Raw milk lecture series wraps up

Food science professor explains the nitty gritty of risk assessment.

By Loren Muldowney

Editor’s note: This is the final installment of a four-part series that has appeared monthly in the nutrition section of our website. It is based on the recent lectures in a Seminar Series on Raw Milk hosted by The New Jersey Agricultural Experiment Station at Rutgers University. See part I: The truth about milk, part II: Lactose intolerance and part III: Got raw Milk?.

Donald W. Schaffner, Ph.D.

Donald W. Schaffner, Ph.D., is an Extension Specialist in Food Science and Professor at Rutgers, The State University of New Jersey. Dr. Schaffner's research covers quantitative microbial risk assessment and predictive food microbiology. He is a prolific scientific author and has an international reputation as a food-safety expert.

Dr. Schaffner's April 7 talk was entitled "A Risk Assessor Takes a Look at Raw Milk." Schaffner began his talk by summarizing the series to date (see links in “editor’s note”). He attended the Cox and Gebhart talks and listened to the audio of the McAfee talk, commenting that it is fair to say that all three come down on the side of "raw milk is good for you." Several critics of raw milk were invited to speak but for various reasons each was unable to accept. Since the series was originally envisioned to include pro and con points of view, Dr. Schaffner was left with a dilemma because he had expected to take a neutral point of view and address issues raised by each viewpoint.
Dr. Schaffner did not take the “con” side, but he did present a slide with links to the Food and Drug Administration (FDA) position (unequivocally negative), an Internet site called "barfblog" based at Kansas State University, and the "marlerblog" on the website of an attorney who makes a living suing food producers. With that out of the way, Dr. Schaffner proceeded to cover the subjects he thinks are important to an informed reading of the literature.

Contamination can happen

He began with a recently published study of diarrheal illness associated with a Pennsylvania dairy selling unpasteurized milk. The study appeared in the November 9, 2007 "Mortality and Morbidity Weekly Report" (MMWR), a publication of the Centers for Disease Control (CDC). Using the MMWR data, Dr. Schaffner explained why he thought the evidence by association was very strong in linking the cases of illness to the individual farm, in particular the isolation of the outbreak strain of the Salmonella organism from affected people and repeated isolation of the same strain from the farm's bulk milk tanks, coupled with the fact that the individual strain identified is relatively rare. The method used to identify the strain, pulsed-field gel electrophoresis (PFGE), provides a “genetic fingerprint” of the organism, he explained. Furthermore, he said, the timing of the illnesses paralleled the times when the dairy was selling product and stopped when the sales were discontinued. Schaffner expressed concern that some people appear to believe that raw milk has never been linked to illness, so he wanted to present one case to the contrary where he found the evidence compelling.

This led into a discussion of "case control studies," a tool used in epidemiology to identify factors associated with, and possibly responsible for, some conditions. Dr. Schaffner introduced the terms "odds ratio" and the "confidence interval" associated with the odds ratio, using a fictitious example to illustrate how an odds ratio is computed. He brought his example to life with anecdotes on how difficult it is to remember what you ate or did not eat a week ago; his point being that it is difficult for the investigators to get accurate information following what appears to be a cluster of people becoming ill, so it is to be expected that the data will be messy and require statistical analysis to extract meaning.

Regarding the statistics of sampling, Dr. Schaffner explained that a pathogen which contaminates some food is seldom equally mixed throughout all the portions, so not everybody who eats the food gets sick. Individuals also differ in their sensitivity to illness, he said, also allowing that the passage of time can and does cause the population of a microbial pathogen to change, sometimes increasing, sometimes decreasing, at a rate generally related to temperature.

Testing has its limits

Staying with the fictitious foodborne illness, Dr. Schaffner asked the audience to imagine that a food is “implicated” by a case control study, that the unconsumed food is tested for the pathogen isolated from the sick people, and that the pathogen is not found to be present. What should be concluded, he asked, following that statistical correlation does not prove causation or explain any mechanism of transmission. Maybe the food is not responsible, he said, or maybe everyone ate all the contaminated portions. The pathogen population may have decreased greatly since the food was consumed, he said, since it usually takes several days for anybody to realize that an outbreak is going on. Perhaps, he suggested, the detection method in place cannot find low levels of pathogens, which are nevertheless high enough to cause illness. The take-home message from Dr. Schaffner was that “there are limits to the usefulness of microbiological testing,” with the result that a negative pathogen test does not equal proof of safety.

Obviously tongue in cheek, Dr. Schaffner said that the best, most sensitive method of detecting foodborne pathogens is feeding people and observing the results (although he does not recommend this bioassay method).

Dr. Schaffner illustrated two different ways of looking at risk: on a population basis or on a per-serving basis, making the point that a relatively risky food that hardly anybody eats makes fewer people sick than a relatively safe food that everybody eats. Using bar graphs, he showed how the two ways of looking at risk can lead to opposite conclusions about which food is “safer.” On a population basis one can be safer, but that same food can carry a much higher risk on a per-serving basis.

Regarding the mathematics of pasteurization, Dr. Schaffner explained that some time interval is sufficient to inactivate 90 percent of the initial pathogen. Since the initial numbers are so high (in the billions), logarithms (powers of 10) are used in the literature. In response to a question from the audience, Schaffner acknowledged that the technology is indiscriminate: not only pathogens but all susceptible organisms are reduced. While very effective, he said, pasteurization will not kill spore-forming organisms. Dr. Schaffner believes that situations described as failures of pasteurization are more likely the result of recontamination after the pasteurization than of the pasteurization itself not working.

Finally, responding to a question about competitive interactions between normal benign bacteria found in raw milk and some potential pathogenic organisms, Dr. Schaffner said that studying the microflora of raw milk would be enough to keep a large team of food microbiologists busy for many years. That subject is “very complicated,” and much remains unknown, he said.

Loren Muldowney holds a bachelor's degree in biochemistry and a master's degree in soils and water from Rutgers University, where she is continuing studies in sustainable agriculture. Her backyard is home to nine compost units, organic gardens and is a certified wildlife habitat.