A Minister of Health was responding to epidemic disease. He said:
‘Too many people are dying. We need to save more of the confirmed cases.’
He gave orders for a comprehensive survey of patients’ health conditions.
‘We need to know the diagnosis for each and arrange the best possible corresponding treatments.’
A survey team visited hospitals and identified the main categories of illness for every patient.
The Minister ordered that every patient would be able to access the best practice treatment for their condition.
A junior doctor who had studied statistics said: ‘They should have done their survey in the morgue. They should have surveyed causes of death and identified treatments that would have prevented each. By mistake, the best practice treatments were applied to conditions that had a prognosis other than death. They have surveyed survivors, when they should have surveyed those that succumbed. This would have saved more lives.
Is it a paradox that treatment needed might be inferred from a sample who can cope, rather than from those who have failed to cope?