Before falling into a coma or losing all decision-making abilities, very few people sign what is known as an “advance directive”, which specifies what type of medical care they would prefer. Doctors must then rely on the decision that relatives and loved ones make on their behalf. If a simple formula could predict how these patients would want to be treated in dire medical situations as accurately as their loved ones could, then computers may one day be able to assist the surrogate decision-makers to better assess the wishes of people in a coma or with severely impaired mental abilities.

A study designed to estimate how accurately the wishes of these patients can be predicted was run by a team of the National Institutes of Health in Bethesda, Maryland, USA. Another aim of the study was to determine how this accuracy could be improved, as surrogate decision-makers had previously been shown to be right in only 68% of the cases. A parallel study of the information collected by pollsters and scientists about the attitudes towards medical care held by the general US population showed that, should there be at least a 1% chance of full recovery, people would want to be treated, whereas they would prefer to die if this chance was below 1%. Using this information and a new, easier to understand questionnaire, it was found that surrogates predicted the patient’s wishes with an increased accuracy of 78%.

A computer algorithm could then be designed, taking into account not only the chances of full recovery, but also the ethnic and cultural characteristics of the patient and relatives, with data on medical preferences collected on a much wider scale. Such a tool would certainly relieve some of the pressure on those who have to make a decision on whether or not a life support machine should remain switched on.

Obviously, a delicate question of ethics is raised. As a computer is not able to make life-or-death decisions, it should simply assist but not replace the human element in this extremely subtle equation.

Source: PLoS Medicine

Share This With The World!
  •  
  •  
  •  
  •  
  •  
  •  
  •