We'll get through it. I have no disclosures. Like you, I'm a practicing orthopedic surgeon. I'm a team physician. We just have to know this stuff. This is the stuff that we're going to be covering today. I'll be putting in real life examples from the sports medicine literature to make it real. And whether you are actively performing your research or consuming research, which all of us do when we read our journal articles, you just need to know this stuff. And I'll try to make it clear and easy for you. So whether I am designing a clinical trial, which I do often, or consuming the literature, reading it critically, there are a number of factors that I think about, and I'm going to share those with you. The first is called the expected size of effect. We all want our surgery, our intervention, our research to be the shot heard around the world. What do I mean by that? Is your surgery, is the effect that you bring into someone's life strong enough to have a detectable effect? Now some things are going to be statistically significant, but clinically irrelevant. And we'll talk about that in a minute. Here's an example from the literature. This is Christian Gerber, preeminent shoulder surgeon from Switzerland, who published a landmark paper on latissimus dorsi transfers for irreparable cuff tears. And one of the primary outcome measures was external rotation. And you can see that the external rotation increased from 22 to 29 degrees with the P value of less than .04. This was a statistically significant change. And if you focus on me, look here at my left shoulder, if you can't see much motion, it's because there isn't much. This is about seven degrees of rotation. So you may say that this is statistically significant. I might argue this is clinically irrelevant. So for example, for that reason, I haven't adopted this particular operation in my practice. But there are ways for us to determine if something is clinically relevant. And I'm going to tell you how. This thing called the MCID, sometimes called the MID, the minimal clinically important difference. It allows us to distinguish between statistical significance and clinical relevance. We know, as I just mentioned, there's some very obvious treatment-induced statistically significant changes that are completely irrelevant clinically. So the MCID is defined as a threshold. It's the absolute smallest change in somebody's status that would be considered relevant by both the patient and the caregiver. And it's a statistic. You generate it after looking at a population of patients for a given outcome measure. And this is how we apply it. This is real data from my rotator cuff registry. The orange line represents 600 patients, roughly, who have had rotator cuff repair surgery. The blue line represents patients with known rotator cuff tears who didn't have surgery. And they're different populations. The non-operative patients are a little bit older and may have some comorbidities that made them not surgical candidates. But what I can tell you is both populations, in a statistically significant way, improve over time, from baseline to final follow-up. So I said this is an MCID slide. How do we talk about that? So the MCID for the ASES score is 20 points or 20%. So we start at the baseline here, around a score of around 50. And I add my 20 points. That's the threshold. And there's my threshold line. And you can see that this population, over time, quickly exceeds that and stays over that. So this is a statistically significant improvement that is also clinically relevant, based on this threshold measure of the MCID. Now contrast that to our non-operative patients. I'm telling you they are statistically significantly improving by p-value, if that's important to you, over time. But you can see, same thing, I add 20 points. They never cross this threshold of MCID. So here's a real-life example of something that's statistically significant, but clinically irrelevant. And remember the MCID. For each outcome measure, you know, there'll be one. They tend to hover around 10 to 20%. You don't need to know values for exams, but know the concept. Okay, the next item for studying or evaluating a research project is, is it applicable? Will it change your practice? If you assume this intervention, is it going to change what you offer your patients? Here's an example from the European literature. An indwelling femoral nerve catheter following routine ACL reconstructions. So you have to ask yourself, is pain control after routine ACL surgery a big enough problem in your life that you'd either want to study it or even read the journal article? If it is, great. If it isn't, move on and don't waste your time or resources. So what we do and what we study needs to be applicable. The next concept is that of external validity. That means that whatever findings that you have or I have, we're hoping that they're relevant for the population at large. So in my rotator cuff research, I'd love to be able to study every rotator cuff tear in the country. That would be represented by the big box, you know, millions of people with cuff tears. That's not feasible. It's too expensive. I can't do it. So what do I do? I take a study sample. And that's a small population from the big population and study that population and hopefully make inferences that we can then apply to the population at large. And when you can do that, it's called external validity. Often there's a lot of threats to external validity. You know, I work in a small kind of weird college town and maybe my findings have nothing to do with you and your practice in the coasts, for example. And so external validity is very relevant. Here's a manuscript out of the European literature, again, looking at the relationships between hamstring injuries and knee strength. I take care of a major league football team. You know, this stuff is relevant to me, but I have no idea in the world what Irish Gaelic footballers do. So this one may not have external validity in my world. Moving on to control groups. So whenever we have an intervention or a study, we like it to be comparative, you know, treatment A versus treatment B. And one of the two treatments is going to be the study group and the other one is going to be the control group. And in the best study, the two groups should look alike in every way possible and differ only in the intervention rendered. So for example, if you wanted to test the latest and greatest interference group, you know, wouldn't it be nice to have the barber twins here because they're identical twins. That's a perfect control group. Maybe some ethical concerns, but my point is control groups should look similar in every way except for what you're doing to them. And there are a lot of different ways to generate control groups, and this is listed from worst to best. Often, when you don't have other data, you can look at the literature, look at historical controls. The problem is that there's chock full of biases. And as we'll see a little bit later, the gold standard for a control group is a random assignment, and we'll discuss that later. There are a million outcome measures in orthopedic surgery, some good, some not good, some validated, some not validated. This talk's not going to touch on this at all. I just want you to be aware that when you read the literature for your own purposes, know the outcome measures that are relevant to you, know their strengths or weaknesses, and maybe know their MCID now that you know how to apply it. So study size is an important thing, and making a study big enough is relevant for a lot of reasons that we'll review right now, one of which is if you don't have enough subjects in your study, any statistical test you apply may be completely irrelevant. So it's kind of this Goldilocks theory, too big, too small. If you don't have enough subjects in your study, if your sample size is too small, you likely will just fail to provide reliable answers. There may be a difference. You just won't find it. That's power. We'll talk about that. If your study is too big, not only will you be wasting a lot of time, resources, and effort, but you, by virtue of statistics, you may find something that's relevant just by virtue of a 5% error rate, and it's completely unrelated. So don't study too many things, and we should establish your standards before you start an operation. So this is an example from the literature of a sample size gone bad. Gary Gartzman, an American shoulder surgeon, was going to provide us with the mother of all, father of all research studies to determine whether a hemiarthroplasty is better or worse than a total shoulder arthroplasty for arthritis. I think most of us have a sense now that total shoulder arthroplasty has some real benefits over hemiarthroplasty, but it's never been studied. So he designed a very well-designed, randomized clinical trial to look at that question. It was so good, the patients didn't even know what operation they were going to have until they were randomized in the time of surgery. And he did his homework, and he determined, based on some things we're going to talk about in a second, that each group needed 35 subjects in it. Well, when the dust settled and the patients published, instead of having 70 subjects, he only had 50 subjects, and guess what? Too small of a sample size, we were not able to discern a difference. So this would have been the perfect study that just lacked what we call power. There was a difference, probably, we just didn't look hard enough for it. So what is power? Power is testable. This one shows up because it's a really easy concept. Power is the probability of identifying a difference when there really is one. So in our literature, by convention, we accept a power of 80%. So that means that 20% of the time, there may be a difference between two groups, but our studies aren't designed to detect it. So that means we accept a 20% false negative rate. And here's the mantra you should remember, power in numbers. How do you improve power of a study? You study more patients. And there'll be a scenario in a test. This test only studied 40 patients. Didn't show any difference. Suspect there's a difference. How do you improve this? You study more patients. There are power in numbers. So this is very closely related to errors in statistical analysis. There's only two types of errors. One is a false positive and one is a false negative. So type one is known as a false positive. That means your experiment suggests that there actually is a difference or is an association between two things when there really isn't. Type one error, false positive, or a false negative type two. Your experiment indicates there is no association or no difference when there really is. That's called a false negative. And that's usually an underpowered study. We accept a 5% type one error rate in the literature. And as I mentioned, a 20% type two error rate. So let's look at all the other factors, and only two of them, that affect our outcomes when we're looking at sample size. We've already talked about the two errors. The other one is effect size. So how much difference do you think you're going to see or do you want to show between two populations or two interventions? The other one is variance. How much variation is there in your population under study? So we'll go through a quick learning module. I do rotator cuff research, as I mentioned. So let's say we want to look at a randomized trial looking at arthroscopic versus open rotator cuff repair. My outcome measure is going to be this thing called the Western Ontario score, the WARC. And I know well enough to know that in this population, the mean should be, let's say, 1,000. And the standard deviation, which is a test of variation in this population, should be about 450. So that sets the standard here. So basically I need to define two things to determine how big of a study I want to have. One is the effect size. Again, how much of a difference do I want to see between these two surgical interventions? And two, based on either my population or previous public populations, how much variation is there in this rotator cuff population anyway? So here we are, effect size, and this is what you need to know. The greater the effect size you like to detect between the two treatment groups, the smaller the sample size. Bigger effect, fewer patients you need to see. And this is real data. So based on standardized power, type 1, type 2 errors, if I want to find a 10% difference between arthroscopic and open, that's a small difference, I need to study a lot of patients. I need to have 319 per group, so 640 patients minimum. If I think the arthroscopic is going to be so much better than the open, maybe 50% better in my outcome score, then I only need to study 14 subjects per group or 28. So you can see how treatment effect here controls sample size. And here's the other one, variance or variation, we measure that by standard deviation. And here's a take-home one, the greater the variation, the greater the variance there is in your population, the larger the sample size you need. So all else being equal, if you have a standard deviation of 100, that's a pretty small variation, you need very, very few subjects, 4 per group. If you have a population that's all over the board, a standard deviation of 700, then you need over 125 per group. So you can see how very, and these are real numbers, just plug them into a calculator, these are real numbers. And then lastly, I talked about power, but if your lifelong research program hinges on something that's a subtle finding, you don't want to miss it. So how do you improve your power? How do you find that difference that really exists, albeit small? You study more patients. So the greater the power, the larger the sample size you need. And you can see here now, as I march from 70 to 80 to 90% power, because I just don't want to have that 20% false negative rate, I need more subjects per group. And this is kind of how it looks in real life. More in numbers. And then, as an aside, we determine based on these thoughts that I need 100 subjects per group, then there's a lot of other things. How much money do I need to do this? We always think we do a lot of one thing, and we try to study it, and we aren't generating data, so we collaborate with our friends, and then we have attrition, we lose patients to follow up. And if this stuff interests you, here's a link to a free software online where you can plug these numbers in and see how they change with differences of variation, effect size, and things like that. Okay, we're going to move ahead and talk about study designs. And I'll try to categorize these clearly. Studies are either observational or experimental. Observation is you're observing a population and describing what you see. An experimental study is one in which you have an intervention. A laboratory study is a good example of that, or a randomized trial would be an experimental one as well. Things are either retrospective, meaning whatever you're studying happened before today, and you're looking back in time, or prospective. We generate a study, we start it today, and we collect data forward over time. Tends to be a little bit better quality data. When we talk about observational studies, there are several kinds. We know what the case series is, it's the most common type of study that's in our literature. It tends to be a low level of evidence study, and we'll talk about that later on the level of evidence part. Then we talk about a case control and cohorts, and those are two important study types, and I'll explain those in greater detail. So let's look at the case control study. Let's go back in time over 100 years ago, 150 years ago, when it wasn't clear that smoking caused cancer. So what would happen, you're interested in studying this problem, these blue guys are cancer patients, and you've noticed that you've seen a lot of patients coming in with lung cancer and you just don't know why. So when you have a rare disease, a case control study is the way to go. So this is your case group, the cases are your cancer patients, and now we need a control group, a comparative population. You pick a group of patients that don't have cancer, but in this study you want this group to look like this group, and they differ only in the blue guys have cancer and the red guys don't. So you match them for age, gender, occupation, geographic, you match them for as many things as you think might be relevant. So the difference might be teased out in the study later. So what do you do now? You have your cancer patients, you have your non-cancer patients that look alike, other than cancer, and you take histories of both populations. And what do you find? The group that has cancer, of the five cancer patients, three of them said, oh yeah, I used to smoke a lot of cigarettes, and of the non-cancer patients, none of them were smokers, so you compare the histories statistically and you draw conclusions. And this is how you would determine, for example, that cancer is caused by smoking. So case control is an excellent way to look at rare diseases, and it's a retrospective type of study. This differs from a cohort study, and I think you're probably more familiar with this one. A cohort study is good at looking at rare, what we call, exposures. Rare diseases for case control, rare exposures. So let's say that we live by Three Mile Island, there is a radioactive leak, and we just don't know what the health effects are, or in this case we can use smokers. These five guys were exposed to radioactivity, or they were smokers, and we follow them over time, and then we have a comparison group that were not exposed to this thing, we follow them over time, and guess what? The smokers, or the Three Mile Island guys, develop cancer at a rate of three out of five, and our comparison group doesn't. So you can see, by comparing the groups, you can see that there's a different distribution of disease based on exposure, so then you can look back and make some inferences. Exposure causes cancer, and that's done usually in a prospective way. So that's cohort. And then when we talk about experimental studies, the gold standard is a randomized clinical trial. I mentioned that in the control groups, the best control group is randomization, because it eliminates biases to the best of our ability. So let's say we want to do a randomized trial, the millionth one, unfortunately, looking at hamstring tendon autograft versus patellar tendon autograft, for whatever reason we think it's still relevant, and we have a group of patients with ACL tears. So what we're going to do is randomly assign. This group gets patellar tendon, this group gets hamstring, and we follow them up over time for whatever outcome we want to measure. Let's say arthritis. Well, we think we know what outcomes or what patient factors affect rotator cuff outcome. Sport level, obesity, age, presence or absence of chondral or meniscus issues at the time of initial surgery. We know that stuff, but we don't know what we don't know. So there may be a thousand other things that are relevant that we're not even thinking of, really. So the power of randomization is it takes all those unknown variables and distributes them evenly between the two groups, and essentially, they should cancel each other out. So when the dust settles, hopefully, we're seeing the effect or the difference between hamstring and patellar tendon alone and nothing else. So randomization diminishes bias in a study. We're going to move ahead and look at—we looked at observational, experimental. Now a very common type of study is one that evaluates the accuracy of diagnostic tests. They're pretty common in our world in terms of physical exam findings, x-ray findings when two surgeons are trying to determine if they can agree on something, their agreement tests. So this is the basis for that. It falls under the category of screening tests. Now it's easier to think of cancer, for example, but we think of the natural history of disease. So there's a spontaneous mutation back here. The red line represents time. There's a spontaneous mutation. Disease is initiated. Disease marches along its natural history. It comes to a point where it's detectable by screening, whatever the screening test would be, blood test, something like that. And then it marches along, and then the patients are now able to come to you and say, I'm having clinical symptoms. And then ultimately they develop complications, and in this case it may be death. So the purpose of a true screening test is one that identifies disease prior to the onset of clinical symptoms so you can have an intervention. So in orthopedic surgery, not so much, but I'm just using this as an example. What we'd like to do, though, is in order to do this, we'd like to be able to distinguish patients who have disease or a finding or an MRI abnormality from those who don't. And we do that every day in our practice. And so in order to understand that, you also have to understand that biologic variation exists in human populations. Most things are either distributed bimodally, and I'll show you what I mean by that in the next slide, or unimodally. So many of us work in institutions where you have to get tuberculosis tested, so the PPD, in this case it's the other test, the BCG, is a test to determine if we've been exposed to tuberculosis before. And this is a bimodal distribution. These subjects have no induration, they've never been exposed. These subjects have a lot of induration at their injection site. And you can see that this population, just by looking at the graphical data, is very different than this population, so it would be very easy to create a screening test for this. You could just say, well, anyone who has more induration than, let's make a threshold of 9, probably is exposed. So orthopedic surgery example. Dr. Daniel introduced us to the arthrometer, the KT, now it's in a far generation, but KT-1000 was very common in our practices before. And you can see, this is ACL sufficiencies, normal ACLs, these down here are patients who have ACL insufficiency, and this axis is the side-to-side millimeters of difference. And if you superimpose this on this, you can see a bimodal distribution. Normal ACL tears. Now sometimes things are unimodal, and this would be systolic blood pressure in this population sitting in this room. And what you can see is there's basically a bell-shaped curve, and so it'll be really hard to know here what determines disease state and not disease state, so you'll have to set off a cutoff point that's more biologically or disease-driven than something that's very visual like the KT difference. Okay, so we've established that we're looking for differences in populations, and there's ways that we assess these tests. Remember this could be your Lachman test for ACLs, it could be your MRI finding. How good is your test? We describe validity, and we test that by sensitivity and specificity, we'll go through each of these. Should I use the test, yes or no? We measure that by efficacy, and we'll use the example of positive predictive value. And then how much repeatability or reliability is there in the test? If I do the test, or if you do the test, does it make a difference? So the validity is the ability to distinguish between those who have disease and those that don't have disease. And epidemiologists love these two-by-twos tables, and we'll go through this in just a minute, but basically we compare your test result, MRI finding, physical exam finding, it's either positive or negative, with a gold standard. It could be a biopsy, it could be a surgical finding, whatever we determine the gold standard is. And so this is truly testable. So there will be a test question where they say, calculate the sensitivity, calculate the specificity, or calculate the positive predictive value of this test. And if you understand this two-by-two table, it's really, really easy. So sensitivity is the ability to detect true positives. That means the probability that someone who actually has a disease is detected or defined that way by the test. Specificity is the opposite, the ability to detect true negatives. The probability that someone who doesn't have the disease tests negative. And this is how it works in a table form. Remember, we have a gold standard on whatever it is, let's say biopsy. And patients either have disease or don't have the disease. And then we have whatever test you have under question, a physical exam finding, for example. And your Lachman's going to either be positive or negative. And then these get populated. So the sensitivity, I said, is the ability to determine people who really have the disease. So who has the disease? This column. We know they have the disease, it's the gold standard. A and C have the disease. So of all the people that have disease, who tests positive? That means A over A plus C. So the people that test positive over all the people that have disease is our sensitivity. Very simple. Specificity is for the patients who don't have disease, who tests negative? So who doesn't have disease? This column. These patients don't have disease, so B and D. And these patients in this row test negative. So of all the patients who test negative B and D, how many actually test negative D? So it's D over B plus D. Look at it a few times. It should make sense. And here's an example. I want to know what the sensitivity and specificity is for the diagnosis of ACL tears. I'm going to say that the gold standard is arthroscopy. You're looking at it. You're feeling it. I want to know how good MRI is in determining ACL tears. So here's this four by four table. Let's calculate the sensitivity. Arthroscopic is the gold standard. So who has a torn ACL? 421 patients. We know that because our gold standard says that we determine 421 actually have the disease. Of those 421, how many tested positive by MRI? 394. So the sensitivity is 394 over 421, 94%. So how do you interpret that? Is 94% of patients who have ACL tears will be detected by MRI. Let's move along to efficacy. We're going to use the same two by two table, same concept. We like to use our data to be predictive as well. So a positive predictive value is the probability that someone who tests positive actually has the disease, and the negative is the corollary. Someone who tests negative doesn't have the disease. So we go back to this two by two table, same concept, gold standard. These people have disease. The rows are people who test either positive or negative. So the positive predictive value is as follows. Of all the people who test positive, so now we're going to go across the row, A's and B's test positive, how many of those patients actually have the disease, the A's? So A over A plus B is your positive predictive value, and let's go back to our ACL data. So we know that we had 426 positive MRI reads, those tested positive. Of those, 394 we know are true positive. So the positive predictive value is 394 over 426, 92%. So you can say if someone has a positive MRI, I'm sorry, you can say that if someone has an ACL tear, there's a 92% chance that if it's positive on MRI that they really have it. So just look at the two by two tables and get a sense of how to play with them, because there's usually a test on that. They'll give you all the information you need to construct the table, sensitivity, specificity, or usually positive predictive value. Okay, sorry for hitting you in the head with that one. This one usually isn't tested, so take a break. Reliability, the ability of a test to produce consistent results when it's repeated. So do we have repeatability? Do we get the same test when you do it and when I do it? And there is an intra-subject variation, meaning I can get different results on Monday, Tuesday, and Wednesday, and there's inter-observer variation. I do a test and I get a result, and Dr. Gill does a test, and we may get the same result or we may not get the same result. And we use what's called the Kappa statistic to describe agreement. So in an agreement study, again, you don't need to know this, but get a sense that if we agree statistically more than 75% of the time, that's considered excellent agreement, and anything less than 40% is considered poor agreement or that would be equivalent to chance alone. So the reliability, the reproducibility of a test is communicated by the Kappa statistic or agreement. All right, we're going to move ahead to epidemiology. Until now we talked about study design and study types. Epidemiology refers to the study of populations, health factors, entire populations. These two terms show up every time, and it's a good thing because they're really easy. You need to know the difference between prevalence and incidence. Prevalence is a snapshot in time. It's the proportion of individuals within a population that have a disease at a given point in time. So the prevalence of obesity in American adults in 2001 was 21%, okay? Point in time, what's the disease burden? Incidence, on the other hand, is a rate. It's a moving target. And it describes a proportion of new cases that appear over a period of time. So here's an example of an incidence. Let's say you're interested in ski injuries. You sit down at the bottom of a ski hill for a season, and you note that 56 skiers that passed by tore their ACL over a six-month period, and a total of 500 skiers went by you. So the incidence of ACL tears is 56 skiers out of 500 over half a year. So the incidence here is 0.22 injuries per person per year. So the year here introduces a rate. So prevalence is a snapshot. Incidence is the number of new appearances of a disease, for example, in a time period. And that's all you need to know. It's a little bit relevant moving forward when we talk about the different tests, but I don't think you need to know that. So how do we describe association between a risk and an outcome? One way we describe it is called the relative risk. That estimates the magnitude of association between an exposure to a risk and an outcome of interest. Let's go back to smoking and cancer because it's easy. So the relative risk is the incidence of cancer—the incidence of outcome, cancer—in exposed smokers over the incidence of outcome in unexposed. So incidence of cancer in smokers over the incidence of cancer in nonsmokers. So we're looking at exposure to risks, and that's your relative risk. And how do we interpret that? If a relative risk equals 1, that means the denominator and the numerator are the same. That means there's probably no association between the two. If the relative risk is greater than 1, then obviously the numerator, the smoking, dominates and is associated—there's a strong association. And the greater it is—the more it is greater than 1, the stronger the association normally. Now what happens if the denominator is bigger? It's going to be—simple fraction—it's less than 1, and that means whatever that exposure is on the bottom is probably protective in some way. So that's how we interpret relative risk. An odds ratio is exactly the same thing. It's either 1, greater than 1, or less than 1. And we use an odds ratio when we are looking at retrospective studies like a case control. You don't need to know that. But you need to know that a relative risk or an odds ratio produces a ratio that's either 1, greater than 1, and less than 1. And it helps you interpret the association between an exposure and an outcome or disease. Evidence-based medicine has been a buzzword now for quite some time. And what it means is, you know, we're kind of obligated to use or apply in our practices the best evidence in decision-making for our patients. And I think you're all aware of this tree of level of evidence, and it looks like this. This is as defined by the JBJS. Level 1 would be a high-quality, randomized clinical trial, controlled trial, or a systematic review of such, all the way down to level 5, which would be expert opinion. And you can see as you move down, the evidence is diminished in value. And I'll give you an example. I hate this example, but, because it's been repeated too many times in literature, patellar tendon versus hamstring ACL reconstruction. Let's look at levels of evidence. Let's just say you want to design a study in your practice. Let's look at the different ways you can study this based on levels of evidence. So if you wanted to have the highest level of evidence, you would start, you know, Monday in your practice, and if a patient comes in with an ACL tear, you randomize them either to patellar tendon or hamstring, and we talked about why that's good, because all the biases and all the things we know and don't know about are distributed equally. That's the best type of evidence. Let's move down the chain a little bit, a prospective comparative study. There's no randomization here, but we'll start tomorrow, and we'll start collecting data prospectively. Some of your patients are going to get patellar tendon, some are going to get hamstring, you decide, but we follow them prospectively. There's a lot of bias in this study, but it's still pretty good. This is kind of a cohort study, you're comparing two things, and you're looking at it prospectively, which is usually a little bit better. Prospective is not quite as good, so we could go back to the office tonight and say, let's look back at 10 years of our graphs and see how they did. And it's okay, but you can see there may be even a little more bias introduced here. Why? Because that high school football player 10 years ago, he may have given a patellar tendon because he was a high school football player, and the 45-year-old sedentary guy gets a hamstring, but 10 years later, you're not teasing that out, so there's a lot of bias when you look at the retrospective stuff. A case series is not comparative. You simply say, my hamstrings did this, or my patellar tendons did this, you're just reporting on a population. And lastly is expert opinion, you know what that is, you give me your opinion, and it's worth what it's worth. Expert opinion is really important when there's no other data. So you'll see these consensus statements coming out by the academy, and they can be very frustrating. I've been involved with them. But if there's no high-quality data, then all you can do is assemble a group of experts and say, we think that Patel or Tennant is better or not better based on expert opinion. And then hopefully over time, we have more and more studies and creep up this level of evidence ladder. So that's what level of evidence means and how it works. All right, let's get to the meat of this thing. I'm going to make this very clear and very simple and not onerous. And this is somewhat high yield. There'll be a question or two on this, and I'll try to make it as simple as possible. There's two types of statistical analyses, I would say. One is descriptive statistics. You're basically describing a population. What does your population look like? Mean, median, mode, standard deviation, and I'll show you some examples. The second part of statistics is inferential statistics, basically applying tests. And I think when we think about statistics, that's what we think about. And what we're trying to do there is determine whether the difference between the Patel or Tennant and the hamstring was real or due to chance alone. And that's all we're doing with our statistical tests. So let's just talk about some statistical descriptive statistics. How do we describe populations? I think you know what these words mean, median and mode. So if I asked everyone to take out your wallet and tell me how much cash you had, so the mean would be, we add up everyone's money and we divide it by the number of people in the room, and the mean value is that, say $40. Now, the median is a little bit different. We have everyone with their cash in their hands, and we would line you up from my left to my right against that wall from least cash to most cash. And if there's 100 people, we'll go right to the 50th person and say, how much money do you have? That's the central value of the population. That's the median. And the mode is the one that is the value that is the most common. So out of 100 people, if 23 of you had $45, then $45 is likely to be the mode. So you can see on the left, that's a perfect bell-shaped curve where mean, median and mode are exactly the same, but it's very uncommon that populations look like this in our world. There's usually a little bit of skew, and the mean and the mode and the median kind of fall a little apart from each other, but by knowing what those values mean, you can get a sense of how they can tease this bell-shaped curve one way or the other. The other term you should know is the standard deviation, and that's simply the statistical way to describe how much variation there is in a population. So if you look at enough people, you should get a bell-shaped curve, okay? And the bell-shaped curve could be nice and broad like this. That's typical, but if we studied something that was very, very homogeneous, for example, actually the test result from this test you're going to take, if you're taking the test, is a very steep curve. So it'll be a very, it's a bell-shaped curve, but it'll be a very tight spike. Still bell-shaped, but just a different distribution. But what's interesting about the standard deviation is by definition, if you take the mean here, zero, and march one or two standard deviations this way, two standard deviations here, and then one or two here, so if you look at two standard deviations on either side of the mean, that represents 95% of the data by definition. And that is a concept you should understand because that's how our p-values and our statistical tests are determined. So two standard deviations from the mean on either side represents 95% of the data. We can also describe that as a confidence interval, okay? So what does that mean? Let's go back to our, look at this image here. We took everyone's wallets, it turns out that the mean is $40, and we describe the confidence interval, $37 to $42. That means that I am, if, I am, how do I do that? I took the mean and I determined the standard deviation and I added two standard deviations this way, and two standard deviations this way, and that created our 95% confidence interval. So what can I do with that? I can say that if next year, assuming the population that attends this meeting is similar in makeup and it often is, next year that if we did the money test and took everyone's money and analyzed it, that I am 95% confident that the average dollar amount in the wallet of this population is going to fall between $37 and $42. And that's how we use confidence intervals. And we accept a 5% error rate, type 1 or a false negative rate. So there may be here next year we get some high rollers and the mean ends up being here, but that's only going to happen 5% of the time. So we accept that. So understand what the confidence interval is. And then, in concluding, in a slide or two, a very common type of test question is, here's a scenario, what statistical test would you apply? And it seems daunting, but I'm going to walk you through it and it's going to be really, really easy. So you should be able to get this question easily. Alright, so the first thing is, what type of data do you have? Data is either continuous, a continuous variable, things like age, weight, something measures strength, something actually is measurable and has a wide range, an infinite range. The other type is categorical or discrete data. It's categorized, male or female, yes or no, an outer bridge score of either 1, 2, 3, or 4. That's categorical data. So you need to know what type of data you have and then you have to know, is your data normally distributed or not? And they're going to tell you that in the test. They won't tell you what kind of data you have, you kind of have to be able to identify the difference, but they almost always will give you normalized data. So what does that mean? A normalized data is bell-shaped curve and usually small populations may not be bell-shaped curve, so that would be non-normal. And don't worry about the non-normal because they won't test you on it. But just know that for every test that you have for normal data, there's a cousin test for abnormal data. So the normal tests are called the parametric tests, please don't memorize that, and the non-normal data tests are called the non-parametric tests and they have funky names and I'll show them to you in a table in a minute. But the common question is, what test do I apply? This is what you should know and it really makes sense. So what type of data do you have? Continuous or discrete? And if continuous, is it normally distributed, yes or no, and they always give you that one, and then how many groups are you going to study? And if you use this simple algorithm, you know 99% of your statistical tests. So here's an example. What analytical technique would be used, or what statistical test would you use to test the hypothesis of equal, and I would say put in a continuous measure here, equal weight in two groups tested in the laboratory? And they tell you the data are distributed normally. So it's basically saying, of these five statistical tests, which one would you use for normal data? And we're looking at two different groups, we're comparing two groups. So let's go through this. It's continuous data, because we're looking at weight, it tells you it's continuous. It tells you it's normally distributed, and now it's just simple of how many groups are you comparing? In this case, we're studying two groups, so you just come down here, and for this test, let's go back, we go for the t-test. So just follow that line, sorry it went backwards. Let's do another one, very simple. You're testing the stiffness of four reconstructions. So stiffness is your data, stiffness is continuous, correct? It can go from zero to infinity, and we're testing four different techniques. So four techniques, continuous data, what test do we use? Don't worry about all these listed here, because you only need to know one of them, so you go down here. Our data is continuous, it's normally distributed because it always will be in these tests, and how many groups? It's not one, it's not two, it's four, so it's greater than two, so we want to use an ANOVA test here. And you don't even have to know what an ANOVA test means, just circle the answer and move on. Leave that to me. And then the last example, you're comparing two groups of patients who have had some kind of treatment, and each patient is graded as poor, good, or excellent. So all of a sudden, ding, ding, ding, ding, this is not continuous data, this is categorical or discrete data, and this is the freebie. Why? Because it would only be one thing. If it's not continuous data, if it's categorical or discrete, you always use chi-square. So that's a gift when you can see categorical data. And that's what you should, just look at this, you know, the day before the test, and you really need only to memorize one or two things here. And I mentioned, you know, for each parametric test, there is a non-parametric test with a really funky name, please don't memorize them. If they give you an example of a non-parametric test, I'll be shocked, just pick a name that looks kind of funky out there that you don't recognize. And then I'm going to finish up on this issue of a p-value. What does actually p-value mean to you or me? So by convention in our literature, we determined the p-value of less than .05 is statistically significant. What does that mean? That means there's less than 5% chance that the observed difference between two populations was due to chance alone. So if it's not due to chance alone, then we assume it's due to your intervention. So that's all that it means, a 5% chance that there's a real association between an intervention and an outcome, 95% chance. And we accept that 5% false positive rate. For those of you who really need more of this, I hope there are very few of you, but the AJSM in the late 90s published a series of primers written by orthopedic surgeons for orthopedic surgeons. So I just include the reference list. They're really well done if this interests you, if it doesn't, move on. But these are some good references and I thank you very much for your time today. Thanks, Bruce. That was really great. Everyone feel like they understand statistics a little bit better? I know I do. And I agree that along the lines of the brachial plexus, as far as just drawing it, like looking at it five minutes ahead of time, do exactly what he said. Draw that little two-by-two table, because you are going to get something on that two-by-two table. Remember, I think 6% of the questions will be stats. So do that. And then that little flow sheet that he did on the T-test, ANOVA, there's going to be something on that as well. So just draw those as well. So this was the talk that was moved from Saturday. Does everybody have it? Everyone can't find it. Or from Friday, sorry. This is the rotator cuff, muscles and loss of motion. It might be hard to tell. I'm not Dr. Adelin. I'm not the spine. All right? So this is the list of things we're going to go over. Does everybody have it yet? Want to go over it? Page 335. Thank you. Anyone not find it yet? All right. We're good. So here's a list of things we're going to go over. We're going to start with adhesive capsulitis. One of the things to remember, the difference between all loss of motion is not adhesive capsulitis. Adhesive capsulitis has to do with the capsule itself. So you have the articular cartilage of the humeral head here. You have subscap coming up and around. So you want to always decipher clinically, is this an intraarticular capsular problem? Or you might get a question about stiffness post-instability surgery, which especially if it's an open surgery where the subscap was taken off or they give you a picture of something with an old staple in there from an old putty plat or something like that, you want to differentiate between the two different types of stiffness. The vast majority of questions on stiffness will be adhesive capsulitis as opposed to the post-traumatic. So just to look at what is adhesive capsulitis, it's a thickening of the capsule itself. It gets contracted. I like to use in my patients the analogy of shrink-wrapping something, the best way to look at it. The association, and by the way, I put in this talk, as you might expect, everything that's been asked, everything you really need to know is in yellow. So the association, as soon as you hear adhesive capsulitis, the next thing you ought to be thinking about is, are they going to ask me something about diabetes? Okay, because they go together and it is a big risk factor for it. In addition to diabetes, hypothyroidism, and females more than males. The other thing that will come up is the essential lesion, if you will, for adhesive capsulitis. So just like with shoulder instability, we talk about the essential lesion being a bankart or labral tear, it is a contracture of the coraco-humeral ligament, the CHL. And when that is contracted, that causes a loss of external rotation. Coraco-humeral, that coracoid is the anterior part of the shoulder, from the coracoid to the humerus, a structure in the front. If that's tighter, if you're tight in the front, you can't rotate externally. That's the way to remember it. I like to think of adhesive capsulitis as the three phases that were initially talked about. Carter Rowe and others initially started talking about it. But when you see your patients, sometimes they'll come in and they're very painful and stiff. That's the painful freezing phase. Most people would say do not operate on people when they're in the painful freezing phase because it's just going to scar up again. The frozen phase, where there's a loss of motion but they're not so painful, that's typically if you're going to operate when you operate. And then the thawing phase. And the key thing is look down here at the length of time, 0, 6, 12 months, 18 months. The natural history of adhesive capsulitis, if left untreated, is about 24 months to resolve. And I use resolve in quotes because a lot of people still have some pain and some stiffness. So remember, adhesive capsulitis is not always global loss of motion. It depends what part of the capsule is most involved. You certainly can have the entire capsule involved, but it just stands to reason if it's the front of the capsule that's involved, you're going to be losing external rotation at the side. If it's the back of the capsule, it's going to be internal rotation. So depending on what part, axillary pouch, kind of more elevation. Beware of when they give you a vignette of it looks like it sounds like it's going to be like a diabetic 50-year-old woman with loss of motion, and by the way, she fell off her bike. Sometimes they will show you a picture, and looking at that x-ray, it will be a dislocated shoulder. And they want you to jump in. The first choice is going to be adhesive capsulitis, do a capsular release or something like that. But make sure that they're not trying to show you a dislocated shoulder. Arthritis is a loss of motion. That will be easy to see. But with a dislocated shoulder, it may come up. And the goal of the plain films, you want to look for arthritis, high riding heads, osteopenia, and just make sure you have an axillary view. The initial treatment of adhesive capsulitis, unless they give you disclaimers, has been in PT for a year, has had three injections, something like that, it's non-op, non-op, non-op, until they tell you some, again, some modifier. If they do go to surgery, typically they do not want you to answer manipulation alone. I'm not suggesting what we do in practice, anyone who's doing that alone in practice is right or wrong, but for the purpose of the test, the answer is not closed manipulation under anesthesia, because it's been shown to have a higher risk of fracture, cuff tears, dislocations, axillary nerve injuries. The answer is arthroscopic capsular release. And then you can do your manipulation kind of after that. And potential risk of a capsular release is the fact that the axillary nerve at the 6 o'clock position is a little over a centimeter away. Post-op, sometimes you'll ask a question about how to treat these patients post-operatively. You never want to, say, leave an indwelling catheter in anymore. It's been shown to have chondrolysis. They might give you a vignette that shows a 22-year-old with a trashed shoulder with arthritis, loss of motion. The most likely cause of this complication is, and it'll be the glenohumeral catheters for pain management. And I'm going to be repeating this because it comes up so often. Diabetes, diabetic patients have a worse outcome than everyone else for capsulitis. So a 47-year-old woman, painful stiff shoulder for three months. The other thing for adhesive capsulitis questions is look at the duration of time. Three months, stiff. She's had anti-inflammatories. She's had an injection, painful, normal x-rays. Management at this time, they kind of want you to jump to capsular release. But remember, it's three months. So it's a physical therapy program. Recalcitrant idiopathic capsulitis. We've already said this three times. What's the answer? As soon as you hear recalcitrant, association, it's diabetes. Fifty-year-old man had a slap repair. As a general rule, the tests are not big fans of slap repairs. And so they're going to give you a lot of post-op complications from them. And therapy for nine months after surgery. So at this point, they're trying to say that they no longer want you to kind of go down the non-operative road. Still stiff, 110 degrees, 30 degrees ER. What's the best treatment? And it's going to be a capsular release. Stiff shoulder, post-intervention, nine months. And they're not going to say four or five. That gets too controversial. The first question I gave you was three. This one is nine. So just look at the lengths of the time. Fifty-year-old tennis player, PT, on and off for three months with no relief. Globally limited motion, positive impingement, and active compression test. Remember, active compression is for a slap tear, but it's a 50-year-old. So the test is never going to want you to do something about a slap in a 50-year-old. Medical history, unremarkable. In other words, she doesn't have diabetes. Strength test is symmetrical, so her cuff is okay. What would you do next? So three months, PT was only done on and off. Cuff's intact. Not looking for surgery at three months with a PT program that's been on and off. Remember, that's not just a throwaway. These are the little buzzwords in the questions you want to look for. The fact that they added that means they want you to take that into consideration. Forty-seven-year-old, fell on an outstretched arm six weeks ago, immediate pain. All right, so this is the classic. This is one of the questions I talked to you about. Forty-seven-year-old fell. So yes, limited motion, 120 degrees of elevation, zero degrees external rotation. But 47-year-old men, first of all, do they tend to get a lot of adhesive capsulitis? Not really. Usually they're going to give you kind of a middle-aged woman. Fell, outstretched arm. What happens with falling on an outstretched arm? Dislocations, right? So here's the different options. Here's the X-ray. So AP, you look at it quickly, and you say, well, maybe that's not so bad. So let's go back to the choices Here's the X-ray. Axillary radiograph. All right, let's hold on to that. MRI arthrogram. They seldom want, after a trauma, you to have the first test be an MRI. ORIF. Nope. Sling immobilization. Well, they're already stiff, so that's not it. PT with range of motion. All right, so maybe E. Maybe they want to get the motion back. But if you look here, you're looking at this, and it's a little bit funny. Maybe there's a little overlap. This is a patient that dislocated their shoulder, right? And here's a picture of the axillary. It's a locked posterior dislocation. So be aware of that vignette. But again, the vignette was a little different. It was a 47-year-old man who fell. So that's going to be more of a traumatic injury than an adhesive capsulitis. AVN, the anterolateral branch of the anterior circumflex humeral artery, or the arcuate artery of Lange. Associations in yellow here. These are the ones you're going to get. It's going to be a vignette either with chronic steroid use, alcoholism, or sickle cell. I've got the rest of them listed there, but those are the three that are going to be part of the test. Everyone got that? Last thing to remember about AVN. It's very simple. If the head is concentric, so if it's still round, they want you to save the head, whether it's with a core decompression, whether it's with PT, whatever it is. But if the head is round, save the head. If the head is not round, if it's got a crescent sign or collapse, they want you to replace the head. Don't do a core decompression. Don't do something else. So round head, keep it. Not round, replace it. There's five different types. Same with femoral AVN of the hip, but the main issue is in 1s and 2s on MRIs, you see a little bit of the signal here in stage 2, the FECOD classification, but there's no collapse. You start to see a crescent sign in phase 3, which is the collapse of the head. So as long as the head is basically concentric early on, we want to do non-operative first. For the purpose of the test, they do feel that core decompression, although controversial a little bit, but in stage 1 and 2 does give good results. But also, remember, in sickle cell, core decompression does not alter the natural progression of the disease. Stage 3, remember, stage 3 is where you have the crescent sign, you're starting to get collapse. While some people are doing it for the purpose of the test, you don't want to answer core decompression as your first line of treatment. It is a little bit more successful in the shoulder than the hip in stage 3, only because the shoulder is not a weight-bearing joint. But again, for the test, don't answer core decompression. If it's 3, 4, or 5, so any kind of collapse, this is the one time when hemi-arthroplasty still gets an answer on the test. And there can be no glenoid involvement. Now obviously in practice, that's evolved. A lot more people are doing totals now for it, but hopefully if they want you to say hemi, they're going to show you a grade 3 or a collapsed head, and they'll say the glenoid is okay. So if they say the glenoid is okay, you can go hemi. So, 35-year-old man with alcoholism, with shoulder pain. So alcoholism, younger man, I'm thinking AVN is it. Decent motion, and had a cortisone, no relief. And past several years. So they don't want you to do several years of pain, had an injection, had an intervention, no relief. They're looking for something surgical. X-rays show no collapse. So again, no collapse, and MRI shows AVN. Here they want you to do the core decompression. 39-year-old right-hand laborer, increasing shoulder pain over the past two months. Short time here. Despite PT, NSAIDs, cortisone. He's had issues with stiffness in the past, but something's changed recently. Medical history, significant for hep C. And palpation, tenderness over the greater tuberosity. So looking at the MRI, he's got edema, he's got evidence of AVN, he's starting to get some collapse of the head. So because he is young, again, glenoid okay, they're looking for hemi-arthroplasty. Laborer, they don't want you to do total shoulders in laborers. So they really should be giving you that type of indication if they want you to go hemi over total. If you look at arthritis, the key thing for all your arthritis questions and management is the status of the rotator cuff. If the rotator cuff is intact, we're going to head down the total shoulder route. If it is a concentric head with an intact cuff, then they might want you to do a capsular release. Remember that people with post-stabilization arthropathy, so someone that's had an open bank card or putty plat or something in the past, when the subscap has been violated in the past, typically they get arthritis and they have stiffness and external rotation that's going to push the head posteriorly and you're going to have bad posterior glenoid wear. So when they show you an x-ray, make sure you look at the version of the glenoid, where the humeral head is, is it wearing down the posterior side preferentially, because they might want you to talk about, and I know I'm not a big fan of this for the test, but it's been on there, talk about eccentric reaming as something you have to modify when you do your total, and I'll show you an example of that. When someone's had previous shoulder fracture, so post-traumatic arthritis, or an osteotomy, or if they've had arthritis due to AVN, the results of arthroplasty are not as good. There's more stiffness in patients who get an arthroplasty after AVN. Post-capsuloraphy arthritis, a couple things to think about. They get arthritis, but you always want to check the status of the subscap, make sure that hasn't been pulled off along the years in an old procedure. Here's a Latter Jay or Bristow procedure, but again, characteristically you get the posterior glenoid wear and an internal rotation contracture. And definitely in these cases, and this is the other thing we'll go over and over, for arthritis now, do not answer hemi-arthroplasty. It's very well accepted now. Total shoulder replacement gives you better pain relief than hemi-arthroplasty. Contraindications, this is the other one that's going to come up, is the absence of a cuff, or if you have an axillary nerve injury and your deltoid doesn't work. Those are the main ones, there's other ones, Charcot joints and the like, but the ones you'll be tested on will be the lack of a cuff, and that's because of so-called rocking horse deformity. The advantages of a humeral head, if they are, the only advantage of a hemi is that you don't get the biggest problem with a total shoulder, which is glenoid loosening. The advantages of a total shoulder is that there's actually less revisions in total shoulders than there are in hemis, and usually because the hemis wear down the glenoid and cause more pain. The other thing, if they really get picky and talk about peg versus keel glenoids, hopefully they won't, but if they do, there's no difference between the two of them. Again, total shoulders outperform hemi. Complications, glenoid loosening. It's really all you need to know with a total shoulder hemi. Always pick total over hemi if you have an intact cuff, and the biggest problem with totals, glenoid loosening. What about arthroscopy? Again, you need a concentric humeral head if you're going to start talking about arthroscopic capsular releases. So a 65-year-old woman underwent a total shoulder on the left a year ago, now she can't tuck in her shirt behind her back with her left hand. So reaching behind, this is internal rotation, that subscap, she had a total shoulder which had a subscap takedown, so I'm going to look here for something that has subscap involved, and there it is. A 65-year-old man with arthritis is supposed to have an arthroplasty, and here's a CT scan. So remember I said, don't forget to look at the version of the glenoid. So this is a patient that's got posterior wear, and they want you to say, you don't have to do a reverse, again that's only for a lack of a cuff. We're not going to bone graft, it's not so bad that we're going to put a big bone graft, and they're not going to ask you about bone grafting glenoids on this test, or building it up with cement, or osteotomies, they're not going to want you to do a posterior. So it's really going to be eccentric reaming of the glenoid to center it. And by some of these I have on here, I'll let you read through them later, the explanations of the answers, and they're worth reading because they give you good extra info on the subject. So here's one of those pictures I showed you before. As soon as I see staples in here, in arthritis, this is probably someone that had some kind of post-capsuloraphy arthritis. So a 50-year-old athlete, they even say an athlete, severe pain and loss of function for the past year. So if it's been for a year, they don't want you to send them to physical therapy. And here it is, he underwent a surgery for recurrent dislocations 22 years ago. Management, well we're not going to fuse the shoulder on an athlete, we're not going to do a glenohumeral debridement, and so it comes down to resection or arthroplasty, obviously not manipulate them, obviously not, they want you to do a total. 55-year-old left-hand dominant woman, worsening pain, has already had lots of PT and injections, and so that's not going to be an answer, non-operative treatment. MRI, joint space narrowing, well why are they telling us joint space narrowing on an MRI? That's not how we go about assessing arthritis, but they want to be able to say concentric wear on the glenoid, intact cuff. So intact cuff means total. Don't go down the humeral head replacement route because pain relief is much more predictable after a total. So osacromialy, two things to remember about osacromialy, one, don't operate on them unless you have to for the test, two, it's between, and I don't know why but this question does come up a lot, metachromium and mesochromium is the junction on where it comes up, and the way you look at that is on an axillary x-ray. So here's a picture of the os here, 56-year-old man with a rotator cuff tear is undergoing a decompression and a repair. They look at this, what's the best way to evaluate it? The best way to evaluate an osacromialy is an axillary view of the shoulder. 22-year-old man, here's another osacromialy right here. What's the likely cause of an osacromialy? So we know it's a failure of ossification, so I'm looking for something of failure of effusion, so we can get, it's not going to be a fracture, they're not going to test you on a chromial fracture as I promised you. So it's either going to be, we've narrowed it down now to two answers, if you don't remember you can guess, but just meta and meso. Pec major and triceps problems. Remember there's two parts of the pec, the clavicular head and the sternal head. They rotate about 90 degrees and they attach lateral to the biceps groove innervated by the medial pectoral nerve and the lateral pectoral nerve. It's powerful internal rotation and adduction and all muscles, not just the pec, the quad, the hamstring, all muscles get injured eccentrically. Anytime you have a muscle injury question, look for the word eccentric, it's going to be the answer. They get in the eccentric phase of contraction. The fact that this is in the shoulder, when do you get an eccentric injury to a pec? Well it's probably going to be the throwing motion, right? And it's going to be late cocking, let's say contracted, it's going to be as you're going into increasing more late cocking. That's when you're going to have, because your pec's trying to slow down some of that external rotation. Pec major ruptures, always going to give you a vignette of bench press. So if you see a question with someone bench pressing, the next thing you want to think about is, is this a pec major question? And there's going to be a loss of the axillary fold. MRIs are very, very hard. And I bet even for Bill Palmer, MRIs for pec tears are very difficult to find. You really need kind of dedicated chest wall views. Here is the tip right here of the pec treatment. They're seldom going to ask you, especially on a sports exam, to opt for a non-operative treatment of a pec. Typically they're going to want surgery because of the marked loss of strength in the pec, about 50% loss of strength without the surgery. Even if it's after six weeks, you can still do a primary pair sometimes with a graft. During a bench press, pec question, and here it is, when is the pec major insertion at greatest risk of rupture? So any muscle, throwing, late cocking, bench press, when do you get eccentric contraction of the pec in a bench press? Well, if bench press is going to get the bar away from you, it's going to be as the bar starts to come down. That's when you're going to see it. So it's going to be during downward deceleration. That's when you get the eccentric contraction of the pec. 23-year-old athlete, injured his shoulder during bench press. What's the best pre-treatment? Bench press, athlete, tourist peck. I don't even have to look at the MRI. I'm going to look at it and pretend that I know that the peck's over here. And then I'm going to fix the peck, because it's a bench press question. 43-year-old firefighter. So this is a bunch of the Boston firefighters. We'll see this a lot, coming with peck ruptures from fire hydrants, acute anterior shoulder pain, echemosis on the anterior aspect of the shoulder, loss of the contour. Here's the fluid. Here's the peck tendon. Fix the peck. Now, we haven't talked about it yet, but if you do get a question about pitchers and the latissimus, the answer is non-op. Just something to don't spend a lot of time studying it. Just if it's not one of the major, major muscles, like peck major, if it's teres major, if it's latissimus dorsi, generally, they want you to treat those non-operatively, because they're a little esoteric. But it doesn't mean they're not going to test you on them. So pitcher tears his lat, and they want you to do non-operative management. Pitcher felt a pop in the back of his shoulder this time. So it's not going to be a subscap tear, and has an asymmetry when he looks at the back of his shoulder. Now, that's the picture of vanity that all baseball players have, is that they want to check out their back in the mirror. But he looks, and on exam, he's got full but painful forward elevation, and external rotation is a bit limited due to pain. Internal rotation, normal. So subscap, again, is OK. Muscle testing, 5 out of 5 strength. So this is not a cuff tear, per se. It's not a supra or an infra tear, because they tell you that elevation and external rotation are OK. There's some pain with belly press and internal rotation. So belly press is subscap, but wait, this is posterior pain. So what else kind of works with subscap, or if I had a chronic subscap, what other muscle can kind of substitute for that function, is the way to start thinking about this. On exam, no asymmetry at the pec, and there's a subtle bulge at the inferior border of the scapula. So what muscle is at the inferior border of the posterior scapula? If you come down, you have supra, infra, teres minor, teres major. So best initial manage, they want you, here's the MRI. It's low. When you start looking at an MRI image like this with the arm, you know this is going to be very low, because it's the teres major. It's a strain injury of the teres major, so we want to do gradual motion, just PT. Triceps, if there's an association of bench press with pec injuries, there's an association of triceps teres with steroids. Inability to extend against resistance, you can feel the defect. Sometimes you can see this flex sign. They'll give you an x-ray sometimes, a little flex sign, this little piece of bone that evulses off of the olecranon. The treatment, if it's a partial rupture, treatment that they're looking for is non-op. But if it's a complete rupture, it's going to be surgical, and a variety of different ways to do the procedure. I promise you, you're going to get a question on Little League shoulder. Obviously, from the AOSSM standpoint, we've got a big emphasis on youth sports, throwing shoulders. They want you to know about Little League shoulder. Any pediatric question, as a rule in orthopedics, they want you to kind of get comparison x-rays. If you just get the x-ray A here, you may not notice that there's fissile widening. But if you get the contralateral shoulder, you can see that there's fissile widening here. So the two buzzwords for pediatric thrower's shoulder, or Little League shoulder, fissile widening, and injury to the hypertrophic zone. I've had that on a couple of my tests. Where is the injury in Little League shoulder? And it's in the hypertrophic zone of the cartilage formation. Treatment for Little League shoulder, don't throw. And it's three months, don't throw. It's not three weeks. It's not three days. It's three months of don't throw. So 12-year-old baseball pitcher comes in with his dad. Mild pain for a month. And here's the key. Playing on two teams is also a big focus now overuse of adolescent sports injuries. Plays on two teams, school team, traveling team. How do you minimize the son's risk of injury? Simple, limit the number of pitches. He's probably throwing six innings a week on both teams, which means 12 innings overall, which is too much. 16-year-old collegiate baseball pitcher, that kid must be smart if he's 16 and in college. But he said, posterior shoulder pain, decreasing velocity and accuracy. So it's going to be pretty rare for a 16-year-old to have a cuff tear. So probably not going to be that. When the abducted shoulder is in the supine position, it's 120 degrees of external rotation with 40 degrees of IR. And he's got less external and more internal on the non-throwing. But again, that's pretty typical. And the total arc of motion is pretty similar, 160 versus 170. Remainder of the exam is unremarkable. And the MRI shows a small, partial, articular-sided infraspinatus tear. Another word for small, partial, articular-sided infraspinatus is internal impingement. As a general rule, don't answer operate on internal impingement. And as a general rule, don't answer operate on a thrower shoulder unless they give you descriptions of the vignette that they've done three years of physical therapy, hasn't gotten better, can't return, what's the next option. Especially in younger people, they don't want you to do this. So as a result, they're going to say, well, he's got internal impingement because he's got an internal rotation contracture as well. So they probably want to do posterior capsular stretching. When they ask you how to treat thrower shoulders, the answer is almost always posterior capsular stretching. Proximal biceps tendon. Just take a minute just to look at the ones that are elevated or that are highlighted here. You're always going to see these together. When you start thinking of proximal biceps, think subscap, this biceps interval, think subscap, long head of biceps, coracohumeral ligament. Because I promise you there's going to be a question that has something to do with this trio here. If they show you something that has a biceps dislocation, they want you to recognize that almost by definition, not almost, by definition, there's a subscap tear as well. If they show you a subscap tear, they want you to think about, could there be a biceps subluxation or dislocation? They always go together. So for the purpose of the test, because it's rare to have biceps pathology in isolation. And this is on multiple, multiple self-assessment exams, OITEs, et cetera. And I've got this listed multiple times in your handout. You can look at it. Physical exam for biceps. Point tenderness is going to be the most sensitive. But speeds and Yergeson's are also involved. The thing to remember about biceps, it's somewhat controversial about whether you operate on them or don't operate on them, depending on the age. So hopefully, they shouldn't be asking, what do you do with a 45 or 50-year-old biceps tendon that is having problems? Most of the questions should be worded as, what's the difference in outcome between tenotomy or tenodesis? And the answer is, there's basically no functional defect, no functional difference between the two. But tenotomy sometimes get pop eye cosmetic deformities. No functional difference, cosmetic difference only. Tenodesis helps prevent that pop eye deformity. Again, cosmetic deformity. So denervation, the other thing you're going to see, I promise you you're going to get a question about cysts, paralabral cysts, and suprascapular nerve. So when you see an MRI like this, a cyst, it's going to be in the spinal glenoid notch. Remember, if you have a cyst in the spinal glenoid notch, it's going to knock out just the infraspinatus on the suprascapular nerve. If it's approximately at the notch, you'll get both the supra and the infra. So I'm looking at this notch. It's going to have infraspinatus. It's going to affect external rotation. I'm going to look over here, and it's external rotation. But wait, there's two external rotations. Adduction and abduction. Well, remember hornblower sign, hornblower's test, that's abduction, external rotation. If there's that lag sign, that's for teres. So if you have a positive hornblower's external rotation, loss of that in abduction, it's teres, minor. If it's in adduction, it's infra. So in this situation, it's going to be loss of external weakness and external rotation in adduction. In addition to arthroscopic debridement of the joint for an irreparable cuff tear, what does a tenotomy most likely produce? Well, I'm looking for a Popeye deformity. So here's Popeye deformity, but if you don't read the whole question, it says plus significant strength deficit. No, there's no problem with strength, 5% to 10% at most. So it really is, what does a tenotomy do? Good pain relief. So you've got to read, don't jump to the quotes in the Popeye deformity, read the whole answer before you answer that. Cuff tear arthropathy. Cuff tear arthropathy oftentimes is going to be one of two things. Either, again, massive cuff tear, or somebody did an aggressive decompression debridement for a massive cuff tear or took down the CAR to the CA ligament. Patients will often present with pseudoparalysis. Pseudoparalysis, as defined by, really can elevate past 60 degrees or up to a shoulder level. You can get this kind of fluid sign, this big swollen bursa in the shoulder. The treatment for cuff tear arthropathy, you can do injections. You can do physical therapy for the anterior deltoid, and maybe the teres minor. But treatments are as follows. Typically, they're going to want to talk about surgical treatment. And as we talked about earlier, this rocking horse deformity. What a rocking horse, remember those old horses that kind of go back and forth? This is the way the humeral head replacement should sit on your glenoid replacement. Without a cuff, when you elevate, it's either going to shift posteriorly and point load the posterior part of your glenoid component, or shift anteriorly and point load the anterior part of your component. Back and forth, that rocking horse technique is going to, with time, loosen the polyethylene with the cement. So don't answer total shoulders for cuff tear arthropathy. I would tell you that historically, the test has looked at hemiarthroplasty as an acceptable answer. Even since I first took the test and first started with this course, I would say over the past five to seven years, reverse is what people are looking for. Reverse has become an acceptable answer, or the acceptable answer, for cuff tear arthropathy. A couple things to know. If they do want you to do a hemiarthroplasty, they might give you a cuff tear, but say it's like a 40-year-old or something like that, that maybe you don't want to do a reverse in that situation. But as a general, and if they say the patient's got great motion and is young, that's the only time to consider hemi. Otherwise, the vignettes will usually be 70 years old, and you're going to go with the reverse. The thing to know about a reverse for the purpose of the test, the buzz word for complication is there are two things, scapular notching and dislocation. Those are the two complications that they want you to think about. If there is scapular notching, really you don't do anything about it, you want to observe it until it becomes a problem. The problem with a notching is it can lead to dislocations. So here we go, here's a reverse, here's a dislocated reverse. So which factor is associated with that? So one of the things that I look for when I first was reviewing this was, okay, is there a answer of scapular notch? No. So now it's just a question of you just have to know what are the risk factors for dislocation in someone who's had a hemi, and we've listed those. I'm sorry, in someone who's had a reverse, and they're listed here. Patients with a primary diagnosis who's had either a previous humeral non-union, had a previous fixed glenohumeral dislocation, or failed prior arthroplasty. Those are the patients that are most likely to dislocate a reverse. The way to remember that is just think back to what happens if you've got a non-union, that humeral head is not going to be rotating much. You're going to get a contracture, you're going to have a lot of soft tissue contractures, and when you do your reverse, you're going to be doing a lot of releases. Same thing with fixed dislocations, you get contracted soft tissues, failed arthroplasty, post-surgical scarring. All of these things affected the soft tissue envelope before you did your reverse. So they're going to be more likely to cause problems after the reverse. 71-year-old with night pain can't raise her head above shoulder level, right? So this is kind of a quasi-pseudo-paralysis. PT's not helping, so they don't want you to do any kind of non-operative treatment. What surgery gives her the best chance to restore function? Well, we talked reverse, maybe hemi. But for the purpose of the test now, reverse has become the answer. 82-year-old, one-month history of shoulder pain, active elevation 150. So decent active elevation, but she's painful. She can't sleep. So if you don't read this quickly, what do you say? Well, it's pretty simple. She needs some kind of, and here's her x-ray, right? So they want you to use some kind of replacement. I'm thinking total shoulder, she's got good motion, maybe her cuff's okay, maybe reverse. But wait, if you don't read the question, what's the one thing we didn't talk about? One month history. They never want you to operate on an 82-year-old with a one-month history of anything. Okay, so non-operative first, you gotta read it closely. Underline that when you're reading through the question, I usually underline those little buzzwords. One month, 82, sleep, and then go down from there. And notice how they're also really nice, they put that as the last choice when you could just jump to the other ones. 76-year-old, already had two injections, already had anti-inflammatories, no relief, all right, so we're not going to treat this non-operatively. Exam shows 60 active and 120 degrees passive active forward elevation. 30 of external rotation. No, no, no, that's not external rotation. Where is this word? External rotation lag. Why are they telling us? All these other questions I've taken for the first 100 questions, they talked about external rotation, not lag. So I'm going to put a big circle around lag when I'm reading this question. And a positive hornblower sign. What's a positive hornblower sign? Remember we just talked about that. Abduction external rotation is Terry's minor. So I'm looking through these answers, non-surgical, no, all right, it's already done all the non-surgical stuff. HEMI with partial cuff repair, no, you're almost never going to answer HEMI on this test. Reverse with lat transfer, maybe, because remember, pseudoparalysis, if she's got pseudoparalytic, the chance of you being able to do a primary cuff repair and have a good functional outcome for the purpose of the test is extraordinarily low. A cuff repair with no acromioplasty preserving the CA ligament. Again, if you're pseudoparalytic, you're not going to be able to do a cuff repair. Limited goals debridement, all right? So in this situation, the key is the external rotation lag. If you get a question about arthroplasty in the setting of a big lag or lack of external rotation, for better or worse, they want you to do a latissimus dorsi transfer. I know it's kind of a niche type of question, but just remember that. So it's going to be a reverse with a lat transfer. So cuff tears and different types of impingement, rotator cuff and footprint anatomy. You got the supra, infra here. Remember that the function of the supra is it initiates abduction and has some external rotation as well. Infra is going to be primarily an external rotator and you test that in adduction. Most commonly torn supra, next most commonly torn infra, then subscap, then teres. Arthroscopically, if they show you an arthroscopic picture, they show you the bare spot and there's a tear above it. The bare spot frames the infraspinatus on an arthroscopic view, just the other thing to keep in mind. I doubt they'll ask you that, but just in case. From a physical exam standpoint, we examine the subscap with a liftoff test, belly press test, or bear hug. Remember what we talked about before, you can't talk about the subscap without talking about it inserts on the lesser, it's right next to the long head of the biceps, and its fibers blend with the coracohumeral ligament. Basic science type two cartilage, and the cuff is a dynamic stabilizer of the glenohumeral joint in the mid-range of motion, whereas the ligaments are in the extreme of motion. From a functional standpoint, you have to have your supra intact to keep your humeral head centered in the glenoid as the deltoid abducts the arm. Without it, you get proximal migration of the head and obligate with more than 30 degrees of abduction. When they ask you about what to do with cuff tears, look to see if it's partial or full thickness. Typically, they want non-operative treatment, and especially if they say less than 50%, but it's going to be non-operative treatment initially for partial tears. If they say someone's got a partial tear, it's been nine months, they're not getting better, it's going to be convert to full tear and repair it. For the purpose of this test, a selective partial rotator cuff tear repair is not going to be the answer, all right? It's going to be converted to a full and fix it. The other thing, whether we agree with this or not, is that they like us to think about, they give you a vignette where it is a completely asymptomatic patient. And they show you an MRI and tell you they've got a cuff tear. If they make a point of telling you they're completely asymptomatic, they want you to say initial non-operative treatment. And I'll show you an example of that in a minute. There's intrinsic and external causes of cuff tears. Basically, the only intrinsic thing to think about is the watershed area, relative avascularity where the supra tears. Externally, compression from the arch, or they'll show you a big type three acromion with an osteophyte. If they give you a story that says, I have a 51-year-old who dislocates their shoulder, they want you to think cuff tear, right? Anytime they give you someone and they make a point to tell you the patient is over 40 and they dislocate their shoulder, the biggest complication is not recurrent instability, it is a cuff tear. And until proven otherwise, that's what you should be assuming. For outlet impingement, this is an outlet view. You can see you've got this kind of hook. In that situation, they're going to want you to talk about necromioplasty. The most sensitive test for impingement, bursitis, partial tears, is the Hawkins test, all right? So the near sign, straight forward elevation, Hawkins, 90 degrees elevation, internal rotation, adduction. All right, that's Hawkins test. A painful arc occurs between 60 and 120. And just for, remember the acromion types, I hate these questions, but they do show up from time to time. Just remember, flat, curved, hooked. They're not going to ask you a question about a curved. They might show you, though, say they have a smooth acromion and want to know what you're going to do. Don't answer, with someone with shoulder pain, don't answer subacromial decompression. Typically, they'll show you something like this with a hook and want you to recognize that a decompression might help there. Best imaging for partial versus full, Dr. Palmer's going to go over how to read cuff tears with you as the next talk. Non-operative treatment, the treatment for cuff tears non-operatively, injections, PT. But again, they have to be completely asymptomatic in the history to answer that. Treatment-wise, we've talked about a little bit already, the indication for doing a lat transfer. So in that case that we showed, where they wanted you to do a reverse and a lat transfer, you have to have a functioning subscap to do a lat transfer. If you have a partial thickness tear, remember, theoretically you can debride it if it's less than 50% torn. But you should repair it if it's greater for the test purposes. If you have a full thickness rotator cuff tear, you want to fix it in the acute, young, painful shoulder, or the older athlete. And you can treat them non-operatively, if they give you an asymptomatic 80-year-old with a cuff tear, and it's chronic, then you can treat that non-operatively. Complications, the biggest complication of rotator cuff repair is a recurrent tear, or it doesn't heal. That's going to come up time and time again, that's all over the literature. It doesn't mean necessarily that you have to answer a revision repair, because even with a non-healed cuff tear, function and pain relief can still be very, very good. The other thing is, whether it's an arthroplasty question or a sport shoulder question, shoulder plus infection equals P-acnes for the purpose of the test. And I've had that as well, so in my mind, what's the most common organism for infection? They want you to say P-acnes. And by the way, for P-acnes, it's got to be a long incubation in the lab, 21 days. Decompression, indications for a decompression, this type three hooked acromion, this guide for the cuff repair, be careful about answering isolated decompression for anything on the exam. As a matter of fact, as we all know, when you're coding, you don't get reimbursed anymore for a 29826, an isolated decompression. This test is not going to want you to ever answer, take a patient to the OR and do a subacromial decompression. All right, so if that's an answer, you can cross it out. Dysclavical excision, almost always it's going to be a vignette. So remember, we got bench press for PEC, for dysclavicular osteolysis, weight lifter. A vignette, I have a weight lifter, atraumatic onset, superior shoulder pain is going to be AC joint dysclavicular osteolysis. Point tenderness, pain with cross body adduction, good response to an injection. Unless they say the patient's got tenderness at the AC joint and they show you an MRI like this, I don't recommend surgery based on imaging alone. Almost everyone's got some kind of AC joint inflammation on their MRI. Make sure they tell you they're tender to palpation at the AC joint if you're going to recommend a dysclavical excision. All right, on the T2MR, you're going to see bone marrow edema. And the edema correlates better with a symptomatic AC joint than arthritis alone. I don't think they'll get too much into this, but as a general rule, if you look at single versus double row repairs, the biomechanically double row suture bridge constructs have less gapping and more strength than straight double row repairs using just anchors, which are stronger from a gapping standpoint than single row. That's all I'm going to say about single and double row, really, and that's based on some of Gartzman's work. The key, though, is just like in tenotomy versus tenodesis, there is no difference in functional outcome between single and double rows, all right? So, remember on the biceps question, it was one had a better functional outcome? That is not the case with single and double row. Post-op, the key thing to remember is just don't do strengthening too early. They say if someone starts strengthening at six weeks, that's too early, all right? Twelve weeks to start strengthening. Question of what prevents cuff tears from healing? So, first and foremost, I always thought it was old age, but in reviewing with this, it's not. It's, I'm sorry, I always thought it was smoking as the key answer, but it's not. It's old age for the purpose of this test is the highest correlation with failure to heal. It's usually the avascularity and the poor tending quality. Big tears, muscle atrophy, retraction, smoking, isoprosis, diabetes, even cholesterol. All of these affect healing, but it's old age that is the worst. Remember that massive tearing does not mean irreparable tears. You need these sagittal obliques that are going to show lack of muscle. Jumping over to calcific tendonitis, the key thing to remember about calcific tendonitis for the test is that, first of all, the bone, the deposits are not contiguous with the humeral head, all right? They're going to be in the muscle itself, sorry, in the tendon itself. For the pearls for this exam, they want you to think about the formation phase versus the resorptive phase, which is basically like an adhesive capsulitis when it's painful versus when it's frozen, all right? It's non-op, non-op, non-op with injection until they tell you they've failed injections. Only then do you take them to surgery. Do not answer shockwave therapy for calcific tendonitis. This is going to be cortisone and PT. Operatively, the indications for surgery, so they failed non-op. You've got these big deposits and continued pain, in which case it's a subacromial decompression with or without a cuff repair, depending on how much tendon gets taken away when you do the debridement. So, six-week follow-up visit after a cuff repair, not gone to PT yet. What's going to be the outcome at one year? This is the same thing as a disc herniation, all right? They want you to say that, although they might be stiffer early and PT has been shown to help early motion, there's no difference at one year between early PT and not early PT. Twenty-four-year-old javelin thrower arrives late, can't warm up. First throw, at the end of his follow-through, he has a sharp pain in the right shoulder and he can't keep throwing. On exam, he's tender over the biceps, all right? So, you're thinking, could it be slap? Could it be biceps? But don't forget, when you think biceps, what's the next thing you got to think? Subscap, okay? So, think of those two things. Positive O'Brien's test, okay, so possible slap. Weakness of internal rotation, ding, ding, ding. Positive lift-off test, same thing. Hornblower sign, so it's not his teres, it's a hat being involved. And here's the MRI, and if you go down, it's going to be primary repair of the cuff because he's evolved his subscap tendon. 30-year-old pitcher, onset of weakness, and so he's lost control and he's weak. His velocity is down. He's got slight tenderness lateral and anterior to the acromion with a full range of motion, but he's got a lot of weakness in his external rotation when it's adducted and abducted to 90 degrees, all right? So, adducted at 90 is what? It's infra. Abducted at 90 is teres. So, which muscles are going to be involved? Here's the, it's going to be infra, okay, because it says lateral and anterior to the acromion is tender with a humerus adducted and abducted to 90. So, but the MRI shows that the teres is intact. And you have the ganglion cyst right here at the spinal glenoid notch. So, even though you don't see a retracted cuff, here's your cyst and you're going to think infra because teres is innervated by the axillary nerve. 55-year-old woman, pain in the right shoulder for five days, which keeps her up at night. So, five days keeping her up at night, this is usually something that you're going to think calcific tendonitis, quick onset, severe symptoms, 55, she's clutching her arm. People with cuff teres don't clutch their arms like that. And she denies any recent injury. She has positive impingement with decreased motion and some weakness. So, capsule release, no, it's only been five days. You're not going to jump to an MRI after five days. You're not going to manipulate her after five days. You're not going to do a breeze blind. It's going to be a steroid and PT because if you look here, here's this big calcific deposit. 52-year-old, prior cuff repair, persistent pain and limitation of motion, big atrophy, all right, lots of atrophy of the infra. Active overhead elevation of 140 degrees with significant weakness of external rotation. So, at this point, you've got no functioning infra or teres. So, what do you have to do when you have no functioning? You've got good elevation. They're not going to want you to do a reverse if there's good elevation. All they want you to do is deal with the external rotators with a lat transfer. So, always look at this. Is it a pseudo-paralytic patient? If that's it, then it was going to be reversed with external rotation transfer. If there's good elevation, leave the elevation part alone and just fix the external rotation. 80-year-old tennis player falls for a first time dislocation. Right away, what do you think of? Well, fracture, but you're going to think cuff tear. Complains of difficulty elevating his arm. So, tennis player wants you to say that person's in good shape, all right, they're athletic, it's cuff tear. And you can see how many times this has been on exams right here. 56-year-old laborer, subcoracoid dislocation as a result of falling off a scaffold and now can't raise his pain. Cuff tear, 56-year-old laborer fell. 80-year-old tennis player fell. Cuff tear, cuff tear. Which risk factor? Again, I jumped, when I was first taking this myself, I got it wrong because I jumped to smoking. You just have to remember, risk factor for failed cuffs, it's going to be old age. That's why I had it start in your handout, circle it in your handout, circle it on this. Old age is the most, and that's been on exams. 48-year-old arthroscopic cuff repair. Infra is mobilized to the footprint, but supra can't be. What should you do? Remember that partial, there is a benefit to partial rotator cuff repairs. So in this case, leave it as a partial cuff repair and don't jump to an arthroplasty. And I'm going to, just for the purpose of time here, I'm going to move along. Now, here's the one thing that we can argue about clinically, but you just have to know for the test. 68-year-old weakness, I'm looking at the MRI, cuff tear. All right, so I want to fix this right away. We're surgeons, we want to fix it. But no pain. Exam shows atrophy in the supra and infra. Cuff tear. Weakness in empty can. Weakness in external rotation. Cuff tear, I want to fix it, I want to fix it. What's the most, so I'm looking to fix it. But the first time I read through it quickly, just like before when it said one month of pain, external rotation lag, circle it, when it says no pain, circle it. Because they want you to treat this patient non-surgical. It doesn't matter if we agree with it or not. It doesn't matter if we fix it in our practice or not. The test, when it says no pain on a cuff tear, they want non-surgical. And they even give you, because they know people are going to argue, this is from the academy's clinical practice guidelines where this is published. Is the opinion of our work group. And just wrapping up now, I'd be surprised if you got a question about subcoracoid impingement. But just in case, you should know what it is. It's when the lesser tuberosity impinges under the coracoid. Normal distances are there. You can do a coracoplasty, but I can't imagine that's going to be ever asked about on the test. Oftentimes, if they might talk about subcoracoid impingement, but it's in the setting with a lesser tuberosity impingement, what's going to tear? Typically, it's going to be the subscap. And we've gone through this a fair amount already with the subscap. Think subscap if they give you a small anterior supraspinatus tear. If they give you something like this with a biceps dislocation, the subscap's going to be torn, all right? If you've had a dislocation of the shoulder on every single test that I've ever taken, there's a question that they want you to think about the fact that a subscap has been torn. A physical exam, increased passive external rotation of the arm, and a positive belly press or lift off sign. An external, or sorry, internal rotation lag sign, so that's either a lift off test, belly press. To look at it, you want to look at the axial images. Here's the end of the subscap here. It's a perched biceps tendon. And the treatment for it is going to be a repair. So I'm going to stop there. The rest of the questions are in your handout. I'm going to ask Bill Palmer to come up now. And go over the radiographic findings for your cuff questions, okay? Yes. The question is, for a subscap's tear and biceps dislocation, the answer is operative, correct. And this will be our last correlation. But Bill, come on up. Yes. The age cutoff for slap or slap? The slap cutoff for slap repairs? They, if they come anywhere, they hate slap repairs. Doesn't matter if we like them. I can't remember seeing a, the only time, put it this way, the only time they're going to want you to answer slap repair is if they show you a picture of a paralabral cyst, okay, in someone with a big slapped hair. And they might say, fix the slap and decompress the cyst. But the only time, don't hold me to it, but the only time that I've seen slap repair on a test be an answer is in association with a paralabral cyst. All right, Bill.

adhesive capsulitis

frozen shoulder

shoulder capsule

loss of motion

diabetes

hypothyroidism

coraco-humeral ligament

external rotation

physical therapy

steroid injections

surgery

rotator cuff tears

impingement

calcific tendonitis

surgical intervention