The Braist Study

[hdugger]

Googling back, I'm finding snippets from the editorial, but not any definite ratios. What I see is this:

"Thus, "the bracing indications described are probably too broad, resulting in what may be unnecessary treatment for many patients," Dr. Carragee and Dr. Lehman said. Although there do not appear to be physiological side effects to bracing in adolescent idiopathic scoliosis, "it carries financial, emotional, and social burdens that need to be considered."

"The challenge in the field going forward is to identify children who are most likely to benefit from bracing and those who are unlikely to benefit," they concluded."

which I think is absolutely true and the other main conclusion of the study. You'd like to see (well, not *like* to see) 100% progression in those offered but refused a brace an 0% progression in those braced. I'm hoping that, now that they're past wondering whether bracing works at all they can focus on who actually benefits and who doesn't. They've wasted a lot of years dithering around with this.

The thing that really surprised me about this study is just how effective braces were in the kids that wore them for a long time. I'd seen something from Dolan earlier talking about a review of the braces they did early in the study and how ill-fitting, etc some of them were. Can it really be true that you can wear even an ill-fitting brace for 12 hours a day and have it keep you from progressing to surgery? That just doesn't make any sense at all. Maybe as part of the study they made sure that the braces actually fit right? Or, maybe, knowing that they were being watched over, everyone really made sure that the braces were fitted right? I don't know, but 90-93% is *really* effective - much more than I expected.

And, how does this tie back to the whole Scoliscore thing? If bracing just does not work at all for a defined group of genetic patients, why is there a dose-response? Presumably, those kids are scattered throughout the different hours-of-wear groups and they'd fail bracing at a steady rate no matter how many hours a day they wore the brace. Or, maybe, they have to wear for 12 hours a day for it to work? Don't know. But this study is throwing a lot of what I thought I understood into question.

[Pooka1]

Hi hdugger,

The 9 for 1 figure I quoted was from the full-test of the editorial when it was first published. Unfortunately it has now been changed to a 100-word free preview and you can not view the full-text editorial unless you want to pay $15.

This is the same over-treatment factor that Newton and his buddy calculated in their study.

A curious comment in the original Braist article that most people seem to miss is that 42% of the children in the brace group who never wore their brace were successful. These non-compliant kids were nontheless counted as bracing successes, since they counted all children in the brace group based on intent to treat, rather than focusing on those who actually wore the brace. So those 42% of the braced kids whose brace was under the bed or in the back of the closet were counted as bracing successes! How is that possible?

I hope they follow the "successes" out over time. I have surmised that braces may only delay surgery and this cohort would be good test of that. They should also follow the W&W group for later surgery. Anyone in either group who is north of 30* has no guarantee of avoiding progression, possibly to surgical range. In that sense, I wonder if there is a single true "success" in the braced group. I'd like to see an exit interview on the percentage of kids who think the bracing has let them dodge surgery for life. I bet it's most of them.

[Pooka1]

I don't know, but 90-93% is *really* effective - much more than I expected.

Would it still be really effective if these kids in this "90-93%" had curves north of 40* and even north of 45* after bracing?

And just how many kids wore the brace that long and how many happened to be lumbar cases? It could be a handful. I'll see if I can reconstruct the numbers from the table data.

[hdugger]

Googling back, I'm finding snippets from the editorial, but not any definite ratios. .

Ah, I see the review paper that it's based on. They say:

"The NNT (number needed to treat) for routine scoliosis bracing is about 9 patients for each surgery prevented. The NNT for patients highly compliant with bracing is about 4. "

But they then go on to say:

"We caution that these NNTs are derived from nonrandomized cohorts [they did a survey review of studies to date], and the true values from quality randomized controlled studies may be substantially different. "

The BRAist is that randomized controlled study. The figures from the Braist study - 3 NNT for 1 (for bracing > 6 hours) or 2 NNT for 1 (for bracing > 12 hours) is a more reliable figure than the 9 to 1.

[hdugger]

Would it still be really effective if these kids had curves north of 40* and even north of 45* after bracing?

And just how many kids wore the brace that long and how many happened to be lumbar cases? It could be a handful. I'll see if I can reconstruct the numbers from the table data.

We could lay out any number of scenarios based on data which we simply do not have access to yet. If every braced child ended up at 49 degrees bracing would seem much less effective. If every unbraced child ended up at 49 degrees while the braced kids all stayed at 30, bracing would seem much more effective. And an infinity of points in-between. There's no point in evaluating that data until it's available.

They set out to study one thing, and that was "Does bracing keep kids from progressing to surgery." And they answered that question. The answer is a definitive yes. Right now, that is the only thing we know, and we have to guide ourselves with that one piece of information until they publish the rest of the data (which they're planning on doing). It's maddening, but that's just how it works.

So, if you want to increase your odds of keeping your kid off of the operating table, you'll brace, and you'll do it for more than 6 hours a day. If you'd rather not brace, you can gamble that your kid is the 1 out of 2 kids who won't progress to surgery. And if you're aiming for any other end point, you'll just keep waiting until the data is available. Welcome to the snail's pace of science.

[Pooka1]

I think they will eventually publish the final curve measurements but it will be averages (not medians) and it will be binned in a certain way. We won't see the individual trajectories which would almost certainly show the same thing that Katz et al. (2010) showed, namely that you can't predict the outcome for any particular child no matter what they do or don't do. Kids agreeing to bracing probably don't get that.

And if the final curve measurements were stunningly in support of their conclusions, these measurements would have been in the first published paper in my opinion. Like Dr. McIntire says, we are asked to just trust them and it isn't clear why they published the beginning curves. Why bother? Just give the final percentage for "success".

[hdugger]

Measuring risk

I'm pretty sure that risk is measured by comparing the reduction in risk to the risk (and not adding to the non-risk).

So, if not bracing has 50% progressing to surgery and bracing has 75% not progressing to surgery, than bracing has reduced the risk by half (from 50% to 25%). I think that's the correct way to state it, and not that it's increased the chance of not progressing by a third.

It would make more sense if you did it with something where the numbers were less similar. So, if your lifetime chance of getting breast cancer is 10% (and, therefore, your chance of *not* getting it is 90%). Then something that reduces your lifetime risk to 5% has cut your risk in half (from 10 to 5). It hasn't just added 5% to your chance of not getting cancer. I mean, it has, but that's not how it's normally stated.

[hdugger]

I think they will eventually . . . but it will be . . . and it will be . . . We won't see . . . which would almost certainly show . . .

And if . . ., these . . . would have been . . . in my opinion.

Hold on, Mildred, we're taking a sharp turn into conjecture alley!

I suppose any outcome is possible. I'm going to hold any further conclusions until I get some information to support them.

[Pooka1]

I'm pretty sure that risk is measured by comparing the reduction in risk to the risk (and not adding to the non-risk).

So, if not bracing has 50% progressing to surgery and bracing has 75% not progressing to surgery, than bracing has reduced the risk by half (from 50% to 25%). I think that's the correct way to state it, and not that it's increased the chance of not progressing by a third.

I assume you are correct about how these types talk about risk (their field, they decide) but both statements are mathematically correct. The second is not "wrong" other than maybe in not using the terms that these guys use. In contrast, we have seen terminology where it is actually wrong to not use the correct terminology in the correct way ("environmental" factors in scoliosis). This isn't like that because it's math.

It would make more sense if you did it with something where the numbers were less similar. So, if your lifetime chance of getting breast cancer is 10% (and, therefore, your chance of *not* getting it is 90%). Then something that reduces your lifetime risk to 5% has cut your risk in half (from 10 to 5). It hasn't just added 5% to your chance of not getting cancer. I mean, it has, but that's not how it's normally stated.

I think you are right about how these things are commonly stated but the math is the math.

Here's how I look at it...

1. there's only about 25% of the braced group where we can say bracing affected the outcome of the study. That's because we know from the control group that half the braced group would not have hit 50* at maturity anyway no matter what they did or didn't do with the brace. And we know about 25% of the braced hit 50* before maturity but based on the dose-response curve, those kids were likely non-compliant and should not be considered "braced" in my opinion. Including them makes bracing look worse than it probably is as someone mentioned above. Also, including connective tissue cases also makes bracing look worse than it is apparently. Who knows what else is a "buzzkill" that should be excluded.

2. this 25% is probably in the sub-group who wore the brace the longest (dose-response curve). I would like to see every point in this smallish(?) group and see if the final measurements are that different from the unbraced group or the 50% of the braced group who were effectively unbraced (didn't wear the brace much if at all).

3. I hope they include the numbers of patients in each bin of final curve measurements. I'd like to know how many kids managed to wear the brace for each time bin. It is a result if only a handful managed it. That's why these guys publish intent to treat stats as far as I know.

I'm going to crunch some numbers on this.

[Pooka1]

Hold on, Mildred, we're taking a sharp turn into conjecture alley!

I suppose any outcome is possible. I'm going to hold any further conclusions until I get some information to support them.

It is nerve-wracking to talk about another field of science that is not my own. I am hedging my words constantly because of this. I KNOW that I don't know.

In contrast, we have some "google scholars" with no sense of knowing that they don't know. These people google a topic for 5 minutes and claim to have found the answer to long-standing scientific questions. These are the people making the dramatic, definitive truth claims.

Because you have relevant training, I don't see you making wild truth claims like these people. The more relevant training someone has, the more they hedge on claims in a field that isn't theirs. That's my observation.

[hdugger]

I assume you are correct about how these types talk about risk (their field, they decide)

It's not a technical term - when people talk casually about risk, like when they report it in the news, it's always reported this way. They always say something like "smoking doubles your chance of lung cancer" and not "smoking decreases your chance of not getting lung cancer by 2%". So, it's the way that risk is commonly (not just technically) understood.

I'm going to crunch some numbers on this.

I'm not sure it's necessary. This is at least the third report I've read that talks about keeping kids from surgery by following a certain protocol in bracing. Shaugnassey (phonetic spelling ) at Mayo clinic I thought had the strongest argument. He used to see a bunch of kids progress to surgery even while braced, and then they changed the bracing protocol to a) make sure there was 50% correction in brace and b) make sure the kids wore it long enough (can't remember the time frame now) and suddenly no one was progressing to surgery. The people at Scottish Rites also saw a big decrease in the kids who followed the protocol.

Not that you can't crunch the numbers, of course, but I really think there's enough data in to make an informed decision. The open question has always been "are bracing experiments just so flawed that they make bracing appear to be effective when they're not." This study answers that question. Bracing really is effective in keeping kids from surgery.

With that nailed down, you can finally leave this question behind and move on to the other ones that have been woefully neglected while they dithered around with that one. What's the best protocol? What's the *minimum* number of hours you can have a kid wear a brace and still have it be 90% effective? Which kids benefit most from bracing? etc.

Once you get the data back on this study, you might be able to parse it out even more. How much did different groups of kids advance *in* brace? Were there differences in protocol/type of curve etc that might help you find a method that not only kept kids off the operating table, but really held their curve at exactly where it was when you put the brace on.

And so on and so on.

But, even without that information, for parents, this has been a really helpful study. Now they know (roughly) how likely their kid is to advance with and without bracing and they can use that information to make a decision. To me, that's huge.

[flerc]

But, even without that information, for parents, this has been a really helpful study. Now they know (roughly) how likely their kid is to advance with and without bracing and they can use that information to make a decision. To me, that's huge.

Hdugger, you believe that because you have a science background, you know about statistical and Maths, but we all must to wait that Pooka1 finish the demosntration she began here:

Quote Originally Posted by flerc View Post
And how explain your not only limited but also ignorant, as it seems to be, science version this difference, mainly when it is similar to other studies as you are saying? This is what I asked you

Do you understand the concept of having a control group? What do you think the ~50% "success" rate in the control group means?

Maybe Maths is not what we believe it is.

[Pooka1]

But, even without that information, for parents, this has been a really helpful study. Now they know (roughly) how likely their kid is to advance with and without bracing and they can use that information to make a decision. To me, that's huge.

This particular study is about reaching 50* at maturity, not avoiding surgery for life. That is why parents who take the percentages at face value and don't have the final curve measurements would be potentially fooling themselves.

Do you count the kids who are >40* and >45* at maturity as likely being "successful" in terms of avoiding surgery for life? Isn't that what patients and parents care about as opposed to 50* at maturity?

If all the successes were 49* would that be "successful"? I don't think they are but the published percentages would be entirely consistent and correct with this outcome.

[hdugger]

In contrast, we have some "google scholars" with no sense of knowing that they don't know.

Is this a reference to Dingo? I've said it a few times before, but it's probably worth saying again. Dingo made the single biggest difference in my son's life of anyone we encountered on our scoliosis journey. His postings on the efficacy of exercise started us on our course of finding good PT/massage/etc, and that *completely* changed how my son looked and how he felt about himself. That's a huge gift for a kid with a stiff obvious curve and no great treatment options.

I appreciate Dingo's effort to help others, and I really marvel at the effort he's put into making his kid's life better. IMO, that makes him a stellar support forum participant. The rest of it - how he interprets studies, what he think about genetics vs. environment - is really no nevermind to me.

[Pooka1]

The rest of it - how he interprets studies, what he think about genetics vs. environment - is really no nevermind to me.

Well I guess I am coming from a point of view where people study for years and dedicate their lives to helping sick kids and are pained to constantly identify what they don't know, not to overstep the data, having an appreciation for the limitations of their studies, etc. etc. In contrast, we are treated on a routine and on-going basis to being told these researchers are wasting their time, going down blind alleys, being stupid, etc. etc. because a lay person googed something in no time flat and didn't understand a damn thing they read.

The situation is so uncanny that it almost can't be parodied... whatever you might make up, the actual claims from the google scholars are even more bizarre.