Implicit ethical choices in ultra-rare therapy development: researchers’ perspectives on patient selection and communication, therapy value, and accessibility Authors: Mariia V. Maksimova, Rosan L. Lechner, Felicia V.Y.L. Lee-Kong, Johannes J.M. van Delden, Ghislaine J.M.W. van Thiel Published: 2026 ABSTRACT BackgroundThe development of new highly technological and highly personalized therapies for ultra-rare diseases is one of the fast-advancing areas in biomedicine. Yet such an approach is experimental, resource-intensive, and accessible mostly in affluent countries, thereby provoking ethical concerns. From a constructivist perspective, those therapies are not neutral tools, and ethical concerns are not merely added to science but embedded in the very choices researchers and technology developers make. Early research decisions about how therapy goals are defined, which patients are included, and which diseases are prioritized already include implicit normative decisions that set the further therapy development trajectory. Nonetheless, the perspectives of researchers working on these individualized therapies remain underinvestigated. In this study, we aim to address this gap and explore how researchers developing such therapies understand and enact ethically loaded choices in early therapy development in practice.MethodsWe conducted 15 semi-structured interviews with researchers developing a standardized platform for discovering personalized ASOs for ultra-rare neurodevelopmental disorders. Interviews were subjected to inductive thematic analysis, focusing on semantic themes. We report this study according to the Consolidated Criteria for Reporting Qualitative Research.ResultsFrom the analysis, four main semantic themes were developed inductively. Researchers described patient and disease selection as a matter of feasibility, focusing mainly on the technical aspects of therapy development. They saw emotions as a potential source of bias in the selection process and thus advocate for selection committees and decision-making algorithms that reduce personal involvement. Communication with patients and their families was perceived as an ethically fragile area due to false hope phenomena and the emotional pressure of personal communication. Perceptions of therapy’s value varied, ranging from symptom relief and reduced family burden to doubts about whether the intervention was justified given the high translational uncertainty. Finally, researchers identified various barriers to therapy accessibility, yet described them as parts of the landscape rather than as sites for personal actions.ConclusionsThe early stages of individualized therapy development for ultra-rare diseases include implicit ethical choices that are often seen as purely technical. Explicating the ethical component of these choices can help make the process of therapy development more transparent and fair.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12910-026-01463-x. Background The development of new highly technological and highly personalized therapies for ultra-rare diseases, often severe and early-onset conditions, is one of the fast-advancing areas in biomedicine [1]. Among available modalities, antisense oligonucleotides (ASOs) are currently considered one of the most feasible therapeutic options for ultra-rare diseases [2]. For instance, 38% of all neurodevelopmental disorders could be amenable to ASO approach [3]. ASOs are synthetic nucleotide sequences designed to selectively modify RNA and regulate protein production or function, without altering DNA. Because the body gradually eliminates them, ASO therapies do not provide a permanent cure but require repeated invasive administrations [2]. ASOs are increasingly explored for personalized therapies, offering hope where no treatment previously existed. At the same time, ASOs are labor-intensive, costly, and cannot fully cure the underlying disease; they are likely to remain accessible mainly in wealthy countries [4, 5]. This combination of scarcity, experimental nature, and high costs of personalized therapies raises ethical questions about their true therapeutic value, patients’ and families’ expectations, and fair and equal access. From a constructivist perspective, those therapies, as biomedical technologies, are not neutral tools, and ethical concerns are not merely added to science but embedded in the very choices researchers and technology developers make [6–10]. Early research decisions about how therapy goals are defined, which patients are included, and which diseases are prioritized already include implicit normative decisions that set the further therapy development trajectory. To address these issues, we need to explicate the perspective of therapy developers, in this case, researchers. So far, however, little research has been done to investigate the perspectives of researchers directly involved in the early development of individualized therapies for ultra-rare diseases. The main focus of existing qualitative studies is on the views of clinicians, patients, their families, and caregivers. Several studies covered opinions of clinicians and clinical geneticists involved in individualized therapies or rare disease trials [11, 12], as well as expert opinions on ethical dilemmas in access and assessment of high-cost orphan drugs [13]. Other empirical studies have explored academia experts’ perspectives on systematic challenges in rare disease drug development [14] and the social impact of the high-priced orphan drug [15]. Additionally, some studies examined stakeholder perspectives, including researchers’ views, on some aspects of rare genetic research or individualized trial approaches, like methodology, study designs, or consent [16, 17]. Yet, to our knowledge, no empirical study has examined how researchers reason about early decisions in individualized therapy development, such as fair patient selection, establishing therapy value criteria, and managing family expectations, which collectively shape the therapy development trajectory. To address this gap and explore how early therapy development choices are understood and enacted in practice, we conducted qualitative interviews with researchers in the PSIDER TAILORED consortium. The PSIDER TAILORED is a Dutch interdisciplinary consortium developing a standardized platform for discovering personalized ASOs for ultra-rare neurodevelopmental disorders. Ultimately, TAILORED seeks to use this platform to achieve successful therapy. Parallel work in our project investigates the perspectives of patients and caregivers; those findings will be reported elsewhere. Methods Design Ethical work in TAILORED is guided by the Value Sensetive Design (VSD) methodology, which combines conceptual, technical, and empirical investigations. In VSD, semi-structured interviews are considered one of the core methods to study the stakeholders’ values [18].Thus, we conducted semi-structured interviews with researchers working within TAILORED. The interview script contained questions about the respondents’ research goals, expected benefits and risks, general questions about ethical perspectives, and social impact (see Additional File 1). The script was revised after the first three interviews. One section was removed to avoid redundancy and improve respondent engagement. We report this study according to the Consolidated Criteria for Reporting Qualitative Research (COREQ) [19]. Sample of respondents Participants were recruited via purposive sampling, and the inclusion criterion was affiliation with the TAILORED project and direct experience in research decision-making. All leaders of the consortium’s Work Packages (WPs) were personally invited via email (excluding GvT, who took part in the design and data analysis of the interviews herself). Additionally, PhD candidates and postdoctoral researchers within the consortium were invited to participate during the consortium meetings and through a group email call. To mitigate potential hierarchy effects, junior researchers were recruited and interviewed separately, and interviews were conducted individually and confidentially by a medical student and a PhD candidate. Data collection Interviews were conducted online via Microsoft Teams in three waves, with time in between to reflect on data from May-August 2022, December 2023, and May 2024. The first wave focused on principal investigators and served to identify key issues early. The iterative timing also followed ethics workshops and regular ethics sessions, allowing revisiting insights after ethical deliberation. Ethical workshops were part of the consortium’s ongoing governance activities. This approach is consistent with the iterative characteristic of VSD, in which empirical investigations are repeated at different moments to capture emerging value considerations [20]. Thus, the data capture not merely or exclusively a baseline of researchers’ “spontaneous” pre-existing moral intuitions, but the ongoing articulation and negotiation of value considerations within an ethically mediated real-life process of individualized therapy development. The interviews were conducted primarily by MM and RL, who completed training in qualitative research and in-depth interviews. MM is a full-time PhD candidate at the Department of Bioethics and Health Humanities at UMC Utrecht. RL is a PhD candidate and. medical doctor affiliated with the department of Clinical Genetics and the department of Pediatrics, Erasmus MC. Additionally, GvT (PhD, associate professor) and FLK (student intern), both affiliated with the Department of Bioethics and Health Humanities at UMC Utrecht, assisted in conducting interviews. The study’s purpose was clarified both in emails and before interviews. Interviewees gave oral informed consent. Interviews were conducted in a trusted atmosphere and were video-recorded via Microsoft Teams, then transcribed verbatim by MM and FLK. Recordings were deleted after transcription. Field notes were taken during interviews and discussed with participants individually at the end of each session to ensure key points were accurately captured. Transcripts were not returned to participants. Data analysis We performed inductive thematic analysis following the guidelines by Braun and Clarke, focusing on semantic themes [21]. A preliminary coding tree was developed through independent coding by GvT, MM, and RL of a subset of 7 WPs interviews. Preliminary coding tree of themes and subthemes from these interviews was shared with participants for feedback. The coding tree was refined iteratively by MM and FLK. All transcripts were coded using Nvivo 14 software. GvT and HvD independently reviewed two major themes coded for all 15 transcripts to ensure coding credibility. MM finalized and reported the thematic structure. Results In total, 15 participants agreed to participate in the study (7 WP leaders, 4 PhD researchers, and 4 postdocs). The interviews lasted from 44 to 85 min, with a median duration of 57 min. Characteristics of participants are summarized in Table 1. Most were more senior researchers, the majority worked primarily in wet-lab research, and just over half had some degree of experience in interaction with patients or caregiver involvement. This data was derived from interview transcripts and represents our interpretation of participants’ accounts, rather than self-identifications of interviewees. Notably, “meeting patients or caregivers” in Table 1, refers to participants’ reported professional experience of direct contact and meaningful communication with patients or patients’ families (e.g., through clinical work, diagnostic consultations, or patient-facing research-related activities), and not to a formal part of their consortium role. Table 1Characteristics of participants in the interviews Career level PhD Candidate4 Postdoc/Assistant professor4 WP7 Field of work Wet Lab Research9 Clinical & Translational5 Computational & Theoretical1 Meeting patients or caregivers Yes8 No7 Table 1Characteristics of participants in the interviews Career level PhD Candidate4 Postdoc/Assistant professor4 WP7 Field of work Wet Lab Research9 Clinical & Translational5 Computational & Theoretical1 Meeting patients or caregivers Yes8 No7 Characteristics of participants in the interviews From the analysis, four main semantic themes were developed inductively. These four overarching themes, with their subthemes, were identified early and were repeatedly reported in the next interviews, as shown in the saturation grid (see Additional file 2). This suggests saturation at the level of the main thematic structure. Later interviews mainly added to existing subthemes. However, several subthemes within Theme 4 related to therapy access appeared less consistently, likely reflecting differences in how participants perceive their impact on therapy accessibility, which we elaborate on further in the paper. Theme 1: Disease and patient selection decisions Most of the interviewees were concerned with the process of selecting patients and diseases for the development of individualized therapy. Many described this process as pragmatic. They primarily focus on “technical” criteria such as genetic eligibility, measurable outcomes, and alignment with regulatory requirements. Disease severity (as mentioned by interviewee: families do not always agree, yet the common ground among researchers is that diseases should be severe) and disease progression are also considered selection criteria for study participants. According to expressed opinions, the more severe cases justify taking a heightened risk of experimental therapies. Resource scarcity in academia (e.g., money, capacity, personnel) was also named as a valid reason to limit eligibility to the most severe cases. The disease should progress rapidly enough so that researchers can demonstrate the impact of therapy on symptom decline. At the same time, the condition should not progress to the point where treatment becomes pointless:"if you have a specific disease, just to use blindness as an example, if the patient is already fully blind, there’s no point in treating, because you will not make him or her unblind." "if you have a specific disease, just to use blindness as an example, if the patient is already fully blind, there’s no point in treating, because you will not make him or her unblind." Interviewees said that patients’ personal stories and emotional requests from the families, or even their personal optimism, may influence the decisions to choose this specific patient for further therapy development process. This subjectivity is described simultaneously as a challenge and an inevitable part of interacting with families who faced rare and severe diagnoses. As the solution to reduce the impact of personal emotions, interviews mentioned developing decision algorithms and assembling decision-making boards:"what we’re currently doing is we are designing algorithms to automatize that whole process to take the human element away.""I think we are trying to take out the personal involvement in the sense of we have these multiple steps and multiple people that are looking at these cases. So it’s not a single person that decides and it’s not one driving factor." "what we’re currently doing is we are designing algorithms to automatize that whole process to take the human element away." "I think we are trying to take out the personal involvement in the sense of we have these multiple steps and multiple people that are looking at these cases. So it’s not a single person that decides and it’s not one driving factor." Another value-laden level emerges in the discussion of the role of disease prevalence in the selection process. Some interviewees support the development of truly N-of-1 therapies as they find them more regulatory feasible and scientifically valuable. Others doubt the ethical and economic justification of investing in N-of-1 research, as investing in more prevalent rare diseases could benefit more patients."… if it’s a less common disorder, does that mean that the patients who have this disorder are less deserving of a cure? No, right? But targeting those patients will cost a lot more[…] if you have to choose to spend your 10 million, and you can either help 100 people with your 10 million or only one, then you will choose 100. I think most people will." "… if it’s a less common disorder, does that mean that the patients who have this disorder are less deserving of a cure? No, right? But targeting those patients will cost a lot more[…] if you have to choose to spend your 10 million, and you can either help 100 people with your 10 million or only one, then you will choose 100. I think most people will." Several suggested that the main focus would be to understand how to develop these therapies and establish therapy development pipelines that could be scaled to more prevalent diseases later. Yet interviewees noted that scaling therapy development pipelines may raise additional ethical issues, such as reassessing acceptable risks or setting a new threshold for what is considered a rare disease. Lastly, interviewees mentioned how families who are well-informed and understand risks well may be preferred for selection. As one interviewee pointed out, it helps to mitigate the risks of false hopes and therefore reduces ethical tensions."I’d rather pick those patients […], for which I know that patients have been counseled by a clinical geneticist who have had a full analysis of all of the genetic data before. […] So consent-wise and legislation-wise, everything is in place. I’d rather do that. So that’s my additional ethical, moral standards." "I’d rather pick those patients […], for which I know that patients have been counseled by a clinical geneticist who have had a full analysis of all of the genetic data before. […] So consent-wise and legislation-wise, everything is in place. I’d rather do that. So that’s my additional ethical, moral standards." Theme 2: Patient communication and managing expectations The interviews show how researchers regard communication with patients, participating in research, their families, and caregivers. Most interviewees specified the role of tone and transparency in communication as a tool to manage expectations. According to some, families “listen differently” than scientists and interpret any information through personal experience and emotions. Interviewees repeatedly emphasized the importance of clearly explaining the risks and boundaries of what is possible, rather than giving false hope. The false hope was mostly explained as patients and families interpreting early or partial research signals as promises of a cure. For example, believing that preclinical results equal a developed therapy, or abandoning proven standard care because of the imagined cure just around the corner:"So they think “well, it doesn’t matter that he slows down quicker now without the steroids because CRISPR will undo." "So they think “well, it doesn’t matter that he slows down quicker now without the steroids because CRISPR will undo." Interviewees highlighted that patients’ limited understanding of genetics or preclinical work may be a barrier, as even publications or mentions in the media can be treated as a promise of a therapy. The practices of other colleagues can also create inflated expectations in families: “We had one researcher [in the field]… who would tell parents. “What I need is a skin biopsy or a muscle biopsy from your son, and then I will make the treatment”. So they would ship samples … and he would design the oligo in cultured cells and then he would send them the results. Then these people would go to their clinician and say “look, there is a drug for my child. I want this”. “We had one researcher [in the field]… who would tell parents. “What I need is a skin biopsy or a muscle biopsy from your son, and then I will make the treatment”. So they would ship samples … and he would design the oligo in cultured cells and then he would send them the results. Then these people would go to their clinician and say “look, there is a drug for my child. I want this”. In the described case, clinicians declined the requests, noting that it was just a research result from an in vitro experiment, not a ready-to-use therapy. According to interviewees, such cases place clinicians in a difficult position, as they must deny requests and are then perceived negatively by patients. This example illustrates how false hope can undermine trust and complicate responsible communication with families. Most interviewees saw their role in responsible communication as bringing a realistic picture to patients, and explaining to them as clearly as possible why such promises can be unrealistic. Overall, they saw an obligation to present a clear picture of the current possibilities of biomedical science to patients, families, and to facilitate such discussion on a broader public level. The common thread was that explaining biotechnology in advance could help reduce misconceptions. Interviewees mentioned the tension in the mere fact of initiating communication with families. They described concerns that early patient involvement may be felt as a promise that could generate risks of false hope, especially if the project fails. The careful choice of language and communication style was seen as a way to mitigate these risks. Simultaneously, communication with patients in the research is seen as a duty. Several interviews described regular feedback to patients in research as a form of paying respect and building trust. The absence of feedback was described as unethical and even exploitative. However, as mentioned by interviewees, this feedback should be delivered in a less personal and more objective manner, focusing on the experimental context. One interviewee explained that researchers should avoid casual lab-meeting sayings like cell line performs well or badly, or referring to the cells as “your cells”, as patients might feel personally connected to these cells. All who had personal experience communicating with patients at the time of the interviews described this experience as motivating and insightful. At the same time, a few noted that it may cause emotional pressure, and dealing with it requires specific frameworks:"We all talk about patient participation, but you also get emotionally personally involved with the families. You know, there are families that live here in my town. And whenever there’s a meeting, we always travel together. And I see his daughter. She’s getting sicker, … year by year. That’s just horrible. But still, you have to keep it separate from your work because the experiments in the lab will not go any faster if I feel pressure or guilty." "We all talk about patient participation, but you also get emotionally personally involved with the families. You know, there are families that live here in my town. And whenever there’s a meeting, we always travel together. And I see his daughter. She’s getting sicker, … year by year. That’s just horrible. But still, you have to keep it separate from your work because the experiments in the lab will not go any faster if I feel pressure or guilty." Finally, one of the interviewees emphasized that only by talking with others can researchers discover what they might miss, making communication and openness to different perspectives essential for knowledge building. Theme 3: Factors defining therapy value Researchers described individualized therapy for ultra-rare diseases not as unquestionably good, but framed its value as plural and sometimes ethically ambivalent. We could not identify a clear correlation between their view and the background of the respondent, their role in the consortium or reported patient contact, as the saturation grid shows no clear clustering within the reported theme (Additional file 2). Additionally, all the interviewed researchers work within the same consortium focusing on ultra-rare neurodevelopmental disorders. Some described the value as relief of particular symptoms, e.g., fewer seizures or improved behavior, which could lessen the burden on caregivers and improve the quality of life for the patient’s family. Yet they noted that, even with these symptoms, some patients may be affected more, others less, and framed the symptom alleviation as related to personal experience of the disease. For example, one interviewee described a possible disconnect between what researchers may think patients need help with and the symptoms that actually cause the most burden to patients and their families:"You might want to think they are super happy, you know, when you find the treatment so that they can walk again. But actually this is not their real problem, because they also have difficulties with swallowing and with speech." "You might want to think they are super happy, you know, when you find the treatment so that they can walk again. But actually this is not their real problem, because they also have difficulties with swallowing and with speech." Even at very early stages of therapy development, as mentioned by one interviewee, an intensive diagnostic assessment can itself be valuable for the patients’ families, particularly in paediatric neurology, as it can give parents more prognostic information about the disease, and support future reproductive planning. Another interviewee explicitly mentioned the value of therapy in lessening the burden of close family members caring for a patient: “… if their child is doing better, the parents are feeling better, [it] takes away some of the everyday care that they have to take. They might be participating more in society”. “… if their child is doing better, the parents are feeling better, [it] takes away some of the everyday care that they have to take. They might be participating more in society”. At the same time, interviewees questioned whether certain types of “improvement” are necessarily valuable for the patient. A few questioned whether a developmental gain, for example, “going from a neurodevelopmental age of 18 months to an IQ of 60 or 70”, can be truly meaningful or if it results in higher levels of distress for the child. In certain cases, interviewees described how families might benefit more from non-experimental treatment or supportive care and assistance. There are families who do not want experimental interventions for their child. One interviewee recalled a first encounter with the parents who rejected such an intervention for their child: “ …I met the first parent that said, “ I really like what you’re doing, you should be doing it, but it will never be for my child because I don’t want it” It was really a surprise to me… I really thought that everyone was really thinking that we need to improve quality of life or find a treatment or cure. I had not envisioned that people said “I think things are perfectly fine as they are”. “ …I met the first parent that said, “ I really like what you’re doing, you should be doing it, but it will never be for my child because I don’t want it” It was really a surprise to me… I really thought that everyone was really thinking that we need to improve quality of life or find a treatment or cure. I had not envisioned that people said “I think things are perfectly fine as they are”. Several interviewees emphasized that experimental interventions may not result in the improvement hoped for, and the impact is often slow or indirect. Some emphasized that additional value can lie in giving families a sense of possibility, even if the therapeutic result is uncertain. Interviewees also highlighted that therapy value depends on the burden and characteristics of the treatment modality itself. ASOs were seen as more feasible, reversible, and “normal” in comparison with gene therapies, resembling familiar therapies, but required repeated intrathecal injections, which were acknowledged as a heavy burden :"It might [be that] the treatment will never cure disease…So it’s an injection in your spinal cord that’s also quite high burden on patients, and then what is worse? Is the treatment worse than or the administration worse than the disease?" "It might [be that] the treatment will never cure disease…So it’s an injection in your spinal cord that’s also quite high burden on patients, and then what is worse? Is the treatment worse than or the administration worse than the disease?" Uncertainty was repeatedly described as an inalienable characteristic of ultra-rare therapy development impacting its value. Many mentioned translational uncertainty, meaning that strong results in cell models may not translate into improvement for the patient. Others highlighted the uncertainty about what may constitute the success of therapy. Several also acknowledged the uncertainty of unintended harm, the possibility that a therapy could worsen a patient’s condition. Interviewees reflected on when such an uncertain medical intervention is ethically justified. The common narrative was that therapy should only be used in severe cases or where the expected benefit-risk ratio is strong. Some of the interviewees described a feeling of moral duty to act if preclinical results look promising. Others doubted whether treating such disorders is actually in the patient’s best interests and questioned where the line is that is not to be crossed:"I’m very convinced that there’s a lot of things we can do that we really shouldn’t." "I’m very convinced that there’s a lot of things we can do that we really shouldn’t." Finally, the benefit of ultra-rare therapy development was seen as extending beyond individual patients. Streamlining the process via platforms or pipelines was seen as creating value for future patients. The possible benefits also include a greater understanding for parents and societal gains, like reduced care burden and advancements in ASO methods. Thus, across the interviews, therapy value was described as a trade-off between (i) symptom priorities and lived burden of the disease, (ii) individual and family-centered interpretations of meaningful benefit, (iii) treatment burden and reversibility, (iv) translational uncertainty and risk of harm, and (v) longer-term scientific goals of establishing platform for treating future patients. Theme 4: Explicit and hidden barriers in therapy accessibility Interviewees discussed the accessibility of individualized therapies at two levels: the patient level and the infrastructural (system-level) context. Most interviewees mentioned well-recognized geographic, socioeconomic, and language barriers that shape therapy accessibility at the patient level. Therapy development and clinical trials are mostly centered in affluent Western countries, and patients from low- and middle-income countries or remote regions face limitations and a lack of access. Among other patient-level aspects of accessibility, interviewees highlighted social capital and active advocacy from parents. Active, knowledgeable families with connections have increased chances of getting into clinical trials or getting early access to therapy. However, this exact mechanism increases inequality. Families with fewer social resources or less education may find themselves as disadvantaged. At least a third of interviewees expressed concern about the financial pressure on vulnerable families who are forced to participate in crowdfunding or “opportunity shopping” in the commercial market in order to access individualized therapies. Beyond individual circumstances, interviewees pointed to system-level barriers to the ability of healthcare and research infrastructures to accommodate new, individualized types of therapy development. Even in wealthy countries, resource scarcity and infrastructural limitations force systems to choose which therapies to support. A few interviewees described academic-based therapy development for ultra-rare diseases as an emerging alternative to traditional commercial models. They regarded it to be more ethical than commercial models, as academic institutions operate on a non-profit basis and can price therapies based on actual costs. One explicitly proposed the design of a sustainable funding model for academia-led therapy development:"The idea is… that we could get reimbursement for the development of one of those personalized treatments, and then we can use that reimbursement to develop the next personalized and then sort of make a sustainable model." "The idea is… that we could get reimbursement for the development of one of those personalized treatments, and then we can use that reimbursement to develop the next personalized and then sort of make a sustainable model." However, there were also doubts about whether the academic laboratories could support large-scale therapy development. Interviewees emphasized disparities between large, well-funded labs, which can allow expensive projects, and smaller labs with limited resources, as well as challenges associated with transferring pipelines and knowledge internationally due to diverse regulatory environments. Furthermore, researchers point out insufficient clinical infrastructures, such as shortages of trained medical staff and treatment facilities, as limitations for broader access to therapies. Some interviewees suggested collaboration with the industry as a solution, citing its potential to develop individualized therapies. At the same time, interviewees’ concerns include industry reluctance due to financial risks and data sharing issues, as well as commercialization leading to the abandonment of less profitable therapies. Additionally, some worry that companies may make promises to patients that are hard to keep for profit: “So there’s a company in the US that asks for three million dollars to make your personalized antisense oligonucleotide. [Patients] will say yes to it. [Company] will just want the money. That is something that I see coming up where I get very uncomfortable because they are… they’re making promises they won’t be able to hold.” “So there’s a company in the US that asks for three million dollars to make your personalized antisense oligonucleotide. [Patients] will say yes to it. [Company] will just want the money. That is something that I see coming up where I get very uncomfortable because they are… they’re making promises they won’t be able to hold.” Discussion All interviewees identified one or more ethical aspects of the dilemmas described in each theme. They saw different barriers in access to therapy for ultra-rare neurodevelopmental disorders. They wanted to treat patients in a fair manner, yet gravitated to feasibility criteria. Researchers recognized the emotional impact of personal stories and seek to create tools to minimize emotional and personal bias. Therapeutic value was described by researchers as a value constructed through technical aspects, true patient needs, family expectations, and possibilities for future patients. Without naming as such, interviewees often referred to ideas central to patient and public involvement. Such ideas strongly resonate with several well-developed bodies of literature on how benefit is interpreted in rare diseases, via an emphasis on symptoms important for everyday and social functioning, and the alleviation of caregiver burden [22, 23]. Interviewees’ doubts that certain developmental gains are not necessarily bringing a positive impact align with the literature on treating intellectual disabilities. Scholars in disability studies have questioned whether treatments that partially improve intellectual disabilities may be problematic, as they may reshape patients’ and families’ lived experience while not necessarily improving well-being [24]. This concern intersects with the disability paradox, a phenomenon demonstrating that people with disabilities can achieve good quality of life through adaptation to their current state [25]. Therefore, our data show that researchers appeared to be reasoning in line with contemporary perspective: successful development of personalized therapy requires well-conducted conversations about meaningful treatment goals for the patient and their family [26–29]. Moreover, the patients’ meaningful input is increasingly expected by regulators. The U.S. Food and Drug Administration (FDA) provides guidance documents on the approach to define a meaningful aspect of health and justify a fit-for-purpose COA (Clinical Outcome Assessments), and the European Medicines Agency (EMA) calls for early and systematic considerations of patient experience data throughout therapy development [30, 31]. Yet, interviewees did not describe how patient perspectives are used in research decision-making. Taken together, this suggests that, in highly individualised ASO academic development not always aligned to a clear regulatory pathway, this regulatory logic has not yet been operationalized in research decision-making practices. Further work should develop practical tools that link patient input to concrete decision points regarding the therapeutic benefit of individualized therapy development. While interviewees’ concerns about therapy accessibility were situated in an ethical context, they were rarely extended to personal responsibility to intervene. The one notable exception was their active role in taking actions to build a non-profit academic model of individualized therapy development, which was regarded as more ethical and sustainable in comparison with commercial drug development. The pattern resonates with the paper by Niemansburg et al. [32], where professionals recognized societal consequences of regenerative medicine technologies, yet focused mainly on measurable “hard” impacts like risk and cost-effectiveness. Justice, equity, and lifestyle effects were rarely discussed and were externalized [32]. Our findings resemble what Swierstra and Rip [33] called organized irresponsibility in research and developmental practices, in which actors acknowledge injustice, but responsibility is distributed so widely that no one feels positioned to respond or able to act against it. Political theorists describe a similar broad pattern as structural injustice, in which disparities are visible to actors yet seen as background context, resulting from widely accepted institutional norms, rather than as sites for individual interventions [34]. One possible explanation for this observed pattern is that ethical responsibility in therapy development process is often described as collective and role-distributed, including regulators and governance structures. Thus, some participants could have located accessibility concerns at the level of team governance structures, including ethics WP within the consortium. However, this should not be interpreted as evidence of a lack of ethical reasoning among researchers. It illustrates that consortium-based therapy development relies on explicit distributions of agency across designated roles. We therefore suggest that such consortia require roles of embedded ethicists who should also support researchers in reflecting on how and which therapy development decisions may shape therapy accessibility. Emotions in our data emerge as something often treated as bias to be managed. Patient and disease selection is described by researchers as a pragmatic task under uncertainty, with attention to “technical” criteria, i.e., genetic eligibility, measurable outcomes, and regulatory alignment. Researchers use committees and protocols to promote fairness, depersonalize decisions, and create ethical neutrality. They also describe building decision algorithms and multi-person boards to reduce the impact of emotions and to standardize the decision-making process. By analogy to Frank’s discussion of “moral enhancement” and nudging technologies [35], these boards and algorithms can function as moral load management tools. They lower the burden of inner moral struggle ( in our case, caused by tension between felt emotions and professional distance) by shifting deliberation to procedures and groups. Similar to what Shah described as “division of moral labor” in research regulation [36], here, moral work is redistributed across procedures and people. Worth noting, what might appear as “procedural distancing” can also be understood as a chosen ethical strategy, rather than an avoidance of moral complexity. Procedural approaches to justice, like A4R (Accountability for Reasonableness) as proposed by Daniels and co-authors [37], are widely used in healthcare. A4R has been implemented in various healthcare and research contexts to guide priority setting or selection [38, 39]. Additionally, a view on emotions as a potential for bias appears in how interviewees described communication with patients. Emotional narratives and requests are seen as creating a risk of false hope and misunderstanding, which some try to mitigate it by clear counseling. In short, emotional involvement is experienced as ethically risky. Professional distancing from emotions is chosen as a reasonable response to enhance consistency and achieve perceived fairness. Yet, as Roeser [40, 41] argues in the context of deliberation about technological risks, moral emotions like empathy, guilt, anger, and compassion can function as morally informative signals rather than mere biases. These emotions provide data on personal and social values and highlight areas of attention in the decision-making process [42]. In our context, such emotions may help highlight ethically salient aspects of cases that remain invisible when decisions are framed as purely technical. Thus, emotions can also serve as an additional source of morally informative signals. A more constructive approach could involve redesigning selection frameworks to actively accommodate structured ethical reflection, including empathy and recognition of emotions as moral signals of injustices. For instance, if a selection committee had a structured way to express how they feel about proposed cases and their narratives (e.g. ‘I feel this case was treated fairly or unfairly,’ or ‘I feel a strong bias here because this story seems very tragic’), they could open a discussion about underlying characteristics that triggered those reactions. At the same time, patient selection decision-making in rare diseases happens amidst broadly acknowledged resource and manpower scarcity [4, 5]. It can also be characterized by high opportunity costs, where selecting one patient may come at the “cost” of another patient not receiving therapy development, especially as the disease may progress to a stage where treatment is no longer feasible by the time the research team can return to the case [43]. This combination of factors directs the overall discussion toward the field of allocation ethics in healthcare and raises explicit questions about selection under scarcity: “Who and why should get access to therapy development first?” Although this paper does not aim to answer this specific question, our empirical findings reveal the absence of an explicit ethical framework or principle guiding patient selection in practice. This mirrors an absence of thorough consensus on suitable principles for priority setting in health care [44]. The need-based approach or alternative severity criteria are proposed to set healthcare priorities in contexts where cost-effectiveness is difficult to establish [44–47]. In situations of scarcity and equally strong moral claims, lotteries were suggested to resolve indeterminacy in who should be chosen [48]. However, it is unclear how these healthcare priority-setting approaches can be applied to patient selection for the development of individualized therapies in ultra-rare diseases. This reinforces the need to design a practical decision-making framework to support such choices, including explicit moral justification for which considerations should legitimately inform patient and disease selection. Taken together, our findings suggest that in the early stages of therapy development, researchers already engage in ethical decision-making, including some form of priority-setting under scarcity. Yet this ethical process is implicit and often embedded in technical criteria, procedures, and committees, rather than made explicit through normative frameworks. Strengths and limitations Our study provides a unique empirical view on how researchers within the TAILORED consortium shape and explain their choices in ASO-therapy development. Our position as embedded ethicists within the consortium helped build open communication. This facilitation allowed for the collection of rich input about how researchers reason about diseases and patient selection, communication with families, and the value and accessibility of developed therapies. Our data fills in the almost empty niche in the literature of the ethics of ultra-rare diseases. At the same time, our embedded position in the consortium and the iterative VSD approach, including ethics workshops between interview waves, may have influenced how participants of later interviews were aware and reasoned about ethical issues. Thus, results may present articulated moral reasoning within a context that actively stimulates ethical reflection. While we would interpret this as a desirable outcome of ethically-informed technology development, it limits our ability to disentangle pre-existing practices from those shaped by ethically mediated reflection. Additionally, this study is based on purposive sampling of interviewees from the TAILORED research project. It can potentially also affect the generalizability of the study’s results. Conclusions Overall, our study has shown that even at the early stages of individualized therapy development, researchers make ethical choices that are often seen as purely technical. Decisions about which patients and diseases to include, how to communicate with patients and family members, and how to conceptualize the therapeutic value of individualized therapies reflect implicit assumptions, emotions, and infrastructural limitations. Understanding these implicit aspects can help make the process of therapy development more transparent and fair. Supplementary Information Additional file 1. Title of data: PSIDER TAILORED Script “Mock Ethics Review” semi-structured interview. Description of data: This file contains the semi-structured interview script used to guide data collection, including an introductory prompt and topic-specific questions. Additional file 1. Title of data: PSIDER TAILORED Script “Mock Ethics Review” semi-structured interview. Description of data: This file contains the semi-structured interview script used to guide data collection, including an introductory prompt and topic-specific questions. Additional file 2. Title of data: Study Saturation Grid. Description of data: This file is a theme-occurrence matrix showing which subthemes appeared in each participant’s interview (P01–P15) across three interview waves. The “x” marks the presence of a subtheme in a given interview, and the “new” marks where each subtheme first emerged. Additional file 2. Title of data: Study Saturation Grid. Description of data: This file is a theme-occurrence matrix showing which subthemes appeared in each participant’s interview (P01–P15) across three interview waves. The “x” marks the presence of a subtheme in a given interview, and the “new” marks where each subtheme first emerged.