This project further develops a new EEG-based prognostic technology for comatose patients, identifying relevant socio-ethical values and how to pro-actively take them into account in the innovation process.

The research team investigates the socio-ethical values embedded in current practices of care for comatose patients and the effects the technology might have on the realisation of these values. It is subsequently explored how these values can pro-actively be embedded in the technology and the social practices surrounding its usage. Preliminary results confirm the high predictive value of the new technology and its potential for cost reduction.

Outcome prognosis of patients in coma after cardiac arrest is crucial for decision making on (continuation of) treatment. A recent technological innovation is visual classification of continuous EEG, which enables ultra-early, high-quality prognosis of poor outcome within 24 hours. This innovation, especially when not visually interpreted but translated in a quantitative index, may however also have negative impacts. It may:

• deepen existing controversies on the expected quality of life of surviving patients;
• aggravate tensions between personal values (life of relatives) and societal concerns (cost reduction);
• create new controversies regarding timing of the prognosis and subsequent decision making.

Responsible development

The project is a collaboration of researchers with expertise in empirical ethics, medical technology, health technology assessment and philosophical-conceptual analysis. Together they explore:

• How EEG-enabled prognosis might affect socio-ethical values in care for comatose patients, comparing the Netherlands, the USA and South Korea. The researchers take a practice-based approach in which as a first step the values embedded in current prognostic practice are investigated. For example, what do the values of beneficence or justice entail in this context? The results are subsequently used as the starting point for subsequent reflection and deliberation on how the new technology might influence the realisation of these values.
• How these socio-ethical values can be taken into account to ensure a responsible development of EEG monitoring technology. The researchers take a value sensitive design approach, meaning that they investigate how values can pro-actively be embedded in the material (hard- and software) and social (clinical practice, regulation) components of prognostic practice. In other words, they establish how values such as beneficence or justice can be assured by making the right choices in the design of the technology and usage practices.

Finally, the researchers will explore the usefulness of the insights gained for other types of coma prognostics innovations and for responsible innovation more generally.

International comparative perspective

This project performs a stakeholder-informed ethical analysis in the Netherlands, and then investigates whether these conditions would also be applicable in the USA and in South Korea. Since technological innovation in the biomedical domain travels easily from one country to another, Responsible Innovation in this domain should be defined in a way that transgresses national boundaries. The case of innovation studied in this project lends itself very well to such an international approach, since the technological hardware used (EEG) is already available on many Intensive Care units worldwide.

Preliminary results

The research team already:

• Carried out studies (a prospective cohort of 430 patients in post-anoxic coma after cardiac arrest) that confirm the high predictive value of visual analysis of rough EEG data for the prediction of both good and bad outcomes.
• Performed a preliminary cost analysis, concluding that EEG based prognosis and treatment decisions can lead to cost reduction.
• Optimised computer-assisted, quantitative analysis of continuous EEG data for the discrimination of patients with a good or a bad outcome, using a ‘random forest classifier.’ This was done by exploring the possibilities of a ‘deep learning approach’, in which the computer ‘independently’ learns to interpret EEG patterns (simulating a neurophysiologist); an original approach which seems quite promising.

---

coma prognosis, comatose patients, brain injury, coma recovery, coma recovery, continuous EEG, prognostic indicators, medical practice, design, cost reduction, prognostic practice

Official project title: