Send us Fan Mail A “normal temperature” can be the most dangerous kind of comforting. We sit down with Drs. Kimberly Woodruff and Sahar Rostami to unpack what happens when ocular thermometry and rectal thermometry show similar diagnostic performance for detecting abnormal body temperature in dogs, and why that still doesn’t mean you should relax after a normal reading. We translate sensitivity and specificity into day-to-day clinical decisions. High specificity means these methods are fairly good at confirming normal temperature when a dog truly is normal, but low to moderate sensitivity means they can miss dogs with abnormal core temperature. That’s where the nuance lives: an abnormal result is often meaningful and deserves follow-up, while a normal result shouldn’t outweigh your clinical intuition when the patient looks unwell. We also discuss why the shelter population matters for interpretation, including how rare abnormal temperatures can widen uncertainty around sensitivity estimates. Then we get practical. We walk through combining ocular and rectal measurements using parallel testing (any positive) versus serial testing (both positive), and how your choice depends on whether you’re trying not to miss cases or trying to be confident before labeling a dog abnormal. We also explore why ocular screening may be a better fit in shelters and some clinics: less handling stress, faster workflows, and potentially lower risk of disease spread than moving a thermometer between animals. Along the way, we touch on measurement technique, eye-region considerations, and why rectal temperature isn’t a perfect gold standard, including how Bayesian latent class analysis can better reflect reality. If you care about smarter canine triage, better shelter medicine protocols, and making test results match what you see in front of you, hit play. Subscribe, share this with a colleague, and leave a rating and review wherever you listen. AJVR article: https://doi.org/10.2460/ajvr.26.04.0185 INTERESTED IN SUBMITTING YOUR MANUSCRIPT TO JAVMA ® OR AJVR ®? JAVMA ®: https://avma.org/JAVMAAuthors AJVR ®: https://avma.org/AJVRAuthors FOLLOW US: JAVMA ®: Facebook: Journal of the American Veterinary Medical Association - JAVMA | Facebook Instagram: JAVMA (@avma_javma) • Instagram photos and videos Twitter: JAVMA (@AVMAJAVMA) / Twitter AJVR ®: Facebook: American Journal of Veterinary Research - AJVR | Facebook Instagram: AJVR (@ajvroa) • Instagram photos and videos Twitter: AJVR (@AJVROA) / Twitter JAVMA ® and AJVR ® LinkedIn: https://linkedin.com/company/avma-journals
Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
Transcript
From the show. Tap a time to play from there.
SPEAKER_00You show up for pets every day. It's because care is your calling. And it's ours too. We know it's what drives you to make a difference in the health and well-being of every pet. That's because at Zoetis, hundreds of veterinarians are part of our team too. Working with you to provide science-backed medicines, technology, and practical support to help elevate your standard of care. Learn more at care is your calling.com.
Lisa FortierWelcome to Epic Award-winning Veterinary Vertex, the Avium Made Journal's podcast where we delve into a behind-the-scene look with manuscript authors. I'm editor-in-chief Lisa Fortier, joined by Associate Editor Sarah Wright. Today we're discussing how ocular and rectal thermometry shows similar diagnostic performances in detecting abnormal body temperature in dogs with Kimberly Woodruff and Sahar Rustami.
Sarah WrightThank you for joining us, Kimberly and Sahar. Thank you for having us.
SPEAKER_02Thank you for having us on here.
Lisa FortierKimberly, why don't we start with you? What was the impetus? What was the idea behind comparing eye and rectal temperature specifically rather than other non-invasive methods?
SPEAKER_02Well, um, we started this idea just as a way to screen for problems, um, disease problems in shelters. And it turns out there had already been some work done with thermography. And so we had a little bit of a starting point, a little bit of some background and um some ideas of some things to try. It also gives us a little more freedom than some of the other methods. For example, if you remember during COVID, we have things that we shine at our forehead, um, infrared thermometers. And those are really tricky in dogs because you have to be a certain, you have to be very close.
SPEAKER_02You can't, we don't have four fur on our forehead. So that plays a role with dogs. Um, coat color could play a difference, make make things different. Um, sorry, I should say, coat color can impact the reading. Um, and this gives us a little bit of freedom to play around with different sites, um, to be different distances from the dog. And so it just seemed like a better fit for what we needed it to accomplish.
Sarah WrightSahar, both tests had high specificity but low to moderate sensitivity. What does that mean practically for a clinician using either method?
SPEAKER_03Well, as you mentioned, we reported that both of them have similar diagnostic pair formats. Regarding specificity interpretations, we can say that if the dog has a normal core body temperature, both of these tests work pretty well at confirming that. However, regarding that sensitivity interpretations, if the dog has an abnormal core body temperature, there's a fair chance that both of these tests will miss it. So a normal reading from either of these two tests doesn't necessarily mean that dog's normal core body temperature is normal.
SPEAKER_03And in here, I just want to briefly add some clarification regarding our sensitivity estimation from epidemiological standpoint. The specific shelter that we conducted the study, uh, the number of dogs with abnormal cold body temperature was quite rare. This factor affected our sensitivity estimation precision. And we also talk about it in the paper as well, and that's why we got a wide credible interval for our sensitivity, and its precision is kind of uh lower in comparison to the precision that we got for specificity.
SPEAKER_03So that's why we also be cautious when we interpret the sensitivity estimation.
Lisa FortierSahara, if you boil that down clinically, you found that a normal reading isn't all that reassuring, but an abnormal reading is more meaningful. Why does that asymmetry exist?
SPEAKER_03Regarding this asymmetry, uh, this really comes down from the killing capture before we measure the temperature. For example, a negative reading or a normal reading from either of these two tests provides little extra rollout beyond the clinician's initial assessment. For example, if a dog looks healthy and we're not worried about abnormal temperature, a normal reading is reassuring. But if a dog looks sick, we're worried about abnormal temperature, then a normal reading in this condition, it's not quite reassuring and helpful.
SPEAKER_03However, on the other hand, when we get a positive results or an abnormal reading from either of these two tests and we're concerned about abnormal temperature, in this case, that positive results or abnormal reading results are highly likely to be correct and they are much more informative and need further attention.
Sarah WrightLet's say that a clinic is interested in using both of these methods together. Sahari, you looked at combining both tests using any positive versus both positive rules. What's the practical difference between those approaches for a clinic deciding how to use them together?
SPEAKER_03This any positive, both positive, these are the ways that we combine the tests. When we want to decide which one to choose, it depends around a specific question of interest and the specific trade-offs that we're willing to have. For example, regarding that any positive or as we call it parallel testing, in this case, we consider a dog abnormal. By abnormal, I mean having abnormal temperature. We consider the dog abnormal if either of these two tests give us positive reading or abnormal temperature reading. In this case, we're maximizing sensitivity and also maximizing the case detection.
SPEAKER_03With parallel testing, we're just trying to ensure that we detect most of the individuals that are truly abnormal temperature dogs. However, the trade-off here is that in this parallel testing, we have more false positive. As I mentioned, in this uh study, we reported that both of these methods have low to moderate sensitivity. And as I mentioned earlier, parallel testings maximize the sensitivity. So one of the ways for us to increase that sensitivity could be by combining these tests are using this parallel testing.
SPEAKER_03We also can put the other type of uh health assessment tests alongside them in parallel testing to increase the sensitivity. And that's one of the ways to help that for the test that have high a specificity but how low sensitivity. And also in the paper, we reported that for any positive, we got better diagnostic performance results. And for post-positive or serial testing, it means that we consider a dog abnormal if both of the tests give us positive reading, or both of the tests give us abnormal reading.
SPEAKER_03In this case, we are much more confident of the abnormal reading that we get, or calling the dog sick. However, here is the trade-off also, and in this case, we have more false negative. So it depends on what are what is our next step and how much we care about this false positive false negatives. If we just want to be more confident of our abnormals, we should go for serial testing. But in these cases that we have low sensitivity and we just want to be careful not to miss any individual with abnormal temperature, it's better to go with parallel testing.
Lisa FortierYeah, it's a lot of sensitivity and specificity. Kim, let's go to you for a minute. If it would be boil this down to real practical, which is really important in shelter, but in in all of our clinical settings, given that these two methods performed relatively similarly, is there a case for preferring ocular over rectal speed, stress of the animal?
SPEAKER_02Yeah, we really sort of developed this idea with a shelter in mind because catching a dog that has an abnormal temperature is maybe one of our first indicators of disease. And so if you're thinking about trying to treat that animal and trying to get that animal out of general population, a very good indicator is abnormal temperature. But it's not feasible to go take every dog's temperature twice a day. That's you just can't do that. But it is feasible maybe to just put a camera on them once or twice a day and just screen for something abnormal. And so we really kind of thought about it in the um in a time-saving kind of thing and a feasibility thing.
SPEAKER_02But there's also um also situations you may have a dog that is uh doesn't like to be handled, or that is in a hospital situation too painful to be handled, or a dog that is aggressive, or um or some situation where it's just not easy to put your hands on that dog. Um and also in a shelter situation and in a clinic situation, this is a way to be able to monitor temperature that carries less risk of disease spread. So if you're monitoring sick animals with a thermometer, we do the best we can to clean it.
SPEAKER_02But really, it's safer to just not ever have to touch that animal with or and move between animals. So that's kind of was our thinking and developing this question.
Lisa FortierSo screen ocular.
SPEAKER_02Yes, yeah. We um that is kind of the method that we developed because dogs um have different lengths of hair coat and different colors of hair coat, and there's all sorts of things that um that may affect a reading otherwise on the body, but if you can do an ocular screening, that uh might give you the best results.
Sarah WrightThis is rifting a little bit. When I'm thinking about eye color, like thinking about our Siberian breaths that have those lighter blue eyes sometimes, or our Australian shepherds, things like that, compared to like the German shepherds and like have like brown eyes. Did that did you guys look at that at all in your study?
SPEAKER_02Well, that's what Sahar is gonna do for our next PhD, right? Sahar? No, we did not look at that. I don't Sahar, can you anecdotally did you have any huskies or blue-eyed dogs?
SPEAKER_03I don't think we even uh you know the method that we use in here was also an easy one in comparison to other papers. We just look at both eyes, highest temperature, maximum temperature in either eye, and we really didn't have the specific region of interest at the corner or other parts of the eyes, which this method makes it easier even for the shelter staff to conduct it. And most of the times the highest temperature that I got was in this carnical area, which is not much affected by the eye color and so on. So I cannot give an absolute answer because we didn't record the eye colors.
SPEAKER_03We have it, but we didn't record it and test it. But that would be something interesting to look into further. But since most of the time uh this carnal area, this tissue has the maximum temperature, I'm not sure how much will it affect it.
Sarah WrightWell, I look forward to hopefully seeing that next study across our plate someday. I like how you said this is time saving. I also like too how it sounds like it could be safer for both people and animals potentially. If a dog has a normal reading, but a clinician still suspects something's wrong, what should happen next? And how should that normal reading factor in?
SPEAKER_02Yeah, so this is a case of parallel testing, right? This is um we are going into this, um considering what we know about the dog, what our expectations are of that dog, do we think that it's healthy? You know, we do that in a clinic every day. When we do a diagnostic test, we already kind of have in the back of our head what we're expecting that diagnostic test to be. And so um, this is no different. And so if you have a dog that um that for whatever reason you expect to have an abnormal temperature and it doesn't, it's still grounds for um diving deeper into that.
SPEAKER_02At the same time, you know, there it doesn't, we talked about earlier pairing it with erectile temperature as a parallel test, but a really good way to do this would be to pair it with some totally different kind of diagnostic test, which is even something as simple as, well, did that dog have an appetite today? Or is that dog lethargic? Or is there just something off about that dog? So our visual cues of that dog is also a great diagnostic test. And so if any of those things are a positive result for indicating a sick animal, we act on that and we dive deeper into that.
SPEAKER_02And so this is just one test to help us um in a whole um whole gamut of things that we can use to get information on that dog and to make our final decision.
Lisa FortierOh, yeah, the good old physical exam.
SPEAKER_02Yes, always go back to that. What do you see?
Lisa FortierUh, I feel like both of you said this, but I just want you to distill it down to one thing. So we'll start with Kimberly. Uh, for the listeners, it what would be one takeaway you'd want a clinician to walk away knowing about how much to trust a normal ocular temperature reading?
SPEAKER_02Well, I think um maybe we put a little bit too much in that. Like I think I think in a situation we take a dog's temperature and um it's normal and we feel good about it. And maybe the takeaway is that's not always a a you know, a free pass that their temperature is good, so everything's okay. And so always to interpret that, um, you know, we do that on the other end all the time. We have a puppy that comes in that's bouncing around and happy and excited, and we get an an abnormally high reading and we say, uh, that's probably not a big deal, but we don't really do it on the other end.
SPEAKER_02And so maybe this is a um a charge to us to think more about how really, how do we really trust a normal temperature reading in a dog?
Lisa FortierYeah, that's a great point. Sahara, what would you like to add to that?
SPEAKER_03Regarding uh, you know, in epidemiology, we recognize that most of uh the common diagnostic methods and tools that we do in as a first step to go, most of them are usually not that perfect gold standard. And then, you know, they're not always correct in the results that they give us. And actually, that's one of the reasons that in the study we used Bayesian latent class analysis, because we didn't consider rectal thermometry as the perfect gold standard to tell us the cold body temperature in the animal. And in retinary, we want to make diagnoses that this dog is sick or healthy, we're always uh piecing together the result of multiple tests that usually most of them are imperfect tests.
SPEAKER_03And in this case, this temperature measurements. But as we mentioned, we got low to moderate sensitivity, and it just tells us that if uh our clinical intuition tells us that there's something might be wrong with the temperature in the animal and we get a normal reading, we shouldn't factor that normal reading in our final judgment too heavily. We should trust our clinical judgment and as mentioned earlier, uh conduct more uh deeper diagnostic methods and workups.
Sarah WrightKimberly and Sahar, thank you so much for joining us today. We appreciate you being here, and I learned so much more about your study. Thank you.
SPEAKER_02Thank you for the time. I also just want to say that we're representing a team. Um, we were joined by Dr. David Smith, who is a colleague here in um, and we also had a wonderful student worker, Abby McBride, who helped collect the data. So we're here kind of on behalf of the whole team. So thank you to them, and thank you to the two of you for spending time delving further into this.
Sarah WrightIt's great to hear that it was a team effort. For our listeners and viewers, you can read Kimberly and Sahara's article in AJBR. I'm Sarah Wright with Lisa Fortier. Be sure to tune in next week for another episode of Veterinary Vertex. And don't forget to leave us a rating and review on Apple Podcasts or whatever platform you listen to.
Send us Fan Mail A pair of suture scissors can feel like the definition of “low risk” until you look at what’s actually living on them between patients. We sit down with Kathryn P. Spivey, returning guest and AJVR...
Send us Fan Mail A calf can look “fine” and still be losing ground in the lungs and that’s exactly why BRSV keeps costing farms money. We sit down with Dr. John Ellis to unpack a rare head-to-head comparison of two...
Send us Fan Mail Dogs with chronic enteropathy can make even seasoned clinicians feel stuck: antibiotics help, symptoms fade, then everything flares the moment treatment stops. We sit down with Drs. Moshen Hanifeh and...
Send us Fan Mail Septic arthritis in a neonatal foal is one of those diagnoses that can change everything fast: the foal’s comfort today, survival to discharge, and even long-term athletic soundness once cartilage gets...
Send us Fan Mail One strange beat on a resting horse ECG can spark a surprisingly high-stakes question: are we seeing a normal sinus arrhythmia, or an atrial premature complex that changes how we think about risk and...