Re: For sore throats, seven days of penicillin is not superior.
Both Dr Fleetwood and Prof Cantekin point up a curious anomaly in
life-table analysis of recovery from acute illness. If the outcome measure
chosen is the day on which symptoms last occurred, rather than the end of
the period of continuous symptoms, the kind of pattern observed in the
study of Zwart et al is almost unavoidable.
Imagine 2 patients, Bob in the 7-day group, Paul in the 3 day group.
Both get better after 24 hour but Paul, unsurprisingly, relapses on day 5
and recovers on day 7. On the life table Bob appears to have recovered
within the crucial first 3 days while Paul does not. Hence the early take-
off of the recovery curve in the 7-day group. It does not in any way
indicate allocation bias between the groups.
Perhaps the lesson is that life-tables should be used with caution
for data of this kind.
Competing interests: No competing interests