Medicine's ‘Basic’ Mistake: How the Molecular Revolution Displaced an Entire Category of Inquiry

ABSTRACT

The integration of molecular biology in medicine represents a remarkable achievement of modern science. Nonetheless, the success of molecular methods reshaped medicine's epistemology over time such that biological explanation came to mean explanation exclusively at the cellular or molecular level. This review argues that medicine committed a philosophical error—molecular investigation came to determine epistemic status, displacing person‐level clinical research as an entire domain of rigorous inquiry. Person‐level clinical research encompasses clinical observation, psychosocial and behavioral medicine, patient‐reported outcomes, and the biology of biography. Each tradition satisfies established criteria used to define foundational science. For example, KRAS mutation testing shares its epistemic structure with biographical constructs such as chronic neuroendocrine dysregulation and allostatic load. Both traditions identify specific activated pathways, predict therapeutic response, stratify patients into biologically distinct populations, and operate upstream of pharmacologic interactions. The biological significance of a pathway depends on the person‐level context in which it is activated. Two patients sharing the same mutation but carrying profoundly different neuroimmune histories, inflammatory set‐points, and stress exposure are not the same biological system. Precision medicine, as currently practiced, is often precise at the wrong level of analysis. Evidence supporting biography as mechanistically foundational has not been refuted; it was marginalized by self‐reinforcing institutional forces. Federal funding concentrated in molecular science created financial dependencies that transformed an epistemological preference into a structural imperative. Tenure structures systematically protected molecular faculty whereas clinical scientists were moved to contingent appointments. Successive generations of trainees oriented rationally toward the molecular model. The displacement of person‐level clinical research was not a scientific decision. It was a historical accident compounded by the financial and administrative architecture of modern academic medicine. Restoring this domain to foundational status is not a humanistic plea; it is a scientific corrective—and its costs, if deferred, are borne by patients.

Medicine congratulates itself, and rightly so, for having rejected the Cartesian dualism of mind and body. [1, 2] Credible contemporary scientists would dismiss a proposition that these two entities are separate ontologically. Yet the institutional architecture of clinical science—as expressed in its departmental organizations, funding structures, curricula, journal publications, and research awards—continues to reproduce that separation in practice. The contradiction is not incidental; rather, it is structural.

1. Background

The molecular revolution transformed clinical science [3] and remains one of the important intellectual achievements in modern medicine. For example, the ability to isolate or identify signaling pathways, receptors, mutations, immune mechanisms, and pharmacologic targets has changed medical diagnosis and treatment. The problem is not that medicine embraced molecular biology. Mechanistic rigor matters, and molecular biology deserves its central role. Over time, however, the extraordinary success of molecular methods reshaped medicine's epistemology. A molecular focus became not merely a foundational component of biomedical science but synonymous with what is foundational. Biological explanation came increasingly to mean explanation only at the cellular or molecular level. What could not be elucidated there was progressively treated as secondary, contextual, or scientifically “softer.” [4].

What was demoted was a coherent domain of inquiry that had been central to the scientific foundation of clinical medicine for centuries. That domain can be described as person‐level clinical research. It can be informally said to encompass clinical observation and natural history, psychosocial and behavioral medicine, patient‐reported outcomes, and the biology of biography [5]. Each discipline was producing rigorous, reproducible, and clinically consequential knowledge at the time it was being displaced.

This review argues that medicine made a philosophical error that hasn't been fully recognized. Disguised as scientific progress, the molecular level of investigation came to determine epistemic status. Person‐level clinical research as a class satisfies the same criteria medicine already uses to define foundational science. Exclusion of such research from “basic science” is not a rigorous conclusion. It is an institutional and philosophical stance that became normalized over time and was compounded by the financial and administrative architecture of the modern academic medical center.

2. What “Basic” Actually Means

In physics, basic science includes Newton's laws of motion because they are foundational—upstream of downstream phenomena, generative of predictions, and structurally necessary for a system to make sense. The main criterion for “basic” in science has historically been explanatory priority—not scope, tractability, or level of investigation.

Medicine gradually drifted from this definition. The molecular revolution generated discoveries of enormous importance, including receptor pharmacology, the genetic code, cellular signaling cascades, oncogenic drivers, and immunologic checkpoints. Over time, however, the tools that enabled these discoveries came to define what counted as rigorous science itself. “Basic” increasingly came to mean what could be measured at the molecular level.

This transition was understandable but regrettable. Molecular approaches produced extraordinary explanatory power at a time when person‐level constructs were less developed methodologically. But methodological achievements are not equivalent to ontological sufficiency, and methodological reductionism does not require ontological reductionism. Person‐level phenomena can be biological mechanistically yet not reducible to isolated molecular events.

Georges Canguilhem argued decades ago that pathology is intelligible only relative to the individual organism in which it occurs [6]. Disease is not merely a molecular event; it is a disturbance occurring within a living system whose biological norms are themselves shaped over time. This premise is not a humanistic sentiment. It is an epistemological claim about where clinically relevant causal architecture resides. Context is not external to mechanism; it is part of mechanism.

3. What Medicine Taught Itself to See

The educational architecture of medicine is not merely a reflection of its epistemological hierarchy; the structure represents a primary mechanism of reproduction.

The first 2 years of medical school, still called the “basic science” years, are organized around molecular and cellular biology. Disciplines including biochemistry, cell biology, pharmacology, and molecular pathophysiology constitute the cognitive foundation that precedes clinical training. This architecture encodes the epistemological hierarchy before a student has ever seen a patient. Molecules are signals; biography is noise. Students absorb such aphorisms as curriculum, not ideology.

The physician‐scientist was supposed to be the corrective. The MD‐PhD was conceived as a bridge between the molecular laboratory and the clinical encounter. The idea was sound, but the execution failed. In most institutions, the MD‐PhD became another route into basic molecular science. Physician‐scientists were trained to address laboratory questions using laboratory methods, reviewed by laboratory scientists, and evaluated by laboratory metrics. The “physician” modifier became a credential rather than an epistemic orientation. The bridge was built, but only from one side.

The result was a generation of “scientist‐physicians” whose clinical observations of patients who defied molecular prediction were experienced as puzzles to be resolved by better molecular tools, rather than as evidence that the explanatory framework was incomplete. Clinical observation as the contribution a physician‐scientist was trained to make was not recognized as a scientific product.

Reforming this architecture requires more than adding a social determinants lecture or a narrative medicine module. It requires placing stress physiology, neuroendocrinology, developmental biology of adversity, and the science of patient‐reported outcomes alongside biochemistry in the preclinical curriculum—not as contextual supplements, but as mechanistic foundations. And it requires redesigning physician‐scientist training so that clinical observation is recognized as a scientific act, and questions generated at the bedside help to drive the next generation of laboratory inquiries.

4. The Displaced Domain: Person‐Level Clinical Research

What was displaced was not the absence of person‐level clinical inquiry but its recognition as foundational. Each tradition in this domain was demoted in a characteristic way.

Clinical observation and natural history comprise a tradition running from René Laennec's diagnostic innovations through William Osler's humanistic medicine to Alvan Feinstein's clinical epidemiology [7]. These advances clarified how diseases actually behave in “whole” human beings over time. The work involves rigorous empirical inquiry at the level of the patient [8]. Although that tradition did not disappear, it was reclassified as “clinical description” rather than basic science and became valued as raw material for laboratory‐based hypotheses rather than as a foundational knowledge base in its own right.

Psychosocial and behavioral medicine established that social circumstances, relationships, behavioral self‐regulation, and psychological states are among the strongest predictors of disease onset, progression, recovery, and mortality [9]. The mechanisms linking social isolation to cardiovascular risk, depression to immune suppression, and social support to cancer survival were being characterized at the pathway level. Yet this research was also routinely categorized as “soft.” Its constructs were regarded as confounders to control for, rather than causes to investigate.

Patient‐reported outcomes, functional status, quality of life, and the patient's own account of illness each occupy an epistemically irreducible position [10]. They are not approximations of some more fundamental data awaiting molecular specification. Fatigue accompanying cancer, dyspnea in heart failure, and pain in rheumatoid arthritis are clinically real and in many contexts are the primary therapeutic focus.

The biology of biography is the most mechanistically elaborated domain of person‐level clinical research [1, 2, 5]. It provides the strongest case for restoration as a basic science precisely because it meets molecular biology on its own terms—by identifying pathways, describing signaling mechanisms, linking experiential inputs to measurable biological outputs, and generating experimentally testable hypotheses. To the extent that this domain has been systematically excluded from foundational status, the explanation is not methodological inadequacy. The reason is institutional epistemology operating below the threshold of explicit argument.

5. The KRAS Mutation Test

Consider what makes a molecular finding foundational. As an example, the test to identify a KRAS gene mutation at codon 12 is foundational because it (a) determines specific signaling pathways that are constitutively activated, (b) predicts therapeutic response, (c) stratifies patients into biologically distinct populations with divergent outcomes, and (d) is upstream of the pharmacologic interaction the clinician is attempting to engineer [11].

Now apply those same criteria to the biological‐biographical construct of chronic adversity associated with sustained hypothalamic‐pituitary‐adrenal (HPA)‐axis dysregulation and elevated allostatic load. Such a construct identifies specific neuroimmunological pathways that are chronically activated. Sustained cortisol dysregulation alters inflammatory tone, suppresses CD8+ lymphocyte function, reshapes autonomic balance, and modifies a tumor microenvironment in ways directly relevant to disease progression and therapeutic response.

Patients with high allostatic load demonstrate systematically different immune baselines, endocrine dynamics, and treatment trajectories. The process is upstream of pharmacologic interactions [12, 13]. The epistemic structure is identical to the KRAS test. The difference is not causal status; the difference is experimental methodology and tractability.

Biographic variables are often multidimensional, temporally distributed, and more difficult to isolate experimentally. But difficulty of measurement does not diminish mechanistic reality. The history of science repeatedly demonstrates that tractability and causal importance are not synonymous. Medicine has mistaken the feasibility of measurement for a hierarchy of causation.

Even relatively simple measures of stress exposure have demonstrated clinically meaningful associations, including in non‐small cell lung cancer and malignant melanoma [14, 15]. Two patients who share the same KRAS G12C mutation but possess profoundly different neuroimmune histories, inflammatory set‐points, and stress biology are not the same biological system. Precision medicine, viewed from this perspective, is often precise at the wrong level of analysis.

6. The Evidence Was Not Refuted

The biology of biography does not rest on metaphor or intuition. It rests on a growing body of mechanistically grounded evidence linking lived experience to measurable biological regulation.

Chronic psychosocial stress activates the HPA axis and sympathetic nervous system, producing sustained glucocorticoid signaling, inflammatory dysregulation, altered immune surveillance, endothelial dysfunction, and changes in metabolic and neural regulation. These pathway‐level findings have been replicated across species, methods, and disease domains [16].

Preclinical studies in lung cancer models have demonstrated that chronic psychological stress accelerates tumor progression through glucocorticoid receptor signaling and protumor macrophage polarization [17, 18]. These findings provide a mechanistic account wherein experiential adversity becomes tumor biology. An ongoing STRESS‐LUNG prospective cohort study was designed to test this hypothesis and reported that emotional distress attenuated the response to immune checkpoint inhibitors in non‐small cell lung cancer [14]. In parallel, results from the MoTrPAC molecular exercise project revealed that many of the same signaling pathways implicated in stress‐mediated injury are involved in exercise‐mediated resilience and repair [19]. Accordingly, the biological significance of a pathway depends on the person‐level context in which it is activated. These findings are evidence that molecular events cannot be interpreted independently of the bio‐biographical state in which they occur.

The placebo adherence findings from the Coronary Drug Project provided an early signal indicating that adherence itself might carry independent prognostic weight, apart from the pharmacologic activity of the assigned treatment [20]. Subsequent causal reanalyses of that trial—applying inverse‐probability weighting to extensive post‐randomization covariate data—found that the crude survival advantage of adherers was substantially confounded (attenuated) by time‐varying health‐related behaviors. When this “healthy adherer” phenomenon was taken into account, several adjusted estimates were compatible with a null effect [21]. In contrast, a multivariable analysis of data from the Beta‐Blocker Heart Attack Trial (BHAT)—adjusting for clinical severity, sociodemographic status, psychosocial features, and smoking—found the adherence‐mortality association was essentially unchanged [22]. Of note, the BHAT reanalysis did not employ time‐varying causal‐inference methods later applied to the Coronary Drug Project.

The disparity in results of the two reanalyses highlights the fact that a covariate's causal role is not a fixed property of the variable but instead is a function of the causal structure assumed. Contemporary causal‐inference methods, including inverse‐probability weighting, treat adherence as the exposure and its correlates as pre‐treatment confounders to be accounted for, and by that logic adjusting for them is warranted. But if those same correlates are instead the joint downstream expressions of a specific biographical mechanism, rather than representing independent confounders, then adjusting for such factors does not remove bias from an estimate of the effect of adherence. Instead, adjustment conditions on the observable mechanism itself, leaving little by which its effect could be detected.

Both scenarios cannot be evaluated using the same statistical approach, given that the disagreement is not about adjustment procedures but rather is about what the covariates represent. The inclination to treat adherence as a problem in confounding—rather than a difference in conceptualizing the biographical mechanisms involved—is an example of the epistemological issue this essay describes.

The broader phenomenon of understanding how biology and biography co‐constitute disease risk and treatment response has been marginalized because the institutional architecture of medicine is not organized to recognize biography as mechanistically foundational. The cost of this narrowing is both scientific and professional. Scientifically, medicine developed a sophisticated molecular science without a comparably sophisticated science of the biologically embedded human state. Resilience, vulnerability, recovery capacity, inflammatory regulation, autonomic flexibility, and treatment responsiveness are cogent bio‐biographical phenomena that cannot be adequately modeled at the molecular level alone [23, 24, 25].

7. How the Molecular Hierarchy Became Self‐Reinforcing

The hierarchy that favored molecular explanations did not emerge through conspiracy or malice; it arose through success. Research institutions committed early to molecular biology, and their productivity was genuine. Study sections, training programs, journals, and award structures then evolved around that framework. Investigators trained in the dominant paradigm reviewed grants written within it and introduced subsequent generations to its assumptions.

Mechanistic adequacy gradually became defined mainly in molecular terms alone. Proposals focused on stress biology, neuroimmune regulation, or whole‐person causation were frequently judged “insufficiently mechanistic” by evaluative criteria already shaped by the hierarchy conducting the evaluation. The dismissal of person‐level research was framed as methodologically justified—patient narratives were unreliable, psychosocial variables were imprecise, biography was difficult to quantify.

But measurement variability is endemic throughout molecular biology. It is seen in antibody inconsistency, batch effects, cell‐line drift, assay instability, and interoperator variability, and their occurrence is acknowledged as routine technical realities. Molecular biology responded by refining its tools. The claim that biographical data are inherently unscientific is not an empirical conclusion. It is a methodological prejudice that confuses complexity with illegitimacy.

8. The Institutional Economy and Its Workforce Consequences

Persistence of the current hierarchy requires an economic assessment alongside an epistemological one. Federal funding to academic medical centers is not merely a measure of scientific productivity. It is a primary revenue stream that funds building maintenance, faculty salaries, graduate programs, administrative support, and other institutional overhead.

Among research‐intensive universities, the concentration of funding is striking. Institutions including the University of California at San Francisco, University of Pennsylvania, Stanford University, Washington University, and Yale University each received more than half a billion dollars annually in research support from the National Institutes of Health for fiscal year 2025 [26], the majority flowing through basic science structures organized around molecular and cellular biology. These grants do not simply reward scientific excellence; they constitute the financial architecture of the modern academic medical center.

This situation transforms an epistemological preference into a structural dependency. An institution that derives a substantial fraction of its operating revenue from molecular basic science grants is not in a neutral position when evaluating whether to invest in person‐level clinical research. Reorienting even a portion of a research enterprise toward clinical epidemiology, psychosocial medicine, or bio‐biographical science would mean competing for a smaller and less reliably funded pool of grants, generating less recoverable overhead, and training fewer graduate students.

The financial incentives run mainly in one direction. Institutions that lack the molecular faculty depth to compete at the highest level nonetheless organize their research missions around the same model, sacrificing a potential comparative advantage in person‐level clinical research for a competition they are structurally unlikely to win.

The epistemological hierarchy did not remain confined to grant committees; it restructured the academic medical workforce. A 2022 survey of 155 U.S. medical schools documented what four decades of institutional prioritization had produced [27]. The percentage of full‐time clinical faculty holding tenured or tenure‐track appointments had fallen from 59 percent in 1982 to 18 percent in 2022. Over the same period, tenure rates for PhD faculty in basic science departments declined only modestly, from 78 percent to 64 percent. The asymmetry is not incidental. It is a quantitative indicator of which category of faculty, and by extension which category of science, the institution considered worth protecting.

Tenure is not merely a credential. It is an institutional commitment indicating that the work a faculty member pursues is sufficiently important to warrant long‐term investment and protection from shifting priorities. When clinical scientists are systematically placed on non‐tenure tracks, the institution communicates legibly and durably to every trainee that their science is contingent rather than foundational. When tenure‐track positions, named chairs, department leadership roles, and training grants flow toward molecular scientists, the rational response for an ambitious trainee is to orient toward molecular science. The corresponding workforce is not the product of scientific consensus about what the most important questions are. Rather, the workforce is the product of a personnel architecture built on an epistemological assumption that was never explicitly defended or evaluated.

9. Conclusion

Medicine has not failed by embracing molecular biology. It has, however, lost something of immense value by narrowing its definition of foundational science to what could be isolated most easily at the molecular level. The resulting hierarchy did not materialize from a philosophical debate. It developed from the extraordinary productivity of molecular methods and then became institutionalized through funding structures, journals, curricula, and prestige systems.

Scientifically, medicine developed a sophisticated disease science without a comparably sophisticated science of the biologically embedded human state. Non‐response to therapies that perform well in aggregate populations has too often been treated as statistical noise when it may represent biologically meaningful heterogeneity invisible to current trial architecture.

Person‐level clinical research as a class satisfies the same criteria medicine already uses to identify foundational science. Clinical observation and natural history explain how diseases unfold in living people. Psychosocial and behavioral variables stratify outcomes and predict therapeutic response with a reliability that, in many disease domains, exceeds currently available molecular markers. Patient‐reported outcomes are the primary endpoints in whole therapeutic categories and are variables that directly measure what treatment is meant to accomplish. The biology of biography modifies the substrate on which molecular events unfold, reshaping immune tone, endocrine dynamics, inflammatory regulation, and tumor microenvironments in ways that determine whether molecular interventions succeed or fail [28].

The two lines of evidence are complementary rather than duplicative. One focus is on biography's role in disease risk and progression before treatment begins, and the other focus is on biography's role in modifying response once treatment is underway. These distinct mechanistic claims together define the scope of the bio‐biographical framework [5, 28].

The future of clinical science will require research capable of integrating molecular and personal/bio‐biographical levels simultaneously. Longitudinal bio‐biographical phenotyping, AI‐assisted narrative phenotyping, and clinical trials that treat experiential factors as mechanistic variables rather than confounders are not speculative ambitions. They are the logical extension of evidence that has been accumulating for decades.

The question no longer is whether person‐level clinical research is scientifically foundational. That was settled by the evidence, though the institutional hierarchy chose not to organize based on the findings. The pending question is whether medicine and the academic medical centers that train its future workforce are prepared to recognize that the displacement of an entire class of rigorous clinical inquiry was a historical accident, not a scientific verdict.

Precision medicine, as currently imagined and practiced, is precision along an axis the medicine ecosystem has most thoroughly implemented. Patients who fail to respond as expected are not statistical outliers. Many of them are people whose biographical biology has never entered into the therapeutic equation. The cost of ignoring the full range of basic science is paid, ultimately, by patients.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.