Why one mortality number isn't enough
Producers regularly collapse "mortality" into a single percentage, then steer decisions by it. That hides as much as it shows. A farm with a tidy 4% overall mortality rate can still be losing half of every animal that catches a particular disease — the headline number stays low only because few animals are getting sick.
Three measures pull the signals apart. Overall mortality answers "how many animals did I lose this period?". Cumulative mortality answers "what was the risk of dying during the outbreak window?". Case fatality answers "once an animal got sick, how lethal was it?". Each points at a different piece of management — biosecurity, treatment protocols, or general husbandry.
The three formulas
Deaths in a period are the residual once you account for every animal that moved through the stock:
deaths = opening_stock + newborns − sold − closing_stock
mortality_rate = deaths / (opening_stock + newborns)
cumulative_mortality = deaths / closing_stock
case_fatality = deaths_from_disease / cases
opening_stock— Head count at the start of the periodnewborns— Animals born during the periodsold— Animals sold, slaughtered or removed aliveclosing_stock— Head count at the end of the periodcases— Animals diagnosed with the specific disease being measured
A worked comparison
Two pig farms with 200 head each, hit by the same outbreak. Farm A reports 53 cases and 18 deaths; Farm B reports 19 cases and 9 deaths.
- Farm A: mortality rate 9.0%, case fatality 33.9%
- Farm B: mortality rate 4.5%, case fatality 47.4%
Farm B looks healthier on the headline number, but its case fatality is materially worse — the pigs that got sick there were more likely to die. Two different problems: Farm A has a containment issue, Farm B has a treatment issue.
Common ways the numbers get misread
Mortality figures are only as good as the records behind them. A few pitfalls show up repeatedly in farm audits.
- Pick a tight observation window — For case fatality especially, the period has to be short enough that deaths from unrelated causes don't pollute the numerator. A stretched-out window inflates the rate.
- Count cases at first diagnosis, not at death — Counting only animals that died as "cases" forces case fatality to 100% by construction. The denominator must include every animal that showed disease, recovered or not.
- Separate baseline mortality from the outbreak — Newborn losses and old-age deaths happen regardless of disease. For outbreak analysis, strip them out so what's left reflects the actual event.