“Good Health Records”: A Necessity for Outcomes Driven Health Management

Current dairy health records lack the consistency needed for quick and easy summary and evaluation by a computer.  They are not ‘computer-friendly’.

The example below is from a 5000 cow dairy using Dairy Comp 305®.  It depicts a sample from a list of cow’s MAST events and the remarks recorded for those events.



Over the 15 MAST events shown, quarter, treatment, culture result and pen data appear to have been recorded and presumably are of interest to the dairy’s management team.  However, these remark entries clearly lack consistency:

1.  Data contained in a remark:

  • 14/15 (93%) contain quarter data
  • 8/15 (53%) contain treatment data
  • 5/15 (33%) contain culture result data
  • 5/15 (33%) contain pen data
  • 1 contains CMT data

Since data are missing from so many remarks, none of the data are useful to summarize at the herd level.  All remarks could easily contain all data if the system simply prompted for each separately during MAST event entry.

2. Abbreviations in a remark:

  • Notice 2 quarter combinations exist like BH and BL for both hind and both left quarters, as well as AL for all quarters, however cow 48 had 3 quarters affected and BH and RF was recorded.
  • Treatments are usually 2 characters, but sometimes 5 (NL and NOLVA for Nolvasan).
  • Culture result is variable as well (Colif and EC for E.coli).

The ‘Protocols’ function in Dairy Comp 305® can help standardize abbreviations used in remarks on a limited basis.  However, if you had 3 mastitis treatments and 4 quarters, 12 different protocols would be needed for each treatment quarter combo to ensure consistent entry.  Add culture result on top of that and your protocol table runs out of room for just MAST events! Again, simply prompting the user to enter needed data using pre-defined dropdown lists would make ALL abbreviations consistent ALL the time.

3. Order of data in a remark:

  • Quarter comes first in 93%, however, missing data contributes to 9 different data orders for the 15 remarks.  To be computer-friendly, the data need to be in the same order for 100% of remarks.

The dairy never recognized these inconsistencies as a problem because they were only evaluating the MAST events on an individual cow’s record to make individual cow management decisions. However, when they summarized the data as shown above toevaluate MAST events in the herd and make herd-level management decisions, the inconsistencies were obvious and is was clear such evaluation was not possible.  These inconsistencies make it impossible for computer software to be used to quickly and easily evaluate the data.  Is this why Dairy Comp 305® has a BREDSUM command but no MASTSUM command?

Dairy health records are focused on recording that a disease occurred and how it was treated, but not on the resulting outcomes.  If repro records were kept like health records, breedings would be recorded, but preg checks wouldn’t be recorded and conception rates and pregnancy rates never calculated.  Percent of cows in the dry pen and number of cows calving in a month would be used to gauge the success of the reproductive management program.

To drive home the importance of remark consistency, try this:

1.  Draw a single vertical line through the remarks in the list below to separate quarter from all the other data.



  • You will notice that the first column created usually has quarter data, but notice cow 48 has BH for both hind and RF for right front.  The RF would be excluded from the quarter data column instead showing up in the column to the right.

2.  Now, draw a second, vertical line to further parse out the data.

  • You will notice just about anywhere you draw the second line, the data are so misaligned that there are little useable data in either of the new columns created.



Until current dairy management software makes data entry system-defined rather than user-defined, you can achieve consistent remarks by following the 3 Simple Rules of Good Recording, specifically the 3rd Rule: Record the same information, in the same order, using the same abbreviations every time.

When the same herd developed written Health Data Entry Protocols detailing how MAST event remarks should be entered, their records became computer-friendly.


Now notice the following:

  1. All remarks contain the same information in the same order.
  2. No treatment is recorded as ‘NT’ so it is clear that an antibiotic was not given and remark order is maintained.

Each quarter affected is recorded as a separate MAST event.

  1. Culture result and appropriate treatment can be recorded for each quarter, rather than ‘prioritizing’ what gets recorded.
  2. Tracking outcomes at the quarter level (re-treatment, recurrence, lost quarter) is possible.




Unfortunately, current dairy management software programs do not provide the ability to evaluate the outcomes of disease episodes even if data are accurate and consistent.  For the herd above, we were able to monitor quarter or cow level outcomes by culture result, treatment or lactation and pen-level incidence of clinical mastitis, examples of which are shown below:


Percentage of clinical mastitis episodes in a month that required retreatment to achieve a clinical cure by milk culture result: coliform (COLI) or environmental Streptococcus spp. (ENS). A clinical episode was considered to have been retreated if a subsequent episode was recorded in the same quarter within 14 days of the date of that episode.





Percentage of clinical mastitis episodes that had a recurrent clinical episode (same cow, same quarter) 15-60 days later by milk culture result: coliform (COLI) or environmental Streptococcus spp. (ENS).


Percentage of cows with clinical mastitis in a month that were removed within 14 days of the event by milk culture result: coliform (COLI) or environmental Streptococcus spp. (ENS).


Percentage of cows in a pen with clinical mastitis in a month by pen.  Pen 17 and 18 are identical open-lot pens.  Fresh cows with salable milk are moved into one or the other pen based on pen space, thus the make-up of each pen is the same.  Though not perceived by management at the time, clearly the risk of clinical mastitis was much greater in pen 18, August through November 2010.