Toward demographic robustness in digital patient twins: Addressing the gender data gap

Area of Science:

• Digital health
• Biomedical informatics
• Health equity

Background:

• Digital patient twins (DPTs) are emerging tools for precision medicine, simulating individual health.
• Their effectiveness and fairness hinge on representative data and unbiased algorithms.
• Existing DPT initiatives frequently exhibit biases, overrepresenting certain demographics.

Purpose of the Study:

• To argue for sex- and gender-based equity as a core principle in DPT design and implementation.
• To highlight how current DPTs perpetuate health disparities due to underrepresentation of diverse populations.
• To propose solutions for creating equitable and clinically valid DPTs.

Main Methods:

• This opinion piece synthesizes evidence from cardiology, endocrinology, mental health, and medical device research.
• It analyzes documented examples of bias in DPT-related research and applications.
• The authors draw on established knowledge of structural biases in clinical research.

Main Results:

• Current DPT initiatives disproportionately represent male, white, and affluent populations.
• Women, gender-diverse individuals, and people of color are underrepresented, leading to misdiagnoses and inaccuracies.
• Examples include under-recognition of heart failure in women and overestimated oxygen saturation in darker skin tones.

Conclusions:

• Structural biases in clinical research contribute to DPT inequities.
• Addressing these requires balanced data, sex- and gender-informed design, participatory approaches, and subgroup validation.
• Equity in DPTs is essential to avoid exacerbating existing health disparities and realize personalized medicine's potential.