MG Testing Methods for Poultry

Can the current government testing programs for MG (Mycoplasm gallisepticum) infections in turkeys help us?  Is there adequate cost-benefit for this testing?  Is MG testing being used as a smoke & mirrors distraction that mis-leads customers, thereby helping the #TurkeyMafia maximize profits?

In previous Blog postings (see MG in Ontario and MG (Micoplasm Gallisepticum): Poultry Enemy #1 ), we discussed a number of issues on how the #TurkeyMafia have failed to properly address this infectious disease.  Is MG infections an indicator of the turkey industries success in biohazards and prevention?

A research paper describes the  MG testing methods typically used today in North America.  The paper says there were 3 houses of about 15,000 birds per house, so we calculate 45,000 birds total on this farm.  To test for MG, 150 serum samples were taken, using an unknown sampling method, and an unknown sampling source (ie. which of the three barns).  Were samples taken from the one house, all three houses, the first chickens they were able to grab, etc.?  We don't know.  The testing found there were 10 positives out of 150 samples, or 6.66% prevalence.

The section of the paper titled "MG isolation and culture" details a different sampling, this time with 10 birds per house.  It appears this occurred after the initial 150 bird sampling.  At least we know geographically where the samples were taken.   However, we are still left to guess who and how did they pick 10 birds out of a house flock of 15,000 birds.  That is 0.067% sample size.  To me, with such a small sample, you had better be VERY careful on how you choose that 10 bird sample, or the results and conclusions are likely may be misleading and spurious.  Small samples like this are always suspect, dangerous, and of questionable value.

Statistical Sampling Methods
I suggest this scientific paper demonstrates poor reporting, or poor sampling methodology, or both.  Poor sample selection leaves significant doubt about the quality, believability, and usefulness of everything else that follows.  Very unfortunate.  A statistically valid sampling protocol which is valid, useful, feasible, fast, cheap, and effective means to get a 150 bird sample collected may be as follows.

Choose a random number between 1 and 30 from a random numbers table (found in most statistical textbooks), which will be the starting bird #.  Pass the birds through an isolation chute from one area to another, counting the birds as they pass.  When you get to the bird # which is the same as the random # chosen, take a serum sample.  Thereafter, take a serum sample from every third bird passing through the chute without forcing any particular bird into the chute or specific sequence (ie. allow the birds to haphazardly enter the chute).  Continue sampling until 300 samples have been obtained.

As close as possible, every bird in the flock should have equal probability of being selected for sampling.  That's the purpose of the randomized starting point, then the consistent sequence thereafter. 

I hope to get the informed opinion of a commercial turkey farmer, or a DVM, or a statitician as to the feasibility of this sample selection methodology, so leave your comments below. Be sure to fully explain your reasons for your opinion.

NPIP-2009
The US developed the National Poultry Improvement Plan, the latest version being passed into prescriptive law in 2009.  Since about 18% of the US chicken flocks get exported, the US would be at a competitive disadvantage if they were hotbeds of MG or other diseases in their poultry flocks.  By testing and isolating infected flocks, the spread and damage from these infectious diseases is minimized.

According to NPIP-2009 on e-CFR, Section 145.43 (c) US M. Gallisepticum Clean states [emphasis added by me]:

(c) U.S. M. Gallisepticum Clean. (1) A flock maintained in accordance with the conditions and procedures described in §147.26 of this chapter, and in which no reactors are found when a random sample of at least 10 percent of the birds in the flock, or 300 birds in flocks of more than 300 and each bird in flocks of 300 or less, is tested when more than 12 weeks of age, in accordance with the procedures described in §145.14(b): Provided, That to retain this classification, a minimum of 30 samples from male flocks and 60 samples from female flocks shall be retested at 28-30 weeks of age and at 4-6 week intervals thereafter.

Section 145.11 states:

§145.11   Supervision.

(a) The Official State Agency may designate qualified persons as Authorized Agents to do the sample collecting provided for in §145.14 and may designate qualified persons as Authorized Testing Agents to do the sample collecting and blood testing provided for in §145.14.
(b) The Official State Agency shall employ or authorize qualified persons as State Inspectors to perform the qualification testing of participating flocks, and to perform the official inspections necessary to verify compliance with the requirements of the Plan.
(c) Authorities issued under the provisions of this section shall be subject to cancellation by the official State agency on the grounds of incompetence or failure to comply with the provisions of the Plan or regulations of the official State agency. Such actions shall not be taken until a thorough investigation has been made by the official State agency and the authorized person has been given notice of the proposed action and the basis therefor and an opportunity to present his views.

Section 145.14 states [emphasis added by me]:

§145.14   Testing.

Poultry must be more than 4 months of age when tested for an official classification: Provided, That turkey candidates under subpart D of this part may be tested at more than 12 weeks of age; game bird candidates under subpart E of this part may be tested when more than 4 months of age or upon reaching sexual maturity, whichever comes first; and ostrich, emu, rhea, and cassowary candidates under subpart F of this part may be tested when more than 12 months of age. Samples for official tests shall be collected by an Authorized Agent, Authorized Testing Agent, or State Inspector and tested by an authorized laboratory, except that the stained antigen, rapid whole-blood test for pullorum-typhoid may be conducted by an Authorized Testing Agent or State Inspector. For Plan programs in which a representative sample may be tested in lieu of an entire flock, except the ostrich, emu, rhea, and cassowary program in §145.63(a), the minimum number tested shall be 30 birds per house, with at least 1 bird taken from each pen and unit in the house. The ratio of male to female birds in representative samples of birds from meat-type chicken, waterfowl, exhibition poultry, and game bird flocks must be the same as the ratio of male to female birds in the flock. In houses containing fewer than 30 birds other than ostriches, emus, rheas, and cassowaries, all birds in the house must be tested.

Section 145.33(j) states [emphasis added by me]:

(j) U.S. M. Gallisepticum Monitored. (1) A multiplier breeding flock in which all birds or a sample of at least 30 birds per house has been tested for M. gallisepticum as provided in §145.14(b) when more than 4 months of age: Provided, That to retain this classification, a minimum of 30 birds per house shall be tested again at 36 to 38 weeks and at 48 to 50 weeks at a minimum: And provided further, That each 30-bird sample should come from 2 locations within the house (15 from the front half of the house and 15 from the back half of the house). A representative sample of males and females should be sampled. The samples shall be marked “male” or “female.”

Section 145.14 Testing  states, in part:

"...Samples for official tests shall be collected by an Authorized Agent, Authorized Testing Agent, or State Inspector and tested by an authorized laboratory, except that the stained antigen, rapid whole-blood test for pullorum-typhoid may be conducted by an Authorized Testing Agent or State Inspector. ..."

OMAF has stated that they have adoped NPIP.  I have therefore asked OMAF to disclose who is authorized to collect samples, their training and certified capability (ie. did they learn anything in the NPIP training they attended, if any), and their random & representative sampling methods that adopt CFR 145 for NPIP-2009 as an official policy, guideline, or program of OMAF.

Section 145.14(b)(4) of NPIP states:

(4) Any drug, for which there is scientific evidence of masking the test reaction or hindering the bacteriological recovery of mycoplasma organisms, shall not be fed or administered to poultry within three weeks prior to a test or bacteriological examination upon which a Mycoplasma classification is based.

Does OMAF's MG program ensure that this restrictive requirement is met in all cases?  How is this issue confirmed and documented in a positive manner?  For example, an extra draw of serum can be drawn on a random basis while drawing the main sample, and these secondary samples could be screened for drug and probiotic residues and metabolites.  Does OMAF have any management, inspection, or auditing program to independently and objectively verify that this requirement is met?  If not, why not?

Statistical Acceptance Criteria
Section 145.43(c) states a flock can be designated "U.S. M. Gallisepticum Clean." upon negative results from testing all birds, or testing 300 birds for flocks of more than 300 birds. 

Using a Binomial distribution, alpha=0.1 (ie. a 10% probability of rejecting a good lot; a typical industry-wide standard value) and beta=0.05 (ie. a 5% probability of accepting a bad lot, a typical industry-wide standard value), we calculate the AQL (Acceptable Quality Level) as 0.1186% of birds are MG infected and the UQL (Unacceptable Quality Level) is 0.765% of birds are MG infected, thereby creating the statistical test of passing the flock is no birds are positive, and rejecting the flock is just 1 bird tests positive.

I wonder if anybody at OMAF knows what a binomial distribution, AQL, UQL, and Type I and II errors are, and how these factors significantly affect MG testing?  If they don't, it's no wonder that we still have spurious outbreaks of MG throughout Canada, including Ontario.  Is OMAF satisfied with this sampling plan and its resulting AQL of 0.1186%, and the UQL  of 0.765%? If not, what are OMAF's comments on these calculated quality levels?

Testing Lab Capabilities
Once the samples are collected, somebody has to test them to see if MG can be detected.  Can a Chicken Farmer, or anybody else just look at the glass vials of blood and accurately and consistently see the MG in the blood?  No, we need a laboratory to test the blood.  It must be done so it's accurate and repeatable.  The above calculation assumes that the lab test is 100% accurate and sensitive.  Where is the data on the sensitivity, selectivity, accuracy, repeatability, and reproducibility of the laboratory that does the MG screening tests?

What are the qualifications, certifications, and proven capabilities (eg. accuracy, bias, repeatability, stability, linearity, sensitivity, selectivity, reproducibility, etc.) of all Ontario labs used or authorized to perform and/or produce these MG test results?  Are any of these labs third party registered (under Standards Council of Canada's authority) for ISO 17025?  If not, why not?

Trust & Verification
OMAF scientists have repeatedly suggested that everybody involved in MG testing wants accurate results.

I'm not so sure.

After a career of 25 years as a quality assurance auditor across thousands of organization and almost all industry sectors, throughout North America and the rest of the world, I am continuously reminded that "people work the way they are measured".  In other words, they do what they have to, to make their measuring sticks look good.

For example, during the Chineese Melamine Scandal, melamine plastic powder was repeatedly and purposefully added to fluid milk, infant formula, and other protein based foods in China, because the supply chain was measured on % protein content in the food, and melamine showed up as protein by the "standard test".  Supposedly, testing for %Protein in milk prevented the diluting down of the milk with water.  A solution almost always has unintended consequences.  Perhaps that's why it's best if the governments of the world stop trying to help.  In China, with the misleading %Protein tests, food samples always looked great, with 100% passing all %Protein tests, until they were caught adding melamine after some babies died.  Eventually, a total of 19 babies died, 300,000 people were poisoned acutely.  In the long term, almost everybody who survived has serious chronic health consequences.

Why did the Chinese Melamine Scandal occur?   Mainly greed, and because people work the way they are measured.

For MG in Canada, I suggest that 0% to 5% of breeder/growers will get MG tested, even if the government said No!, even if it cost them big money.  They'd do it because it was the right thing to do.  For the next 40%, they'll get tested if it's a reasonable commercial price to do so and there are business case advantages for them on a selfish 1-sided basis (save your breath on "save the world" reasons).  The next 40% will do it if it's free or heavily subsidized by the government and your pressure  them to comply.  The last 15% will refuse, or do it only while the government is standing there watching them, and won't leave them alone until they do it.  How does this compare with your reality?

I suggest the majority of turkey farmers don't want MG testing, they just want the MG certificate so they can export their birds and get their cash.  For many, if they get their profits with or without a pretty MG-certificate, they're fine with it, regardless of the true MG-status of their flocks.

Is a MG-free certificate for Canadian turkeys the same as a melamine-induced %Protein certificate for Chinese milk?  In other words, is the government issued MG-free certificate a tool for enabling fraud to be perpetrated onto the #TurkeyMafia's customers?

What are OMAF's plans for when a customer of a Canadian turkey farm buys an alleged MG-free flock with a pretty government issued MG-free certificate (CFIA or OMAF), but some birds they bought are soon discovered to be sick &/or dead on shipment arrival, and the customer's Vet reports MG-contamination in the flock?

What if soon after the first MG surprise, it happens again?

Soon, everybody is testing for MG in spite of the pretty government issued MG-free certificates which were heavily subsidized by the government for the nice boys on the Canadian Supply Management turkey farms.  Remember, these Supply Management Turkey farms and their Marketing Agency are totally self-funding; except for MG testing, and many other exceptions that aren't even listed in the fine print.

One day soon, in this atmosphere of misleading MG-free certificates, there will be a sudden realization of a poultry epidemic at hand.  Suddenly, in a panic on Bird Flu or some other issue, what they have repeatedly turned a blind eye to before, will suddenly get the full spotlight.

Suddenly, they no longer trust Canadian turkeys, nor will they trust the pretty government issued MG-free certificates, so these customers will ban all Canadian turkey imports.  Other countries soon follow suit.  A world-wide embargo similar to Mad Cow Disease might soon be imposed on Canada's MG turkeys and misleading MG-free certificates, and Canada is at the epi-center. They will accuse Canada of knowingly giving out dubious MG-free certificates so as to unfairly and callously facilitate exporting, a breach of trust by Canada of their duty to other countries.

I believe that day is coming soon.  OMAF and the #TurkeyMafia can get ready now, or wait till it blows up like Mad Cow, Avian Flu, etc.

Today, there are 531 turkey farms, 9 hatcheries, and 20 turkey processing plants in Canada which produce 160.2 Million kg of turkey each year, which adds $377.8 Million per year to Canada's GDP.  Of that, 13.1% gets exported.  Could all of that be put at risk because of phony, misleading MG certificates?

Once Mad Cow was identified as causing CJD, and prions were in the food system in 1986, how long did it take for the British beef industry to be destroyed?  I suggest it took less than a year.

Look at China and SE Asia today with its Bird Flu and the ongoing losses and poultry industry decimation.  Look at pork in the US today with their virus.  Is the #TurkeyMafia truly 100% confident that they are completely immune to MG, and all similar epidemics?

Remember, Ontario's Premier Wynne want to Double our agri-food exports before 2020, not cut them in half.  We can't double if nobody trusts us and our disease-free certifications.  If we were found or suspected of having mis-leading MG-free certificates, this will likely tarnish all government certifications, not just MG.

All of this is occurring during the current government's watch.  Are they aware?  Are the mandarins watching? Are they competent and effective if we need them to handle an outbreak?

For Ontario, we have 186 turkey farms, and 3 hatcheries; all of which produce about 63 Million kg per year of eviscerated turkey meat, which is about 45% of the Canadian total production, with a farm gate live bird value of $160.8 Million per year.  With an average live bird weight of 8.509 kg., and an average eviscerated to live weight of 77.39% (ie. an average eviscerated weight of 6.585 kg,), the average Ontario quota-based turkey farm raises 9.567 Million turkeys per year are raised in Ontario.  With an average commercial grow-out period of 95 days, there are 3.84 grow cycles per year.  Therefore the average Ontario turkey farm has 13, 395 birds on-range at any one time.  The 2012 average farm gate live price was $1.98/kg., so the quota-based farmer has $26,522 at risk to an MG epidemic at all times during the year.  If MG hits, the recommended practice is to remove birds, clean, sanitize, and sit empty for 6 weeks (ie. 1.56 grow cycles). Loss of opportunity plus the costs thrown away give a net total MPPD (Maximum Probable Property Damage) of $67,896 per MG incident.

If they don't clean & sanitize properly after an MG incident, they are likely to suffer a second incident immediately following the first.  I suggest they don't need too many MG incidents for their banker to notice a significant change in the financial status of the #TurkeyMafia.

Cascade an MG incident through the entire economy, for an MG incident will affect hatcheries, feed mills, vets, suppliers, turkey meat processors, and retail; justifying a 4 fold multiplier.  Therefore, it is estimated that one MG incident has a $271,585 hit to the Ontario economy.

What is the effect as we cascade an MG incident through the entire economy of Ontario?  An MG incident at a farm will also affect hatcheries, feed mills, vets, suppliers, turkey meat processors, and retail; justifying a 4 fold multiplier.  Therefore, it is estimated that one MG incident at a turkey farm in Ontario has a $271,585 hit to the Ontario economy.

The Avian Flu incident in Fraser Valley, BC is estimated to have cost in excess of $300 Million.  Only $65 Million of this was a direct expense to the government under Health of Animals Act.  Is this not a subsidy of the Supply Management system, contrary to their insistence that they are 100% self-funded?

Isn't there significant Moral Hazard associated with these obsolete government programs?  Today, the #TurkeyMafia takes the risk as they please.  If they get caught, the public pays the cost.  If the #TurkeyMafia's gamble works for them, they get to keep 100% of the resulting profits.

What is your opinion about this $271,585 per incident MPPD calculation?  When you multiply that single average loss by the average number of MG incidents per year (confidential data to OMAF), we can only guess the total Province-wide annual loss.   Where does that loss sit on OMAF's priority list?

StatsCan reports in CANSIM Table 004-0004 that in 2011, there were 2,767 turkey farms with an aggregate total of 8,221,500 turkeys, and therefore had an average of 2,899 turkeys per farm.  We conclude that the SM turkey farms are 4.6 times bigger than the average Canadian turkey farm. 

Turkey Farmers of Canada reported in 2012 that they had 531 turkey farms under the Supply Management quota system.  We therefore calculate that there are 2,236 Small Flock quota-exempt turkey farms in Canada, which is 80.8% of the total number of Canadian turkey farms.  Again, the Supply Management system has constructed another apartheid system of the minority to overpower and oppress the majority.

We will have to wait for OMAF's answers, and whether any of the dystopian predictions above actually come true.