Last week at the 74th U.N. General Assembly, the Rockefeller Foundation announced a $100 million Precision Public Health initiative to integrate data analytics and data science tools such as machine learning into the community health systems of low- and middle-income countries. During the announcement, Rockefeller noted the significant impact of data science innovations in the most privileged communities, explaining its hope to replicate this effect to improve healthcare for people globally. The initiative aims to save at least 6 million women and children’s lives by 2030.

The task is not an easy one. There are three key challenges anticipated over the next 10 years:

  1. Low availability of high-quality data
  2. Lack of responsible data use policies
  3. Deficiency of trained human resources

As the specific details of the initiative become available in the coming months and years, we at DAI’s Center for Digital Acceleration always keep an eye on these efforts to see what concrete solutions emerge. We have several questions regarding the future of this initiative, building off the gaps that the foundation has identified:

1. How will the initiative ensure that the most marginalized individuals and communities are represented in the data?

Quality and availability of data is the first challenge on the Foundation’s list. Rightfully so, as data science only works when you have good data. Good data is difficult to obtain, in high-income and low-income countries alike. It requires data processes and standards for collection that often do not exist. And when the data does exist, it frequently excludes populations that are more difficult to reach—such as poorer communities, women and girls, or rural populations. As a result, these groups are being left behind in the digital health revolution and suffering from more severe health issues because of it. Therefore, it is imperative for stakeholders involved in creating health initiatives to be motivated and committed to pursuing better healthcare for everyone.

2. What steps will be taken to protect health data and confirm that it is used in an ethically responsible manner?

As mentioned by a participant in one of our cybersecurity workshops in Ukraine, “cybersecurity is a process, not a product.” While responsible data use policies are important in securing health data, legislation on its own is not the solution. Cybersecurity is as much about people practicing safe digital activities as it is about having the appropriate technical hardware and software. Creating and enacting policy is one step in the right direction but policy itself does not change people’s behaviors. All stakeholders in this initiative, from frontline health workers and community members to tech partners to government bodies, need to be invested in the entire cybersecurity process for responsible data to be considered successful.

3. Will gender equality be incorporated into the initiative? And if so, how?

In a world where women and girls are underrepresented in STEM —including the technology companies in high-income countries that the Foundation is seeking to partner with—gender equality must be a critical component of this initiative.

chartoftheday_4467_female_employees_at_tech_companies_n (1).jpgSource:

Equal representation of genders not only empowers a historically marginalized community, but it brings a more diverse set of opinions and voices into an inherently biased field of study. Data science tools take on the assumptions and beliefs of those who create and work with them. This means that more diverse data science talent will provide more comprehensive and useful solutions to the health issues of the world.

Machine learning, artificial intelligence, and deep learning are hot topics in healthcare. Nevertheless, to date, there have been few successful applications of tools on a large scale anywhere around the world. Sometimes it’s because the available data is insufficient. Sometimes it’s because the algorithms aren’t accurate. In this case, however, the technical aspect is not the only component for measuring success. The other is accessibility and inclusivity. This initiative can only succeed if it take steps to represent all people in its work. Without it, we are left with biased data science tools that only augment health inequalities and outcomes. With these questions and concerns in mind, we are excited to follow and contribute in our own way to the Rockefeller Foundation’s 2030 goals.