How vaccine coverage ends up above 100%

Every so often a district reports that 112% of its children were vaccinated. It's not good news and it's not quite fraud. It's almost always a denominator problem, and it's one of the clearest examples I know of why the bottom of a fraction deserves as much care as the top.

How the number is built

Administrative vaccine coverage is simple on paper. The numerator is the number of doses given, counted by health facilities. The denominator is the target population, usually the number of babies expected to be born or survive to their first birthday that year. WHO and UNICEF describe it as reports from providers on children vaccinated, divided by reported target population data.

The numerator is counted. The denominator usually isn't. It's projected, often from a census that might be ten years old, with assumptions about births and growth.

Ways the denominator goes wrong

Old census, wrong growth rate. Small errors compound over a decade.
People move. A district with a good clinic near a border or a city may vaccinate children who live elsewhere. Their doses count, their births don't.
Boundaries change. Districts split and merge, and the population figures don't always follow at the same time.
Births aren't fully registered. If the expected births are estimated low, coverage looks high.

The numerator has its own issues too, like double counting or doses recorded for the wrong age. But the denominator is the one that tends to get less scrutiny.

How common it is

It's not rare. A study of target population data reported to WHO and UNICEF from 2000 to 2016 found coverage above 100% in 11% of national reporting events for BCG and 6% for DTP3. In 9% of events, the reported number of surviving infants was larger than the reported number of live births, which would imply a negative infant mortality rate. Reported births also tended to be lower than UN population estimates, which pushes coverage up.

At district level it's even more common. A Journal of Global Health analysis found at least one district with DTP3 coverage over 100% in 76 of the 96 countries it reviewed for 2016.

Why it matters

A coverage figure over 100% is obviously wrong, which is helpful. The bigger problem is the one at 85% that should be 70%. The same denominator issues that push some places over 100% quietly inflate others to numbers that look plausible. Those are the places that might not get the attention or resources they need.

This is why the official WHO/UNICEF estimates don't just take the reported figure. They weigh it against household surveys and other context, and they document the reasoning country by country.

The general lesson

This isn't only a vaccine problem. Any rate you report has the same weakness: conversion rates, readmission rates, percent of customers who did something. People check the numerator because it's the interesting part. The denominator gets copied from wherever it was last time.

So when a rate surprises you, good or bad, check the bottom of the fraction first. Where did it come from, when was it last updated and does it cover the same people as the top?