Risk Assessment of Raw Caprine and Bovine Milk Produced and Consumed in the Willamette Valley of Oregon

Introduction:  Asymptomatic cows and goats sporadically shed zoonotic microbial pathogens into milk leading to numerous outbreaks of foodborne illness linked to raw milk consumption.  Despite these known risks, the demand for raw dairy products remains high and state laws regarding retail sale are becoming increasingly lenient.  Effective testing strategies to detect contaminated milk coupled education efforts targeting raw milk producers are necessary to reduce risks associated with raw milk.

Purpose:  Understand attitudes, knowledge, and production practices of raw dairy producers to identify opportunities for extension education and determine daily changes in microbiological quality and safety of raw milk from bovine and caprine dairies. 

Methods:  Surveys were conducted to assess producer knowledge, attitude, and production practices using commercially available software (Qualtrics, Provo UT). Two caprine and two bovine dairies in the Willamette Valley of Oregon provided sequential daily raw bulk tank milk samples for 30 days.  Milk samples were analyzed for aerobic and coliform counts and for the presence of L. monocytogenes, Salmonella spp., and Shiga Toxin-producing E. coli (STEC) using a multiplex real-time PCR assay targeting invA, stx1, stx2, and iap genes coupled with cultural methods. 

Results:  Thirty-four Oregon dairy producers responded to the survey, 16 organic and 18 conventional farms with diverse responses for record keeping, sanitation, and motivation to produce raw milk. From the 30-day sampling period, two samples from a single bovine dairy were positive for STEC and one sample from a caprine dairy was positive for L. monocytogenes. 

Significance:  Production practices, attitudes, and knowledge among raw milk producers in Oregon are variable. Clinically healthy animals sporadically shed zoonotic pathogens confirming a risk to raw bovine and caprine milk consumers.   Standard testing frequencies are insufficient to detect sporadic shedding events.