Confidence Intervals
                        ====================

Confidence intervals for a proportion when the numerator is either zero or
"n" are the easiest kind of all to calculate, using Hanley's "Rule of 3's".

Upper limit of confidence interval = 3 / n

Thus, if no complications are noted in 100 consecutive patients given a
specific treatment, the 95% confidence interval for the incidence of this
complication ranges from 0%-3%.  In your specific case, if no false negatives
are noted in 100 diseased subjects, the 95% confidence intervals for
sensitivity would range from 97%-100%.

See:  Hanley JA, Lippman-Hand A:  If nothing goes wrong, is everything all
right?  JAMA 1983; 249:1743-1745.