Ethical considerations for conducting research with transgender and gender-diverse patient registries: Balancing our need for evidence with patient vulnerability

Abstract

Most, if not all, medical centers have patient registries for quality improvement/quality assurance projects and patient outcome data, which can also be used to conduct population health studies through secondary research. Such registries offer significant public health benefits, such as facilitating better care, guiding resource allocation, and improving patient outcomes. Because patient registries promise possibilities for collecting patient outcome and longitudinal cohort data, researchers in transgender and gender-diverse (TGD) health have advocated for the creation of patient registries to increase the evidence base for the efficacy of gender-affirming care and population health outcomes. However, there is an absence in the literature on the ethical dimensions of creating and using these registries for research. Existing TGD patient registries were likely developed during the Biden administration, a period of expanding legally protections and social acceptance before the current wave of widespread political attacks on TGD healthcare. Using the same patient registries for research now presents different ethical challenges than when they were initially developed. This discussion provides recommendations for conducting ethical research, ultimately advocating for the creation of TGD patient registries in which participants are fully aware of how their data will be used and stored with robust mechanisms in place to ensure confidentiality and security.

Background

Research conducted using transgender and gender-diverse (TGD) patient registries has contributed to our understandings of the efficacy of gender-affirming care for treating gender dysphoria, best practices for gender-affirming surgery, and social drivers of TGD health, such as unemployment.^1–4 Moreover, these studies have provided valuable insights into biopsychosocial challenges that lead to higher rates of mortality in TGD populations. ⁵ For example, one longitudinal study in Denmark identified that TGD populations have higher rates of all-cause mortality in addition to the well-known disparities associated with suicide. ⁶ This study importantly demonstrates that social drivers and biopsychosocial dimensions of health impact more than just the mental health of transgender, and perhaps the broader TGD communities; these dimensions of health create all-around lower life expectancy. While researchers and clinicians have mostly advocated for TGD patient registries to increase the evidence base for the efficacy of gender-affirming care (including but not limited to gender-affirming hormone therapy and gender-affirming surgery), registries could also be implemented to monitor healthcare maintenance goals, healthcare utilization patterns, mental health outcomes, and preventative care.^7–9

As the evidence base for transgender health outcomes expands, researchers in the field have historically advocated for the development of TGD patient registries to better collect longitudinal cohort data.^10,11 By systematically collecting health data specific to TGD populations, researchers have argued registries could help fill gaps in research where these populations have historically been underrepresented, contributing to more inclusive healthcare policies and practices. This potential is crucial given the documented inequities in mental health and access to care, plus experiences of discrimination faced by TGD people.

Research on TGD communities often faces barriers related to participant recruitment, small sample sizes, or use of a subgroup of data not focused on TGD individuals, challenges which could be reduced with the construction of patient registries, particularly when those registries combine multiple locations. Indeed, O’Bryan et al. emphasize patient registries as one of the most powerful research tools for studying underrepresented and/or vulnerable populations, such as TGD youth. ¹² They also cite creating a national patient registry as one of the top research priorities for transgender health. ¹² Kimberly et al. similarly asserts the need for a comprehensive registry to track physical and mental health outcomes to develop ethically sound gender-affirming care provision for TGD youth. ¹³ As antagonists to gender-affirming care have called the evidence base into question, continuing to strengthen existing research findings with longitudinal cohort data is a research priority in the health of TGD youth. ¹⁴

Most, if not all, medical centers have patient registries for quality improvement/quality assurance (QI/QA) projects and patient outcome data, specific to populations with conditions like breast cancer, diabetes, and heart disease. ¹⁵ Such registries offer significant public health benefits, such as facilitating better care, guiding resource allocation, and improving patient outcomes. ¹⁶ As medical centers typically include a collection of ancillary sites throughout urban centers and into rural centers, they comprise a vast collection of population health data, which could potentially be used to fill the gap of needed longitudinal data on TGD patients and their healthcare. Registries of TGD patients could be created for traditional QI/QA initiatives. In addition, TGD patient specific registries could then be used for secondary research purposes detailed above.

However, establishing patient registries, along with their use for secondary research, raises significant ethical considerations for TGD patients. While these registries are often created within institutions to improve patient experiences or conduct QI/QA analyses, conducting original research using patient registries may make their existence known to external parties, creating potential harm. For example, registries created without patient consent could create a situation where TGD individuals feel that they did not consent to participate in such registries, undermining their perceived autonomy in their health care decision making. They may also fear that their protected health information could be disclosed to state or federal governments. While researchers have advocated for the creation and use of TGD patient registries, there is an absence in the literature on the ethical dimensions of creating and using these registries for research.^10,11 It is also important to note that while this discussion focuses on TGD individuals, the ethical implications also pertain to other vulnerable populations such as intersex individuals. Similarly, while our discussion focuses on legislative and sociopolitical changes within the United States, similar changes are occurring across European countries where TGD patient registries have been used for research. Thus, many of the ethical considerations and recommendations we put forth have broader global applications.

This discussion attempts to put forth an ethics of research using TGD patient registries during a time of political uncertainty and increased vulnerability, addressing risks for using both identifiable and de-identified data. This research could pose more risk during a time in which transgender health is under scrutiny by the federal government and the Trump Administration has made calls for research invalidating gender-affirming care. ¹⁷ Legislation regarding TGD health is currently ever-changing; as such, these recommendations are intended to help researchers reduce harm and navigate ongoing sociopolitical uncertainty, rather than respond to any single executive order or legislative ban.

TGD patient registries

As illustrated, TGD patient registries promise the possibility for monitoring health outcomes in TGD populations and informing the evidence base for best practices in TGD healthcare, gender-affirming care and otherwise. Despite this potential, the process for conducting research using these registries is not well documented. Even within published research utilizing patient registries, it is often unclear how exactly these registries are constructed and used. There are three parts to conducting research with TGD patient registries: sexual orientation and gender identity (SOGI) data collection, data curation, and secondary data analysis. Ethical research conduct informs each aspect of the process.

SOGI data collection

To address the structural barriers to health care faced by TGD communities, health care systems and major institutions have advocated for the collection of SOGI data. The specific data collected varies by institution and the electronic health record (EHR) system used, but SOGI data generally includes sex assigned at birth, gender identity, legal sex, sexual orientation, and organ inventory to accurately understand anatomical reality for cancer screening, STI testing, etc. Patients can complete this information themselves through EHR access in applications such as patient portals or in pre-appointment paperwork, though EHR data updates may also occur during clinical appointments. Just as healthcare systems have shifted from using observation methods toward self-report to more accurately document demographic data (race and ethnicity), clinicians must not update SOGI data based on their assumptions about a patient’s sexuality or gender. The Department of Health and Human Services provided guidance to the Centers for Medicare and Medicaid Services and Center for Disease Control on systematically collecting data on sex assigned at birth, sexual orientation, and gender identity, although SOGI questions were removed from Medicare enrollment forms in February 2025 to comply with Trump Administration executive orders related to sex and gender.^18,19 As few population-based surveys collect SOGI data, utilizing EHR information may offer a valuable source for studying TGD health and related disparities. Moreover, the collection of SOGI data benefits individuals across genders, creating an opportunity to address gaps in knowledge for cisgender, intersex, and TGD individuals alike.

It is worth noting that with recent anti-LGBTQ+ state legislation and executive orders from the federal government focused on TGD individuals, self-reported SOGI data may no longer accurately reflect patients’ lived identities. Many TGD patients, increasingly fearing discrimination or loss of insurance coverage, may be cautious about disclosing their gender in the EHR through either self-disclosure or as reported to their clinician(s) that can be accessed by providers, institutional administrators, and third parties such as insurance companies. TGD populations already experience higher rates of unemployment and uninsurance than their cisgender peers, likely due to employment and insurance discrimination. ²⁰ Demographic data, including SOGI data, also suffers from a high rate of missing information. Therefore, using only SOGI data derived from the EHR does not give an accurate account of all TGD patients, particularly those who are the most vulnerable.

Data curation

SOGI data has been used to identify TGD individuals and is typically the first step in using EHR data to identify TGD patients for QI/QA and secondary research studies. ¹² For systems not including SOGI data, other algorithmic methods based on gender dysphoria diagnosis, gender-affirming hormone therapy prescription, and/or gender-affirming procedures has been used to identify potential research participants. ²¹ Natural language processing and artificial intelligence tools have also aided in development of registries with EHR, but best practices require the data to be cleaned by a researcher or clinician familiar with the patient population to ensure validity. Registries developed for QI/QA historically focused on a shared disease (e.g. cystic fibrosis) while those built with SOGI data are based on the assumption of a shared identity, since not all TGD individuals experience gender dysphoria and intersex individuals who do not identify as TGD may be included. ¹⁵ For instance, registries to examine gender-affirming surgery outcomes may require a TGD identity marker or a gender dysphoria diagnosis for the study population, but the registry is constructed around a surgical intervention. Additionally, a registry constructed around gender-affirming care as the medical intervention may not include all TGD patients, as not all TGD patients choose to receive gender-affirming care, such as gender-affirming hormone therapy and surgery.

Secondary data analysis

 Research conducted with information collected for a different primary purpose—in this case, clinical care, are considered secondary data analyses, including those conducted with existing patient registries. Observational studies using secondary data analysis, such as a cross-sectional study utilizing a clinical dataset at a single timepoint, have long been considered the standard for patient outcomes research as it captures the totality of a population while minimizing selection and recruitment bias. Studies based on secondary data have been considered optimal for TGD populations because they not only allow researchers to compare TGD health outcomes to their cisgender counterparts when appropriate, but they additionally reduce potential study bias. Among vulnerable populations with a history of medical mistrust, studies that rely on participant recruitment can introduce substantial selection bias—patients with established, trusting relationships with providers are more likely to enroll, potentially skewing results and excluding the experiences of patients without established or affirming care. Patient populations disproportionately affected by social determinants of health are also less likely to enroll in research studies due to financial and structural barriers (i.e. not being able to miss work or not having access to transportation). ²² Researchers have advocated for the formation of TGD patient registries with which to conduct observational studies partly so that marginalized members of TGD communities, often those with intersectional identities, are not excluded from health outcomes research. ³

When submitted to an Institutional Review Board for compliance with research standards, secondary research is generally labeled as exempt from ongoing Institutional Review Board oversight, as it is considered low risk to the study population. If the study could be completed with the use of a fully de-identified dataset from the EHR, the Institutional Review Board would determine the research is not be legally defined as human subjects research, as it does not involve interaction with human subjects or private identifiable information. ²³ If the study team directly accesses a patient registry (or otherwise receives identifiable data) then it is often eligible for “expedited review” with a waiver of participant consent, because it is treated as posing a “minimal risk” to the individual data subjects. Thus, despite the fact that secondary research with TGD patients’ EHR may engage with vulnerable populations’ data, it is often not subject to a full ethics review. As a result, research teams do not benefit from the same expertise of institutional review boards that are provided to other vulnerable populations (e.g. incarcerated individuals) and hence deciding how to ethically conduct research with TGD patient registries is the sole responsibility of the research team. Researchers working with TGD patient registries may not recognize unforeseen risks, such as exacerbating mistrust with healthcare systems among TGD individuals and potentially greater chances of re-identification.

TGD registry data risks

Malicious actors

Many TGD patients fear even being identified as TGD in their EHR, fears that have been exacerbated by medical centers disclosing patient records to anti-LGBTQ+ state governments. ²⁴ For example, Seattle Children’s Hospital was pressured by the Attorney General of Texas to overturn records for patients receiving gender-affirming care, but Seattle Children’s ultimately blocked the release of this information to protect their patients. ²⁵ While Seattle Children’s sued the Texas Office of the Attorney General to prevent disclosure of their patient records, Vanderbilt University did not do the same. Vanderbilt University Medical Center in Tennessee complied with requests for transgender patient records from state attorney generals in Tennessee, Indiana, Missouri, and Texas by turning over 65,000 pages of data on transgender patients, as well as information on providers offering gender-affirming care. ²⁶

In situations like these, TGD patients may feel that the safety of their patient data depends on the bravery and goodwill of the institution housing it. While neither the Vanderbilt nor Seattle Children’s cases explicitly included requests for patient registries, the looming threat of state attorney generals requesting patient and/or research records would likely make any TGD patient wary of disclosing their gender identity in EHR and certainly worry them about hospital and/or researcher constructions of TGD patient registries. When researchers conduct studies with patient registry data, it creates additional data sets that could be accessed by malicious actors. More generally, TGD individuals may fear that their data is vulnerable to hacking or misuse by actors who mean harm to TGD communities.

Visibility of a registry

While patient registries often exist in medical centers for QI/QA tracking, the public and the patients included in them are generally unaware of their existence. Conducting and publishing research with a patient registry makes its existence visible. Because TGD populations have been made more visible to the public because of anti-LGBTQ+ legislation, the visibility of TGD patient registries now looks different than when they were created. Existing TGD patient registries were likely developed during the Biden administration, a period of expanding legally protections and social acceptance before the current wave of widespread political attacks on TGD healthcare. Using the same patient registries for research now presents different ethical challenges than when they were initially developed. This difference puts a significant ethical onus on the research and healthcare institutions to anticipate and respond to shifting political dynamics. When TGD patients have not been consented for their inclusion in a hospital-based QI/QA registry that is then used for secondary data analysis, learning of its existence through publication of research findings may undermine their trust in the medical system.

Risk of re-identification

De-identified data is not necessarily anonymized data when working with unique and small populations like TGD patients; true de-identification of data can be very difficult.^23,27 While the removal of direct identifiers may render research low risk under standard human subjects research guidelines, this may be insufficient for TGD individuals. For instance, in a given medical center there may be only a few Latina transgender women receiving vaginoplasty. Creating a registry of transgender women who received this gender-affirming surgery and reporting outcomes by racial demographics could inadvertently enable re-identification of individual patients, thereby compromising confidentiality. Common methods protecting people represented by small sample sizes, such as using <5 or <10 in cell sizes, for both institutional and public facing reports may need to be expanded for outcomes from TGD registries.

Recommendations

Community stakeholder engagement

Structural barriers, along with medical mistrust, limit TGD communities’ access to healthcare services, and their willingness and ability to participate in research requiring their active engagement. Because population health research on TGD communities raises a social risk for stigmatizing and/or pathologizing gender identity, research done using TGD patient registries should include input from members of the community, acting as a proxy for community consent. Community consent, whether through a full opt-in model based in community-based participatory research principles or a community advisory board model, is essential for ethical research conduct.^28,29 In order to mitigate possible harms, community members, whether actual patients included in the registry or representative community members in an advisory board, should help guide research questions to ensure research addresses community priorities with culturally relevant methods.

This approach may also include partnering with community organizations, Federally Qualified Health Centers, and providers who serve TGD patients to increase health engagement and research participation. The All of Us program’s strategic community outreach to oversample marginalized, vulnerable populations provides an apt model for building community capacity and investment in research. ³⁰ TGD advisors should be compensated for their time and expertise, educated on the risks, benefits, and ethical considerations of the research process, and engaged in building methods for participant recruitment and data collection. ²⁹

Community feedback is also crucial and essential in shaping how data is interpreted and disaggregated in research results. Community members are experts on their communities and typically have a better understanding than researchers of community vulnerabilities that should be considered in data interpretation and presentation. Their feedback thereby provides harm reduction for risks associated with stigma, pathologization, and re-identification. Including community stakeholders, moreover, can help build rather than erode trust between medical institutions and TGD patients.

Obtaining consent

The Gender Wellness Center (GWC) Pediatric Patient Registry was established by clinician researchers in 2017 to allow them to conduct longitudinal, systematic research on physical, mental, and quality-of-life health outcomes for TGD youth. ¹² It was constructed to answer 10 primary research questions on topics such as prevalence of self-harming and disordered eating behaviors, adverse effects related to gender-affirming treatments, and mental health outcomes while receiving gender-affirming care. As an ideal example of obtaining informed consent for registry inclusion, the GWC Pediatric Patient Registry used rigorous informed consent procedures with youth assent and parental consent to recruit patient participants. It countered typical approaches to registry construction by implementing a transparent consent process with voluntary participation. For example, participants were importantly re-consented once reaching the age of 18.

The GWC Pediatric Patient Registry participants were recruited by their clinician during a routine clinical visit. The recruitment process included an explanation of the registry and its research goals with interested patients then given the opportunity to meet with a research coordinator to learn about research participation, risks, benefits, and any other aspects of the study. This process led to a 98% recruitment rate, demonstrating the importance of research education when engaging TGD communities in the creation of a registry. ¹²

A fully-consented registry is not without its own ethical considerations. Although patients may initially consent to their data being used for research, a shifting political climate could later prompt TGD individuals to request the removal of their data from registries. They may also want their data to be removed due to life changes or other personal reasons, such as shift in their gender identity (e.g. identity changes from binary transgender to genderfluid). Such a request creates challenges insofar as choosing to withdraw will still require maintaining links between the patient and past registry data. Ideally, any fully consented registry informed consent documents include details to ensure that participants fully understand that withdrawal may not be as simple as removing their name from a registry.

During a time of immense political uncertainty and vulnerability for TGD people, a fully consented registry prioritizes psychological safety and patient autonomy. ³¹ Paired with community outreach and engagement, this approach may promise a more accurate account of TGD patients through collection of SOGI data aligned with patients lived experiences. Conversely, without consent, transparency, and awareness of how their data may be used, TGD patients are likely to delete and/or change their SOGI data to re-align with their sex assigned at birth. Missing data tends to increase when patients do not feel safe. When combined with community input, a fully-consented registry has the potential to create research that empowers its TGD participants through their active participation in research that benefits their communities and is thus our strongest recommendation.

The data used to develop TGD patient registries was more trustworthy when patients felt safe during a period of greater social acceptability and adoption of good faith practices like the collection of SOGI data. Moreover, perhaps the best data should not be the priority when TGD patients feel imminent threats to their safety and healthcare access. Transparent conversations about the benefits and potential risks of research increase patient autonomy and demonstrate respect for the use of their personal data. A registry constructed in this way promises greater feelings of patient safety and accuracy of the data collected.

Transparency

Community stakeholder engagement and processes of informed consent increase research transparency and as a result, trustworthiness for both researchers and medical centers. Additionally, researchers could work with medical centers to provide patients with education on the benefits of QI/QA research. Researchers could also collaborate with medical information systems to create greater transparency for the use of their data within EHR. For example, the EHR demographics form could include an option for patients to opt out of the use of their data for QI/QA purposes.

Other protections

In all cases, data protection safeguards, secure data management practices, and ethical researcher conduct are critical. Researchers should be sure that their institution has secure data protections for patient registries before conducting research with them, including access control, encryption, and firewall protections to mitigate possible hacking. ³² While Zayhowski et al. developed their privacy protection recommendations for SOGI data, many of them are also relevant for TGD patient registries, particularly the following ²⁴ :

1. Enhanced confidentiality: TGD registry data should be protected by stricter institutional protocols for confidentiality. Data encryption, limiting access, and review of who uses this information can prevent misuse by bad actors.

2. Explicit consent: TGD patients should have say over which aspects of their data are shared with researchers.

3. Training and awareness: Institutions should develop training for both researchers and information system administrators on the respectful and ethical use of TGD patient registry data.

Lastly, researchers should be careful about how they report research results so that individuals cannot be re-identified through participant characteristics like demographic markers. Specifically, when research results are published researchers should consider how results are disaggregated to make sure that intersecting demographic information (race, age, gender, etc.) could be triangulated to re-identify marginalized patients. Oftentimes for small sample sizes and the people representing these groups are protected by using <5 or <10 in table cells.

Conclusion

TGD patient vulnerability due to the current sociopolitical climate raises significant ethical concerns for conducting research with patient registries. Yet, as the federal administration calls into question the scientific value of collecting gender data (as opposed to only sex assigned at birth) through a flurry of executive orders and the cancelation of federally funded research on LGBTQ+ populations, their demand for evidence for the efficacy of gender-affirming care intensifies. ³³ These tensions create an imperative for TGD health researchers to continue growing the evidence base of research that both affirms the value of studying gender and adds to the evidence base for gender-affirming care efficacy, particularly for youth populations, all while protecting the sensitive protected health information of their participants.

Ethically conducted and sufficiently anonymized secondary research could present one such solution, especially research including fully-consented individuals. As the government ends federal funding for this research, medical center patient registries offer an important resource for secondary data analysis. Rather than suggesting research with TGD patient registries is too ethically fraught to continue, this article hopes to provide recommendations for creating a safer research environment for TGD patients through consent considerations, privacy protections, and ethical data analysis. We ultimately hope for the possibility of TGD patient registries in which participants are fully aware of how their data will be used and stored with robust mechanisms in place to ensure confidentiality and security.