|
|
Re: Diagnosis and Palpation in Manual Tx
|
Logged in as: Guest
|
|
Users viewing this topic:
none
|
|
Login  | |
|
Re: Diagnosis and Palpation in Manual Tx - October 6, 2005 8:47:00 AM
|
|
|
JLS_PT_OCS
Posts: 1684
Joined: January 30, 2005
From: USA
Status: offline
|
And yet, it is still a useful tool in determining treatment choice as determined by the other statistics like NNT...
J
_____________________________
Jason Silvernail DPT, OCS, CSCS
"It isn't what you're able to do that requires your courage but rather what you have come to understand and are willing to express." - Barrett Dorko,PT
**I no longer post on RehabEdge**
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 6, 2005 9:14:00 AM
|
|
|
steve
Posts: 454
Joined: May 14, 2003
From: victoria, bc Canada
Status: offline
|
Agreed, I think that is what is the really interesting part about it....
Steve
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 6, 2005 3:58:00 PM
|
|
|
Jon Newman
Posts: 1697
Joined: April 24, 2004
From: Amherst, WI
Status: offline
|
Hi Jason,
Thanks for posting that section. I'm thinking that my main concern might have been missed however. My concern may be overblown, but here it is again. Those values are from another study. We don't know if the clinicians in this study were much better or much worse. We are left to "mentally extract the (nonexistent) data" as SJ suggests.
I don't expect every study to prove that the people performing tests and measures are up to snuff when the measure has long standing and widely accepted reliability. But I don't think that can be argued in this case. Add to that that data dredging is prone to type 1 errors and we are left with information that has statistical interest but perhaps not so much clinical interest.
I'm also concerned that it fails the "Does it make clinical sense" question, at least for me. For example, of the 15 people with hypermobility many got better with the "wrong" treatment while many failed to get better with the "right" treatment. The treatment cannot affect the other variables in the prediction rule except perhaps for hip internal rotation (which also fails the "Does it make clinical sense?" question).
What are hypo and hypermobility telling us? The implicit message seems to be that a hypomobile segments needs to be manipulated and a hypermobile segment needs to be stabilized. But does that actually make sense? For example, it seems far fetched to me to think that PA mobes that reveal hypermobility are detecting muscular weakness instead of ligamentous laxity or similar problem. And if indeed it is ligamentous laxity, this will not be solved by muscular splinting unless the muscles are 'on' all the time, right?
This doesn't even take into consideration the fact that we may not even be able to detect the condition reliably in the first place. I've been at the point for a while now that testing vertebral mobility is not unlike the other symmetry testing in that it isn't particularly important, especially as a stand alone palpation skill. I'm not sure but I think determining beta weights would have been a more informative post hoc analysis. I don't understand why these aren't being reported in these CPR's.
jon
_____________________________
[URL=http://www.sonymusic.com/clips/selection/30/064887/064887_03_03_30.wav]Evidence[/URL]
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 7, 2005 3:35:00 AM
|
|
|
JLS_PT_OCS
Posts: 1684
Joined: January 30, 2005
From: USA
Status: offline
|
Jon-
I think you lost me, man.
On the reliability and data analysis point:
This study as well as the one it comes from (the CPR study) has some reliability discussions in it and they do reference a previous reliability study:
Maher, Latimer, Adams. An investigation of the reliability and validity of posteroanterior spinal stiffness judgments made using a reference based protocol. Physical Therapy 1998.
Abstract: Background and Purpose. The reliability and criterion-related validity of protocols for rating posteroanterior (PA) spinal stiffness using reference values for comparison have not been investigated. In this study, mechanical reference stimuli for points on an 11-point rating scale were used to determine whether this approach may be feasible. Subjects. Five different raters took part in 2 studies in which they rated 40 subjects who were asymptomatic for low back pain. Methods. The interrater reliability of ratings was evaluated with intraclass correlation coefficients (ICCs) and standard errors of the measurement (SEMs). Criterion-related validity was evaluated by correlating judgments of PA spinal stiffness assessed manually with measurements of PA spinal stiffness provided by a mechanical device, the "Stiffness Assessment Machine" (SAM). Results. Although the reliability indices were generally high, with ICCs reaching .77 and with SEMs as low as 0.72 points, the evidence for criterion-related validity (ie, the ability of the examiner to judge spinal stiffness levels) was less convincing, with correlations reaching only .56. Conclusion and Discussion. The reference-based protocol allows for more reliable measures of PA stiffness judgments than previous protocols have; however, the human ratings are not highly correlated with the SAM measures. The protocol will have clinical value if judgments made using it are shown to be reliable in clinically relevant subjects and to have validity for clinical management of patients. [Maher CG, Latimer J, Adams R. An investigation of the reliability and validity of posteroanterior spinal stiffness judgments made using a reference-based protocol. Phys Ther. 1998;78:829-837.]
I think that last sentence says it all. Reliability is not the only factor to consider here.
The "Does it make clinical sense" question:
It is true that many people with "wrong" treatment got better and many with "right" treatment did not improve. But looking at the NNT data, this piece of the puzzle was a good signpost in the road. It's not a map with a GPS system attached, but a signpost is a good bit better than the sundial we've seen before. CPR or not, those are good NNT numbers.
In making decisions about what treatment to apply to what patient, this is one of the few good pieces of evidence we have (speaking of this plus the CPR).
On 'what's it telling us":
I don't know what it's telling us for sure. While you may think it far fetched that the PA mob could reveal muscle weakness, i'm not sure that weakness per se is the impairment treated by a stabilization intervention. According to Panjabi's biomechanics treatises and the work of Jull et al in the stabilization area, it was mild clinical instability and muscle activation patterns along with weakness that was proposed to be helped by the stabilization routine.
I don't think they ever argued that the proposed mechanism of stabilization was that of halting the clinical micro instability, just that it helped reduce a pattern of repetitive mechanical deformation and load intolerance. Given more recent literature on the subject it is also reasonable to think what many of us have also, that the training has a CBT and pain modulation effect separate from any theoretical physical effect.
I don't think anyone is asking us to depend on this particular palpatory finding alone in determining treatment, just that it is a useful piece of information.
I am not up to the challenge of your last question, I don't remember enough about those weights from stats in school, and we haven't got to them yet in my DPT curriculum... :)
J
_____________________________
Jason Silvernail DPT, OCS, CSCS
"It isn't what you're able to do that requires your courage but rather what you have come to understand and are willing to express." - Barrett Dorko,PT
**I no longer post on RehabEdge**
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 7, 2005 4:10:00 AM
|
|
|
Jon Newman
Posts: 1697
Joined: April 24, 2004
From: Amherst, WI
Status: offline
|
Hi Jason,
I'm simply asking, does anyone think it is important that no reliability testing was done with these particular clinicians in this particular study considering the shaky reliability demonstrated (referenced in this study) in other studies?
If they did do reliability testing in this study, I keep missing it on my read through, which can certainly happen. It could be one of those "right in front of your nose" things.
I think we could answer the "what does it tell us?" question if we had the beta weights. From Portney and Watkins
"It is useful to present beta weights in addition to regression coefficients in a research report, to provide the reader with a full and practical interpretation of the observed relationships."
Instead of not knowing why or having some numbskull suggest problems with reliability, we could learn how much relative importance the variables are in recovery using a particular technique. If it is a small amount then it is no surprise that many didn't get better using the "right" technique.
What structures do you refer to when you state:
"it helped reduce a pattern of repetitive mechanical deformation and load intolerance"?
jon
_____________________________
[URL=http://www.sonymusic.com/clips/selection/30/064887/064887_03_03_30.wav]Evidence[/URL]
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 7, 2005 8:28:00 AM
|
|
|
JLS_PT_OCS
Posts: 1684
Joined: January 30, 2005
From: USA
Status: offline
|
Hey Jon,
On the structures I was referring to: lots of them. Any structure in the low back area - the disc material, facet joint, nerve root, etc etc. Since we can't accurately establish the "pain generator" I can't specifically address your question.
Since stabilization, like manipulation, is not terribly specific, then I don't think anyone knows the answer to that question. Least of all me. :)
On page 1750, they go into further detail about the low reliability and yet high percentage agreement.
"The paradox of high agreement but low [kappa] values is an anticipated phenomenon when the prevalence of positive ratings by each examiner is substantially greater or less than 0.5. .... It is likely that these prevalence values [hypomobility .71, hypermobility .12] have deflated the [kappa] values reported in this study and in previous studies."
Clearly, I have a lot to learn from these authors about stats, as well as clinical practice, b/c they left me behind at "paradox". :)
They did mention another study in which they looked at reliability issues with the palpation measure, in addition to the CPR study and the one I mentioned before:
Fritz, Piva, Childs Accuracy of the clinical examination to predict radiographic instability of the lumbar spine. Eur Spine Journal (in press)
Also page 1750, right hand column:
"Although counterintuitive to the traditional admonition that reliability is a precursor to validity, other studies have reported diagnositic tests with low reliability coefficients but clinically useful accuracy for predicting a certain disorder or outcome."
Cool.
Hope that helps, thanks for the good discussion.
J
_____________________________
Jason Silvernail DPT, OCS, CSCS
"It isn't what you're able to do that requires your courage but rather what you have come to understand and are willing to express." - Barrett Dorko,PT
**I no longer post on RehabEdge**
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 7, 2005 10:12:00 AM
|
|
|
Jon Newman
Posts: 1697
Joined: April 24, 2004
From: Amherst, WI
Status: offline
|
I'm feeling like a dunce here. So be it, but what were the kappa values generated by this study?
_____________________________
[URL=http://www.sonymusic.com/clips/selection/30/064887/064887_03_03_30.wav]Evidence[/URL]
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 7, 2005 11:36:00 AM
|
|
|
steve
Posts: 454
Joined: May 14, 2003
From: victoria, bc Canada
Status: offline
|
Hi Jon,
If I understand things correctly, the CPR for manip published in the Annals of Internal Medicine did not look at reliability of testing. This was looked at in the 2002 study where all patients were manipulated and those variables that correlated with a successful outcome where used for the validation study.
Steve
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 7, 2005 12:18:00 PM
|
|
|
Jon Newman
Posts: 1697
Joined: April 24, 2004
From: Amherst, WI
Status: offline
|
Hi Steve,
That was my understanding also. I'm not sure if anyone sees a problem when statements about predictive validity are being made in a study in which no reliability testing was done (of the clinicians in that particular study). It may not matter, that's why I'm asking. As I stated numerous times, my concern might be overblown. I'm learning, with help from others.
Another question:
What do you suppose the prevalence of hypo and hypermobility is in an asymptomatic population?
jon
_____________________________
[URL=http://www.sonymusic.com/clips/selection/30/064887/064887_03_03_30.wav]Evidence[/URL]
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 7, 2005 12:29:00 PM
|
|
|
steve
Posts: 454
Joined: May 14, 2003
From: victoria, bc Canada
Status: offline
|
Jon,
I have no problem with what the authors have done with the study data, think of the first study as determining what factors should be looked at and what kind of reliability they have, the second study validates that these factors are able to identify our target group. No need to revisit the inter rater reliability as it was dealt with in the prior study. I almost would think of the two studies as being attached and that comments on predictive validity would then be acceptable.
Excellent second question, my guess is that the numbers would not be much different from a symptomatic patient. Don't know if it has been researched.
Steve
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 7, 2005 12:38:00 PM
|
|
|
Bournephysio
Posts: 576
Joined: April 25, 2002
From: Calgary
Status: offline
|
Jon, reliability is necessary for validity. If a study finds that a test has validity (as in the Childs rct and previous study) then it necessarily has reliability. If you then do a study on reliability of the test and find that it is not reliable then either the validity study or the reliability study is wrong.
I have spent a fair bit of time looking at reliability of continuous variables for my masters. The biggest problem with reliability studies is that they usually rely on an arbitrarily set value of what is reliable (i.e. ICC > 0.7 is good reliability). In actuality, acceptable reliability is going to depend on a number of different factors including what other tests are being used, the consequences of the test being wrong, how the test affects the treatment decision, etc. It seems to me that physiotherapy research has concentrated on reliability and hasn't paid enough attention to validity. You can have a completely reliable test that has no validity.
Doug
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 7, 2005 5:05:00 PM
|
|
|
Jon Newman
Posts: 1697
Joined: April 24, 2004
From: Amherst, WI
Status: offline
|
Hi Steve,
I don't think reliability testing for determining hyper or hypo mobility was performed in the Flynn study either. I think they only checked reliability on SI tests, none of which made the cut anyway.
Still, if we are assuming reliability from one set of clinicians to another, it shouldn't matter where we are getting the data from I guess.
I've been thinking that perhaps it is not the judgement of the therapist that is unreliable as much as it is the patient's response to the testing. That is, the patient's behavior may be one of "tensing" in a manner that produces "hypomobility" while others are rags dolls and seem to be hypermobile. Since the behavioral response of the testee could vary from moment, this could explain the lower reliability. This might even explain the difference in success somehow. Perhaps those who guard benefit from some treatment component of the manipulation group such as the manipulation itself, the relaxation needed to be manipulated (whatever amount that might be) or the AROM associated with that treatment regime.
The rag doll folks may be more ready and able to activate those postural muscles. Obviously I'm speculating here but I'm trying to make more sense of something that is statistically true but doesn't make sense, to me anyway.
Another question: For those who perform diagnostic PA mobes, do you find that it typically changes to "normal" after the appropriate treatment?
jon
_____________________________
[URL=http://www.sonymusic.com/clips/selection/30/064887/064887_03_03_30.wav]Evidence[/URL]
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 7, 2005 8:34:00 PM
|
|
|
steve
Posts: 454
Joined: May 14, 2003
From: victoria, bc Canada
Status: offline
|
Jon,
I think that your theory that the patients response to PAs in the form of "Tensing" likely has some merit in it. Although I am certainly not one to subscribe to the biomechanical model, I certainly thin that passive mobility testing has its place if it is used in conjunction with a number of other assessment tools (Ie. Fear avoidance measures, neuro testing etc.). I have found that it does return to a decreased tone feel, particularly with manipulation (Although it can be accomplished with any type of manual/exercise/education based treatment). Again, I believe that this is a neurological effect and the evidence suggests this but does not prove it. Walter Herzog has published interesting research in this area.
I would like to hear further explanation from Doug on how the kappa levels can paradoxically be deflated in situations of high agreement. I thought it was the preferred statistic with continuous variables and took into account chance agreement which reliability measuring is unable to account for.
Steve
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 18, 2005 3:18:00 PM
|
|
|
Bournephysio
Posts: 576
Joined: April 25, 2002
From: Calgary
Status: offline
|
Steve, that would be a great question to ask Paul Stratford at the course He, Kris Head and I are teaching in November.
I was hoping got get some time to read up on kappa values but I haven’t had a chance. Basically there are two types of reliability statistics (a simplification), relative and absolute. Current recommendations are to report one of each when measuring reliability since each has its flaws. Relative reliability can be considered how well a straight line fits the data (say test vs retest). ICC (which I have a lot more experience with than kappa values)is a relative statistic. I believe that kappa values are as well. Percentage agreement would be an example of an absolute statistic. The following graph shows a problem with only using a relative statistic. The values are all clustered very close together and look very reliable but it is very hard to fit a line to it. This appears to be what happened with this study. An absolute value would give a much better result. The graph below it shows what would happen if you increase the range drastically. You would get a much better line and much better relative reliability. If these are outliers you would have an artificial increase in reliability. If the values are outside the range that you would measure clinically the reliability would be different than in clinical practice. This is one of the reasons I don’t like reliability studies that are just looking at normal subjects. You are artificially decreasing the range and thus the relative reliability. I would imagine that when you only have three different values (hypermobile, normal, hypomobile) the consequences of having a limited range would be much greater since any difference between examiners would be at least 50% of the entire range. I’m sure this wasn’t clear at all.
ignore the dots the "x"s are the points
|
|……………………………….x……..x
|………………………………….x x
|…………………………….x……x……x
|
|
|
|
|
|
|
_____________________________________________
|…………………………………………x
|……………………………….x
|……………………………………x x
|……………………………….x …….x…….x
|
|
|
|………..x
|
|
|………x
_____________________________________________
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 18, 2005 4:15:00 PM
|
|
|
Jon Newman
Posts: 1697
Joined: April 24, 2004
From: Amherst, WI
Status: offline
|
Thanks Doug,
That gives me plenty to think about. A point of clarification: When you state,
[QUOTE] This appears to be what happened with this study. [/QUOTE]Which study are you speaking of?
jon
_____________________________
[URL=http://www.sonymusic.com/clips/selection/30/064887/064887_03_03_30.wav]Evidence[/URL]
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 18, 2005 7:24:00 PM
|
|
|
Bournephysio
Posts: 576
Joined: April 25, 2002
From: Calgary
Status: offline
|
Whichever one reported the low kappa values but excellent percent agreement.
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 19, 2005 9:55:00 AM
|
|
|
steve
Posts: 454
Joined: May 14, 2003
From: victoria, bc Canada
Status: offline
|
Doug,
Thanks, just when you start to get a handle on statistics and evidence based practice.....
Where is your course being held at? Paul Stratford is great, he realizes that statistics are boring for most people and is very good at poking fun at himself as he lectures. He has done some outstanding research.
Steve
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 19, 2005 10:09:00 AM
|
|
|
Bournephysio
Posts: 576
Joined: April 25, 2002
From: Calgary
Status: offline
|
The course is in Calgary. Its mostly on shoulder stuff "clinical application of research principles surrounding evidence-based diagnosis of the shoulder." Paul's section of the course is more general. I just found out last night that the course is full but there is discussion of holding it elsewhere in the future.
Doug
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 19, 2005 10:59:00 AM
|
|
|
JLS_PT_OCS
Posts: 1684
Joined: January 30, 2005
From: USA
Status: offline
|
I want to mention a good website that I am cruising through as part of my DPT curriculum:
http://www.sportsci.org/resource/stats/
"A New View of Statistics"
Nice explanations of some complex stats stuff.
Jon-
I'm not sure what to tell you about what happens to the PA mobs after treatment. To be honest, I don't use ythem in making decisions about whether someone has improved, I think that's a different question. I think it's reasonable to think that there are a wide variety of responses in asymptomatic subjects, I would think this is a lot like the hip rotation part of the manipulation CPR.
I think in the discussion section of the paper, they did make a good case that even unreliable findings can be useful in the clinical examination, when high percentage agreement is found and the distribution of findings is nonstandard (in this case, hypermobility was found far more often than hypermobility). We should all keep in mind as we wrestle with stats issues that we are making assumptions about normal distributions for these variables which does not seem like a reasonable thing to do, especially considering the nonstandard distribution.
I think Rob Wainner made a good case about reliability in his editorial "Reliability of the clinical exam: how close is close enough?" in JOSPT a few years back. It's worth looking up and reading. He talked about just this sort of issue -- something does NOT need a high level of reliability to be useful in making clinical decisions. Especially if it's just one piece of the puzzle. I thought the example the authors gave in this most recent study really helped see the picture better.
J
_____________________________
Jason Silvernail DPT, OCS, CSCS
"It isn't what you're able to do that requires your courage but rather what you have come to understand and are willing to express." - Barrett Dorko,PT
**I no longer post on RehabEdge**
|
|
|
|
|
Re: Diagnosis and Palpation in Manual Tx - October 19, 2005 11:27:00 AM
|
|
|
Jon Newman
Posts: 1697
Joined: April 24, 2004
From: Amherst, WI
Status: offline
|
Hi Jason,
Thanks for the new link.
I've read that editorial and I agree with the concept. My struggle is to make sense of things. There seems to be a message developing that we should "stabilize" the "hypermobile" and "manipulate" the "hypomobile". If these are, in fact, important to the biological rationale of the use of the technique, I'm hoping that a reliable method for detecting mobility comes along. Also, I would expect a change in the presence of the finding if it is significantly contributing to the problem, reliable or not. As it stands, there is no explicit biological rationale for the use of the technique being posited that I've found.
Perhaps someone will give me a reference.
jon
_____________________________
[URL=http://www.sonymusic.com/clips/selection/30/064887/064887_03_03_30.wav]Evidence[/URL]
|
|
|
|
 New Messages |
 No New Messages |
 Hot Topic w/ New Messages |
 Hot Topic w/o New Messages |
 Locked w/ New Messages |
 Locked w/o New Messages |
|
 Post New Thread
Reply to Message
Post New Poll
Submit Vote
Delete My Own Post
Delete My Own Thread
Rate Posts |
|
0.109
|
Printable Version
