The Right to Interpret

In March 1987, Larry Kramer addressed a meeting at the Gay and Lesbian Community Services Center in New York. Two days later, three hundred people formed the AIDS Coalition to Unleash Power. By 1988, ACT UP had become the most consequential patient activist organization in American history.

ACT UP's interpretive claim was specific. The patients dying of AIDS, the people who loved them, and the clinicians treating them understood what was happening with the epidemic more accurately than the FDA approval process, the NIH research priorities, and the Reagan administration. The interpretive claim was not framed as a complaint about access. It was framed as a competence claim. The patient population had the more accurate read of the situation. The institutional read was wrong, and the wrongness was killing people.

The October 1988 FDA action was the consequential moment. The agency that had taken eight to ten years to approve drugs of AZT's class, that had refused to release other promising drugs because they produced side effects patients were willing to accept, that had structured its clinical trials around exclusion criteria that disqualified most AIDS patients, faced a coordinated occupation of its headquarters by people whose accumulated knowledge of their own disease the agency had not seen fit to consult. Anthony Fauci, who became director of NIAID in 1984 and was for years the target of ACT UP's most pointed criticism, eventually credited Kramer and the broader movement with transforming the relationship between activists and the federal research apparatus.

The institutional changes that followed are documented. The FDA established the accelerated approval pathway in 1992. The NIH adopted formal patient consultation in trial design. The treatment IND program expanded compassionate use access. The AIDS movement's interpretive claim, made by patients about their own bodies and their own deaths, was eventually integrated into the regulatory and research architecture that had refused to consult them.

The gallery's curatorial position is that ACT UP is the modern-era founding artifact of organized patient interpretive authority. The wing's later exhibits all, in one way or another, descend from the operation ACT UP demonstrated: that a patient population could, by organizing, force an institution to recognize what the patient population already knew.

Susan Love's Breast Book was published in 1990. The book treated women diagnosed with breast cancer as adult readers capable of understanding their own disease. By 1991, Love had helped found the National Breast Cancer Coalition. The Coalition's organizing model was explicitly adapted from ACT UP's playbook, with the political register adjusted for a different patient population and a different cancer.

The NBCC's interpretive claim followed the ACT UP precedent. Women diagnosed with breast cancer, and the women around them who had watched the disease, understood the experience and the stakes more accurately than the research apparatus that was, at the time, allocating breast cancer research at a fraction of the rate the burden of the disease warranted. The Coalition organized to change the allocation directly. By 1993, the Coalition's lobbying had produced congressional appropriation of more than one billion dollars in breast cancer research funding routed through the Department of Defense's medical research program.

The Coalition's other consequential move was educational. By 1995, NBCC's Project LEAD program was training breast cancer survivors as scientific advocates. The trainees learned the molecular biology, the trial design language, the regulatory framework, and the funding architecture. They then served on grant review panels at the NIH, on FDA advisory committees, on hospital institutional review boards, and on the steering committees of major clinical trials. The Coalition's argument was structural. If patient experience was going to be consulted in research priorities, the patients consulting needed to be fluent in the language of the institutions consulting them. The Coalition built the fluency.

The gallery's curatorial position is that NBCC documents a second move in the patient-interpretive operation. ACT UP forced institutional recognition through confrontation. NBCC consolidated the recognition through expertise. The combination, confrontation that produces a seat at the table, and expertise that holds the seat once it has been won, becomes the playbook the subsequent patient movements inherit. Love died in 2023. The NBCC continues, with the Project LEAD program in its fourth decade.

The Association of Cancer Online Resources was founded in 1995. The trigger was a single patient's experience. The founder's wife had had a cardiac event during pregnancy. The clinical encounters that followed convinced him that patients with specific conditions needed direct communication with other patients with the same condition, unmediated by institutional gatekeeping. The mailing list architecture of the early internet made this possible at scale for the first time.

By 2000, ACOR was running more than one hundred and fifty disease-specific mailing lists with tens of thousands of subscribers worldwide. By 2010, hundreds of thousands of cancer patients had passed through the lists. The lists were patient-moderated. The subscribers exchanged clinical experience, treatment outcomes, drug side effects, surgical recovery details, and emotional support, across a population that no single hospital, no single medical society, and no single research institution had ever previously assembled in one conversational space.

The institutional response was, initially, suspicion. The medical literature of the late 1990s questioned whether patient-to-patient communication was safe, whether it would propagate misinformation, whether it would interfere with the physician-patient relationship. By the mid-2000s, the literature had begun to update. Studies showed that ACOR participants were more, not less, adherent to clinical recommendations; were more likely to participate in clinical trials; were more likely to seek second opinions when warranted; and were more likely to understand their own diagnoses with the precision the clinical record required. The patient community was, by every measure the literature could devise, improving clinical outcomes for the patients who participated in it.

The gallery's curatorial position is that ACOR documents the moment patient interpretive authority became routable across geography. Before 1995, a patient's interpretive community was the patients she could meet in person, in waiting rooms, in support groups, in hospital corridors. After 1995, her interpretive community was, in principle, every patient on earth with her diagnosis. The shift is the central enabling fact of every patient movement that follows in this gallery.

The curator of this museum is the founder of ACOR. The acquisitions colophon and the museum's about page disclose this. The wing displays ACOR as artifact, not as personal accomplishment. The 1995 archive of the mailing lists, much of which has been preserved, is the museum's largest single body of primary-source documentation of patient interpretive activity in the early internet era.

The slogan has a layered history. The Latin form, Nihil de nobis, sine nobis, appears in the 1505 Polish Nihil novi constitution. The English form entered the disability rights movement in the early 1990s, traveled from Eastern European activists to South African disability activists Michael Masutha and William Rowland, and was popularized by James Charlton's 1998 book of the same name. By the mid-1990s, the slogan had become the central principle of organized disability self-advocacy.

The autistic self-advocacy movement, organized in formal terms beginning with the founding of the Autistic Self-Advocacy Network in 2006, inherited the slogan and made it operational for a population whose interpretive authority the broader disability rights movement had not always fully accepted. The argument was specific. Autistic people understood their own neurology better than the non-autistic researchers, parents, clinicians, and educators who had until then dominated the institutional conversation about autism. The standard cognitive-behavioral framework, which had described autism in terms of empathy deficits and theory-of-mind impairments, was inverted by the autistic self-advocacy literature: the deficit, the self-advocates argued, was located in the empathic capacity of non-autistic people attempting to read autistic communication, not in the autistic people whose communication the non-autistic researchers had been unable to read.

The argument was, for a decade, treated as activist position rather than research finding. The 2012 Milton paper documented in Gallery II changed that status. The autistic self-advocacy movement's interpretive claim about itself, that the impairment-of-empathy framework had the direction of the impairment backwards, entered the peer-reviewed literature and is now the dominant framework in serious autism research two decades after the self-advocates first articulated it.

The gallery's curatorial position is that the autistic self-advocacy movement is the wing's clearest case of a patient population correctly diagnosing the structural error of the clinical literature about itself, decades before the literature accepted the diagnosis. The institution did not catch up by reading the literature. The institution caught up by reading the patients, who had been writing the diagnosis down all along.

Dana Lewis has Type 1 diabetes. In 2013, the commercially available continuous glucose monitor she relied on had an alarm too quiet to reliably wake her when her blood sugar dropped overnight. She contacted the manufacturers. The manufacturers responded that the alarm was loud enough for most people. Lewis built a louder one.

The louder alarm became #DIYPS. The #DIYPS system became a personalized predictive algorithm that forecasted blood sugar trajectories hours in advance. The predictive algorithm became, by December 2014, a hybrid closed loop: an artificial pancreas built from off-the-shelf hardware, open-source code, and Lewis's existing insulin pump. The system did the work of a pancreas. It read continuous glucose monitor data, calculated the appropriate insulin response, and dosed automatically. It worked. Lewis and Scott Leibrand then made the system open source, under the hashtag #WeAreNotWaiting, so that other people with Type 1 diabetes could build the same system themselves.

By 2016, the OpenAPS community had logged more than 250,000 hours of real-world use across more than one hundred self-built closed-loop systems worldwide. By 2018, the community had presented poster sessions at the American Diabetes Association Scientific Sessions, peer-reviewed letters in the Journal of Diabetes Science and Technology, and case reports in mainstream diabetes endocrinology venues. In September 2022, the New England Journal of Medicine published a randomized controlled trial of an open-source automated insulin delivery system using the OpenAPS algorithm. The trial's results were on par with those of the commercial closed-loop systems the FDA had been approving in parallel.

The gallery's curatorial position is that OpenAPS is the consequential 21st-century artifact in the wing. Previous patient communities had organized to demand that institutions act. The OpenAPS community organized to build what the institutions had not yet built and to share the build with anyone who needed it. The interpretive claim was no longer that patients understood their condition better than the institution. The claim was that patients could, when necessary, build the clinical technology themselves, and that the institution would, if the build was good enough, eventually validate it. Both elements of the claim turned out to be correct.

The wing's closing exhibit is the movement currently in progress at the time the wing opens. The hashtag #PatientsUseAI began appearing in patient-community discussions in 2024. By 2026, the hashtag had become a recognizable shorthand for the practice of patients using large-language-model AI tools to interpret their own medical situations, to research conditions and treatments, to translate clinical language, and to prepare for clinical encounters with the institutional vocabulary in hand.

The institutional response has been mixed and is documented at length in Wing IV. The American Medical Association passed resolutions discouraging patient AI use. State legislatures have introduced bills to restrict it. Clinicians have published op-eds warning of the practice. The patient communities have, predictably, continued the practice. The pattern is the same one ACT UP, NBCC, ACOR, the autistic self-advocacy movement, and the OpenAPS community demonstrated before: the institution warns against a practice the patient population has already determined improves their own outcomes, and the patient population continues the practice while documenting the outcomes.

What distinguishes #PatientsUseAI from its predecessor movements is that the underlying interpretive technology is, for the first time, broadly available to any patient with internet access, without specialized training, without community membership, and without organizational mediation. The interpretive labor that ACOR distributed across moderated mailing lists, that the autistic self-advocacy movement distributed across academic writing and conference papers, and that OpenAPS distributed across GitHub repositories is now, in 2026, also being distributed across individual chat interfaces, each one functioning as a personal research assistant, a translator, and a second opinion that the patient consults privately before, during, and after the clinical encounter.

The gallery's curatorial position is that #PatientsUseAI is the present, not the future. The movement is happening now. The institutional response is now. The wing displays the movement as a continuation of the forty-year arc the previous five exhibits documented. The wing closes with a 2026 movement because the wing's argument is that the arc is unfinished, that the patient communities are still organizing, and that the museum's structural answer to the four wings of dismissal is not a historical event but an active, ongoing practice the visitor can join the day she leaves the museum.

Gallery I documented the slogan as a community rallying cry. This exhibit documents the slogan as a research methodology. The transition was deliberate and took roughly two decades. The patient-community principle that decisions affecting disabled people must include disabled people in decision-making roles became, over time, a peer-reviewed methodological criterion that funders, journals, and IRBs increasingly require.

The methodological adaptation appears in several literatures. In patient-reported outcomes research, the principle is operationalized as the requirement that PRO measures be developed with patients rather than for them. In rare disease research, the principle has become standard practice in natural history studies, which now routinely include patient-organization co-investigators. In disability technology research, the principle has produced explicit recruitment of disabled researchers, disabled co-authors, and disabled IRB members. The 2022 Convention on the Rights of Persons with Disabilities cites the principle directly in its preamble.

The methodological consequence is concrete. A research design that does not include the patient population's interpretive authority in the design phase, the data collection phase, the analysis phase, and the publication phase produces, predictably, research that the patient population cannot read, cannot act on, and cannot benefit from. The patient community's interpretive authority is not an addition to the research. It is a constituent of the research's validity. The methodological literature has, slowly, accepted this point.

The gallery's curatorial position is that the slogan's career documents the standard trajectory of patient-community methodology. The principle is articulated in the community. The community organizes around it. The community demonstrates its application in projects the institutional literature initially treats as activism. The institutional literature, decades later, adopts the principle as methodology and cites the activist literature in passing.

Damian Milton is autistic. In 2012, he published a five-page paper in the journal Disability & Society that did something the autism research literature had not done in the previous half-century: it located the empathy deficit on both sides of the autistic / non-autistic interaction rather than only on the autistic side.

The paper's argument was simple and structural. The cognitive-behavioral literature on autism had, since the 1980s, framed autism as a deficit in theory of mind, in empathy, and in social cognition. The autistic person, in this framework, was the person who failed to read the non-autistic person's mental states. Milton noted that the framework had never been tested in the other direction. When the framework was tested in the other direction, the non-autistic person turned out to be equally unable to read the autistic person's mental states. The empathy deficit was not a property of autism. It was a property of the interaction between two neurologically different populations attempting to read each other.

The paper's argument was, at the time, treated as a marginal critique by an autistic researcher with a methodological grievance. By the late 2010s, empirical studies had begun to test the double-empathy hypothesis directly. Autistic adults communicated more successfully with other autistic adults than with non-autistic adults. Non-autistic adults communicated more successfully with other non-autistic adults than with autistic adults. The "deficit" was, in measured terms, exactly the bidirectional mismatch Milton had predicted in 2012.

The methodological consequence is large. The half-century of autism research built on the empathy-deficit framework had been measuring autistic communication against a non-autistic baseline. The choice of baseline had determined the finding. With a different baseline, the finding inverted. The research had not been wrong about the data. The research had been wrong about the question. The patient population had been articulating the correct question, in some form, since the autistic self-advocacy movement organized in the 1990s. The clinical literature accepted the question after Milton, an autistic researcher, published it in a peer-reviewed journal.

The gallery's curatorial position is that the Milton paper is the wing's clearest single example of patient-community knowledge becoming methodology with receipts. The autistic self-advocacy movement had been articulating the double-empathy claim for two decades. The clinical literature did not accept the claim until an autistic researcher articulated it in the institutional vocabulary, in a journal the institutional literature reads. The vocabulary mattered. The clinical literature could have read the autistic self-advocacy literature directly at any point since the 1990s. The fact that it did not is the artifact this gallery's coda will name directly.

A natural history study documents how a disease progresses without intervention. For rare diseases, where the patient population is small and the academic clinical centers each see only a handful of patients, natural history is the necessary precondition for any meaningful clinical research. Without natural history, the experimental therapy has nothing to be compared against.

Rare disease patient communities discovered, beginning in the 1990s and accelerating through the 2000s and 2010s, that they were the only organizations with access to enough patients to assemble a serious natural history study. The academic medical centers each saw too few. The pharmaceutical sponsors had no economic incentive to fund the studies until a therapy was already in development. The patient communities, often organized as foundations around a single rare condition, began running the natural history studies themselves, frequently as the foundation's central research program.

The methodological literature has, in the past decade, formally recognized the practice. The FDA's 2019 guidance on natural history studies for rare diseases explicitly identifies patient-organization-led studies as a valid source of regulatory evidence. The 21st Century Cures Act of 2016 included provisions specifically enabling patient-collected data to inform FDA approval decisions. Several rare disease therapies approved since 2018 have used patient-organization natural history data as the comparator arm in the clinical trial that established efficacy.

The methodological principle this exhibit displays is not novel in the abstract. It is novel in the specific. Patients organized around a condition that the institutional research apparatus had not prioritized produced the evidence base that subsequent regulatory decisions required. The institutional research apparatus then used the evidence base the patients had produced. The patient communities were not auxiliaries to the research. They were the research's primary investigators, with the conditions, the data, and the analytical capacity all held by the patient community before any academic or industry investigator entered the picture.

The gallery's curatorial position is that this exhibit is the wing's clearest demonstration of patient interpretive authority producing peer-reviewed and regulator-acknowledged evidence at scale. The receipts are concrete. Therapies the FDA approved since 2018 carry, on their labels, the imprint of patient-organization-led natural history studies that no academic center could have run.

The methodological shift this exhibit documents is small in its formulation and enormous in its consequences. A patient-reported outcome is, by definition, an outcome the patient reports directly. The clinician does not interpret the report. The clinician does not translate the report into clinical vocabulary. The clinician does not rate the report's plausibility. The patient's report is the data.

For most of the twentieth century, this was not how medicine worked. The clinician's observation was the data. The patient's report was the input the clinician used to produce the clinical observation. The clinician's translation of the patient into the medical record was the operation Wing I of this wing's lobby would call the standard procedure: the patient's interpretive authority over her own body was, for the duration of the clinical encounter, ceded to the clinician.

The PRO methodology inverts this. The patient's report is, by methodological design, the primary evidence. The clinician's role is not to translate the patient's report but to record it. The PRO instruments, which patient communities helped design beginning in the 1990s, are validated tools that produce data the regulatory apparatus accepts as primary outcomes in clinical trials. The 2009 FDA guidance made this status explicit. PROs are now, in many therapeutic areas, the primary outcome measures the FDA uses to approve new therapies.

The methodological consequence is direct. A clinical trial in which the primary endpoint is a PRO is a trial in which the patient's interpretation of her own condition is the basis of the regulatory decision. The clinician's interpretation is, methodologically, secondary. This is the inversion the wing's other galleries have argued for: the patient's interpretive authority is the primary evidence, and the clinical interpretation is the secondary translation. The 2009 FDA guidance, and the subsequent methodological literature, has accepted the inversion in a meaningful and expanding subset of the regulatory record.

The gallery's curatorial position is that PROs are the methodological move the wing's previous galleries each pointed toward. The patient's account of her own body, treated as primary evidence with regulatory weight, is the structural inversion of the operation the museum's other wings document.

Software is, in one register, a tool. In another register, software is a written argument. The argument is executable. It does the thing it claims will work. If the thing does not work, the software does not function, and the argument is falsified at runtime. This is, in the abstract, the cleanest form of methodological statement a research community can make: a claim that compiles, runs, and produces measurable outputs against the world.

OpenAPS, documented in Gallery I, is a methodological statement in code form. The patient community's claim was that a closed-loop artificial pancreas could be built from off-the-shelf hardware, open-source code, and existing diabetes devices, and that the resulting system would safely manage Type 1 diabetes for the people who built it. The claim ran. It compiled. It produced measurable outputs. The 250,000 real-world hours documented in the 2016 letter to the editor were not described in the institutional literature as a study, because the methodological literature had not yet developed a category for software the patients built and ran on themselves. The 2022 NEJM trial of the OpenAPS algorithm, which produced regulatory-grade evidence, was the institutional acknowledgment that the code had been doing science the whole time.

The methodological category this exhibit names does not yet have a settled term in the methodological literature. Variants in circulation include patient-driven device innovation, do-it-yourself medical technology research, citizen science in medical devices, and lived-experience engineering research. Each of these terms captures part of the practice. The wing's curatorial position is that the practice is, at its core, methodology delivered in code rather than in prose, and that the methodological literature will need to develop the category fully because the practice is now occurring in a growing range of conditions: glucose management, sleep tracking, chronic pain monitoring, mood tracking, post-surgical recovery, and the parallel emerging practice of patient-built AI tooling that #PatientsUseAI documents.

The gallery's curatorial position is that code can carry an argument and that patient communities have been carrying arguments in code for more than a decade. The institutional methodological literature has begun to acknowledge this. The wing displays the acknowledgment as a leading indicator of where the broader research methodology is headed.

A hashtag is, in itself, a methodological move. By 2024, a hashtag had become the standard way patient communities marked their own contributions to the literature on a topic the institutional literature had not yet caught up to. #WeAreNotWaiting marked the patient-built diabetes technology corpus. #ActuallyAutistic marked the autistic self-advocacy corpus. #MedTwitter marked the institutional response. #PatientsUseAI marks the present moment.

The methodological significance of the hashtag is twofold. First, the hashtag is a corpus marker. A researcher in 2030 wanting to study how patients used AI tools in 2026 will be able to assemble a corpus by searching the hashtag across the platforms it appeared on. The corpus will not be complete, will not be statistically clean, and will not be neutral, but it will be available, and it will document a practice the institutional research apparatus did not document at the time it was happening. Second, the hashtag is a community method. Patients use the hashtag to find each other, to compare experiences, to refine the practice, to share prompts that work, to warn against prompts that produce dangerous outputs, and to develop the collective practical wisdom that no single patient could develop alone.

The methodological position embedded in the hashtag is the wing's clearest contemporary statement of patient interpretive authority. The patients are not waiting for the institutional literature to catch up. They are documenting the practice as they go, in the public record, in language the institutional literature can read once the institutional literature has decided to read it. The documentation is happening now, on platforms that may not survive the next decade, in conversations that may or may not be archived, but the documentation is happening. The wing's exhibits in Gallery III will return to specific instances of this documentation against the institutional pushback that Wing IV catalogs.

The gallery's curatorial position is that the hashtag is methodology in its earliest organized form. By the time the academic literature begins to study what #PatientsUseAI documented, the practice will have moved on. The hashtag will, at that point, function as the historical-record marker that allowed the academic literature to know what the practice was when it began. The wing's archive does not include the hashtag's corpus, but the wing's argument does, because the corpus is the present-day evidence that the wing's larger claim is correct: the patient communities are producing methodological knowledge faster than the institutional literature can catalog it.

Wing II documents that the Hoffmann and Tarzian paper is, by a wide margin, the most cited and most thoroughly ignored single paper in the women's pain literature. The paper has been read. The paper has been agreed with. The paper's findings have not been retired. This gallery re-reads the paper against the patient-community response.

The patient communities of women in pain did not need Hoffmann and Tarzian to tell them about the dismissal pattern. The communities had been describing the pattern, in patient-to-patient communication, for decades before 2001. The contribution of the Hoffmann and Tarzian paper was not the diagnosis. The contribution was the citation. The paper made the diagnosis citable in the institutional literature. The communities had been making the diagnosis in the non-citable register the institutional literature did not read.

What has changed since 2001 is not the underlying pattern. The pattern persists. What has changed is the patient-community apparatus for documenting it. The medical gaslighting hashtags, the patient-led survey research, the institutional advocacy organizations (the Endometriosis Foundation, the Society for Women's Health Research, Tribe XX Lupus, the National Pain Advocacy Center), the patient-organization-led natural history studies referenced in Gallery II, and the increasingly direct patient testimony in the patient-facing health media are all forms of documentation the patient communities are now producing about the pattern Hoffmann and Tarzian named in 2001.

The stakes are concrete. Women with endometriosis wait an average of seven years for diagnosis. Black women die in childbirth at three times the rate of white women. The 94% medical gaslighting figure from Wing II's Exhibit 2.6 was measured in 2024. Each statistic is, in part, a measurement of what happened in the twenty-four years between Hoffmann and Tarzian's diagnosis and the present, during which the institutional response was, by the paper's own subsequent telling, modest. The patient communities have been documenting the cost the whole time. The cost is now visible in the institutional literature because the patient communities have built the documentation apparatus to make it visible.

Wing IV documents S7263 as one of several pieces of state legislation that, in 2024 and 2025, attempted to restrict patient use of AI tools. The pattern is recognizable. The institution has identified a practice the patient population is engaging in, has determined that the practice undermines the institution's traditional authority, and has moved to restrict the practice through legislative means before the institution has built the empirical case that the practice produces worse outcomes than the alternative.

The patient-community response is the #PatientsUseAI movement documented in Gallery I of this wing. The movement's position is that patients are already using AI tools, that the patients using them have determined the tools improve their understanding of their own conditions, that the empirical case for restriction does not exist, and that the institutional impulse to restrict the practice repeats a pattern the institution has previously deployed against patient access to medical information itself.

The relevant precedents are documented in Wing IV. The pattern of warning patients away from medical information was named, in the 1990s, with the coinage of cyberchondria. The pattern of warning patients away from genetic information was named, in 2013, with the FDA's cease-and-desist against 23andMe. The pattern of warning patients away from AI tools is named, in 2025, with bills like S7263. Each iteration has, in its time, been articulated by clinicians acting in good faith on a model of the patient-clinician relationship that assigns interpretive authority to the clinician. Each iteration has, in its time, been overtaken by the patient population's actual practice, which has continued regardless of the institutional warnings.

The stakes are specific. If S7263 is enacted in its current form, patients in New York will face a legal landscape in which AI tools that patients in California, Texas, and elsewhere use routinely are restricted. The restriction will not stop the practice. Patients will use the tools across state lines, on personal devices, with VPNs, in conversations the state cannot reach. The restriction will, however, create a class of users for whom the tools are legally complicated to use, which class will skew toward patients with fewer resources, less digital literacy, and the most acute need for the interpretive labor the tools perform. The institutional impulse, in its history, has tended to produce restrictions that fall hardest on the patients with the least capacity to circumvent them.

Wing IV documents the 23andMe cease-and-desist as a single, specific moment in the longer pattern of FDA action against patient access to information about the patient's own body. This gallery re-reads the moment against the patient-community response.

The 23andMe cease-and-desist did not stop direct-to-consumer genetic testing. The cease-and-desist produced a workaround. Customers who had downloaded their raw genetic data before the order continued to interpret that data using third-party tools that the FDA's authority over devices did not reach. Patient-organized genetic interpretation communities formed around specific conditions. By 2017, 23andMe had completed the regulatory process for several health-related interpretations and was authorized to provide them again. By the early 2020s, the practical landscape of direct-to-consumer genomics was substantially what it had been before the cease-and-desist, with a small set of additional regulatory requirements layered on top.

The four-year gap between the cease-and-desist and the partial regulatory restoration is the stakes this exhibit measures. During the gap, customers who would have received health-related interpretations of their own genetic data instead received only ancestry information from 23andMe. They either accepted the reduced information, paid for parallel medical genetic testing through their physicians (which most of them could not access affordably for the same panels), or interpreted their own raw data using third-party tools whose accuracy varied. The institutional position during the gap was that the patients were better off not knowing the health-related interpretation, because the interpretation had not been regulatory-validated to FDA standards.

The patient-community position was, throughout, the position OpenAPS would articulate in code form a year later: that patients who have access to their own data can interpret the data, that the interpretation may be imperfect but is the patient's interpretation to make, and that the institutional alternative of withholding the data is not, in net effect, safer. The FDA's eventual partial reversal acknowledged the patient-community position implicitly. The interpretations the FDA eventually authorized 23andMe to deliver are interpretations the FDA had previously held could not be safely delivered to patients without a clinician intermediary. The position was, in 2013, the institutional position. By 2018, it was no longer the institutional position. The patient communities had won the substantive question without ever having had to make the case in court.

Wing II's Cardiovascular Hall documents that the word atypical is the artifact under most of the gallery's other artifacts. The word treats the male presentation as the baseline and the female presentation as the deviation. This gallery re-reads the word against the patient-community vocabulary that has been developed in parallel.

The patient-community vocabulary inverts the framing. The patient-community literature does not describe women's heart attacks as atypical. It describes them as women's heart attacks. The literature names the specific symptoms women experience, the specific delays they encounter, the specific clinical biases they face, and the specific interventions that would help. The literature does not begin from a male reference patient. It begins from the women who are having heart attacks now.

WomenHeart, founded in 1999, is the primary patient-organized advocacy organization in the United States cardiovascular space for women. Heart Sisters, the blog and book project founded by patient writer Carolyn Thomas in 2009, has become one of the most cited patient-facing resources in women's cardiovascular care. The Sarah Ross Soter Center for Women's Cardiovascular Research at NYU Langone, directed by Harmony Reynolds, MD, operates explicitly on the principle that women's heart disease is a population worth studying as the population, not as a deviation from a male reference.

The stakes are direct. The 2024 Heart journal study documented in Wing II measured worse one-year survival for women after cardiac surgery in England. The differential survival is the cumulative downstream cost of the trajectory the gallery's other artifacts predict. The patient-community vocabulary, where it has been adopted, predicts better outcomes. The Heart Sisters book and the WomenHeart programs are both, in concrete terms, attempts to help women recognize what their bodies are telling them faster than the clinical encounter is going to recognize it for them. The patient communities have been doing the interpretive work the institutional literature was, until quite recently, refusing to do for them.

Wing II's Reproductive Authority Gallery closes with the 44.8 figure, the cumulative outcome the wing's previous galleries predict when integrated across the trajectory of a Black woman's pregnancy in the United States. This gallery re-reads the figure against the patient-community response that has been organized, by Black women, for the duration of the period the rising mortality covers.

The Black Mamas Matter Alliance was founded in 2013 by a coalition of Black women's health advocates. The Alliance's interpretive claim was specific. Black women have been documenting the maternal mortality crisis the institutional literature took another decade to recognize at scale. The Alliance has, since founding, produced policy frameworks, workforce training programs, patient-advocacy networks, and research collaborations that center Black women as the primary investigators of their own community's reproductive health outcomes.

The Alliance's methodological position is direct application of the principle Gallery II documents. Black women are the population most affected by Black maternal mortality. Black women are therefore the population whose interpretive authority over the causes, the experience, and the necessary responses should be primary. The Alliance has consequently been involved in the development of the contemporary midwifery and doula training programs that have produced the strongest measurable outcomes for Black maternal health. The institutional research apparatus, which until very recently treated Black maternal mortality as a problem requiring further study, is now beginning to study the interventions the patient communities have been operationalizing for a decade.

The stakes are measured in lives. The 44.8 figure is a count of Black women who died in 2024 from causes that, in most cases, would have been preventable with earlier recognition, more equitable triage, and clinical care that took the patient's account of her own body at face value. The Black Mamas Matter Alliance has, throughout the period during which the figure was being measured, been building the patient-organized response. The response has not yet, by the numbers, retired the operation. The response continues. The gallery's curatorial position is that the stakes are visible in the persistence of the figure, and the response is visible in the parallel apparatus the Alliance has built. Both are happening now.

The wing's closing exhibit is the museum itself. The museum is a patient-community work. The curator is a patient-community founder. The argument the museum makes, across five wings and several hundred exhibits, is the argument the patient communities have been making for forty years. The exhibits the museum displays are the institutional record. The wall labels are the patient-community reading of the institutional record.

The museum could not have been built before now. The patient-community vocabulary the wall labels use, the methodological principles the galleries draw on, the institutional acknowledgments the artifacts cite, and the technology platform the museum runs on are all products of the patient communities the wing has documented. The museum is, in one sense, a single instance of the broader pattern Gallery II named: methodology delivered in a medium the institutional literature can read.

The museum will not retire the institutional operations the other wings document. The institution does not retire operations because museums display them. The institution retires operations because patient communities organize sustained pressure, because individual patients refuse to accept the institutional account of their own bodies, because clinicians-of-conscience inside the institution take up the patient-community position, because regulators acknowledge the patient-community evidence, and because the cumulative weight of decades of patient organizing becomes too heavy for the institution to continue carrying its previous positions. The museum is part of the cumulative weight. It is not the whole of the weight. The whole of the weight is the patient communities themselves, who are still organizing the day the museum opens.

The wing's curatorial position is that the museum's existence is itself one of the receipts the gallery has been documenting. A patient-community founder has built and opened a museum about the operations the patient communities have spent forty years organizing against. The institutional alternative to this museum is a museum that does not exist, in which the institutional operations the museum documents would continue without being shown. The museum exists. The patient communities continue. The visitor who has walked the five wings is, by the act of walking them, now part of the population the wing's argument addresses. What she does with the visit is the museum's most important exhibit, which the curator cannot display because the curator cannot reach across the screen to read it. The visitor is invited to send her response, as an acquisition, an objection, a correction, or a continuation, to the acquisitions address. The wing is open. The museum is open. The work continues.