Monday, 25 June 2018

The red line of health data

Elena Torrente




The great technological giants (Google, Apple, IBM or Microsoft) are recruiting scientists and experts in medical research. Their commitment to transform medicine is decided, but John T. Wilbanks and Eric J. Topol warn of possible risks in this interesting article published in Nature. Given the difficulties that medical research has always encountered in obtaining health data, the authors believe that it may be tempting for clinical scientists to access large-scale Internet projects, with significant data collection capabilities and algorithms for analysis.

But there is a risk that Google or other smaller companies such as 23andMe, will access health data with the aim of controlling the methods used to offer certain services based on digital profiles. This is what in other sectors of the Internet is called the society of the "black box". There are ads that are only offered to certain groups of people and if these algorithms incorporate health data, the authors say that pre-existing biases in our society would increase along with inequalities.

The article points out that, although the use of electronic medical records is being standardized in many countries, nowadays there are enormous amounts of medical data available outside the health system. Moreover, great advances for the exploitation of this data are being made with machine learning. For example, Google entered into an agreement with the Mayo Clinic in 2015 to analyze health-related searches, such as the most common causes of back pain. Google engineers introduce verified health data to the intelligent Knowledge Graph engine and users get the most accurate information possible about their symptoms or health conditions. Many companies deny citizens access to health data and, in others that do, it’s very difficult to share them for science. The problem, according to the authors, is that closed algorithms can harm scientific progress. They state that the data collected by sensors and the methods employed to make them clinically usable should be a public good.

Regulation could prevent the collection of health data from becoming a profitable business. But authors believe that it would be more efficient to create open models that would facilitate competition. Where appropriate, at Sage Bionetworks where they conduct clinical studies on various diseases, participants have the automatic right to access and download full copies of all their data. They also have the right to offer their data to be reused by qualified researchers. "Openness" is not an easy task because it must guarantee the protection of privacy and allow those who don’t want to participate to get ahead. But if the new era of digital health is accompanied by demands of greater openness from citizens, it will end up transforming the health sector and beyond.

No comments:

Post a Comment