In 2025, Lara Weed and Jamie M. Zeitzer of Stanford University published an article linking the practice of seasonal time changes (Daylight Saving Time) to negative health outcomes, ranging from acute symptoms (heart attacks and strokes) to chronic conditions (obesity). Now, Professors José María Martín-Olalla (University of Seville) and Jorge Mira Pérez (University of Santiago de Compostela), after analyzing the methodology applied in that study, have concluded that "what the world read as scientific evidence against the time change has turned out to be a mathematical illusion."
The same journal that disseminated the controversial article, PNAS (Proceedings of the National Academy of Sciences), has just published a letter signed by Martín-Olalla and Mira Pérez, demonstrating that the study’s conclusions are not supported by actual evidence.
The original article by Weed and Zeitzer gained significant global traction in the fall of 2025 due to its striking conclusions and its use of the PLACES database (Population Level Analysis and Community Estimates). This database, managed by the U.S. Centers for Disease Control and Prevention (CDC), contains information on the prevalence of 29 syndromes and diseases at the local level. The PLACES data were contrasted against a circadian model developed by the authors.
A Critical Error
The work of Professors Martín-Olalla and Mira Pérez reports a grave error in the study's methodological foundations. The original model computes the difference between the rhythm of the biological clock—the circadian rhythm, determined by the time at which body temperature is at its minimum—and the rhythm of the Earth's rotation. According to the original authors, this difference represents the "regulatory circadian shifting necessary to stay synchronized with the outer world."
Global health effects were inferred from the annual sum of these daily readjustments. However, when performing this calculation, the authors consistently accumulated the magnitude of the readjustment, regardless of whether it was positive or negative. "The use of absolute readjustments instead of real readjustments is the critical error," note Martín-Olalla and Mira. They show that this methodology merely captures the "noise" of the model and, therefore, cannot predict net health effects.
Professor Mira explains: "What the authors did makes little sense; it is as if, while driving, we recorded every small adjustment made by moving the steering wheel back and forth to stay in the lane, but then added them all up in the same direction to report a large value instead of allowing them to compensate for each other. By their logic, maintaining a straight course with small left-and-right adjustments (what actually happens) would be the same as a car drifting further and further in one direction until it ends up facing the wrong way. This alone refutes the study's conclusions."
Professor Martín Olalla adds: "We analyzed the 'guts' of the model and saw that the daily readjustment was small—similar to the model's temporal precision—and fluctuating: one direction one day, the opposite the next, with no global trend leading to significant desynchronization, which is exactly how a readjustment should function. Consequently, the annual cumulative total of these readjustments was zero, even with the time change. The metric used appears to have been chosen with the intent of ensuring the current time-change policy yields the worst possible results; the readjustments triggered by the spring and autumn shifts are made to contribute in the same direction instead of compensating for each other. In this sense, the study’s findings resemble a self-fulfilling prophecy. The fact is that the 'absolute cumulative readjustment' they report is approximately 20 hours per year, which is nothing more than an average of about 3 minutes per day, sometimes in one direction and sometimes in the other. Given the information provided in the study, it is difficult to understand how such a weak value (a mere 0.3%) can be epidemiologically linked to the prevalence of disease."
Concluding their letter, Professors Martín Olalla and Mira pose a fundamental question: What prior expectations did the original authors have when they decided to associate global sociosanitary outcomes with the noise of their own model? "We see no prior hypothesis or causal link that justifies the analysis performed in the original study. This invalidates the methodology and, consequently, the reported findings: the authors cannot conclude that eliminating the time change would lead to a decrease in the prevalence of obesity or acute medical events," they affirm.