Raw milk: The risks

After the last post, a colleague suggested that it might be helpful to highlight a few of the dangers of raw milk. So, here's a brief resume.

Drinking raw milk or eating products made from raw milk (e.g., cream, soft cheeses, yogurt, ice cream) can be dangerous if the unpasteurized milk is contaminated with any of a number of pathogens, including Salmonella spp., Escherichia coli, Campylobacter, and Listeria. 

Salmonella spp. bacteria can cause diarrhea, fever, and abdominal pain. Most people recover without treatment, but some require hospitalization. Recent outbreaks of salmonellosis associated with raw milk include those in Chester County, Pennsylvania, in December 2012 and Texas, in 2010-2011. 

Escherichia coli can cause a spectrum of disease, including diarrhea and fever, and can progress to bloody diarrhea, dehydration, and hemolytic uremic syndrome (HUS), a severe complication which can cause kidney damage and death. Small children are especially susceptible to HUS. There were outbreaks from raw milk in California in 2006 and Tennessee in 2013.

Campylobacter jejuni can also cause diarrhea, cramping, abdominal pain, and fever. The diarrhea may be bloody and can be accompanied by nausea and vomiting. Cases of campylobacteriosis associated with raw milk consumption were observed in Pennsylvania in 2013.

In fact, Pennsylvania has had several outbreaks of salmonellosis and campylobacteriosis in recent years, including recurrent outbreaks associated with the same producer dairies. A recent MMWR noted:

Repeat outbreaks from raw milk producers are not uncommon and not limited to Campylobacter. During 2005–2013, Pennsylvania experienced 17 salmonellosis and campylobacteriosis outbreaks associated with retail raw milk. Five producers had more than one outbreak during that period. Bacterial contamination of raw milk can occur even under optimal conditions; seasonal changes in bovine bacterial shedding or inadequate quality control during milk collection might contribute to outbreak recurrence. Findings here and elsewhere indicate that compliance with state regulations and increased producer awareness after an outbreak are insufficient to prevent future outbreaks. 

Listeria monocytogenes is known for causing disease in newborns, pregnant women, the elderly, and immunocompromised persons, though people outside these risk groups can also be affected. In pregnant women, listeriosis can cause miscarriage, fetal death, or illness or death of a newborn. Listeria has been found in raw milk in South Dakota in 2014 and illness and death resulted from products made from raw milk in Maryland in 2014.

Brucella spp. bacteria are a threat mostly outside the United States. Recent brucellosis associated with raw milk and related products has been described in Spain, Israel, and France, among many other nations.

Because of the dangers involved, Pennsylvania requires that menus in establishments serving raw milk and related products carry the following disclaimer:

Raw milk has not been processed to remove pathogens that can cause illness. The consumption of raw milk may significantly increase the risk of foodborne illness in persons who consume it -- particularly with respect to certain highly-susceptible populations such as preschool-age children, older adults, pregnant women, persons experiencing illness, and other people with weakened immune systems.

Other states have similar requirements, but as I argued in the last post, it's important to understand why raw milk proponents don't heed the warnings. Only then can we conceive ways to reach out more effectively.

In the meantime, just say no to that raw milk latte.

(image source: Wikipedia)

Raw milk and name calling

I've seen several bumper stickers recently espousing the virtues of drinking raw milk. Although selling unpasteurized milk is illegal in most states, you'll find a virtual counterculture that has rejected pasteurization if you search around the Internet. Depending on the website, there are claims that raw milk alleviates allergies, remedies digestive problems, and reduces susceptibility to asthma.

Many of these sources attempt to justify such views with evidence and logic that few clinicians, microbiologists, or public health practitioners would find compelling. In some ways, such views are akin to those voiced from the anti-vaccine movement: They tend to latch on to the occasional, single study with limited findings as scientific validation of their beliefs, while discounting a substantial scientific literature identifying the risks. The truth is that the practice of drinking unpasteurized milk threatens the health of anyone consuming it, especially children and pregnant women.

It's critically important to understand these views (to the extent possible) and not simply write off the people believing them as belonging to a lunatic fringe and call them dumb and stupid. There are raw-milk advocates who eventually come to realize the risks they are taking and change; isn't it better to understand that process and avoid alienating people who might ultimately do the same?

I am reminded of a BMJ Quality & Safety article about Ignaz Semmelweis and the birth of infection control. At the end there is a passage that discusses the inadvisability of trying to convince people of anything by using insults, public humiliation, and haranguing. In talking to (and about) those who advocate drinking raw milk, we should watch the rhetoric. There's a difference between communicating risk and alienating people. Good risk communication can be effective, whereas alienation probably reinforces the behavior in need of modification.

(image source: David Hartley)

Infection prevention: One size does not fit all

One size rarely fits all in life, and this appears to be true for infection control and prevention, too. In a 2011 review of hospital epidemiology and infection control in acute-care settings, Sydnor and Perl observed that:

Growing mandates and restrictions on payments have the potential to lead to increased unnecessary antimicrobial use in an effort to prevent infections, lack of time and resources to address other potentially preventable infections, and instances of individuals gaming surveillance systems (i.e., falsifying data) in order to lower reported infection rates. Broad mandates also impose a one-size-fits-all strategy, when in reality local epidemiology varies, and infection control programs need flexibility to address local problems.

That last point is key. Kirkland has described the issue with great clarity:

It seems intuitively obvious that not every intervention that has ever been shown to work must be implemented in every healthcare setting. However, too often, in an effort to identify “best practices,” guideline writers imply that there is indeed one “right size” that will fit all healthcare facilities. Although there probably are a handful of best practices (e.g., hand hygiene before patient care or the use of prophylactic antibiotics just prior to certain surgical procedures), there are many more interventions that could be considered “good practices,” useful in some settings, unnecessary in others. A better fit might be achieved if healthcare epidemiologists were to select from among these to customize their infection prevention programs. Which good practices to choose likely depends on local context.

Since infection prevention programs often involve a complex set of sociobehavioral interventions that depend on where, how, when, and why practices are implemented, guidance is needed on how to determine what interventions to adopt in specific situations. Lacking such a robust, validated decision support methodology, recent studies like that of Sadsad et al, which illustrate the importance of context by analyzing the impact of different interventions in various wards within a hospital, suggest that mathematical modeling can be an important tool. It is intriguing to imagine applying modeling tools to tailor interventions to specific situations. What is needed is a corps of people with the requisite, relevant skill sets (including inter alia epidemiology, surveillance, nursing, and ID) working together to consider the best evidence and tailor, apply, and validate models to inform the local decisionmaking process.

One size cannot fit all when it comes to infection control and prevention. We shouldn't necessarily expect that what works in one hospital or ward will in another. We need tools to help us determine what approach is likely to be best in different cases.

(image source: David Hartley)