Historia de los ensayos clínicos aleatorizados
Source: https://www.sap.org.ar/docs/publicaciones/pronap/2006/PBE-HistoriadelosECA-AAP2006.pdf ↗
A Spanish-language history of randomised clinical trials — from James Lind's 1747 experiment on scurvy to the twentieth-century institutionalisation of the RCT as the gold standard for medical evidence.
Demirdjian writes accessibly for clinicians and the piece is short, but the historical texture it offers is valuable: the RCT is a young practice, contested at every stage, and its extension to non-medical contexts is a cultural choice more than a logical inevitability.
For product direction it is useful background for anyone thinking seriously about A/B testing and causal inference — knowing where the method came from clarifies what it can and cannot do.
A short paper, worth its length.
Central argument
Demirdjian argues that the randomised controlled trial is not a timeless scientific truth but a historically contingent practice, tracing its contested evolution from James Lind's 1747 scurvy experiment to its twentieth-century codification as medicine's evidentiary gold standard. The central finding is that the RCT was resisted, debated, and institutionalised through specific social and epistemic choices — not through the straightforward triumph of reason over ignorance. This framing implies that the method carries embedded assumptions about causality, control, and what counts as valid evidence that are worth interrogating rather than inheriting uncritically.
Critique
Because the paper is a short piece written for a paediatric continuing-education programme, its historical scope is necessarily selective and its analysis remains descriptive rather than critical — it chronicles the RCT's rise without deeply engaging with the philosophical or statistical objections that accompanied it, such as debates over external validity, ethical constraints on randomisation, or the Bayesian alternative tradition. A reader looking for a rigorous epistemological account of what the RCT can and cannot prove will need to go further; Demirdjian provides orientation, not argument.
Why it matters for product
The paper's core implication — that the RCT is a cultural choice institutionalised under specific conditions, not a universal law of inference — should give product leaders pause before treating A/B testing as a self-evidently correct default for all product decisions. Many product questions involve small populations, irreversible changes, or network effects that violate the independence assumptions the method requires, and knowing this history makes it easier to argue for when not to run an experiment rather than defaulting to one as a signal of rigour. It also helps teams distinguish between measuring an intervention's local effect and understanding why something works, which is the gap where most A/B-driven product strategies quietly fail.