Health Keeping.pdf

Problemspace A: Decision between communicating the result of diagnosis via artificial entity or human

Case description

The initial situation is that the artificial entity is sufficiently certain of its prediction, wether this prediction is positive or negative. Obviously, there is still a probability, that the prediction is wrong. The decision to communicate the result to the patient should be based on the patients preference, perhaps with some restrictions in edge cases, such as especially fatal predictions. It is immoral to force communication by the artificial agent upon the patient, as it would restrict the patients freedom of choice compared to the current state of diagnosis communication.

Factors to be considered for moral evaluation of actions

• Freedom of choice (on the patient side)

  • Some patients might prefer communication through an actual human as opposed to an artificial entity. Currently, it is possible to communicate with a human doctor, so the freedom of choice would be reduced, if this option were no longer available. Additionally, multiple ways of this communication taking place are imaginable, perhaps differing in cost or availability. It is questionable, wether it is morally acceptable to offer these differing types of communication, if they cannot be available for every patient.

• Accountability (of the doctor in case of false predictions)

  • In case of an autonomous communication of the diagnosis to the patient, the doctor would not be aware of this action in every case. He could be informed about the communication between the artificial entity and the patient, but with scale, it would be unlikely that the doctor can keep track of every ongoing communication. Perhaps the doctor must be able to decide which level of autonomous communication he allows, in order to be held accountable in case of an unintended case. Perhaps this level must be adjusted over time through input by the doctor in order to gradually fade to an autonomous system

• Impact on the most primary stakeholder (due to weight of the diagnosed disease)

  • The impact the diagnosis will have on the primary stakeholder (the patient) varies according to the weight of the diagnosed disease. This is a factor, that must be taken into consideration when the decision of wether to autonomously communicate the diagnosis. It is apparent, that some sort of classification is required to do so. In some cases, this classification may be fairly simple, as most might agree, that a common cold will not impact the patient as much as a potentially terminal diagnosis, such as cancer, will. But the differences in diseases are much more granular, as different stages, developments, etc. impact the overall weight of the disease fundamentally. These factors must be weighted against preference factors from the patient, as in how much will the individual be impacted by the diagnosis. This evaluation of the patients reaction poses another challenge, as it is fairly difficult to quantify.

  It is essential to clearly define what a problem space is, as these factors could also be considered problem spaces (they contain aspects of the previously defined problem space A and B)! Or should this use case (or rather the problem spaces in this use case) be reconstructed to one problem space, where there are multiple factors?

Report by the Academy of Medical Royal Colleges:

Reports

Ethical Issues Described in Report

• Patient Safety
• The Doctor and Patient Relationship
• Public Acceptance and Trust
• Accountability for Decisions
• Bias, Inequality and Unfairness
• Data Quality, Consent and Information Governance
• Training and Education