Ethical Issues in Democratizing Digital Phenotypes and Machine Learning in the Next Generation of Digital Health Technologies

Abstract

Digital phenotyping is the term given to the capturing and use of user log data from health and wellbeing technologies used in apps and cloud-based services. This paper explores ethical issues in making use of digital phenotype data in the arena of digital health interventions. Products and services based on digital wellbeing technologies typically include mobile device apps as well as browser-based apps to a lesser extent, and can include telephony-based services, text-based chatbots, and voice-activated chatbots. Many of these digital products and services are simultaneously available across many channels in order to maximize availability for users. Digital wellbeing technologies offer useful methods for real-time data capture of the interactions of users with the products and services. It is possible to design what data are recorded, how and where it may be stored, and, crucially, how it can be analyzed to reveal individual or collective usage patterns. The paper also examines digital phenotyping workflows, before enumerating the ethical concerns pertaining to different types of digital phenotype data, highlighting ethical considerations for collection, storage, and use of the data. A case study of a digital health app is used to illustrate the ethical issues. The case study explores the issues from a perspective of data prospecting and subsequent machine learning. The ethical use of machine learning and artificial intelligence on digital phenotype data and the broader issues in democratizing machine learning and artificial intelligence for digital phenotype data are then explored in detail.

The Sociotechnical Ethics of Digital Health: A Critique and Extension of Approaches From Bioethics

“…The first type of contribution to the bioethics of digital health literature that we identify we refer to as “ applying ethical theory .” In this body of literature, scholars adopt the perspective of an existing ethical theory and assess a subset of normatively relevant issues in digital health from that perspective ( 21 , 26 ). The most common is some form of principlist approach, one that relies on a series of bioethical principles to guide assessment of the ethical implications of any given area of human activity ( 23 , 24 , 27 , 28 ). The field of bioethics is dominated by a principlist approach to ethical thinking ( 20 , 29 ), and it is therefore not surprising that the bioethics of digital health would also be dominated by such an approach.…”

“…In this approach, contributors tend not to use elaborate justifications for a particular orientation to ethical theory, but rather focus primarily on applying the theory to substantive issues in digital health. For example, in a paper on the ethics of digital phenotyping for health-related uses, Mulvenna et al simply state that “the four ethical pillars of medicine are autonomy (right to choice), beneficence (doing good), non-maleficence (do no harm), and justice (equal access), and these pillars should not be overlooked when democratizing digital phenotyping” (p. 8) ( 28 ). The authors then proceed to focus specifically on the principle of autonomy as the primary focus in their ethical analysis.…”

Digital Phenotyping: an Epistemic and Methodological Analysis

“…Although information distortions caused by (say) EMAs versus self-consciousness (as just described) are different, they nonetheless roughly fall under the umbrella of "observer effects". 9 Aside from any digital phenotyping observer effect distortions that interfere with diagnosis or prognostication, digital sensing of an individual could itself also partly cause conditions or states to emerge in that same individual (Mulvenna et al, 2021). This could occur in both good and bad ways.…”

“…For example, suppose that the input consisted of a set of pairs (x, y), with x y being numbers between 1 and 100 quantifying two properties. An unsupervised algorithm such as K-means clustering (Mulvenna et al, 2021) might receive this set and determine that there are three clusters in the underlying structure of the data: pairs with low x and y values, pairs with low x and high y values, and pairs with high x and low y values. This is all done with numerically rigorous method.…”

“…In this paper, we adopt a broad philosophical vantage point to investigate the epistemic nature of digital phenotyping in relation to its alleged promise. This epistemological mapping and analysis extends a small but growing body of recent philosophical work on the subject (Birk & Samuel, 2020;Burr & Cristianini, 2019;Burr et al, 2020;Loi, 2019;Mulvenna et al, 2021;Sharon, 2017;Stanghellini & Leoni, 2020;Tekin, 2020). Unlike most of the literature to date on philosophy and digital phenotyping, which has focused on its ethical aspects, this paper focuses on its epistemic and methodological aspects.…”