Episode Transcript
[00:00:02] Speaker A: Welcome to your Cases on Hold, the JVGS podcast hosted by Andrew Schoenfeld and Aisha Adkeen.
[00:00:08] Speaker B: Here we discuss the best of what each issue of JBJS has to offer with the usual dose of entertainment and pop culture.
[00:00:16] Speaker A: Take us with you in the gym, on the commute, and as ever, whenever your case is on hold.
Welcome back everyone to your Case is on hold. Episode 111 if you are listening on the day we drop it is August 4th for the August 5th issue of the Journal of Bone and Joint Surgery.
As always, what you are hearing on this podcast is the best that JB JS has to offer in terms of this first half of the month's publications.
But the opinions are those of myself and my co host and not those of the Editor in Chief, the other constituent editors of the journals, members of the editorial board, board of Trustees, or those who work at JBGS Corporate this episode of JBGS is brought to you by JBGS cme including Clinical Classroom Self Assessment exams. You can go to jbjs.org it is a one stop shopping for everything that you need in CME to maintain your maintenance of certification, to get up to date on everything in terms of the state of the literature, test questions, you name it, JBGS CME has it. So check it out and subscribe if you're not already subscribing. In addition, for those who are new here, we're kind of at the start of a new academic year so maybe you're new to your cases on hold. If you can give us a five star rating, like and subscribe, share with others.
Find us on Stitcher, Castos, Apple, Spotify and if you can't find us on any of those places, Amazon Podcasts, you can always find us at the JBJS website.
I am Andrew Schoenfeld, professor of Orthopedic Surgery and Vice Chair for Education, Harvard Medical School and I have with me
[00:02:11] Speaker B: hi everyone, I'm Aisha Abdeen. I'm Chief of the Division of Hip and Knee Arthroplasty at Boston Medical center and Associate professor of Orthopedics at Boston University.
[00:02:21] Speaker A: All right, so I think that covers all of the mandatory reporting and we can get into this issue in the top of the pile. We have Introducing the JBJS Ambassadors. This is by Dr. Bandari, the editor in Chief with a high. It is the highlight or a highlight and it is permanently free. This is a important new initiative from the JBJS corporate entity about individuals across the globe who are going to be actively working with their local networks to increase the visibility for the journal and promote its utilization.
We then have heading toward a different future. The microbiome, dysbiosis, microbial translocation and beyond by he then Cost margin, mission and Value what every leader Should Know about the Business of Orthopedic Surgery by day this is the other highlight in this 30 days free my lucky break A rare bone condition, a sudden fracture and the Right team by Sando Permanently free Humility the Essential Value Enhancing Patient Care Education, Research and innovation by Rizzo the persistent challenges of Diagnosing Orthopedic implant related infections by Lum Then we have the Extended Lesser Trochanteric osteotomy ext. In Complex revision Hip Arthroplasty an alternative to the Extended Trochanteric Osteotomy by Fareed so that's what's in the top of the pile. We'll now move into the headlines. My headline is Comparison of Large Language Models with Rules Based Natural Language Processing Algorithms for Extracting Data from Orthopedic Notes this study was done by Yang and colleagues. It does come with a comment, so you don't have to take it from me. You can hear what someone else thinks about this paper and is also 30 days free. So no excuses.
Everyone should be able to access this regardless, at least for the next 30 days.
So this study is looking to compare two different approaches to extracting data from medical records, and that is the large language model with a rules based natural language processing algorithm.
The large language models may be a little bit more AI reliant, although I'm sure you can do natural language processing algorithms that using AI programming at this point, when my team and I have done them in the past, they've been a little bit more individual programmer heavy.
It's, it's somewhat of a of an interesting concept and I think it's different when you're looking at these things for extracting data for research purposes versus extracting data for maybe, you know, reporting to a registry or something to that effect.
It's not entirely clear to me exactly which route these authors are going in terms of their intent.
Honestly, you know, what's good for one may be good for the other. The exception being that this does in some ways, you know, essentially cut corners.
The from a research standpoint, my opinion, take it for what it's worth, is if you're abstracting data, nothing beats a good old fashioned chart biopsy.
It's a lot of work. I know, I know it is, but that's going to be the highest. If you're, if you're outsourcing it to something else, that's just going to pull stuff for you based on a large language model and populate something.
It's likely that they're going to be errors. The question is, you know, how bad and how many.
So their rationale is really just asking the question. If you Compare LLM to NLP, which one is better? And they took 1,000 primary total hip arthroplasty cases performed by 47 surgeons over a 23 year time window. And the notes came from three hospitals located, they say at three geographically distinct sites in the United States. But it's still all within the same health system. And this is the Mayo Clinic. So it's not like you're getting, you know, everybody here has the Mayo Clinic imprimatur. It's not, you know, a community hospital where you'll have like a couple of different private practices or something like that. Like everybody here is Mayo Clinic. They just happen to work at, you know, a different location.
They use two medical student annotators to review this for the data points and they say training and confirmation from board certified surgeons. I don't, I don't really love that, that piece either.
Medical students, as we've discussed in the last couple of episodes, are not generally considered in a position to really effectively execute some of these things. And if you're really trying to make this as scientifically robust as possible, it should probably be someone who didn't train someone else as a proxy, but the individuals who are really the subject matter experts doing the data abstraction themselves.
So compared with the rules based natural language processing, this relies on design rules and keywords. The LLMs are more intuitive and maybe in some ways like they will learn themselves. In some respects they generate this natural language text to extract key structured data.
And this is useful because even when they're different phrases, phrases that might be used to express the same concept, the LLM has the flexibility to sort of capture that.
The inability of the NLP algorithm to extract the concept of polyethylene was the biggest issue that they identified here. And this was where many errors arose. There was a misclassification for metal on poly and ceramic on poly that would get conferred to metal on metal, ceramic on ceramic or ceramic on metal. And the large language model was much more accurate in classifying those cases and more consistently and reliably extracted polyethylene as being the actual part of the implant.
The LLM does have hallucinations, which is a well known Limitation, I have done some, not scientific related research, but historical based research with respect to using Google AI, which in some ways is, is analogous to this. And I was checking on something with respect to a genealogy question and the AI responded, giving me the name of an individual that I had never heard of before, that it said was like the child of, of another individual. And I said, what's your documentation for this? Because I don't have that in my record. And it responded back, oh, you don't have it in your record because that person doesn't exist? Which I thought was like, are you trying to trick me here? Or like.
And, and then in sort of shortly thereafter it made a comment about someone who was the King of Portugal not being related to the King of England. And I said, no, this is absolutely the King of England's, you know, twice great grandchild. And I sort of went through the, the rationalization and then it said, oh, you're absolutely right. I was thinking of a different King John of Portugal. It's like, we're not, we're not hanging out. Like we're not.
[00:10:35] Speaker B: That's creepy.
[00:10:37] Speaker A: I was very clear. But so, I mean, those are the kind, obviously not exactly the same, but those are the kinds of hallucinations that you get sometimes, I think in these contexts.
So as always, they say external validation of the LLM will be necessary, but they're saying LLMs may provide a flexible solution for automated registry curation.
But I'm still left with like, with what purpose? Like, for what? Why, why are you using this?
[00:11:09] Speaker B: Go ahead.
[00:11:09] Speaker A: Yeah, so that, you know, in their, in their clinical relevance, they say LLMs have the potential to impact clinical care. I disagree wholeheartedly. These have no potential to impact clinical care in and of themselves. You can. That's like saying propensity score matching has the impact potential for clinical care. No, it doesn't. In it's a model, it does something and then how you interpret it or how you present it. And here, this is not an analysis, this is not a comparison. It's just, it's a tool that it's demonstrating that this one maybe works better than that one. Neither, in my opinion, in research works better than actually doing the elbow grease of pulling the data yourself, reviewing the medical record yourself, looking and making the adjudications yourself.
And then they say that LLMs can simplify, improve and democratize the construction of registries. I don't understand how the democratization feature really plays in here unless, because again, like, anyone could ostensibly do this, right? Like if you have the ability to understand how a registry should be constructed. And you can build a registry or a data set like on an Excel spreadsheet. If you don't, then it's unlikely you're going to be able to really use a NLP or an LLM as a cheat code to create that for you. Like, you still have to have the knowledge about. Right. How it works. So I, I mean there's nothing wrong with the, the, the comparison and I guess, you know, it's potentially useful in the right context, but I'm just left wondering what is the right context? A and B, you know, some of these claims about impact for patient care and democratization. It may be, as we discussed, some of the sort of strategy behind, you know, getting this accepted, which is all well and good, but I mean, you shouldn't say stuff that really.
They didn't even really get into that in the discussion. I was, I was wanting to see more about, I was left wanting more explanation about these clinical relevance points that they're highlighting. They don't really go.
It was very technical. The discussion mostly focused on comparing LLMs to NLP.
[00:13:34] Speaker B: Yeah, my take was similar. I also was left wondering what their intent was to do with this technology.
But my guess is, and their inference about its use in registries was that the senior author is Cody Wiles and extremely prolific researcher in arthroplasty. So my thought was they would be using this for data gathering from their registries for research purposes exclusively. I think it's interesting that they did include medical students because that's how the real world works. That's who's gathering this data typically and having it validated by surgeons.
It's interesting because I typically, like the Mandalorian, have a deep seated distaste or distrust of the droids and everything automated, but it took me some years. I do use a robot in the or, but this whole concept that this automation and the AI version is going to be better than humans. So I'm with you on that one. But I thought it was really interesting. There was one instance where the human abstractor had initially labeled the bearing surface as unknown, but the LLM was able to identify correctly the type of bearing surface because they had access to catalog numbers and they could instantly retrieve the bearing surface because of the catalog numbers. So perhaps in some respects it can be more efficient and more effective than humans. But overall I still think you're right that there's nothing does beat human review of a chart.
But overall I thought this was interesting. But agree with you that I don't think it's going to replace anything in terms of clinical care.
[00:15:06] Speaker A: And also the reality is that when you're talking about they built this LLM, it learned based off of their Mayo Clinic records, which means it Learned that what Dr. Wiles and his colleagues in arthroplasty, how they dictate their notes or how they talk about their implants, you basically would have to do this at Boston University, separate from Brigham and Women's, separate from University of Pittsburgh, separate from Rothman, separate from like this is not scalable in its direct form to other contexts.
[00:15:39] Speaker B: Yeah. And even within their own Mayo Clinic, the accuracy was higher in Rochester versus the Florida and Arizona locations. Right. So you're right. It was designed for their main, you know, for the. The Rochester site, probably.
[00:15:54] Speaker A: All right, so your headline.
Significant Anxiolytic effect and enhanced recovery benefits of Perioperative Low Dose Olanzapine in patients with Anxiety undergoing tha A randomized controlled trial by Jiang and colleagues with a comment. This is the lead article and it is 30 days free.
[00:16:17] Speaker B: This is an RCT. It was performed in the People's Republic of China.
It was a single center study performed to compare treatment for anxiety in patients undergoing primary total hyperthroplasty. The hypothesis was that olanzapine, an atypical antipsychotic drug that antagonizes several receptors including serotonin, dopamine and histamine, and multiple muscarinic receptors, would be superior to either placebo or benzodiazepine.
The inclusion criteria was 18 to 80 year old patients with at least 40 points on the State Trait Anxiety Inventory State so the Stay S Scale.
It is a commonly used measure of trait and state anxiety.
So essentially long term anxiety and momentary anxiety in the moment. And it can be used in clinical settings to diagnose anxiety and to distinguish it from depressive syndromes.
It is divided into two subscales, the state anxiety, which reflects feelings at a specific moment, and trait anxiety, indicating general anxiety tendencies. So the respondents rate their feelings on a four point scale and higher scores suggest greater anxiety.
The patients were randomized to 2.5 mg of olanzapine or 0.4 mg of alprazolam, benzodiazepine or placebo once nightly beginning on the day of admission for a total of five days. These patients were admitted for five days. I don't know if that's standard of care in this particular institution.
The study was double blinded for surgeons and all clinicians, providers and patients.
The patients were admitted prior to surgery for two days and underwent surgery on the third day.
The length of stay was a five day minimum. Patients who met discharge criteria were discharged on the morning of the sixth day of hospitalization.
All patients had general anesthesia and a structured ERAS pathway with a consistent multimodal analgesia regimen. The primary outcome was the stay S score determined on the night prior to surgery. Basically, patients came in on the two nights prior to surgery and had this anxiety survey performed and then it was repeated on post op day 1 3, post op week 4, week 12 and week 24. The secondary outcomes were visual analog scales for pain, visual analog scale during hip flexion, sleep quality for sleep disorder and the Pittsburgh Sleep Quality Index, the psqi.
The PSQI was not prescribed in the trial registry was added before the study was conducted to provide a more comprehensive assessment of sleep quality.
It's a validated instrument for evaluating multiple dimensions of sleep. The authors noted that this addition was made prior to the data analysis and is clearly noted here so that they were ensuring transparency. They mentioned that additional secondary outcomes included the postoperative daily opioid consumption in terms of oral morphine equivalence, postoperative rescue analgesia, post op daily nausea or vomiting, hip range of motion, Harris hip score, hip disability and osteoarthritis or the WHO's and postoperative length of stay. Adverse events related to olanzapine or alprazolam and potential drug dependencies were recorded for the first dose until seven days after discontinuation.
The power calculation was performed for a power of 90% and they needed 38 patients per per arm and the authors increased this to 45 per arm to account for a 10% loss to follow up.
Of 581 patients considered, 429 did not meet the stay S threshold which was a score of 40 or more on the anxiety scale and 13 were excluded for pre existing mood disorders so they had final analysis of 45 in each the olanzapine group, the alprazoline group and the placebo group. There was 44 in the placebo group.
All patients were ethnic Chinese patients and had a predominance of 70% more females in each group.
For the primary outcome of the stay S score they were lower for olanzapine group on the night before surgery, on post op day one, post op day three and post op week four.
However, while this was statistically significant, the scores when you look at them were very similar at all intervals. For instance on the night before surgery the score was 39.4 for the Olanzapine group and 40.9 for the Alprazolam group and 44 for the placebo. So not a huge difference because 20 is considered not anxious and 80 is considered extremely anxious. So these are like less than a 1 point difference between the olanzapine and alprazolam group and about 3, 4 points for the placebo.
So the authors report better sleep quality on Post OP Day 3 measured by the PSQI for the olanzapine group and however, no difference on post op week 4, 12 or 24.
Pain by visual analog score is also better statistically for the olanzapine group compared to the alprazolam and placebo on post op day one through three, but by less than one point difference for most measurements. So again, not sure how clinically relevant this is.
And the authors also noted a benefit for olanzapine could be nausea and vomiting reduction.
Their results were that this was a significant reduction in nausea and vomiting in the ultra olanzapine group compared with the benzodiazepine and the placebo.
So overall I have concerns about sort of the clinical relevance and the applicability of this data. It's a well conducted study. But at the intro the authors indicate the rationale for the study given that quote. Given that the population undergoing total hip arthroplasty is generally younger and more prone to comorbid anxiety symptoms than the population undergoing total knee arthroplasty, identifying these high risk patients and developing individualized anxiolytic strategy may be an effective solution.
But this is a complete non sequitur. What does the fact that total hips are younger and might be more anxious than total knees mean that we need to study this.
Generally speaking, patients going for total hip replacement, I mean any operation, people are nervous on admission. Whether we admit them a couple days before the day of admission, they're going to be anxious.
And so I'm not sure that the relevance and the need, the clinical need to really address anxiety for total hip arthroplasty patients because it seems to be pretty transient and it doesn't seem to be a major problem that we sort of encounter.
Also, the testing identifies these patients as being anxious in this particular state, but for the most part, anxiety may not be prolonged throughout the course of the stay. And also putting the patients on a drug that potentially could cause a lot of other symptoms, such as extrapyramidal symptoms. I just remember my psych rotation from being in medical school and seeing patients on these medications as antipsychotics. Granted, it's A slightly higher dose, but they're not benign medications.
I'm not sure that necessarily prescribing these as an orthopedic surgeons is warranted. I think if we're worried that someone has anxiety about their surgery, we should identify it further in advance of their admission and really have collaborate with psychologists and psychotherapists to see if that can be addressed before their operation rather than putting them on a few week course of a antipsychotic drug. The other very interesting thing about this study is that all patients received general anesthesia, which is typically not the routine standard of care.
It's not the routine, I would say in North America at least, we're moving more towards ERAs protocols that incorporate regional anesthesia. And one of the main features of that is that it reduces nausea and vomiting. So we may not have that issue of nausea and vomiting if everyone didn't get a ga, for instance.
And so that's another area where I think it may be somewhat less applicable in terms of these data across all arthroplasty programs.
What are your thoughts, Andrew?
[00:24:18] Speaker A: Yeah, I had some of the similar concerns regarding. It's very breezy about like hey, yeah, just use Olanzapine.
You know, while they were powered for their outcomes, they're not necessarily powered to detect adverse effects around the, you know, the use of the drug.
And yeah, I agree with you that the outcomes, while they certainly had some statistical findings, they didn't really seem in terms of like the clinical impact to be that it's a, it's a transient phenomenon and it's just there's no way out but through. There's a, it's a necessary threshold through which all patients have to pass, like you said, with, with any kind of surgery.
So from the methodology front, there's nothing to balk at, I think. Yeah, where there's some questions here is probably on the both.
Are there side effects that we're just not seeing because the sample isn't really that large when you're talking about the use of olanzapine. And then also, is there really demonstrable clinical benefit?
[00:25:34] Speaker B: Yep, good points.
[00:25:36] Speaker A: All right, so now we're going to move into the your case is on hold featurette. CMS proposed substantial clinical benefit thresholds correlate with patient reported measures after primary total joint arthroplasty, improvement, satisfaction and willingness to repeat surgery. This is by Wang and colleagues with a comment.
This study is from the University of Michigan, which is where I did my health services research training and I have collaborated with individuals from the Marquee registry. That's the Michigan Arthroplasty Registry, Collaborative Quality Initiative.
That's the study substrate that they were using for this particular investigation.
I believe it's funded by Blue Cross and Blue Shield. It's a registry that is sort of statewide within Michigan and collects contributing cases from participating hospitals for just this kind of research question, large scale health services research with clinically relevant patient reported outcome measures, things of that nature.
So we've discussed these CMS proposed thresholds in terms of health reform efforts that are looming over us.
It's really part of the zeitgeist at this point. There's palpable anxiety amongst the community about what the impact of these are going to be. And we've covered several studies in the recent episodes that we've done here in your cases on hold about various authors trying to argue against where the threshold is for the most part and trying to amplify their case that you know, essentially that there will be patients who are, who are unnecessarily, you know, I'll use the word harmed, not physically harmed, but harmed in the past in the. To the extent that they may not be able to get a total joint replacement that they otherwise might benefit from from.
And this study kind of took a similar approach just using the marquee data. Close to 3,500 cases, 3,465 collected between 2015 and 2023.
They say retrospectively identify patients who underwent TJA at our institution. So they're using the marquee, but it's not the statewide version, it's just what's at the University of Michigan.
And they looked at changes in scores for total knee arthroplasty and then they wanted to anchor these across different thresholds including the substantial clinical benefit and then where the patients were at before the surgery and how the patients felt after the surgery. If they were willing to repeat the surgery. Interestingly, or I thought it was interesting, the CMS is using substantial clinical benefit thresholds for who's junior and who's junior proposed by Lyman Eyal and Steve Lyman was a previous deputy editor for methods at JBJS for I think over 10 years.
So there's some connection there.
And that was using an anchor based method based on a single six point Likert scale question regarding patient improvement. And this was later validated with data from multiple institutions by cms.
The authors are looking at the substantial clinical benefit thresholds and ultimately they found that the change in scores, particularly at one year were predictive of improvement, satisfaction and willingness to repeat surgery, all with AUCs.
So discriminatory properties in the the 0.7, 0.71 for repeat surgery, 0.77 for satisfaction, 0.79 for improvement.
The change in scores for total hip arthroplasty was predictive of improvement with a much higher AUC of 0.85 and satisfaction and willingness to repeat surgery, which was still in the 0.77. 20% of patients did not achieve the CMS proposed substantial clinical benefit threshold. And I thought that was a missed opportunity. I would have liked to have seen the authors delve in a little bit more into that piece to better understand what were the characteristics of those individuals. We have had some other studies that covered it, but they had a large sample and I thought that that would be illustrative and informative because those 20% is non negligible of course and that that would mean, you know, some degree of penalties institutionally I think in the. In the new payment paradigm, so they say And I thought of all the studies that we've covered, this one was the most complementary of where CMS is at actually saying the thresholds closely resemble the substantial clinical benefit thresholds proposed by the cms.
This suggests that the proposed thresholds are generalizable even when a different anchor is used instead of patient reported improvement which was the original anchorage.
Now they say although there is currently no definitive proposal to link reimbursement to rate of achieving the substantial clinical benefit value which I thought was funny because I feel like everyone is acting like there is everyone is acting like this is they say that prudence should be used if a policy like this were to be implemented because it may inadvertently incentivize institutions and providers to, you know, basically exclude certain patients and they may exclude patients who might otherwise benefit from a joint replacement.
The question for me, and the question that we've discussed before is by how much and you know, I've said it before and I'll say it again in these econometric discussions the cut point is intended to be there's going to be some people who don't meet the threshold that are currently getting the so the goal is not to make sure that everyone who wants a total joint arthroplasty in the pre reform effort is still going to be able to get one in some way shape or form in the post that there'd be no point in doing the There'd be no point in exacting a health reform effort or a payment reform effort. If it's let's just make sure that everyone who wants one from before can still get it. That's really, you know, some of the calumnies that were applied against, you know, when they went through like ACOs and bundling and that there were all sorts of ways to just sort of, you know, the thresholds were so low that they were essentially not, not meaningful or that they didn't make, you know, they didn't make any effective change. Over the last 10 to 15 years, we've gone so through so many different iterations of like how are we going to reform this? Or how are we going to, you know, and ultimately, at the end of the day, it all comes down to there's not enough funding to support where we're at in terms of the level of healthcare delivery and health technology and the costs.
So from a just a scientific methods standard standpoint, I think there are a lot of interesting findings here. How much they move the needle isn't really for me to say. You know, they're saying increased medical complexity does not necessarily translate to worse or better. I don't know why it would be better. But patient perceived outcomes, patients with multiple comorbidities still experience perceived improvement.
But this, you know, in every one of these, I've gone back to the college tuition paradigm.
[00:34:08] Speaker B: You'd go there and that.
[00:34:10] Speaker A: And it just, you know, the dude living in the basement who majored in video games at Yale, he might be perfectly happy with the fact that, Right. He might have perceived that as being great, but the people who paid for it may not agree with him.
[00:34:28] Speaker B: Agreed.
[00:34:30] Speaker A: And so, you know, at the end of the day.
They're saying preoperative scores and comorbidity indices fails predictors of patient perceived outcomes. But I don't know why comorbid, like the patient is living with their comorbidities.
So the comorbidities stayed the same ostensibly between the pre and the post surgical.
So they'll probably, you know, if they benefited from the knee replacement there or the hip replacement, they're going to feel that they benefited. They still have, their diabetes is the same. So I, I just thought that was
[00:35:09] Speaker B: a weird yes and no. Unless they got a prosthetic joint infection because of their uncontrolled diabetes or had a prolonged admission because of their coronary artery disease.
And you know what I mean? And then their experience and, and some of these anchor measurements of would they do it again are going to be lower. Right.
And so then I do sort of argue the other opposite end of the spectrum is when with the really infirm patients that will have some benefit clinically, but may not have met the substantial clinical improvement or the MCID because of these comorbidities and complex social problems. I still think that some of these metrics can potentially reduce access.
[00:35:55] Speaker A: Oh yes, no, they absolutely can. If and if applied incorrectly, they will.
And there's yet to be a study that I'm aware of that has not shown that whatever the health reform effort initiative may be ACO Centers of excellence, it has a disproportionate impact on patients of lower socioeconomic status or who are living in socially disadvantaged circumstances. Every study that has been done on that, and I'm not even just talking about in the orthopedic realm, there's also evidence on that for like bariatrics and, you know, maybe carotid artery disease too, or some type of vascular surgery. There's robust literature that all of these health reform efforts, the first people who are hit are those who are already marginalized.
I don't see that these kinds of papers are all essentially raging in the darkness. Sort of.
The other point that I would say is that the willingness to repeat surgery was lower than I really expected it to be for a surgery or performed in terms of its AUC at a level lower than, you know, for what is widely held to be an incredibly successful surgical intervention.
[00:37:25] Speaker B: Yep.
[00:37:28] Speaker A: It all comes down to the cost. At the end of the day, being
[00:37:30] Speaker B: in the 0.7 realm for the majority of the outcomes. I mean that's not. Wouldn't you want them to be excellent above 8.8?
[00:37:39] Speaker A: I certainly, certainly the 0.71 is probably pretty. A pretty glaring find. And yes, I think you ideally would want it to be above 0.8. Yes, I wouldn't balk at 0.77, but the point seven one.
[00:37:57] Speaker B: Yeah, yeah, that was my thought.
[00:38:03] Speaker A: So I think that's all we have for the year. Cases on hold. We're now going to move into the honorable mentions.
The first is MRI assessment of median nerve size in patients with proximate electrodiagnostic studies by Liu With a comment.
This study is a retrospective review of electronic medical record data for patients who underwent both wrist MRI and electrodiagnosis studies within a 90 day interval over 2020-2022. The authors used a logistic regression model to evaluate diagnostic accuracy of the median nerve cross sectional area at the three anatomic levels. In identifying carpal tunnel, there were only 68 patients included with 76 risks.
The authors maintained MRI based measurements of median nerve cross sectional area, particularly at the inlet, suggest that relying solely on these measurements may not be optimal diagnostic strategy for CTs even with MRI and highly standardized measurement protocols, only poor to fair diagnostic accuracy was achieved. This raises questions about the diagnosis of carpal tunnel based on or solely on cross sectional area measurements on mri. And then we have a very interesting study Oocyte Cryopreservation experiences and attitudes among female Orthopedic surgeons. This is by Chen and colleagues with a comment and there's also a visual summary.
This study is predicated on the fact that female surgeons more commonly delay childbearing and experienced higher rates of infertility than the general population of women.
They had 169 participants to assess fertility preservation strategies among female orthopedic surgeons.
Of the 63% reported intentionally delaying childbearing, 34% reported delaying or planning to delay for greater than four years.
91 respondents 54% considered oocyte cryopreservation.
Only 21% had undergone or planned to undergo at least one cycle. A third ultimately decided not to undergo it.
And the authors maintain that the study shows a high degree of interest in oocyte cryopreservation among survey respondents. But there are persistent barriers of financial burden, inflexible scheduling, institutional stigma, and limited fertility knowledge.
Their clinical relevance is that barriers to fertility preservation during orthopedic training directly affect physician well being and the ability to recruit and retain women in this specialty.
So an interesting study there. That's where we're at for this episode.
We're about out of time. We'll try to do better next time. If you like what you heard once again, please please like subscribe Share Spread the love for our your cases on hold listening community.
And if you're not that into what we covered this time, there's a whole new set of material that's going to be dropping on August 19th. And listen for us on the 18th when Dr. Abdien will be taking the helm.
[00:41:30] Speaker B: Thanks everyone.
