“Gender medicine is personalization for beginners”

It has been known for decades that men and women require different therapies in many areas. Nevertheless, standard procedures persist and women are still underrepresented in studies and data sets to this day. So what are the chances for AI tools to avoid distortions in the database?

This was discussed by two gender medicine specialists and a developer in the panel “Digital Health and Gender Medicine - How neutral can AI be?” Dr. Jana Schmidt, a data scientist, bioinformatician and Chief Innovation Officer at Significo, which develops health apps, among other things, emphasized that it is not so much a question of neutrality as of being aware that people are not the same. This starts with hormonal differences between men and women and extends to living conditions and psychological factors.

“Gender medicine concerns us all,” emphasized Prof. Dr. Gertraud Stadler, Professor of Gender-Sensitive Prevention Research at the Charité University Hospital. Gender medicine means taking gender differences into account throughout the entire treatment cascade – from prevention to mortality. That is why she has been dealing with the question of which biases shape medicine for decades. For example, women are still only represented 30 percent in cardiovascular studies. Men, on the other hand, are disadvantaged when it comes to prevention or mental illness.

Data gaps in a wide variety of areas

“We lack different data in different areas,” added Prof. Dr. Irit Nachtigall, head of the Department of Translational Research, Teaching and Cooperation at the Vivantes Network for Health. For example, depression in men is rarely diagnosed because two-thirds of the questionnaires were evaluated by women. Conversely, women are not only underrepresented in studies, but are also less likely to be diagnosed. For example, one-third fewer blood cultures are taken from them. Why is that? Nobody knows. Nor do we know why women are prescribed medication in the same dosage as men, even though they tolerate less alcohol – that is not logical.

Health apps also often have a gender bias, reports Jana Schmidt. Since women are more interested in prevention, many apps are tailored to them as a target group. Men are not sufficiently reached by the tonality and the offers. In any case, the tonality in the apps is rather one-sided. Surveys show that almost 40 percent of users like the usual medical style, but 60 percent would prefer a more relaxed approach.

There is still a lot of room for improvement

In any case, there is still a lot of room for improvement in terms of data quality and in terms of taking into account different social groups, especially people with a migration background, said Stadler. In fact, many more parameters need to be taken into account in medical data – social situation, for example, ethnicity, sexuality. “Gender is just the beginning and relatively simple,” said Stadler. “Gender medicine is personalization for beginners – and means taking the diversity of us all seriously.”