Clinical disease rates

Definition

A clinical disease rate measures how often cows develop a disease that someone saw and recorded (a treatment or diagnosis event), as opposed to subclinical disease found only by testing. There are two different measures, and they are not interchangeable:

Incidence risk (cumulative incidence): the share of cows at risk that had at least one case during a defined period. Each cow counts once. It is a proportion, 0–100%.
Incidence rate (incidence density): the number of cases divided by the cow-time at risk, usually expressed per 100 cow-years. A cow can contribute several cases. It is not a percentage and can exceed 100.

Use incidence risk for diseases a cow normally gets at most once per lactation, tied to calving: retained placenta, metritis, milk fever, ketosis, displaced abomasum. The standard form is lactational incidence risk (LIR): cows with at least one case in the lactation (or in the first 30 or 60 DIM) ÷ cows that calved and were at risk.

Use incidence rate for diseases that recur and can happen at any stage of lactation: clinical mastitis, lameness, pneumonia, injuries. Herd-level monitoring of these across a calendar period is usually per 100 cow-years.

What counts as one case

A treatment event is not the same as a case. A cow treated on three consecutive days for one mastitis flare is one case. To turn events into cases, apply a case definition with a repeat window: the same cow, the same disease, within N days of the previous event, is the same case.

Commonly used windows (based on Kelton et al. 1998):

Disease	New case if
Clinical mastitis	> 14 days since previous event (any quarter)
Lameness	> 30 days since previous event
Retained placenta, metritis, milk fever, ketosis, DA	Once per lactation

Formula

Plain text:

Incidence risk (%) = cows with ≥1 case in period / cows at risk at start of period × 100

Lactational incidence risk (%) = cows with ≥1 case in first N DIM / cows calving in the cohort × 100

Incidence rate (cases per 100 cow-years) = cases / (cow-days at risk / 365) × 100

Cow-days at risk ≈ average cows in the group × days in period

LaTeX:

\mathrm{IR}_{\text{risk}} = \frac{N_{\text{cows with}\geq 1\text{ case}}}{N_{\text{at risk}}} \times 100
\qquad
\mathrm{IR}_{\text{rate}} = \frac{N_{\text{cases}}}{\sum \text{cow-days at risk} / 365} \times 100

Cases are counted after applying the repeat window. For a calendar month, an average of 1,000 cows over 30 days gives about 82 cow-years; 20 mastitis cases in that month is 20 / 82 × 100 ≈ 24 cases per 100 cow-years.

How to pull it in DC305

Disease events are herd-defined. Event names such as MAST, METR, RP, MF, KETOSIS, DA and LAME are common, but spellings, remark codes (e.g. quarter or drug in the remark) and whether a single treatment day or a whole protocol is one event vary by herd. Look them up in the herd’s private dc305-vocab mapping first.

Typical pulls:

EVENTS\2SI FOR LACT>0                  list of cows and events; prompts for the date range and which events
EVENTS\B2SI FOR LACT>0                 the same, including cows that have since left (sold or died)
EVENTS MAST FOR LACT>0\5               mastitis events, table by month
EVENTS MAST:10 FOR LACT>0\5            the same, with a 10-day repeat (GAP) window for this run
EVENTS\6 FOR LACT>0                    events table by DIM
EVENTS\UD                              transition events table (\UDR adds sold/died)
COUNT FOR LACT>0                       cows on hand, for the denominator

EVENTS covers live animals only by default: \B adds animals that have left the herd and \D gives those alone. Event tables cover the last 365 days by default. \D is not a date switch; \S sets the date range.

Repeat window (cases vs events). Each event has a GAP value, set per event in ALTER\9: a repeat within GAP days counts as the same case, not a new one. Override it for one run with EVENTS MAST:10 FOR LACT>0\5. Check the herd’s GAP for each disease event before comparing rates, or export the event list (ID, event, date, DIM, remark) and apply the repeat window outside DC305.
For incidence rate, take the average cow count over the period (daily inventory or average of month-start and month-end) as cow-days.
For lactational incidence risk, start from the cohort of FRESH events in a calving window and look for the first disease event within N DIM.

Typical range

Herd context: US Holsteins, freestall or dry-lot, herd-recorded treatment events.

Disease	Measure	Typical	Goal
Clinical mastitis	cases per 100 cow-years	25–45	< 25
Lameness (recorded cases)	cases per 100 cow-years	20–50	< 20
Retained placenta	LIR	5–10%	< 8%
Metritis	LIR, ≤21 DIM	10–20%	< 10%
Clinical milk fever	LIR	1–3%	< 2%
Displaced abomasum	LIR, ≤60 DIM	2–5%	< 2–3%

US producer-reported reference (NAHMS Dairy 2014, cows affected in 2013 as a percentage of January 1 inventory, operations with 30+ cows): clinical mastitis 24.8%, lameness 16.8%, retained placenta (> 24 h) 4.5%, milk fever 2.8%, displaced abomasum 2.2% (NAHMS trends report, table F.1.a). These are shares of cows affected, not cases per 100 cow-years.

Published ”% of cows affected” figures (e.g. NAHMS) are usually incidence risks over a year, not rates per 100 cow-years. Compare like with like.

Common mistakes

Counting treatment events as cases. Daily treatment entries can inflate case counts several-fold. Apply the repeat window.
Calling an incidence rate a percentage. “35 cases per 100 cow-years” is not “35% of cows had mastitis”; some cows had several cases and many had none.
Wrong denominator for fresh-cow diseases. Metritis per cows on hand mixes in late-lactation cows that cannot get it. Use cows calving in the window.
Reporting the most recent calving cohort too early. Cows that calved last week haven’t been at risk for 30 or 60 days yet. Only report cohorts that have completed the risk window.
Ignoring recording changes. A new fresh-cow protocol, a new person entering treatments, or a switch from paper to software can move recorded incidence more than real disease does.
Comparing herds with different case definitions. Always state the repeat window and whether subclinical (test-positive) cases are included.
Small numbers. In a 300-cow herd, one month’s DA rate swings on one or two cows. Use rolling 3- or 12-month windows.