Distribution of Longevity and Mortality in Learning Health Systems: The Ultimate Structural Injustice

What if anti-aging science made breakthroughs in the next decade like information science did in the last decade?

Learning Health Systems (LHS) rapidly advance knowledge by leveraging clinical data on platforms. Systematic treatment of all patients as research subjects is necessary to maximize learning but fundamentally challenges normative obligations to promote respect, beneficence, and justice when research and care are no longer distinct. Underlying LHS is an assertion that patients are obligated to contribute to learning and the system is obligated to improve care in return.

Paradoxically, U.S. lifespan plateaued and is now declining in the decade since the HITECH Act formally spurred conversion to electronic medical records and rise of the most powerful learning platforms to date. Meanwhile, disparities in mortality are increasingly stark, with inequities themselves identified as key causal factors. Despite expansion of widespread, systemic learning, some receive timely, prescient care, whereas others may receive substandard, outdated care, if any.

Parallel to these health trends are mounting income inequalities and healthcare costs, in many cases attributed to growing data monopolies. COVID — for all its unprecedented intertwining of care and research, urgency, and scale — has been a pressure cooker, accelerating collective investment in learning while exacerbating asymmetries in health and wealth. Simultaneously, paradigm-shifting information technologies have forged an elite class of industry disruptors hungry for health to match their exponential wealth. With unparalleled access to all manner of health data, their well-funded, AI-empowered federated learning platforms have put digital fountains of youth on the horizon.

The yawning gap between those for whom learning may grant unsurpassed longevity vs. those who suffer preventable, premature deaths casts a shadow on our digital enlightenment. Decline in average lifespan and increasing racial and geographic inequities offer harsh assessments of the efficiency, effectiveness, and justice in our current LHS.

The quest for immortality: Hers, Ours, Theirs

Henrietta Lacks suffered a premature death that may have been preventable with routine gynecologic care, even in 1951. Her parting gift, biopsies obtained during her cancer salvage treatment, were the key to her redemption: HeLa cells became the first immortal human cell line. Appropriately, Henrietta Lacks is the paradigmatic symbol for the dual potential of immeasurable benefits and irreparable harms that may arise when specimens obtained during clinical care are repurposed for academic and commercial research without adequate transparency, accountability and engagement.

She often prompts discussion of ethical issues surrounding informed consent, privacy, and racism — shining light on the particularly dark history (and present) of a fraught relationship between Black women and American medicine. Her story is told and retold — as it should be. And yet, the pathognomonic lynchpin maintaining the status quo remains hidden in plain sight.

Henrietta’s contribution to science is outsized and includes both death-defying and birth-enabling treatments, but her prevailing lesson is a commentary on universality of the human condition. We are simultaneously mortal and immortal, with a common humanity, and yet indelibly unique. Meanwhile, Henrietta’s tissue, like that of other patients, was deidentified, ostensibly to protect her privacy while enabling society to benefit from biospecimen research.

Deidentification as objectification, identification as respect. Credit: Jonathan Newton/The Washington Post/Getty, as modified by Marielle S. Gross

Recognizing the untapped potential to harness growing health data, U.S. law created a “safe harbor” for deidentified data to allow ungoverned use of data when privacy risks are minimized. The largest healthcare data platforms to date take advantage of this workaround on similar grounds of patient privacy protection. However, deidentification, which now boasts removal of 18 classes of identifiers under HIPAA, remains as much a slight of hand for the advancement of institutional and researcher interests as when He/La was written on the dish containing a piece of Henrietta’s cervix.

She was reidentified, and so are we. Any of us could be the “next Henrietta Lacks,” and already might be without knowing it. Deidentification lowers barriers to accessing biospecimens for secondary research by eliminating burdens of complying with traditional human subjects research protections (e.g., the obligation to obtain informed consent). Though traditional informed consent may be neither necessary nor sufficient for the current research-care environment, we should also be wary of the recent addition of prospective broad consent to the Common Rule as a substitute for either deidentification or discrete informed consent. The constant, pervasive and evolving nature of research may render such upfront disclosures akin to “accepting the cookies.”

Converting the identified patient into a deidentified data subject also eliminates fiduciary duties we once had to her as a person at the other end of our biopsy forceps. Patients cannot escape the obligation to contribute their health data to systemic learning, yet there is no guarantee that personal, health or commercial benefits of research will be justly distributed. Someone’s antibodies saved countless lives from COVID, while others helped deliver effective vaccines in record time — and they may fair similarly to Henrietta and her family regarding personal, health and commercial interest in their own flesh and blood. This is most concerning when unblocking individuals from knowledge and benefits of research they participate in sacrifices a potentially life-changing health outcome.

This pattern of profound, pervasive, asymmetric, and unavoidable impact on health for different social groups suggests significant and enduring structural injustice in the social arrangements by which we leverage clinical care for learning . Our learning health system is one in which the cookies are unilaterally doled out, and unanimously accepted, even though this fails to achieve its own goal of care that is maximally efficient, effective, and just. Structurally unjust systems do not fail because of bad actors; they are failing despite being full of patients, physicians, and researchers, who are working hard, taking on risks, all to advance learning through care.

Such a wicked problem may only be corrected by addressing systemic arrangements which allowed such asymmetries and marginalization to accrue. Novel ethical, legal, socioeconomic, and technological solutions are urgently needed. Eliminating reliance on deidentified patient data — disconnecting people from their birthright — for the purpose of advancing research increasingly focused on maximizing longevity for the elite while profiteering from the illness of others may be a matter of life or death.



OB/GYN Bioethicist — Focusing on how to leverage cutting-edge technology to promote quality, efficiency and justice in women’s healthcare @GYNOBioethicist

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Marielle S. Gross, MD, MBE

OB/GYN Bioethicist — Focusing on how to leverage cutting-edge technology to promote quality, efficiency and justice in women’s healthcare @GYNOBioethicist