The funhouse mirror: the I in personalised healthcare

Abstract

Precision Medicine is driven by the idea that the rapidly increasing range of relatively cheap and efficient self-tracking devices make it feasible to collect multiple kinds of phenotypic data. Advocates of N = 1 research emphasize the countless opportunities personal data provide for optimizing individual health. At the same time, using biomarker data for lifestyle interventions has shown to entail complex challenges. In this paper, we argue that researchers in the field of precision medicine need to address the performative dimension of collecting data. We propose the fun-house mirror as a metaphor for the use of personal health data; each health data source yields a particular type of image that can be regarded as a ‘data mirror’ that is by definition specific and skewed. This requires competence on the part of individuals to adequately interpret the images thus provided.

Self-tracking, background(s) and hermeneutics. A qualitative approach to quantification and datafication of activity

“…But as our analysis suggests, self-tracking is a field fraught with interpretation and any claim to objectivity dissolves in a chain of references and mediations, without necessarily granting users meaningful, not to mention deepened or enhanced, access to the phenomena that are the substance of their everyday experience. While existing research has already discussed how data doubles can distort users' perception of themselves (see, e.g., Vegter et al, 2021 for an interesting example), we contribute to the debate by showing that similar processes occur in relation to other aspects of users' lived reality, namely the background in which the tracking occurs. Moreover, our findings are guided by the discussion found in hermeneutic philosophy and thus refer to a nuanced and rich theory of interpretation and its influence on our everyday perception of the world.…”

Digital phenotyping and data inheritance

“…Risk variants are identified on the basis of a statistical comparison of data populations, for example, with and without a specific disease. The ‘individual’ risk scores are thus based on the data inheritance from thousands of other individuals, as well as on the procedures for sampling and comparing data populations in relation to predefined clinical criteria (see also Vegter et al, 2021: 5). Moreover, the status of genetic risk variants often changes over time, as more genomes are sequenced, and new knowledge is gained on the clinical relevance or irrelevance of specific markers (Timmermans et al, 2016).…”

“…Moreover, they underscore how digital phenotyping is not just about discovering ourselves. It also involves the experience of being seen and monitored by others (Vegter et al, 2021). Since we cannot control the data legacies we leave behind, digital phenotypes can at the same time extend and reduce the person.…”

“…The performativity of self-observation through digital representation is further complicated by the multiplicity of data doubles, or what she calls mirror mazes . Vegter et al (2021) take the mirror metaphor one step further in describing how personalized health technologies create digital funhouse mirrors . Rather than providing a singular and well-integrated reflection of the individual's health status, digital mirrors present fragmented pieces of information that require interpretative skills to become meaningful.…”

Reply to Letter to the Editor on “Not all biases are bad: equitable and inequitable biases in machine learning and radiology”

“…In societies characterised by strong social inequities, these inequities are also inscribed in medicine, medical data, and data technologies. This means that the datasets and algorithms used for machine learning (ML) are never straightforward representations of people, bodies, and populations [ 3 ], but they are “funhouse mirrors” [ 4 ] that reflect how access to resources are distributed in a population. As a result—and this is what we argue in the paper—concerns about equity should ideally come in already at the stage of deciding whether or not a dataset is biased: The question of what constitutes a bias should be the starting point of the reflection.…”