Well, I’m back in California. Back to work, back to school, back to Amtrak. I participated in my 4th train-based fatality last Monday after the Capitol Corridor smashed into a vehicle near Richmond. I was on the next train in line, and so was booted onto the county bus as Amtrak waited for the medical examiner. Again. I can’t handle this penance any longer, so I’m sucking it up and buying a car. I don’t have the time to commute 6 hours per day, and after 3 years, I’m through….
In other news, the Real Americana Blog is back on track, and features video of a lovely trip to the pumpkin patch and Maize Maze a few weeks ago…
Like pumpkins? See more at Real Americana!
But now, back to Crude and the NRRD Project. I still have an incomplete data set, as Asha and Co. back in Iringa are finishing up the project, and are set to complete it in mid November. Still, I do have some data. 46 Households, and about 237 fecal samples. So I ran some crude calculations of prevalence. Prevalence is simply the proportion of a population with a disease. I won’t know my lab results until late December, but in the field we did observe for clinical signs of diarrhea by looking at calf butts, and also by feel. I think I need not explain further.
Since we recorded the animals with diarrhea in our 237 samples, I can do some estimations of crude prevalence, including some more interesting calculations of categorical prevalence (prevalence by age group, sex, location, etc…).
But first, why the hell would I do this? Maybe to answer some basic questions about trends in the data. For example, how close to my hypotheses does the data fit, or are my assumptions dead-on or miles away from the field reality? Or maybe because I had a meeting with my advisers and needed to show that I’ve been busy.
What am I doing here again? (Photo by M. Richmond)
Remember what my assumptions were? Me neither. But I found them in my proposal.
Assumption 1: Diarrhea is a major cause of calf morbidity in pastoral livestock herds.
Assumption 2: Diarrhea is more dangerous to neonatal livestock due to their naive immune systems.
Ready, set, go with the data…
And now, I’m proud to present the exciting equation for crude prevalence (CP):
CP = n/N * 100, where n is the number of animals with diarrhea and N is the 237 in my sample population. For you math majors, multiplying by 100 gives me a percentage. Mind boggling!
Crude Prevalence for NRRD:
CP = 47/237 * 100 = 19.8%
I have a 19.8% prevalence of diarrhea in neonatal livestock. That’s pretty high. So we check it against our other field measure of disease prevalence obtained from the survey of the pastoralist livestock managers for consistency. We asked them a few questions about diarrheal disease in their herds, the same herds we sampled. This gives us a measure of reported prevalence of diarrhea.
Reported Prevalence for NRRD:
RP = 12/46 * 100 = 26.1%
Here, 12 is the number of households reporting diarrhea in their herds at the time of the survey, and 46 is the total number of households surveyed to date. 26.1% is pretty high, but not to far from the clinically observed crude prevalence of 19.8%. So things are looking pretty good. Diarrhea is indeed a problem affecting roughly 1 in 5 animals, so Assumption 1 seems to hold up to the field realities. We just have to wait for the lab results to see what’s causing it.
Assumption 2 was that younger neonates (calves, lambs, and kids) were more susceptible to diarrheal disease due to their naive and underdeveloped immune symptoms. While sampling, we worked with a wide range of neonatal livestock, mainly calves, based on the composition of the ndama (calf) herd. I was able to do some categorical prevalence calculations to see how crude prevalence breaks down among different age groups. Again, I still don’t have my full data set, but the trend is what I’m interested in observing. Check out this table:
Categorical Prevalence by Age
Age (months) Prevalence (%)
The trend seems pretty clear and makes me happy. The highest prevalence is in the youngest animals, those less than 1 month old. Then, prevalence shows a gradual decline as the animals age (with the exception of the 5 month old group, though this exception may be a result of disproportionate sampling and other biases in the data, something that will be resolved with true statistics and the full data set. Again, these are measures of Crude Prevalence. Give me a break!). In general, the data supports Assumption 2, and makes us happy.
For the next few months I’ll be analyzing in greater depth all of the data, and developing a statistical model that is designed to shed light on the factors contributing to diarrheal disease in the neonates, and that can be used to make recommendations for the prevention of these diseases in the future. It’ll be exciting, I promise.
And check out Real Americana. It’ll make me happy.