Data Sharing in Autism Research: Ethical Tensions and Community Perspectives

Jill Oliver Robinson, MA – Center for Medical Ethics and Health Policy – Baylor College of Medicine; Mary Majumder, JD, PhD – Center for Medical Ethics and Health Policy – Baylor College of Medicine; Robert Cook-Deegan, MD – School for the Future of Innovation in Society – Arizona State University; Amy McGuire, JD, PhD – Center for Medical Ethics and Health Policy – Baylor College of Medicine

Abstract: While data sharing is considered a driver of scientific progress, in autism research, it raises ethical challenges tied to trust, representation, and governance. To explore these challenges, we conducted semi-structured interviews (n= 29) with autistic adults, parents, and researchers. Two themes shaped perceptions of data sharing. First, the framing of autism in biomedical research as a disorder, an identity, or both has the potential to fuel mistrust and shape how stakeholders perceive the risks and benefits of data sharing. Autistic participants acknowledged the medical aspects of autism but expressed concerns about research primarily focused on genetic risk factors and biological deficits. Many feared this pathologizing approach could encourage prenatal screening and selective pregnancy termination, marginalizing autism-affirming perspectives that value autism as intrinsic to identity. One researcher highlighted the inherent “inescapable threat of eugenics” associated with autism datasets, while some parents worried that consenting on behalf of their child could constrain their child’s future autonomy to make decisions aligned with their identity, since data shared and used in a study cannot be withdrawn. Second, participants across stakeholder groups noted that autism research often excludes specific profiles, like non-speaking individuals and those with higher support needs. Researchers identified systemic barriers that restrict participant diversity, including inaccessible study designs, reliance on verbal communication, and restrictive recruitment methods. Such barriers exclude individuals from diverse racial, gender, socioeconomic, and neurological backgrounds, compromising dataset representativeness and quality, ultimately limiting generalizability. These findings can guide data-sharing practices in autism research and disability data governance more broadly.

Keywords: data sharing, disability ethics, autism research