So should we not bother with X-rays, or only use PennHIP? Seems to me a certain vocal list member recently said that she's run into breed type issues breeding only low DI dogs. Maybe we just have to take old dog stiffness as a fact of life. Personally, I have 5 generations of my lines - oldest is over 13 years old right now. No rampant signs of pain when walking, so not sure what to make of this study (all have tested OFA good over the years in this particular line).

Oh boy... This is going to be hard to explain. You have to put on an unbiased scientific hat to read the actual study and understand the results. First of all, this study compared PennHip and OFA for predicting hip dysplasia in an individual dog, NOT in the dog's progeny. PennHip has always been superior to OFA for that purpose. What would be most useful to breeders but also what we don't have in this study is which test is the more accurate predictor for breeding. Secondly, the study makes some unsubstantiated claims that actually go beyond what was proven. This is a population of dogs bred specifically to be susceptible to hip dysplasia - all but one of the dogs in the study eventually developed hip dysplasia. We know that this is not reflective of the general population of Labradors. And that leads to the obvious bias of the study, for example the statement "of interest is that the average DI of the sample of 48 dogs was 0.54 compared to the current average Labrador retriever DI in the PennHIP database of 0.49 (N = 20,500), which indicates that the OA susceptibility of the study sample is not far removed from the OA susceptibility of Labrador retrievers at large (Unpublished data, PennHIP database, 2010)." The study draws conclusions based on non-published data from a non-scientifically selected database, and summarily discounts the fact that the incidence of hip dysplasia in the general population of Labradors is far lower than the study's 98%, no matter how close the average DIs are compared to the PennHip database. If you read this fact from a different angle, you could also claim that since the DIs are so close, and yet the incidence of hip dysplasia is so different in the two populations, perhaps PennHip has too many false positives?

I really want to be able to use PennHip because I believe that it could be a more accurate predictor in progeny, but we do not have a pedigree database that we would need to be able to best make that complex assessment in our breeding program. We have that pedigree database, however flawed, with OFA.

Bottom line, the answer is NO, don't stop doing x-rays. Push for these researchers to stop wasting time and money comparing OFA and PennHip, and to do studies that really help us make better breeding decisions.

I think Penn Hip solicited and paid for the study.

Purina funded the study.

Thank you, Nancy. Just because an article is published does not mean it meets the unbiased and rigorous methods that would make the findings generally applicable across varying populations. (That's a mouthful).

The FIRST question I always ask is WHO is the primary researcher and does that person have an affiliation that might skew the results. Gail K. Smith is associated with the PennHIP organization and would be expected to only author a study that found that method favorable. If the findings were not favorable, I am sure that no article would be submitted to prove the point

Secondly, you have to determine if the purpose of the study was met and the methods selected were specific to that purpose. The statement of objective was pretty general:
To determine the chronology of radiographic signs of canine hip dysplasia (CHD), specifically joint laxity and secondary osteoarthritis (OA).

This presumes that other manifestations of hip dysplasia (angle of femoral neck, femoral head shape, etc.) were not considered - even though they are generally considered markers for abnormal (dysplastic) development. These traits are not specific to PennHIP scoring, so were ignored. I question the conclusion that early signs of dysplasia evident in these other markers are unimportant in predicting later outcomes. However, since the primary researcher is associated with PennHIP, ONLY criteria that could be determined through DI evaluation were included in the study. This seriously limits the applicability of the findings to the general population - and also brings up serious questions about the underlying motives for the research in the first place.

This leads me to question the reason for comparison of OFA evaluations to PennHIP evaluations. This appears to be the REAL purpose of this study. The title would make one think that since all of the dogs were dysplastic at time of death, it is simply marking at what age radiographic signs appeared. That was not the case. It specifically compared the PREDICTIVE status of the methods, not the chronology of development of the actual OA. My take is that the title, the stated purpose and the population selected for evaluation were all "skewed" to represent conclusions that were not generally applicable.

Sorry, PennHIP...... just another of your attempts to deliver a black eye from behind the mask of real science. Be satisfied with having developed a methodology that many find useful for reducing laxity in their dogs. Taking pot-shots at OFA doesn't become you - especially when you spend a lot of public-supported grant money to finance your feud. So many organizations in the dog world would be more respected if they realized that they can be COMPLEMENTARY and need not be COMPETITIVE.

Thank you Nancy and Maureen for adding information that makes us think about the article, rather than just simply accept the findings as facts.

Yes, thanks so much for providing a way to discount a Purina funded study that took over a decade to complete and that was reviewed and published in a reputable journal. Perhaps you should summarize your critiques and write a letter to the editor of the journal with your comments for publication.

Or perhaps I could write a thank you letter to Purina for paying PennHIP to prove that feeding has no bearing on eventual arthritic changes (OA). It just delays the inevitable, it would appear. Maybe that is enough to take home from this sort of cross-purpose pseudoscience.

I would also like to thank Nancy and Just Me for the comments. They echo my concerns about hidden agendas in research. Just because someone printed it does not make it accurate. As a published author, I do know how that happens

http://www.vet.cornell.edu/research/bvhip/trends.cfm

Thanks! The report appears to support using phenotype, genetic history (pedigree data) AND laxity scoring (DI or NA) as a combination of criteria for the best outcomes. As I said earlier - complementing instead of competing. It helps to get the view of researchers that don't have a "dog in the fight" - so to speak.

This is what the report supports...


"The selection index was developed into the Best Linear Unbiased Prediction (BLUP). The BLUP breeding strategy has been used successfully for genetic improvement, particularly in livestock, and has also been applied in closed colonies of dogs with substantial success."

While I am good at statistics, extrapolating a BLUP formula into a working model from a group of hip scores independently is a bit out of my league. I am guessing that is why the authors rolled it into what they call a "breeding value" score for this study.

The portion of the report that I was referencing is from page 5 of the actual study (not the abstract).

Only true if you use the entire BLUP formula.

It is not useful to question the motives of the scientists or their sponsors. Just be happy the studies are being done. Plus, keep in mind, whenever someone does a study, the first thing they need to do is to make a statement of what they are hoping to prove. It is NOT like, "Gee, I'll try this on a bunch of dogs and hope something comes up." A lot of studies are done hoping to prove something, only to learn the opposite of what they expected.

and the researchers/authors of any study report in the research publications the possible limitations of the study and all possible funding conflicts. Are any studies perfect? No, but they can still provide valuable information.

Here is a primer on how to use BLUP:
http://www4.ncsu.edu/~fisik/course/lecture9_%20blp.pdf
Bon Appetit