Where's the Beef II? Confounding the Public with CWD Statistics

Misusing statistics and computer simulations have become par for the course in the Wisconsin DNRs' endless quest to find foundation for their plan to exterminate 25,000 - 30,000 wild deer near Mt. Horeb. Although CWD diseased deer have now been found in four Wisconsin counties and three northern Illinois counties bordering Wisconsin, the DNR continues to perpetrate the myth that Chronic Wasting Disease (CWD) is confined to the 411 square mile zone encompassing Mt. Horeb.

Witness a recent statement of DNR Wildlife Administrator Tom Hauge: "If you got to have CWD in the state, it is relatively good news to have it contained in one spot. That is what we are hoping for."

Doing some plotting on a Wisconsin road map, the driving distance from the eastern most case in Northern Illinois to the western most case in Richland County appears to be about 150 miles. Plot a line through the known disease cases with a 30 mile buffer and you have a "new" disease zone of 4500 square miles! Yet Administrator Hauge claims testing results indicate CWD is "contained in one spot."

There is an important detail of the statewide CWD sampling program Hauge and the DNR publicity machine have chosen not to advertise. The highly touted Statewide CWD Sampling program was designed to detect a disease incidence rate above 1% of the deer population. Setting the threshold at 1% is what would enable the agency to forecast - with only 500 deer samples per county - the CWD health of that county with a 99% degree of confidence.

Lets boil this doublespeak down to some meaningful factoids that the average bloke can understand.

• Based on the more recent DNR forecast of a Wisconsin whitetail population of 1,400,000 deer, the average per county wild deer population is about 20,000 animals.
  
  
• 1% of 20,000 is 200. The testing threshold of disease is an average of 200 CWD cases per county. Ergo we are looking for rates of disease in excess of 199 cases per county.
  
  
• Thus finding no positives in a county random sample of 500 deer means we can be 99% certain - now listen closely here - that there are less than 200 CWD diseased deer in the county!
  
  
• Given that we've been told that one CWD diseased deer escaping into the wild can trigger a massive CWD outbreak, I find little comfort in knowing that there aren't more than 200 CWD cases in a given county. If you're going to nip the disease in the bud- so goes the govspeak line - your goal is zero cases. Isn't that why
  ALL THE DEER have to be killed in the eradication zone versus just bringing the population down?
  
  
• Yet there are serious practical problems with a detection program that aims to truly guage the CWD Health of a county. As the table below indicates, our sample would need to be far greater if we suspected only a handful of deer in the county might have the disease.
  
  For example, if we set the threshold at 4 CWD diseased deer in a county and we wanted to be 99% certain our county had fewer than 5 CWD cases, we would have to sample 13,674 deer! Even if we only wanted to be 90% certain there are less than 5 CWD diseased deer in the county, our program would require 8753 samples from that county.
  
  If we set the threshold 5 times higher than the above - at 20 CWD infected deer per county - and we were willing to be only 90% accurate in our testing - we still would require 2174 samples to determine if our disease incidence is less than 21 CWD infected deer per county. Only counties in the Eradication Zone had that many samples taken. Every other Wisconsin county slipped under the radar screen.

The DNR has made much over the fact they have tested over 20,000+ deer statewide - outside of the CWD Management zone. It sounds impressive, but when you cook the fat out of it, we get an average of 348 deer tested per county. As the table below illustrates, results from such a small sample prevent reliable prediction even for an relatively large population of 100 CWD diseased animals per county. And if you assume the population of diseased animals is smaller - perhaps 10 or 20 per county, 348 samples just don't come close to cutting the old mustard.

Again as earlier, when you take the lid of this CWD Sandwich served to us, you can't find the beef. If it is there at all, it is slathered in hype geared to justify a flawed program rather than present the truth.

The most optimistic and honest statement that can be made from the 2002 Wisconsin statewide CWD testing results is that it is likely that in most Wisconsin counties the Chronic Wasting Disease incidence is below 1% (or less than 200 diseased animals per county). A more candid statement would be that there is both "good news" and "bad news" from the 2002 sampling program. The good news is there appears to be no massive chronic wasting disease outbreaks outside of Southwestern Wisconsin and Northern Illinois. The "bad news" is that due to the chosen sampling methodology, sampling errors, and inadequate sampling in many counties, it is possible there are several yet undiscovered cases of CWD across the state.

Fun With Numbers II

DNR latest population estimate reveals 8 - 12,000 deer
have disappeared from the Eradication Zone!

The Wausau Daily Herald recently reported this discovery: "For months, the Department of Natural Resources had estimated 25,000 to 30,000 deer roamed a 411-square-mile area around Mount Horeb last fall where chronic wasting disease was discovered 13 months ago. . . . The agency said Wednesday it now believes, based on new helicopter surveys, that the population was probably nearly half that size - 16,400 to 17,900 deer. "

The official DNR explanation for this whitetail deer disappearance is "whoops" - apparently our aerial survey of 2002 overestimated the number of deer in the zone. That's quite a mistake - overestimating by as much as 67%. Why after years of experience estimating deer population and with the stakes so high due to CWD was the DNR so inept at estimating deer population in the zone in 2002? And what magical thing happened recently to subtantially increase their skill at forecasting?

People on the ground, living in the Eradication Zone, have a different explanation. The DNR didn't suddenly learn how to count deer accurately. The discrepancy in numbers is real and was caused by the constant pressure of
6 months of endless shooting driving thousands of deer out of the zone. Assuming a disease rate of 2%, the exodus of 8 - 12,000 deer means there are now 160 - 240 NEW CWD cases outside of the eradication zone.

Great job of containment fellas.

Table of Random Sampling Requirements for
the detection of Brucella or TB in Free Ranging Cervidae

From Beal, V.C. 1988 in Regulatory Statistics, Vol. 28 "Considerations about animal disease program evaluation and surveillance theory." 2nd edition, Veterinary Services, APHIS, USDA

Notes:

The above table and reference is drawn from the source document for the DNR Chronic Wasting Disease sampling program. They - not this writer - elected to apply a TB detection strategy to CWD.

The table presumes the use of random sampling techniques that allow each individual within the total population an equal chance of being selected for the sample.

Example. A sampling collection technique that under-sampled selected populated areas within the region would not meet this test as individuals within that under-sampled area would have an unequal chance of being selected for the sample. This would depreciate the confidence level of the results and render meaningless a finding of "no positives."

Grant County provides an example of a questionable sample collection technique that violated the random collection rule. The single sample collection station was located in the extreme northeast corner of Grant County near or on the county line. This location would have strongly favored a skewed nonrandom sample for the county for it is unlikely hunters at the opposite end of the county would have driven such a long distance to donate their heads for the sample. Every deer in Grant County did not have an equal chance of being selected for the sample.

In fact, the voluntary sample donation program itself may have violated the standards for random samples. Which is the more logical assumption: 1) that opposition to the sampling program is evenly distributed throughout the hunter population or 2) that opposition exists in geographic pockets and is of varying intensity from pocket to pocket? If the latter is true, a voluntary sampling program will not produce a random sample. If you want random, then you can't allow an individual hunter to make the decision on donation or not. Sample donation, according to some randomized or stratified formula, must be mandatory.

An average of 348 CWD Samples were taken in the 61 Counties outside of the established CWD Management Zone. Thus, on the average, Wisconsin County samples outside the Management Zone failed to meet the desired 99% confidence standard for a 1% infection rate.

Five critical Wisconsin Counties adjacent to known CWD areas had below average samples:
Rock County - 306 Samples
Walworth County - 130 Samples
Crawford County - 246 Samples
Vernon County - 303 Samples
Jefferson County - 190 Samples

If there were 20 CWD infected deer in each of these counties, the likelihood of such a small sample catching one of these 20 was pretty slim.

Like so much in connection with CWD, the answer to the question of the size and distribution of CWD among wild Wisconsin whitetails continues to be . . . "We don't know." For all the money and manpower expended on testing, the DNR continued to Chronically Waste Taxpayer and Sportsmen Dollars through poor planning and inept execution.

--Ross Reinhold
roscoe@mhtc.net

15 year Harvest Data in 70A measure effectiveness of CWD Management Plan (published April 1, 2004)

Where's the Beef I (published June 17, 2002)

McMangament & McScience (published April 25, 2003)

The Science of CWD & Chronic Wasting Disease

"Where's the Beef II?" The political management of Data and Statistics continues

Chronic Wasting Disease in Deer & Elk: a critique of current models and their application.

Articles and Papers on Wisconsin Chronic Wasting Disease and DNR proposals

Chronic Wasting Disease Symposium at the University of Wisconsin

Chronic Wasting Disease - CWD Links and Other Resources