Enhancing inclusion of neurodivergent participants in medical research: an intersectional inclusion-by-design approach

Neurodiversity describes natural variation in how people think, communicate, learn, and process sensory and social information across the whole population. In this article, "neurodivergent" refers to people whose cognitive, sensory, attentional, communicative or learning profiles diverge from dominant neurotypical expectations, including autistic people, people with attention deficit hyperactivity disorder (ADHD), dyslexia, dyspraxia and related neurodevelopmental profiles. As recognition grows, neurodivergent people are present across all clinical pathways and throughout the lifespan, far beyond specialist paediatric or disability services. Yet medical research still under-represents and under-supports neurodivergent people, treating accessibility as an exception rather than an expected feature of study design (1)(2)(3)(4)(5). This is not merely a representation gap. It is a matter of scientific validity, ethics, and health equity. When design choices systematically filter out neurodivergent participants-through rigid recruitment, inaccessible consent, overstimulating settings, or narrow notions of 'good communication'-findings become less generalisable. The evidence base can appear methodologically 'clean' yet be clinically fragile: interventions may perform well in selected samples but translate poorly to real-world populations where neurodivergence influences engagement, adherence, adverse effects, and outcomes (5)(6)(7)(8)(9)(10). The problem is amplified by the fact that neurodivergence may intersect with poverty, language barriers, ethnic minoritisation, stigma, trauma histories and unequal access to diagnostic pathways, producing forms of exclusion that are not simply additive but mutually reinforcing (11). The outcome is predictable: those who most need research to reflect their experience are least likely to be counted by it (1,4,5).The stakes are high because medical research does not only generate knowledge; it also sets norms about whose participation and outcomes count (6,7). Research methods communicate who is expected to participate, whose accounts are treated as credible, and which outcomes are valued. When research processes assume a neurotypical default-rapid verbal exchange, comfort with ambiguity, tolerance of sensory overload, and the ability to 'perform' calmness and coherence under time pressure-neurodivergent participants can be disadvantaged before a study begins. This is not a neutral inconvenience; it is a systematic source of bias. It shapes who is recruited, who stays, what data are collected, and how results are interpreted. Even when studies "include" neurodivergent participants, inclusion can be superficial if the environment and procedures remain misaligned with participant needs, resulting in rushed consent, masked distress, incomplete responses, or early withdrawal.Communication mismatch is among the most commonly cited barriers, but it is also often mislabelled as a participant deficit. The double empathy problem reframes communication breakdowns between autistic and non-autistic people as a two-way mismatch in understanding rather than a one-sided lack of social insight (12). Many study designs assume that reliable data require a single "standard" communication mode, typically spoken interaction at speed, in unfamiliar settings, with limited time to process. Neurodivergent participants may communicate differently, need longer to process questions, prefer written or asynchronous formats, use direct language, avoid eye contact, or find it difficult to produce immediate answers under observation. When research teams interpret these differences as unreliability or non-compliance, they inadvertently convert communication difference into exclusion. Even when research focuses on mental health experiences, the experiences themselves-and how well typical therapeutic or evaluative frameworks capture them-may differ for neurodivergent participants. This is particularly relevant where sensory overload, discrimination, or inaccessible communication environments shape lived experience (2,(18)(19)(20)(21)(22)(23). If research tools assume a neurotypical model of distress and then conclude that neurodivergent people are "atypical", the methodological problem has been mislabelled as a participant problem.Environmental barriers can be equally decisive and should be approached through explicit design tools rather than ad hoc goodwill. The Autistic SPACE framework-Sensory needs, Predictability, Acceptance, Communication and Empathy, supported by physical, processing and emotional space-offers a practical way to translate accessibility into research planning (13). Clinical or university research settings often contain bright lights, unpredictable noise, crowded waiting areas, strong smells and uncertain scheduling, all of which can be considered in advance through SPACE-informed planning (13,18,19,(24)(25)(26)(27). For participants with sensory sensitivities, these conditions can be exhausting or intolerable, and the energy cost of attending may not be visible to researchers. This invisibility matters because it affects who attends, how they perform during assessments, and whether they return. The same participant may provide rich, coherent data in a calm setting with adequate preparation, processing time and breaks, and minimal data in an overstimulating setting where self-regulation becomes the primary task. When this variation is not recognised, researchers may wrongly attribute differences in attendance, engagement or outcomes to the participant or intervention rather than to the context in which the data were collected.Institutional caution can also become a barrier when it drifts into overprotection. Ethics committees may worry that neurodivergent participants cannot give valid consent or may be at higher risk of distress, and may therefore demand restrictive safeguards or exclude participants outright. The intention is to protect autonomy, yet the effect can be to deny autonomy by preventing participation. This is particularly problematic because it can result in avoidable exclusion despite the availability of proportionate safeguards (9,20,21,(28)(29)(30)(31)(32)(33). When consent is treated as a one-off signature rather than a process that supports understanding over time, neurodivergent people may be harmed either by being excluded "for their own good" or by being included without adequate supports. Both outcomes represent failures of ethical imagination rather than inevitable realities.Stigma and unconscious bias shape research in quieter ways. Some research teams assume that including neurodivergent participants will reduce data quality, increase attrition, or complicate timelines. This belief can influence decisions about who is approached, how information is presented, and how participant behaviour is interpreted. It can also influence what outcomes are valued. When research continues to prioritise outcomes that are meaningful primarily to observers rather than to participants, such as appearing "normal" in behaviour, neurodivergent people may reasonably distrust the purpose of participation. Community accounts frequently highlight that research is welcomed when it addresses lived priorities and system responsiveness, and resisted when it feels extractive or deficit-focused (3,21,22,28).Funding structures compound these barriers. Inclusive practices require time, training, adaptations and sometimes additional staff, including accessible participant materials, longer or split appointments, sensory-adapted rooms, remote participation infrastructure, interpreter or advocacy support, reimbursement of travel and carer time, and paid involvement for neurodivergent contributors. If funders do not explicitly value and resource these elements, inclusion becomes dependent on individual goodwill. For example, a protocol may state that written responses or breaks are permitted, but without funded staff time and flexible room booking, those adjustments may be unavailable in practice. This leads to variability: some teams do excellent inclusive work, while others deliver the minimal acceptable standard. Research that is underfunded for inclusion often externalises costs onto participants and carers, particularly through long appointments, repeated visits, inflexible scheduling and unsupported digital access. This is not simply inconvenient; it can systematically exclude those with limited resources, precarious employment, caring responsibilities, or transport challenges. Survey evidence on healthcare access shows that adjustments are widely valued yet inconsistently available, reflecting a familiar pattern: needs are recognised, but systems do not reliably deliver (11,23,24,29). Research settings risk reproducing the same pattern unless inclusion is treated as a measurable deliverable rather than a moral aspiration.Finally, there is a conceptual barrier: neurodivergence is sometimes framed as a problem to be corrected rather than as a lens through which research environments can be redesigned. This matters because conceptual assumptions become operational decisions: a study that treats eye contact, rapid verbal fluency, sitting still or appearing calm as signs of engagement may misclassify neurodivergent participation. Similarly, a study that defines success as looking more "normal" may measure conformity rather than wellbeing, autonomy or meaningful participation. Interdisciplinary analysis has highlighted tensions in how neurodiversity is conceptualised and the opportunities this creates for more mature research cultures (4,25,30). If researchers assume that the participant must adapt to the research environment, exclusion follows. If researchers instead assume that the research environment must adapt to human diversity, inclusion becomes a technical, ethical and cultural task with measurable outputs.A central methodological point is that rigour and accessibility are not opposites. In many cases, accessibility increases rigour because it improves the fidelity with which participants can express what they intend to communicate, reduces context-driven distress, and decreases systematic missingness. The practical challenge is to distinguish what must be standardised (construct, timing, scoring) from what can be flexible (response mode, pacing, environment, supports). Inclusion-by-design is not "anything goes". It is planned flexibility with transparent reporting. Table 1 summarises a practical inclusion-by-design framework.First, communication should be treated as a design variable rather than an assumption. Where possible, studies should offer more than one mode of participation, such as written responses, asynchronous options, or structured interview formats that allow pauses and clarifications. This reduces measurement error associated with time pressure and high social-performance demands, and avoids conflating communication style with insight or capacity. Evidence that experiences of anxiety and therapy can diverge from neurotypical models, partly due to sensory and communication contexts, demonstrates why method flexibility is not cosmetic but epistemically important (2,26,27,30).Second, predictability should be treated as a core accessibility feature. Many neurodivergent participants benefit from knowing what will happen, in what order, for how long, and with what sensory demands. Providing clear "what to expect" materials, appointment structure, options for breaks, and explicit consent to pause or stop can reduce stress and improve data quality. Co-production work has demonstrated the value of agreed ground rules and structured systems that support collaboration and reduce ambiguity (20,28,31). Research teams can adapt this principle by establishing participant-facing ground rules that explicitly legitimise needs such as time to process, reduced eye contact, or written follow-up, without interpreting these as disengagement.Third, the sensory environment should be planned, not improvised. Research often treats the setting as fixed, yet small changes can have disproportionate benefits: quiet waiting areas, reduced lighting intensity, minimised background noise, and predictable scheduling that avoids long uncertain delays. Where the physical environment cannot be reliably adjusted, offering remote or hybrid options can reduce sensory burden. The rapid expansion of remote methods during the pandemic demonstrated that some barriers can be reduced through telehealth, while also highlighting that remote delivery can introduce new barriers, including increased rigidity, technology challenges, and inequities related to digital access (15,16). The key point is that choice matters, and that options must be supported rather than merely offered.The COVID-19 period sharpened this lesson. Rapid shifts to remote and hybrid methods improved access for some autistic adults, yet they also revealed new barriers linked to digital exclusion, system rigidity, and unequal recruitment. Pandemic-era clinical-trial literature showed that diverse participation did not improve automatically under emergency conditions, with continuing under-representation or inconsistent classification of minoritised ethnic groups. Remote methods should therefore be framed as tools for optional flexibility, not as inherently inclusive solutions (34)(35)(36)(37).Fourth, eligibility criteria and outcome measures should be examined for unintended exclusion. Neurodivergent participants may be excluded by rules that require certain styles of communication or compliance that are not essential to the scientific question. For example, requiring participants to complete long questionnaires in one sitting, attend at fixed times without flexibility, or tolerate sensory discomfort for non-essential procedures can reduce participation and distort samples. Where feasible, micro-adjustments such as shorter sessions, splitting assessments, and allowing support persons can preserve the core measurement while reducing participant burden. These are straightforward methodological decisions with clear impacts on recruitment and retention.Fifth, inclusion requires attention to the full participation pathway, not only the initial encounter. Attrition is often treated as a statistical nuisance rather than as a signal of poor accessibility. If neurodivergent participants disproportionately withdraw, it suggests that study burden, communication mismatch, or environment is misaligned. Rather than simply increasing recruitment numbers to compensate, teams should treat attrition patterns as quality indicators. Open and transparent research cultures can strengthen this work: combining participatory approaches with open practices can improve trust, accountability, and reproducibility, while making accessibility lessons easier to document and share (10,21,29). Transparent reporting of adaptations, and patterns should become a rather than an inclusion is when have a in research priorities and work that research that leads to real-world and meaningful outcomes across the Community also that research can be and when it shifts from narrow deficit and to lived experience This that inclusion cannot be from people are more likely to when they in the research and the will research methods have been as flexible yet sometimes in and those are inclusion improves when teams are explicit about the of the of and the and are methodological not optional cannot deliver inclusive methods if they have not been to neurodivergent to without and to distinguish distress by sensory mismatch from distress by the of is also to avoid bias. For example, if a participant or avoids eye contact, this should not automatically be as training, research teams can inadvertently data through of of work across settings the need to methods with a that and autonomy In research this procedures that do not treat neurodivergent as to be but as context that should be and will remain unless it is in structures that and require Ethics expectations, and reporting norms can either exclusion or accessibility. The task is to treat adjustments in research as core safeguards that protect autonomy and improve validity, rather than is the where and accessibility must be consent are frequently long, and rather than participant For neurodivergent those with processing anxiety under time or communication can meaningful on accessible consent and procedures demonstrates that the is not to but to and processes that understanding is more language, formats, explicit of processing and communication and opportunities to decisions over time accessible consent should be framed as for accessibility often across the whole participant population. therefore a neurodivergence access communication options, sensory adjustments and procedures for without the committees and should from risk by to risk through that barriers should be and the protocol rather than through exclusion. This is particularly important because neurodivergent people can be as it the of participation and evidence The to about consent is not exclusion, but a consent process additional processing time, involvement where and opportunities to consent over This consent to the ethical is not whether a participant can a under conditions, but whether the consent process has been to support a inclusion is when and are from early design through and when and are explicit This is not merely a It affects practical decisions such as scheduling, and If neurodivergent participants they should be and research should time for involvement also that involvement must be planned and not to These to neurodivergent if research neurodivergent people to participate, then the must support their participation in a and cannot be by one with It requires and clear This methods for participant and across It also the ability to offer hybrid participation where Evidence from work shows the and of remote and that flexibility and individual matter more than single technology is that studies should remote participation pathways with the same as pathways, including and is a of When funders treat inclusion as an optional teams will it when When funders require explicit inclusion and accessible materials, participant support, and involvement becomes deliverable rather than practical of this is the for ethnic research participant recruitment. during to improve diverse involvement in clinical including in the COVID-19 it infrastructure, and planning rather than on recruitment to neurodivergent inclusion is participation on trust, and system not on excluded to barriers Evidence on research priorities also suggests that research with priorities matters for and participation therefore shape inclusion not only through but through what and outcomes are also on reporting Community involvement reporting and open research practices are as to improve reproducibility, and The same should to accessibility. Research should be expected to in a standardised what adjustments were available, what were how consent and whether across groups. This does not require it requires reporting. such it is to systematically from what and inclusive rather than inclusion must the between research and Research participation is by healthcare Evidence on barriers to healthcare access how communication mismatch, and and experiences can to and adverse outcomes If neurodivergent participants research with experiences of not being or being by recruitment and behaviour must those is not by a consent it is by repeated that the communication is that will be and that participation will not require work also suggests that design matters, and that participants often support structured approaches when they opportunities to communicate and adjustments Research can from the more explicitly studies participants to state what the more and predictable participation and whether inclusion is or accessibility must be into consent, environment, and inclusion should be treated as a scientific and ethical of medical and the of this is to that operational rather than simply The evidence base a coherent research that improves research cultures benefit from participatory accessible consent can be without autonomy, and flexible methods can reduce barriers while data fidelity The challenge is these into protocol and operational challenge should be in neurodiversity that language, poverty, or diagnostic will still produce inclusive The studies will therefore treat neurodivergent inclusion and recruitment as linked responsibilities, the same to and outcome In own research and work, are this by treating accessibility as an early design rather than an This offering written as well as spoken communication where feasible, allowing processing time and breaks, sensory and digital access when planning data and advocacy or in the of research key risk is that inclusion becomes a This is by practical a of accountability, and transparent reporting. Table 1 is not to for It is to predictable and to teams to inclusion decisions early rather than The work is from a deficit that how neurodivergent people can be to to a design that how research can be to human risk is is not a single and adjustments are not one may This is why choice of participation, predictable while allowing flexibility, and treating adjustments as rather than individual without methodological flexibility is not a to The is to in advance what can why it can and how it will be also has When neurodivergent people are the and interventions that interpret neurotypical behaviour as the and neurodivergent behaviour as When neurodivergent people are included research often from is the to do contexts, and produce disability and Interdisciplinary work has highlighted the value of Community that research becomes more when it to context and systems rather than to is also a Research that neurodivergent participants can inadvertently by producing interventions that are less in The between conditions and real-world conditions is a challenge in Neurodiversity this when assume and that are not across Inclusion-by-design therefore improves not only but interventions are more likely to work when they are in populations that those who will for is to inclusion measurable and can require inclusion evidence of involvement and explicit for accessibility work Ethics committees can consent planning and neurodivergence access as safeguards rather than as optional Research teams can treat attrition patterns as quality and can accessibility consent supports and patterns with and can be a core for research staff, communication and reducing bias The resulting not on it on inclusion is a about whose Neurodivergent participants are not merely of they are in their own experience and can the mismatch between systems and When research is to the beyond single The questions, and The evidence that the of is an ethical and a scientific The of medical research should assume neurodivergent inclusion as the and study quality partly by whether participation for those whose the evidence to If research is about and trust, inclusion-by-design should be treated as a planned in in by by and with explicit attention to and transparent advocacy and related to and at least one who is a neurodiversity as a to the and practical in this that can represent the full of neurodivergent particularly across language, and support The are therefore as a to be and with neurodivergent participants and in