health, Haifa Department of Family Health Stations (Grant number 1234). Its publication is not contingent upon the sponsor’s approval. The authors declare that no funds, grants, or other support were received during the preparation of this article.

Conflict of interest

Irit Chudner, Anat Drach-Zahavy, Batia Madjar, Leah Gelman and Sonia Habib have no conflicts of interest that are directly relevant to the content of this article.

Ethics approval

All procedures followed were in accordance with ethical approval. The study was approved by the Faculty of Social Welfare and Health Sciences in Haifa University Ethics Committee on 28/2/24, approval number 041996.

Consent to participate

Informed consent was obtained from all responders included in the study.

Consent for publication

Not applicable.

Availability of data and material

Not applicable.

Code availability

Not applicable.

Author contributions

All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by IC, AD-Z, and LG. BM and SH contributed to the conclusions and recommendations development and facilitated access to key stakeholders and rabbinical authorities in this closed community, enabling the recruitment of participants. The first draft of the manuscript was written by IC, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.