Raphael Cuomo, Father of Survival Epidemiology, Outlines Key Tenets of Survival Epidemiology in New Article

by Arthur Walker, M.D.

A field built for the moment after diagnosis

A new article published in the Journal of Clinical Epidemiology by UC San Diego epidemiologist Raphael E. Cuomo, Ph.D. argues that modern population science has a blind spot: much of epidemiology is built to explain who gets sick, even though many of the decisions that shape health outcomes happen after a diagnosis. In his new article, Cuomo defines “survival epidemiology” as a new field of science focused on understanding how people live longer and better with established disease, and on why relationships that predict disease incidence often fail to hold once a patient crosses the boundary into the postdiagnosis state. 

The article frames diagnosis as more than a clinical label. It is a causal threshold that reshapes biology, treatment exposures, and risk. Once disease is present, the relevant questions change from prevention oriented choices to survival oriented choices, including recurrence, progression, treatment tolerance, functional outcomes, and competing risks. Cuomo’s central message is that prevention evidence should not be automatically generalized into “what helps survival,” because the postdiagnosis world has different selection forces, different time scales, and different sources of bias.

Why the “same” risk factor can behave differently after disease begins

A major motivation for survival epidemiology is the repeated observation that common prevention associations sometimes weaken, disappear, or even reverse among people already living with disease. Cuomo’s paper points to familiar patterns such as the obesity paradox described in heart failure and chronic kidney disease, reports of higher cholesterol correlating with lower short term mortality in some heart failure settings, and examples where modest alcohol consumption has appeared neutral or modestly favorable for survival in specific postdiagnosis cohorts even though avoidance messaging dominates prevention guidance. Rather than treating these patterns as simple endorsements of a behavior, the paper argues they often signal that postdiagnosis inference is structurally different. 

In Cuomo’s framing, the act of conditioning on disease status changes what the study population represents. Selection on diagnosis can generate collider bias, evolving symptoms and biomarkers can produce time dependent confounding, and defining exposures using information that occurs after follow up begins can create immortal time bias. Reverse causation also becomes common because changes in weight, cholesterol, or behavior can reflect underlying progression or treatment toxicity rather than an exposure that is truly protective. The paper argues they are central hazards that need explicit design and analytic countermeasures. 

This “prevention versus survival nonequivalence” is also described as a routine test the field should adopt for the same exposure and disease pair, so that public health and clinical guidance does not blur distinct causal questions into a single oversimplified narrative. 

The methodological blueprint

Beyond naming the field, the article lays out what credible survival epidemiology should look like in practice. A consistent theme is design first thinking, especially approaches inspired by target trial emulation. That means explicitly stating eligibility, aligning time zero with the clinical decision point, defining strategies as they are actually used in practice, and handling censoring and follow up in a way that avoids self inflicted biases.

The paper highlights a toolkit that includes target trial emulation, marginal structural models, g computation, dynamic treatment regime methods, joint modeling of longitudinal biomarkers with survival outcomes, and survival frameworks that accommodate competing risks and multistate transitions. It also emphasizes that postdiagnosis settings often involve time varying effects and non proportional hazards, because the relevance of behaviors and biomarkers can change across phases such as intensive therapy, recovery, progression, and long term survivorship. 

Just as important, the paper argues that survival epidemiology requires data elements that many general population cohorts do not capture well. These include stage, molecular subtype, residual tumor burden, prior therapy lines, dose intensity, adverse events, performance status measures such as ECOG, and patient reported outcomes. Without this clinical granularity, survival estimates can become averages across fundamentally different disease states and treatment pathways, limiting interpretability and decision relevance. 

What this could change for patients, clinicians, and public guidance

Cuomo’s proposal is a call to build an equally rigorous postdiagnosis population science that matches the reality that millions of people live for years with chronic disease and cancer. The article argues that journals should expect authors to clearly indicate whether claims pertain to prevention or survival after diagnosis and to report key target trial elements, including time zero and how diagnosis conditioned biases were addressed. 

The article also suggests that guideline bodies should separate prevention recommendations from postdiagnosis survival recommendations when survival specific evidence exists, rather than extrapolating across states. That separation matters because advice that is helpful before diagnosis can be counterproductive during treatment, particularly when preserving lean mass, maintaining caloric sufficiency, managing symptoms, and sustaining dose intensity are central to outcomes. Survival epidemiology, in this view, provides a structured way to communicate that what reduces risk of getting sick is not always the same as what improves outcomes once a person is already sick. 

Finally, the paper calls for funders and training programs to support survival focused reporting standards, curricula, and data infrastructure that make postdiagnosis inference routine rather than exceptional. The goal is a clearer, more clinically aligned evidence base for people living with disease, and a public narrative that distinguishes prevention stories from survival stories so patients are not left trying to apply general population headlines to highly specific treatment realities.