Your ‘health age’ is fake

Wearable makers and health-device companies increasingly quantify users’ health as an age — a shorthand that promises to translate lifestyle data into an intuitive number. Apple’s newly announced Health Age will join existing offerings such as Whoop’s "Whoop age," Oura’s cardiovascular age, Withings’ heart and vascular age readings, and Garmin’s fitness age score. The same person can receive very different results depending on which device and algorithm are used. Those differences stem from what each product measures and how it weights signals. The author’s own device readings varied widely: Whoop said 42, Oura’s heart estimate fluctuated between seven years younger and 1.5 years younger over two years (currently two years younger), Withings showed vascular age one year older and heart age four years younger, Garmin returned 36.5, a facial "longevity mirror" gave a skin age of 26 and physiological age of 35, and a mattress topper estimated 38.7 this week — while her true age is 38. Individual algorithms may favor certain behaviors; for example, the author says Whoop’s metric heavily weights strength training, so prioritizing lifting would tip that number downward. As tools for self-improvement, these metrics can be useful: they can highlight where changes in diet, sleep, or exercise might yield gains and can nudge users toward healthier habits. But they have limits. They do not capture the full clinical picture — a reader reported getting a favorable wearable age yet discovering high cholesterol on an annual physical — and some health factors are genetic or outside personal control. High-profile examples show the cultural angle: Apple’s health lead suggested Health Age could be something people brag about, while entrepreneur Bryan Johnson’s public efforts to remap his biomarkers provoked attention about how such metrics are used socially. Because of those social dynamics and measurement differences, health-age readouts are best treated as one, imperfect input rather than an absolute statement about longevity. They can guide behavior changes that genuinely improve risk factors, but they can also be gamed or misinterpreted, and they won’t capture conditions, genetics, or accidents that affect real-world outcomes. For users, the practical takeaway is to use these scores as targeted feedback — for example, to identify weak spots to train — while keeping routine medical care and clinical biomarkers as primary evidence of health.