RightsCon Toronto 2018

As the patient-centered health ecosystem evolves, a large swath of personal health data in the commercial environment is governed by the “terms and conditions” set forth by the various companies who create the technology, instead of any regulatory scheme. This new data is being shared, aggregated, and used to make decisions about an individual's health and well being. Commercial entities use automated processes, like data analytics, that harness statistics, algorithms, and other mathematical techniques to convert data into actionable knowledge. In the context of health, automated systems like these may advocate for one course of medication or treatment over another. These systems also influence consumer health in more subtle ways by mediating behavior, such as through the food or exercise recommendations given by an activity tracker, or through access to information, such as the results given by a search engine on a specific health condition. In addition, datasets of health information can be sold and combined to include new metrics, and subsequently analyzed for patterns that will be used for predictive purposes. Data from commercial health apps is also frequently used in research that informs medical practices and broader public health concerns.

This panel will investigate the data ecosystem of mhealth apps, such as the types of data collected and where the data flows, then delve into a more technical discussion of the most common types of machine learning applications used to deliver outputs to users and the types of recommendations made for users. Panelists will discuss where potential biases and discrimination might be inadvertently introduced in the analytics used on data from these apps, and how this bias might impact different populations. The panel will look toward potential solutions or mitigation strategies for biased outcomes.