Many people seem to have missed my point. ... I am only interesting in the lack of change in the percentages insofar as they point to trouble with the test itself and motivate people to examine whether the test merits use and/or whether the test could be improved to be more useful.

Really, Kate... don't you think MOST of us know what your point is? I certainly do after all your tirades against OFA and support of PennHIP despite the absence of evidence for either side of your comments. Once again, you state there is a lack of change in the percentages - as if that were a fact! That is NOT true and all your misuse of statistics does not make it so. There HAS been improvement - the current rate of ED is about half of what it was before OFA started evaluating elbows. If you would just process the REAL facts, you would see that the OFA tool HAS demonstrated its applicability. That should motivate most breeders who have been loosly utilizing elbow evaluations in breeding choices to CONTINUE that practice. The REAL figures should also lend a bit of confidence to the practice and result in even better outcomes.

I want to know how much of the heritability of elbow dysplasia is measured by the test.

Since the test is for phenotype, that same arguement could be applied to ANY evaluation of phenotype. How much of ear length is heritable by observing the parents? How much of eye color? How much of HIP status - whether OFA or PennHip evaluation? As long as we are measuring complex phenotype, there can not be any direct correlation to explicit genotype for any of the polygenic features. We can LOOSLY extrapolate a correlation betwen selection for phenotype and infusion of genotype for all these traits. That is the basis of hip, elbow, heart and eye evaluations for health screening. It is the basis for breeding to dogs that have done well in some venue in the hope that their offspring will also excel. They are ALL measuring traits that may be inherited very differently than we suspect.

So.... what do we do? Throw out all testing?? Just breed any dog to any dog? If you look at the figures for OFA and PennHIP, you will note that there is some correlation between selection for phenotype traits and what those dogs produce. It is not always dependable, but it is better than nothing. The same is true for elbows - REGARDLESS of the testing authority. Using phenotype evaluations to assist breeding choices results in more consistent phenotype over time. It is a flawed system, but it is the best we have for now.

Breeders demanded this kind of data before adopting the Optigen test.

Baloney. There are still people who won't trust DNA tests thanks to Optigen

The incidence of elbow dysplasia in the Lab population and the trends in the incidence in dysplasia are only incidental information to me.

Then quit harping on them and misusing statistics to misrepresent them. The CHANGE in incidence of ED over the last 18 years with OFA being the only evaluator for most of that time SHOULD be important to you! It should be important to ALL of us! That trend is the primary means of measuring the phenotype output by selection of phenotype input. The figures I gave are very important and show that we ARE significantly altering the breed toward better elbows. You can dislike the methodology of testing - as I do PennHIP - but you can not ignore the outcomes. That is what applied science is all about - not what happens in the test tube, but in vivo outcomes.

Perhaps your doctoral program did not instruct you about test validation.

Test validation in living systems is measured (as I stated in the last paragraph) by outcomes. I see the trend for improvement in ED to be a very clear outcome. We may never know what percentage of each type of ED is revealed by the test, but we can clearly state that TESTING for the overall manifestations of dysplasia in phenotype and using that information for breeding choices DOES show positive change in the overall population phenotype. That is all we can ask of a phenotype test that is used in a polygenic setting.

I can say, however, that my program did teach me how to use statistics correctly and my basic Algebra class gave me the math tools to see the flaw in your ongoing accusation that ED rates have not changed over time. Be a big girl, Kate, and just admit that you were using fuzzy logic and your conclusion was invalid. It would do a lot more for your credibility than these last two posts of yours. We are all human and make errors. Trying to defend your position (and continuing to state there has been no change in ED rates) starts casting doubt on the validity and objectivity of all your other comments.

And here I was hoping, when I read the title of the original post, that this was going to be a fun expose on Cialis or Viagra!!

"I can say, however, that my program did teach me how to use statistics correctly and my basic Algebra class gave me the math tools to see the flaw in your ongoing accusation that ED rates have not changed over time. Be a big girl, Kate, and just admit that you were using fuzzy logic and your conclusion was invalid. It would do a lot more for your credibility than these last two posts of yours. We are all human and make errors. Trying to defend your position (and continuing to state there has been no change in ED rates) starts casting doubt on the validity and objectivity of all your other comments."


Maureen YOU make absolutely no sense. Given a large enough number equaling statistical significance a subset can certainly be included in the whole set.

Given that the total is over 40,000 cases and the percentage is 11.5% affected then taking out 3000 cases in which that total had a percentage of 11.5% affected then the remainder 37000 would also have a percentage of 11.5%. Thus you can compare the two and see that the 11.5% is unchanged.

I will try to break it down so that the logic and math are easy to follow. You may not be the only one caught by what "appears" to make sense, but draws a completely invalid conclusion. Here goes.

The ED figures on the OFA site are reported like this for BIRTH YEARS:
1990 thru 1993 3,492 Labradors evaluated, 11.5% ED
1994 thru 1997 8,915 Labradors evaluated, 12.3% ED
1998 thru 2001 10,703 Labradors evaluated, 10.1% ED

Let's add the extra figures I was given yesterday. For all elbows evaluated on dogs born BEFORE 1990, the RATE was 15 %. The RATE of ED went down for dogs born the next 4 years. Then the actual RATE of ED went up for dogs born 1994-1997 (more were being evaluated as well). The RATE went down again by more than two points for dogs born 1998-2001. The latest figure I was given yesterday for dogs born from 2005-present is 8%. If you want to draw any conclusion about the CHANGE of RATES, you must only compare the subsets. The trend is for steady improvement in RATES in recent years.

The overall composite percentage for all years gives NO information about whether the rate is going up, down or staying the same. You can not compare one subset to the composite and draw any conclusion or make any comparison. The figures are mutually exclusive.

Is it possible, and it seems to me that it is, that both methods of evaluating the data are valid? And that both methods have their advantages and disadvantages as both are imperfect analysis tools?? And further that there really is not yet enough longitudinal data to draw conclusions with high levels of confidence??? Given the amount of current data, I do not see how anyone can be all that authoritative at this point in time. At the same time, I can intuitively respect both hypothesis. Only time will tell....???

I should have written that I can respect each hypothesis both intuitively and logically.

No. You can not compare a subset of years to a composite of all years and draw ANY conclusion about trends over time. It may SEEM intuitive, but it is both logically and statistically invalid. You can only compare the rates longitudinally --- from year (or year group) to year to see how they change over time. That is the only way to determine IF there is a change and the DIRECTION it is moving.

This may help to see how invalid the comparison really is. If you chose another subset of years (1994-1997) at 12.3% and compared that figure to the composite, the composite rate of 11.2% would be less. If you chose to compare it with the next group of data that is 10.1%, the composite rate of 11.2% would be higher. The comparison of any subset to the composite is meaningless and does not allow you to draw any conclusion about the TREND of ED rates over time. Just follow the logic and the math and you will see that there is no common basis for meaningful comparison.

The only valid conclusion based on REAL facts is that the RATE of ED is going down in recent years and we have cut it nearly in half in the last 18 years. I can't figure out why anyone would go out of their way to misconstrue the figures to make it seem that we have not - I am personally pleased to think our breed is getting sounder every year!

I still think what I wrote has some validity. Maureen and Kate are arguing two different things. Maureen is debating the data itself. Although she did initially make an incorrect assumption based on the data, Kate is focused on the data collection procedures (the actual analysis of the x-rays, etc.). No analysis of the data can be correct if the means of collecting that data is flawed. And again, as more data is collected, we can have much more confidence in our analysis. I do not think it is correct so conclude that OFA elbow are either helping or hurting at this point. Personally, at an intuitive level, as my hope would be that over time the processes will improve as we learn more and ultimately this will help the breed.

I had to laugh. This is my horoscope for today, so you will understand why I will not be posting any more responses on this thread for now
*****
The more you try to drive your point across, the more it will be met with harsh opposition. This is a better day to sit back, observe, and collect data. Emotions are running on a high level, and people's minds are not necessarily acting in a rational manner. It will do no good to try to reason with someone who is determined to fight. Save your ammunition for something else.
*****

My ONLY purpose in starting this thread was to determine HOW Kate arrived at the obviously (to me) incorrect conclusion that there had been no improvement in ED scores, to demonstrate the falacy in that assumption by analysis of her method and to present the REAL facts of the subject directly from OFA. At this point I have explained repeatedly how the current statistics on the OFA site were misused to draw an invalid conclusion. I have given a glimpse of the figures that will be published there soon. I have led the readers through the correct logic to the proper conclusion that ED rates ARE decreasing. If people have other issues or questions about OFA, ED, or breeding choices, start another thread. I think we have thoroughly beaten this one to death

"The only valid conclusion based on REAL facts is that the RATE of ED is going down in recent years and we have cut it nearly in half in the last 18 years."

No this is not true. The only valid conclusion based on the percentages you posted is that the rate of ED BASED ON OFA EVALUATIONS has gone down in recent years. It is impossible to say if OFA evaluations reflect the general population since many animals are never evaluated and those that won't pass are often not submitted. I find that personal experience dictates more accurate comparisons.

I don't understand why or how Maureen has to be so authoritative about absolutely everything - my it must be wonderful to know it all and never allow anyone else to heaven forbid try and dispute HER vast knowledge!

Boy, do I hesitate to get involved in this can of worms. But I think one point has te be emphasized. OFA evaluation of hips and elbows are NOT tests, in the sense that the PRA, EIC, and CMN tests are tests. They are evaluations of phenotype and therefore are more similar to evaluation of coat texture, rear leg angulation, etc. The procedure gives us information about phenotype that we can't see from looking at the external structure of the dog. And the information bears directly on the functionality of the dog, which is why most of us think that it is of enough significance to be worth spending some money and taking the results seriously in planning our breeding program. If I care deeply about any particular phenotypic trait, I am going to avoid a dog that doesn't have or doesn't produce that trait. I care deeply about having an orthpedically sound dog, so I avoid breeding to dogs with non-passing hips or elbows. I have not (and I'm knocking on wood here) had any bad elbows in about 10 years since I have been screening for and eliminating non-passing elbows for three generations now, and I started screening in about 1990 after I produced a litter riddled with them. You can get rid of or at least severely reduce the incidence of poor heads, tailsetes etc., by concentrating on eliminating them from your breeding program by the old-fashioned way of selecting parental stock that do not have the trait. Orthopedic problems are no different. You just have to decide what is more important to you, soundness or other characteristics that you are selecting for. I've been fortunate that the dogs I had to eliminate for bad elbows were not the ones I most desperately wanted to keep because of their other traits. I haven't been tempted to use a dog with a grade 1 elbow, for instance, except as a stud dog from someone else's breeding program, and I have decided against several stud dogs because of an elbow that didn't pass even though there may have been an injury involved. Other people have had to make different decisions. But to ignore orthopedic evaluations because they are not definitive tests means that you are ignoring part of the phenotype and telling the rest of us that it is not important in your breeding program.

Maureen, Your scorn for my intellect and background ring loud and clear. I don't generally find that kind of scorn to be characteristic of trained people. I participate in a national board of professionals in which we discuss in great detail how tests can be validated. I teach in a major medical school where we routinely discuss the validation of tests and their data. And none of these people stoop to the kind of personal attacks that characterize your posts to me. If you are so sure you are right, why not just stick to the facts.

I am not harping on my analysis; I have said that it is a reasonable analysis based on the available data on the website. Nothing more. Nothing less. I used the analysis only to point out that these tests have never received the kind of statistical scrutiny that the Optigen test, or even the EIC test has gotten before we adopted them. I'm not interested in defending it; all manner of people use these kinds of comparisons. For example, my financial advisor just sent me a first quarter earnings report and compared it to both first quarter earnings for last year and total earnings for the year. Simple stuff, really.

Peggy, there are many kinds of tests used by many professional groups; they are not genetic tests, but nongenetic tests too can be validated. In fact, I had to validate and publish a paper and pencil test in order to even begin collecting data for my dissertation years and years ago. There are procedures for validating and examining the heritable components of all manner of tests.

My concern is lack of validation for the OFA tests themselves--not the outcome data. Show me the validation of the effectiveness of these
tests. People who pay as much money for these deserve to know how these tests stack up in defining heritability and how much these tests can contribute
to breeding decisions. (And Maureen, this data is available for PennHIP.) Kate

Yep... my horoscope was right on target!

"Maureen, Your scorn for my intellect and background ring loud and clear. I don't generally find that kind of scorn to be characteristic of trained people."

What is Maureen trained in? I didn't think she was in the medical profession.

Why do you 2 have to always take these subjects up on public boards instead of privately? The dart throwing is not appropriate behavior and the length of postings is tedious to read. We have to sort out the private issues amongst the 2 of you before reading about ED.

Can't you take this into email, an instant messenger or on a telephone call?

A small addition to this discussion which has nothing to do with actual "math" talk. (breathing a sigh of relief)

One of the bigger problems I see in trying to effectively evaluate the data from OFA on elbows, are those who are not in the database because they do one of three things, get rid of young puppies who limp (send them to pet homes,) don't send in xrays of elbows if they believe they are going to fail or don't publish results if their dog fails elbows.

It would be interesting to see what percentage of labs have hips and not elbows, or elbows and not hips (I bet this number is pretty small) If we made the assumption that all of these dogs failed elbows (or 75% did or pick some number) and included them in the data, what would we then see for change over the last 15 years?

Is the change, if there really is a change, a reflection in better breeding secondary to better information or is there some other explanation we are missing? Stats are what you make of them, there are many ways to mathematically handle numbers and numbers sometimes lie.

Another problem, albeit a lesser one, that I see with elbows, is that any change in the elbow is read as ED, and I am pretty sure some of those are not true ED but rather other events that led to changes in the elbow that are environmental rather than inherited.

Bonnie

I don't think anyone scorns true intellect; what they scorn is ego, and ego has certainly landed here.

Bonnie, Given the fact that the elbow dysplasia is judged by evidence of arthritis and sclerosis, false positives (e.g., enviromental accidents and overuse/overloading injuries) are inevitable, especially in large bodied and physically active dogs.

Kate wrote: "Bonnie, Given the fact that the elbow dysplasia is judged by evidence of arthritis and sclerosis, false positives (e.g., enviromental accidents and overuse/overloading injuries) are inevitable, especially in large bodied and physically active dogs."

Do you think Kate, if we did our elbow evaluations at 6 months rather than 2 years we would cut down on false positives or would we miss too many later developing elbow issues?

(am hoping you are sleeping well and pictures will be available soon!!)

Bonnie

Bonnie, I really have more questions than answers at
this point about elbows. I'm really trying to figure
out what to do next and gathering information--for myself, my own breeding program. I care about the breed as a whole, but it is my own puppies who really motivate me to be unsatisfied with the status quo.

It would really help if OFA analyzed their available data and could give us some idea about the rate of false positives in their methods and how much their current elbow x-ray really tells us about inheritance of elbow DJD. Clearly, we need to think more broadly about elbows than just the elbow x-ray. I'd just like to have some clues about how much more broadly we need to think.

I know that some orthopedists recommend 6 month evaluations so that you can find correctable problems before arthritis develops. If early diagnosis could prevent a life of disability, a 6-month evaluation seems like a good thing to me. I'm considering it for my litter on the ground.

Maybe we could also lower the false positive rate by taking better measures to prevent pediatric injuries. I'm wondering what canine rehabilitation specialists might be able to tell us about prevention of pediatric
injuries. I'm already following Chris Zink's lead
about exercise for my babies. But if I knew more
about how injuries occur, I might have a better idea about how to prevent them.

Another area for exploration might be the contribution
of shoulder and rear structure to overloading elbows.
Mary Roslin Williams wrote something about the fact
that the elbows of American dogs stick out more than
those of English dogs. I think there might be
some important clues in what she said. She really
talked about judging soundness of the whole dog
rather than parts. Perhaps elbow disorders are
heritable in ways different from what can be judged
strictly by an elbow x-ray.

I'm trying to figure out how I can use her advice
to lower the rate of elbow problems in dogs I breed.
One of the reasons I liked Joshie as a sire was the
fact that he came from a sire and dam who were both
very active dogs with great structures and movement, and who have been sound dogs into adult life. I found the soundness of his parents to be more compelling than his normal elbow evaluation.

So as I'm spending a lot of time observing those
precious lives in my bathroom, I'm thinking about
their little elbows. (Yes, getting a bit more rest
too. Wink.)

We've got some cute pictures coming soon. Hubbie took a course on documentary film work and he seems to be making a video of the puppies for his next class this fall. Kate

Kate, Mary Roslin-Williams made the observation about the elbows on American Labradors compared to the British more than 30 years ago in April, 1976. I think that point should be made clear.

The first solid, factual and sane comment in this thread. Thank you MSW.

"I'm already following Chris Zink's lead
about exercise for my babies."

Kate, could you elaborate on this please, i.e. age, frequency, and type of exercise? Thanks.

Margaret, Do you think that things are different now
than 30 years ago in regard to elbows? I actually have heard judges make similar comments now about "out at the elbows." Of course, I have also heard judges talk about being "tied in" at the elbows and talk about straight shoulders and rears, which I may also play a role in elbow overloading.

But I wasn't actually implying that MRWs comments about elbows sticking out were the "clue" to understanding elbow issues but more that elbow structure in the context of the whole body and even whole dog was the "clue." She has written about evaluating the whole dog several
times and about overall soundness.

Kate

Maria, What age baby are you wanting? At the moment,
my mind is stuck at 3-16 week olds. But I am planning ahead for the next several months in which I have the litter. I talked with Chris about what we're doing, and she thought it appropriate for puppies we're raising.

We have constructed tiny agility equipment (e.g., dog walk, tunnels, teeter boards, stairs for climbing) that has safety built into it by virtue of its size
and safety rails. No falls really possible. We also start swimming lessons in our pond, which we constructed so that puppies could walk into the water easily; we block use of the dock.

Their agility equipment play areas are packed sand, easy for walking but still soft for falling. (They also like to dig in the sand). They also have a grassed play area. No concrete or asphalt anywhere.

Puppies are supervised most of their time outdoors,
and we enforce some rest times so no one can get
overtired.

We continue this approach with the litter for the
four months we have them and then continue with the ones we keep. Weather permitting, we encourage
swimming in order to discourage high-impact activities on the ground.

Until they are about 9 months old, the only leash walking we do is in the house or a round pen. We can
practice the skill of being on a leash without actually walking them. They never need to walk on
leash. Chris recommended waiting until puppies are
at least 6 months old before walking them on leash; we find that 9 month is a better time to start from a training standpoint. No running on leash until 24 months old. We never ever try to
"tire out" a puppy. We restrict the amount of exercise so that the puppy never gets very tired; lots of injuries occur when animals are tired

We monitor puppy weights regularly by weighing and
correcting food amounts so that they don't grow too fast.

We do lots and lots of supervision of play. I don't just turn them out to play alone, any more than I
would let small children play unsupervised on a playground. No playing with large older dogs unless
we are mighty sure about how the older dog will
interact with the puppies. We have one large male
who is exceptionally careful with puppies and helps
us with swimming lessons by letting the puppies
ride him and use him like a life preserver.

Puppies are most vulnerable to injuries before growth
plates are closed (and there are growth plates in the bones of the elbows just like other places) and joints
(ligaments, tendons, etc.) are forming. Growth plates
close at different times throughout the body; my
understanding is that the elbow joint is most
vulnerable in the first 6 months to injury during
growth.

I'm putting together stuff for puppy exercise as
I learn it. I've consulted Chris and another
rehab specialist and a vet orthopedist at the
vet school in our area. I've taken some
weekend courses from Chris as well. My current
area of interest is elbows so I'm going to do another
round of consultations this fall about elbows.

Who do you think is going to get the last word?

Who knows? This whole thread has taught me never to try arbitration again. I have learned my lesson.

Elbow replacement surgery for dogs with arthritic ED.

http://petloverstips.com/ForTheLoveoftheDog/news-updates/dog-undergoes-rare-surgery-for-arthritis-video

Thanks so much for posting this video. Amazing stuff.
I sure hope that others see it because it sounds like
they are still enrolling dogs for the clinical
trials. Perhaps you should post this on another
thread so that it gets the attention it deserves.

Again, thanks for the post. I'm going to call this
guy so that I can talk to him about elbows generally.

Kate