​
Projects
A risk-informed concept of disease
Abstract: To determine what defines the concept of disease and whether that concept is value laden, some authors (Lemoine 2013; 2015) advocate for an inductive approach. Lemoine argues that this examination should avoid considering value-laden disease judgments by looking at scientific descriptions of disease. I examine what medicine considers a scientific description of disease, finding that diseases are increasingly described in terms of risk. But in the face of uncertainty about what level of risk constitutes disease, decision-makers turn to non-epistemic values, like economic payoffs of diagnosis, to establish a biological threshold of risk which determines where a biological variable counts as disease. An inductive examination of diseases described in terms of risk results in a concept of disease that is not value-free. This historically informed examination of the concept of disease demonstrates that the concept is a tool which is sculpted by specific actors to serve the goals of diagnosis.
The History of the Personal Health Device and Concepts of Disease
Abstract: Meaningful properties of the concept of disease can be examined by exploring the technologies which were created in order to identify and explain disease. These technologies increasingly include personal health devices, such as insulin pumps/glucose monitors or smartwatches. These devices track biological data, like glucose, heart rate, or body temperature, alerting the user to abnormal readings. When the patient wears a personal health device, the link between symptoms and disease starts to fade, replaced by a screen that decides when abnormal readings of biomarkers, such as glucose, constitute sickness. The screen of the personal health device becomes a lens for patients to understand what it means to be sick, instead of the symptom.
I explore recent developments in technology for type one diabetes, proposing that the introduction of and reliance on the personal health device in chronic diseases like type one diabetes has led to conceptual change: thanks to the digital biomarker (e.g. glucose) tracking provided by the wearable device, patients with type 1 diabetes look to their device first for information about if they are “sick” (hypo- or hyperglycemic) or if they will become sick. As this conceptual change spreads beyond diabetes, it has clinical implications, introducing an issue of “burden of proof” in diagnosis. Classically, clinicians have been trained to search for biomarkers based on the presence of symptoms. If biomarkers are introduced prior to symptoms, must we rely on the presence of symptoms in order to confirm disease, or are biomarkers all that matter?
Computerization, the International Classification of Disease, and Diagnostic Precision
Abstract: The rise of computerization in the 1960s altered the landscape of medicine. Physicians had long been tasked with the memorization of patterns in order to diagnose a disease, but the practice left gaps for clinical judgment and thus error. Offloading pattern recognition to the computer promised to standardize diagnosis, leading to higher accuracy. But the effect of computerization in disease classification was instead an explosion in precision: diagnostic codes in the 1989 10th revision of the International Classification of Disease (ICD) rose to 5 times that in the 9th. The 1983 move to Prospective Payment Systems in the United States was a financial catalyst for precision diagnosis, facilitated by computerized diagnostic reasoning. Until 1983, hospitals were reimbursed for any cost incurred by Medicare patients. The 1982 Tax Equity and Fiscal Responsibility Act linked reimbursement to diagnosis: hospitals were now reimbursed a fixed sum given the average resource-usage of a patient’s diagnosis. The 10th revision of the ICD was a reaction to linking payment to diagnosis: diagnostic codes became more granular because of the financial incentive for more diagnostic precision. The story demonstrates that a financial motivation, made possible only by computerization, was responsible for a revolutionary change in medical practice.
Carving the Concept of Disease from Clinical Epidemiology to Precision Medicine
Abstract: Medicine has long been plagued with questions about the concept of disease: what is it, and is it value-free or value-laden (Boorse 1975, 1977, 1997, 2011; Kingma 2010; Lemoine 2013, 2015)? If it isn’t value-free, is medicine a science at all? I take a different approach to these questions, turning instead to history. A historical examination of the concept of disease across the 20th century reveals a more nuanced perspective: the concept of disease is a tool used to organize the current prevailing system, or logic, of medicine.
I focus on the change from clinical epidemiology, a dominant system of medicine from the 1950s onwards which focused on defining disease, and indeed, diagnosing and treating it, in terms of risk (Fuller forthcoming), to the rise of precision medicine in the 1990s, which advocated defining, diagnosing, and treating disease in terms of genes (Tabery 2023). This historical moment includes a shift in a conceptual package: a concept of disease (disease is first a threshold of risk, and then a mutation of a gene), a logic of diagnosis, and a treatment. The conceptual package is organized by a concept of disease, which tells clinicians what to privilege (risk or the gene) and how to reason about it in order to diagnose a disease and treat it. That is, medicine harnesses whatever concept of disease is the most relevant and useful in order to bolster the science, or logic, of clinical reasoning. I argue that long-standing questions about whether or not the concept of disease is natural, or whether medicine is scientific, may be misguided, because the historical examination of the concept of disease instead demonstrates that it is simply a tool for medicine to carve out how to be scientific, depending on its broader context and aims.
