ANALYSIS OF SOMATIC-CELL COUNT DATA BY A PEAK EVALUATION ALGORITHM TO DETERMINE INFLAMMATION EVENTS

Increases in SCC are an expression of inflammation events in the udder. Inflammation events are sporadic, of variable amplitude and duration, and can be analyzed by computer programs designed to evaluate pulses of hormone secretion. Baseline values for SCC, which take into account long-term trends, were calculated using the PULSAR peak evaluation algorithm. An inflammation event was defined as an increase of log2(SCC) of at least 1 unit from the preceding data point if the observed value exceeded the baseline value by a threshold value. Incidence rate, duration, and amplitude of inflammation events were calculated. Weekly composite milk SCC from individual cows from two Florida dairy farms were recorded. Data were analyzed for effects of season (summer and fall) and bST in two separate data sets. Incidence rate of inflammation events was higher in summer than in fall (4.31 and 2.91 events per cow x year). In one of two data sets only, duration of inflammation events was longer in cows treated with bST (2.4 +/- .2 vs. 1.6 +/- .2 wk). In contrast, least squares analysis of variance of log2(SCC) did not detect difference due to season or bST. The use of the peak evaluation algorithm led to the detection of seasonal and bST effects on inflammation events in cows.