Do Mortality Rates Spike on Your Birthday? The Evidence Says: Not Really, and Not for Everyone.

The claim is that a person's risk of dying meaningfully spikes on their own birthday — a dramatic effect often attributed to psychological stress, the 'will-to-live,' or some biological rhythm tied to the date of birth. The verdict is partially false: there is a small, contested statistical signal in specific subgroups, but nothing resembling the universal mortality spike the claim implies.

The most concrete evidence in favor of the claim comes from Ajdacic-Gross et al. (2012), published in the Annals of Epidemiology. Analyzing nearly 2.75 million deaths in Switzerland between 1969 and 2008, the researchers found a statistically significant 13.8% excess of deaths on birthdays compared to an average day. That sounds alarming — until you look at where the effect actually lives. The birthday excess was driven almost entirely by suicides (+34.9%), falls (+44%), and cardiovascular disease in men over 60 (+18.6%). Cancer deaths showed zero birthday effect. This is not a general mortality spike; it is a concentrated signal in a handful of causes among specific, vulnerable groups.

The steelman of the claim rests on that Swiss study's large sample size and statistical significance. But here is precisely where it breaks down. Two independent analyses contradict it entirely. Kestenbaum and Ferguson (2002), using U.S. Social Security Administration death records, found no significant birthday effect in all-cause mortality and raised a critical methodological problem: birthday-date reporting bias, where death records may be more likely to record a death on a memorable date like a birthday, artificially inflating any apparent signal. Then Peña (2015), publishing in PLOS ONE, re-analyzed U.S. mortality data and found no statistically significant birthday spike in all-cause mortality after correcting for multiple comparisons and seasonal confounders, concluding that prior positive findings may be methodological artifacts rather than real biological or psychological effects.

Even the original Swiss researchers, in a 2015 response published in Biodemography and Social Biology, acknowledged that the birthday effect is real but modest, concentrated in the elderly and people with pre-existing conditions, and cannot be generalized to the broader population. That is a significant concession: the authors of the strongest supporting study themselves walked back the universal framing. What is genuinely true is that for older adults with cardiovascular disease or elevated suicide risk, there may be a small, real elevation in risk on their birthday — possibly linked to stress, alcohol consumption, or disrupted routines. That narrow finding is legitimate. The sweeping claim is not.

The manipulation pattern here is a classic case of stripping effect sizes and subgroup limitations from a real study and broadcasting only the headline number. A 13.8% excess sounds large, but it applies to a single day out of 365, in a dataset where the absolute number of extra deaths is small, and where the effect disappears entirely in the largest independent replication attempts. Whenever you see a mortality or health claim built on a single dramatic percentage, ask immediately: excess over what baseline, in which subgroup, and did independent researchers replicate it? Those three questions would have stopped this one cold.