and its two-sided 95% CI is:
One can deduce that the actual number of cases has 95 chances in a hundred of being between 242 and 646. The number of cases identified by SR being (127 + 42) — 12 = 157, the reporting ranges between 24% and 65%.
However, the validity of such an estimate requires that reporting to one system or the other be a truly random and independent phenomenon which could be an unverified assumption. For this reason, the safest way is probably to cease to estimate the actual number of cases and to deal with reporting rates instead of incidence rates!
Was this article helpful?