Experimental Design
Dip teats of half of the quarters after each milking in the experimental barrier teat dip being tested. Experimental design can be either a split-udder or split-herd design. Determine the number of new intramammary infections (IMI) in quarters with teats dipped in the experimental product and in quarters dipped in a positive-control germicidal teat dip. Positive-control teat dips should previously been shown efficacious compared with not dipping in either experimental challenge or natural exposure trials.

Selecting Experimental Herds, Cows, and Quarters
Trials should be conducted in at least two herds. Conduct trials in herds where whole-hearted cooperation of managers to comply with experimental procedures can be attained. Monitor milking equipment and milking management practices carefully and regularly to minimize machine-mediated infections. This is especially necessary in commercial herds where constant supervision by the investigator will not be practical. All quarters are eligible except quarters with teats that are deformed due to previous injury. Exclude quarters with teats that are injured during the trial for the remainder of that lactation; such quarters may re-enter the trial after a dry period if the injury has healed.

Teat Dipping
Dip teats of half the cows in the experimental product in a split-herd design. Cows in the remainder of the herd serve as positive-controls. In this situation, take care to ensure that cows are balanced by: 1) parity; 2) stage of lactation; and 3) bacteriological status of quarters. Ensure that the two groups are milked in the same facility. When a split-udder design is used, either dip two diagonal teats or teats on either the right or left side of each udder with the experimental barrier product. The other two teats on each cow are dipped with the positive-control germicidal product. Apply teat dips immediately after milking machine removal.

Sampling Schedule and Procedures
Collect duplicate or two consecutive single quarter milk samples to determine existing infections in the herd at the beginning of the trial. Obtain a third sample when results of the first two samples do not agree. Culture single quarter milk samples monthly during the trial. A second quarter milk sample should be collected within seven daysfrom all quarters in which the bacteriological status of the gland had changed from the previous month. Culture duplicate samples from cows calving and herd additions prior to inclusion in the trial. When any quarter develops clinical mastitis, collect and culture duplicate milk samples from all four quarters before any treatment is administered. Collect and culture duplicate milk samples from individual cows at drying off or upon leaving the herd. Collect all samples consistently either immediately before or after a regular milking using standard procedures (1,3).

Criteria for Diagnosing Infections
Examine all milk samples bacteriologically and identify organisms isolated according to standard procedures (1,3). In determining that a quarter is free of infection when it enters the trial, no pathogens may be recovered from two of the initial samples. Diagnose a new IMI when the same bacterial species is isolated from: 1) both of the duplicate samples taken from clinical quarters or; 2) two consecutive samples taken during the trial. The status of a quarter should be recorded as a bacteriologically-negative clinical case when bacteriological results of duplicate samples from a clinical quarter do not match. Clinical mastitis and IMI diagnosed during the first seven daysof lactation should not be included in data analyses of teat dip efficacy.

An individual quarter is eligible for only one infection per bacterial species during a lactation (i.e., only one Escherichia coli infection per quarter per lactation). Quarters infected in one lactation may be included in the trial in the subsequent lactation if it is determined that the infection was eliminated during the dry period either spontaneously or as a result of therapy.

Data Presentation
The report of a trial should include: 1) duration of the trial; 2) number of quarters in the trial at the onset and on the date of each monthly sampling; 3) number of total new IMI, categorized by bacterial species or type, that occurred in positive-control and treated quarters; 4) the percentage differences in total new IMI between treated and control quarters and for each bacterial species; 5) the number of new clinical cases, categorized by bacteriological status, that occurred in control and treated quarters; and 6) the percentage difference in new clinical cases between treated and control quarters.

Data Analyses
The purpose of the trial must be decided a priori when using a positive-control. The purpose is most often to determine either: 1) if the experimental product is more efficacious than the positive-control or; 2) the efficacy of the experimental product does not vary from that of the positive-control by greater than a predetermined amount.

Environmental pathogens: Although germicidal teat dips can effectively reduce incidence of new IMI by the contagious pathogens Staphylococcus aureus and Streptococcus agalactiae, germicidal teat dips do not reduce rate of new IMI by environmental pathogens (3). Therefore, to claim an experimental barrier teat dip is efficacious against environmental pathogens (i.e., Escherichia coli, Klebsiella spp., Streptococcus uberis), the efficacy of the experimental product should be greater than that of the positive-control germicide. To determine if the efficacy of an experimental product is greater than that of a positive-control, the hypothesis is formulated and tested as if teats on control quarters were not being dipped. Data must express the relation between quarters becoming infected in quarters treated with the experimental teat dip and in quarters with teats dipped in the control product. Differences between percent quarters becoming infected in treatment groups can be tested where t approximates a standard Student's t statistic:

The denominators n₁ and n₂ can be expressed as the summation of either quarter-days, quarter-months, or quarter-years dependant upon the experimental design. The percent reduction in new infection rate in the treated group compared with that in the control group is expressed as: 100 [(x₁/n₁)-(x₂/n₂)]/(x₁/n₁).

Contagious pathogens: The efficacy of an experimental barrier teat dip against Staphylococcus aureus and Streptococcus agalactiae should not be less than that of the positive-control teat dip. Experimental products may be tested to determine if efficacy is "equal" to that of the positive control. Equivalence can be evaluated by constructing a confidence interval on the difference between two proportions. A 95% one-sided confidence interval for the difference between proportions can be computed using the normal approximation (1):

Trial Duration
Duration of each trial should be at least 12 months to include each season of the year. The length of a trial required to demonstrate efficacy of a teat dip will depend on the number of quarters available initially and the rate of new IMI in the control and treated groups. Guidelines for estimating the number of quarters required, the probable duration of a trial and the point at which a trial may be terminated after 12 months are those detailed in (1). These guidelines are applied to total IMI and separately to each species of bacteria against which efficacy is tested.

References
1. Hogan, J. S., D. M. Galton, R. J. Harmon, S. C. Nickerson, S. P. Oliver, and J. W. Pankey. 1990. Protocols for evaluating efficacy of post-milking teat dips. J. Dairy Sci. 73:2580.

2. Current Concepts of Bovine Mastitis. 1996. National Mastitis Council, Inc., Madison WI.

3. Microbiological Procedures for the Diagnosis of Bovine Udder Infection. 1990. National Mastitis Council, Inc., Madison, WI.

Protocol developed by the National Mastitis Council Research Committee, 1997.