The rule that explains almost everything
When a Boerboel is x-rayed for OFA hip evaluation, the application form carries a block the owner initials to authorize release of an abnormal result. That single initial decides what any buyer will ever see.
OFA states the policy without ambiguity. "All normal of-age (>24 months) hip and elbow results post to the web by default. There are no exceptions." Abnormal results post to the web "if the owner or agent initialed the abnormal release block on the application form." An owner who left that block blank can still open the record afterwards, by filing OFA's application to change a result to the public domain, but nothing requires it and the grade stays private until someone does: "Unless the owner has chosen the open database, dysplastic hip grades are closed to public information."
Read that twice, because the consequence is not obvious. Once a dog is x-rayed at 24 months or older, a normal hip result publishes itself whether the owner likes it or not. A bad one publishes only if the owner agreed to release it, normally in advance on the form, before the grade came back. Preliminary evaluations under 24 months work the other way: nothing publishes there unless the owner initialed the release-all block, which withholds good and bad prelims alike. It is the of-age rule that skews the published pool toward good results by construction, because there the default itself depends on the grade. That does not make every published result good news, because dysplastic and borderline grades do appear in the searchable database, and a published dysplastic grade means the owner released it. Read the grade itself, not the mere fact that a grade exists. Seeing nothing at all is what tells you almost nothing, and that is the whole problem.
None of this is OFA quietly permitting concealment. OFA's own policy page says it "recommends owners release all test results to the public domain as it is in the ultimate health interest of the breed." The release box exists because the choice belongs to the owner, and OFA is on record about which way it thinks that choice should go.
How big the hole is in this breed
OFA publishes two things built from the same underlying evaluations: a breed statistics page, which counts the hip evaluations OFA has completed on dogs aged 24 months and over whether or not the grade was released, and a searchable database of individual dogs, which shows only the results that were released. Preliminary evaluations under 24 months sit outside the statistics table, so both figures compared below are of-age only. For Boerboels they disagree, and the direction of the disagreement is the story.
The statistics page counts 704 Boerboel hip evaluations, 58.9 percent normal and 39.2 percent dysplastic. That implies roughly 415 normal grades and roughly 276 dysplastic ones.
The searchable database, exported in full in July 2026, holds 466 of-age hip records: 412 normal, 4 borderline, and 50 dysplastic.
The normal grades reconcile almost exactly. About 415 expected, 412 found, a gap of three records. The dysplastic grades do not reconcile at all. About 276 expected, 50 found.
Roughly 226 dysplastic Boerboel hip evaluations are counted by OFA and are absent from the database you can search. About four out of every five.
That near-perfect agreement on the normal side is what makes the number readable, and it holds grade by grade rather than only in total: the statistics imply about 57 Excellent, 238 Good and 120 Fair, and the searchable database holds 57, 236 and 119. Any mismatch in what the two sources cover, or in how the export was built, big enough to hide 226 dysplastic records would be expected to move the normal side too, and it does not. That is an argument from size rather than a proof. The statistics page counts evaluations and the database carries one final hip record per dog, so a dog evaluated twice can add to the first count without adding to the second. Nothing of that kind comes close to the size of this gap. What is missing from the searchable database is almost entirely the grades that are not normal.
One caveat belongs right here: 39.2 percent is a floor for the dogs OFA actually evaluated, not a measured rate for the breed. Owners also decide whether to submit films at all. A prospective study of clinically normal Golden Retrievers and Rottweilers found that of 93 dogs radiographed for OFA certification, only 53 percent had their films sent in, with normal-appearing films 8.2 times as likely to be submitted as abnormal ones, and its authors concluded that this bias produces lower estimates of hip dysplasia prevalence. That work was not done in Boerboels, and it found submission rates differing between its own two breeds, so it fixes the direction of the error here, not its size.
The elbow result is the tell
Most unreleased hip results are invisible. A dog whose owner submitted hips alone and declined to release a dysplastic grade simply does not exist in the public record.
But hips and elbows are usually submitted together, on one application, covered by one release initial. That is what makes a subset of them visible. In the July 2026 export, 171 of 769 Boerboels that went through a hip or elbow application have a published elbow result and no published OFA hip grade. Thirteen of those have a PennHIP index recorded with OFA, which leaves 158 with no published hip result of any kind.
The obvious innocent explanation is that those owners simply submitted elbows on their own. That does not account for the size of the gap. If they were genuine elbow-only submissions, abnormal elbows would still turn up among them. The release block governs elbows as well as hips, so an owner who left it uninitialed and got a bad elbow publishes nothing at all and never enters this group, which tilts it toward normal elbows under either explanation. What that cannot do is empty it. Owners who did initial the block publish an abnormal elbow whatever the hips showed, and among the dogs whose published dysplastic hip proves they initialed it, 34 percent have abnormal elbows, 14 of 41, in line with the breed's OFA elbow rate.
| Of-age elbow records | Abnormal | Rate |
|---|---|---|
| Elbow published, no hip published | 2 of 158 | 1.3 percent |
| Elbow and hip both published | 47 of 337 | 13.9 percent |
That difference is far too large to be chance. Dogs with a published elbow and no published hip are almost never dogs with a bad elbow. Elbow-only testing does not account for that on its own, because an uninitialed release box suppresses an abnormal elbow whether or not hips were ever submitted. Even after allowing for that, genuine elbow-only submissions would be expected to show an abnormal rate near ten percent on OFA's own breed figures, against the 1.3 percent observed. What is left over is best explained by hip results that went unreleased on a shared application. Cost does not explain it either, since adding hips to an elbow application costs five dollars more.
Two honest limits. This is our derivation, not an OFA statement, and OFA does not publish which tests shared an application. And it is a statement about a pattern across 158 records spanning more than twenty years and over a hundred kennel prefixes. It can never be applied to an individual dog, or used to conclude that any particular dog was dysplastic.
The other test, and why you probably cannot check it
PennHIP is the other way to screen hips. It measures joint laxity on a continuous scale from 0 to 1, called the distraction index, and reports the dog's rank against others of its breed. It can be done from 16 weeks of age, against 24 months for a permanent OFA grade.
Here is what matters to a buyer. The PennHIP report is confidential and goes to the veterinarian. PennHIP results are not publicly searchable. Unless the owner has paid OFA to record the result, nothing about a PennHIP result is checkable unless the breeder hands you the report.
There is one route to make it public. OFA will record a PennHIP result on request for a 25 dollar fee, on a dog that already has an OFA record, and it records "the left and right DI's only, not the percentile." OFA also reserves the right to decline. In the entire Boerboel database, exactly 24 dogs have taken that route.
So a kennel quoting PennHIP numbers is asking to be taken on trust in a way a kennel with a published OFA grade is not. That is a plain difference in verifiability, and it implies nothing bad about the test or the breeder.
Is PennHIP the softer test? No, if anything it is the harder one
There is a widely repeated idea that breeders choose PennHIP because it produces a friendlier answer. The evidence runs the other way.
In the one published head-to-head, researchers submitted the hip-extended view from 439 PennHIP studies to OFA for grading. Among dogs OFA graded fair to excellent, 80 percent had a distraction index in the range associated with susceptibility to osteoarthritis, including 52 percent of dogs OFA graded Excellent. The authors concluded that OFA scoring underestimated susceptibility to osteoarthritis. A later study of 12,496 guide dogs found the same pattern persisting after eight generations of selection.
PennHIP also requires that every case be submitted regardless of result, on pain of dismissal from the program. OFA submission is the voluntary one. And PennHIP's own training manual includes a sample report in which the breed comparison is used to recommend against breeding a dog whose index exceeds the breed average. It is not a test that hands out reassurance.
One disclosure belongs with this evidence: both comparison studies come from the University of Pennsylvania group that developed PennHIP and holds a financial interest in it, and a rebuttal was published in the same journal. Read it as strong but interested.
The percentile trap
The real risk with PennHIP is not the test. It is being quoted half of it.
A PennHIP report carries both an absolute number and a breed ranking. The breed ranking is the part that flatters, because it is relative. Mean hip laxity in Boerboels sits around 0.50, well above the roughly 0.30 level below which osteoarthritis becomes very unlikely. A Boerboel can therefore be genuinely "tighter than most of its breed" and still carry meaningful laxity in absolute terms.
If a kennel gives you a percentile, ask for the number. The report shows both on the same page.
What the registries actually require
A reasonable buyer assumes some registry is policing this. None is.
- AKC registers a litter on proof of parentage. Its registration rulebook contains no health, hip, dysplasia or screening requirement at all.
- The American Boerboel Club, the AKC parent club, says hips and elbows are "highly recommended" to be x-rayed and graded, which is deliberately not the language it uses for its binding requirements. Its one disclosure duty is to "produce documents upon request and/or publish the information as part of any litter announcement," so handing a document to one person who asks satisfies it.
- SABBS, the country-of-origin authority in South Africa, does require a certified hip status for entry to its Development Register, its Stud Register and its OOR Register, and for approved semen donors. It does not require one for the Stud Book Proper Register, which is ordinary full registration, and it does not require one to register a litter. Its records are not publicly searchable.
- NABBA dissolved in 2024. Its US successor publishes no hip rule.
In many breeds the gap is closed by CHIC, which issues its number only when results are released into the public domain. OFA's CHIC page for this breed reads that the breed is not yet part of the program. So no body in a US buyer's path, active or defunct, operates a public-disclosure gate for Boerboels.
Where testing is compulsory, the certificate goes to a registry office. It does not come to you.
Where the breed actually stands
On OFA's own numbers the Boerboel ranks 14th of 219 ranked breeds for hip dysplasia, at 39.2 percent of 704 evaluations, against 14.2 percent across 1,708,918 evaluations for all breeds. Elbows are worse in rank: 7th of 166 breeds, 34.3 percent abnormal of 688 evaluations.
Treat those ordinals loosely. They rest on voluntary submissions, the sample sizes are modest, and neighbouring breeds overlap. The useful reading is directional. This is a breed with a real joint problem, which is exactly why the missing results matter.
What to ask a breeder
- "What is the OFA number for the sire and dam?" An OFA number is issued only for Excellent, Good or Fair. If the kennel has one, you can verify it yourself in under a minute on OFA's public search, without taking anyone's word for anything.
- "If there is no hip result online, was one done?" There are three honest answers: we did not do hips, we did hips and did not release the result, or we used PennHIP. All three are answerable. Across the breed, dogs with a published elbow and no published hip usually turn out to have been x-rayed for both, though that cannot be assumed of any single dog.
- "May I see the PennHIP report?" Ask for the page, not the percentile. It should show an index for each hip, an osteoarthritis assessment and the breed ranking.
- "Will you initial the release box on your next litter's applications?" This is the question that separates a kennel that publishes from one that does not. It costs nothing, and a kennel can also open results it already holds by filing OFA's application to move a result to the public domain.
- Check the age on any result. A hip evaluation under 24 months is preliminary and is read by a single OFA staff radiologist rather than three outside consultants.
- Ask about elbows separately. Boerboel elbows rank worse than Boerboel hips, and they get far less attention.
A breeder who tests and releases everything is doing something that costs them the option of hiding a bad result. That is worth recognising, and it is visible in the record.
What our own kennel numbers mean
Every breeder page in this directory now shows how many of that kennel's OFA-screened dogs have a publicly available hip result.
That number measures disclosure, not health. A kennel with a low count may be testing and not releasing, may be using PennHIP without registering it, or may not be testing. We cannot tell which, and we do not assign motive. The release decision belongs to whoever owned the dog when it was tested, which for a dog sold as a puppy is the buyer and not the kennel that bred it.
What we can say is the thing that matters to you: for those dogs, there is nothing you can check. You can read more about how we source and review this material in our editorial standards, browse the underlying records on our health testing page, or start with what the four core screenings actually cover.