Study shows activity and rumination monitoring systems find fresh cow diseases faster

Tamara Scully for Progressive Dairyman Published on 11 July 2017
cows with activity monitors

According to one study, activity and rumination monitors have the potential to accurately detect fresh cow diseases like ketosis, metritis, displaced abomasum and even mastitis days earlier than clinical signs can be detected.

A recent survey of 60 New York dairy farms, conducted by Dr. Julio Giordano, DVM, Cornell Department of Animal Science, Dairy Cattle Biology and Management Laboratory, showed that while most farms checked fresh cows once per day, 5 percent did no monitoring and 36 percent checked fresh cows at least twice per day. Labor concerns, lack of time and untrained personnel were some reasons fresh cow health checks were not given priority.

Armed with this data, Giordano and colleagues conducted trials aimed at exploring whether activity and rumination monitoring systems could help farmers pinpoint fresh cows that need medical attention. And, if they could, how did the detection rates of automated monitoring systems compare to direct observations of clinical symptoms? Giordano presented some of these findings at the recent Operations Managers Conference, hosted by PRO-DAIRY and the Northeast Dairy Producers Association Inc.

Collecting the data

Automated health monitoring systems have “the potential to reduce the burden associated with a health monitoring program,” Giordano said, and “ideally will tell us which of the cows will need attention.”

In 2013 and 2014, Cornell researchers in Giordano’s laboratory used a commercially available monitoring system (HR-Tag from SCR Dairy) to monitor fresh cows on a local commercial dairy farm with over 1,000 cows. The selected farm had an intensive protocol for monitoring fresh cows, which resulted in a high percentage of sick cows being identified by personnel and promptly treated. This standard operating procedure was continued even after the monitoring system was activated. Cows were monitored for several weeks prior to calving, to set baseline data, and monitored through 80 days in milk (DIM).

The collar-mounted tags generated activity and rumination data, and its software system generated a proprietary health index score, sending alerts when a cow’s overall score fell below the system’s parameter for health. Researchers were thus able to compare the system’s generation of a health alert with the actual clinical diagnosis of disease for a variety of common early lactation ailments – ketosis, indigestion and displaced abomasum, mastitis and metritis.


The monitoring tag was able to identify 93 percent of all cows with metabolic disorders, and do so an average of 2.1 days prior to clinical diagnosis. The rate of identification, and the days prior to clinical diagnosis, was greatest for displaced abomasum, at 98 percent and three days in advance. Indigestion, which was very uncommon in the herd, was detected 89 percent of the time, one-half day sooner than clinical diagnosis. Ketosis was detected 91 percent of the time, and one and a half days prior to symptoms being seen.

“The sensitivity of the system was pretty high,” Giordano said. “It seems like the system was quite effective in identifying cows with metabolic disorders.”

Metritis detection resulted in 55 percent of clinically diagnosed cows being found by the monitoring system, one day sooner than symptoms were flagged by personnel. While the earlier detection was a benefit, the low rate of identifying cows with metritis lead researchers to further investigate the illness. They found that the herd had two types of metritis: severe cases requiring a strong antibiotic and mild cases in which the protocol was to administer a less intensive medication.

The tag had identified 83 percent of the cows with a severe case of metritis, but only 49 percent of those with a mild form. The activity and rumination data of those cows with mild illnesses was compared with those of the healthy cows, and researchers found the mildly ill cows “behaved as healthy cows,” not showing significant deviations from normal levels.

“Were these cows truly affected by metritis?” Giordano asked, speculating that if they were mildly sick, the cows could potentially have recovered without receiving any antibiotic treatment.

Researchers analyzed milk production data for five days prior to clinical diagnosis, to help assess the low detection rates in mild cases of metritis. Researchers found that milk production in cases of mild metritis that were detected by the activity-rumination monitoring system deviated from the norm. But in mild cases where metritis was clinically diagnosed, but not detected by the monitoring system, milk production was very similar to that of healthy cows.

When researchers looked at milk production data for metabolic disorders, they also found that the clinically diagnosed cows that the monitoring system did not find to be ill did not deviate from normal milk production patterns.

“Cows that were identified by the system have more dramatic changes in parameters of interest” compared with healthy cows and with cows clinically diagnosed, but not detected by the monitoring system, Giordano said.

The same pattern was true of mastitis. Mastitis in general was detected in 53 percent of the cows with a clinical diagnosis. Researchers then broke the diagnosis down by causal agent, and found that mastitis caused by E. coli pathogens was detected 81 percent of the time, while that caused by other pathogens was less detectable. Staphylococcus aureus infection was only detected 46 percent of the time.

Overall, cows that were “systemically infected” were better identified by the activity-rumination monitoring system, Giordano said.

Opportunities and challenges

The activity and rumination tags gave very few false positives, at a 2.4 percent rate. There were few issues with malfunctioning or misplaced tags, with a 3.7 percent rate. Accuracy was calculated at 95.6 percent overall for the diagnosis of illness during the 80 DIM period.

For herds with less intense fresh cow monitoring, the activity-rumination monitoring systems could be beneficial in finding and treating sick cows before an illness advances. For those herds where fresh cow checks are intensive, monitoring systems might offer a chance to reduce labor and limit disruptions to cow schedules by reducing the need for human observation.

Early identification of sick cows by activity-rumination monitoring systems can cause some concerns too. If clinical diagnosis has not been made, should treatment occur even without the presence of symptoms? If treatment is given, was it really needed? Would early treatment prevent more severe illnesses from occurring than if treatment were held until clinical diagnosis?

“There is still a lot of research to do in this area,” Giordano concluded.  end mark

Tamara Scully, a freelance writer based in northwestern New Jersey, specializes in agricultural and food system topics.

PHOTO: Photo by Karen Lee.

Before commenting on our articles, please note our Terms for Commenting.