Abstract of a chapter of Mittelstadt and Floridi’s Ethics of Biomedical Big Data. Although accepted, this chapter was never finished. The abstract was written in April 2015.
In this chapter I will attempt to show how the rise of Big Data necessitates a drastic revision of the notion and elements of informed consent. The “Fourth Paradigm” and its concurrent biomedical practices, such as data-driven science, the automation of Evidence-Based Practice, have changed the very foundation of the biomedical sciences, changes that must be also reflected in their ethics.
The most important change of perspective brought about by these new forms in medical practice is the proliferation of statistically observed, evidence-based treatments and protocols that promise great efficiency. In light of the ever-growing flood of data, the question becomes, can the physician provide adequate information for the patient’s informed consent when she does not know the mechanism of action of the treatment under discussion? Or rather, what are the contents of the information that must be provided regarding the treatment? Is that information only composed of data: statistical rates of success or side-effects, or does it necessarily contain some form of (subjective) medical expertise or scientific aspect that is not equivalent to data?
The question “can the physician explaining the treatment to the patient be replaced with graphical representations of available data” is not a rhetorical one after the “fourth revolution”. The ever-increasing amount of automated evidence-based practice will necessarily incapacitate the physician from gaining and providing subjective and expert information on newer treatments. Taking things even further, we will find ourselves faced with the fact that the same acceleration and increment of data will put efficient treatment forever ahead of the scientific, causal explanation of the condition and its treatment and the discovery of its mechanism of action. Whether or not this will render the theoretical aspects of biomedical sciences obsolete and trivial remains to be seen, but one thing is certain: the possibility of confounding has increased as never before. The issue cannot be reduced to a rehashing of the “computers make errors” cliché; the wide-spread forms of “epistemological uncertainty” (Renee Fox) point to a much more paradigmatic problem. In fact the most important reason for worrying about the insufficiency of data-based consent and the possibility of confounding is that there are no reasons to be concerned about it. The point is, the increasingly efficiency/performance based attitude of the biomedical sciences translates into the prioritizing of proficiency and efficiency in prediction over scientific verity.
Given all of these considerations, a revision of the content of the information necessary for the patient’s informed consent is exigent. I believe it is high time we ask ourselves where we stand with the data/information problem, since surely lives are at stake.