How algorithmic bias in health apps reflects systemic exclusion in digital health infrastructure
Original framing: “Why most health apps are biased even before their first lines of code are written” — bing news
The bias in health apps is rooted in a long history of medical exclusion, from the 19th-century eugenics movements that shaped early health metrics to the 20th-century clinical trials that excluded women and people of color. The legacy of these exclusions is visible in today’s algorithms, which are trained on datasets that reflect historical inequities. For example, the underrepresentation of Black women in period-tracking app data has led to misdiagnoses of conditions like endometriosis, a direct consequence of past medical racism. This history is rarely acknowledged in discussions of 'algorithmic bias,' which often treat it as a new problem rather than a continuation of old patterns.
The bias in health apps is not a glitch in an otherwise neutral system but a feature of a techno-medical complex that prioritizes profit, efficiency, and Western biomedical dominance over equity and cultural relevance.
Historical patterns of exclusion—from eugenics to clinical trial discrimination—have been encoded into algorithms, while Indigenous and cross-cultural knowledge systems offer pathways to dismantle these structures. The trickster reveals the absurdity of framing bias as a technical fix when it is a symptom of deeper power imbalances, from corporate control of health data to the erasure of marginalized voices in design. Solutions must center Indigenous sovereignty, participatory design, and public ownership of health data, shifting from 'fixing bias' to transforming the systems that produce it in the first place.