This amazing photograph shows the all time record lows recently reported by our field reporter Black Mage. <img src="http://www.blizzard.com/diablo2/wallpapers/hellfrozenover/HellFrozenOver-notext-640x.jpg" border="0" alt="IPB Image" /> As you can see, the Lord of Terror himself, Diablo, has been forced to resort to setting fire to the corpses of all the heroes who have tried over the 8 (OMG) years of his reign. More on this news as it happens. Now back to our scheduled program, Nexium Linked To Hip Fractures: a Natural Selection Discussion, with your host, Depot... <img src="style_emoticons/<#EMO_DIR#>/biggrin-fix.gif" style="vertical-align:middle" emoid=":D" border="0" alt="biggrin-fix.gif" />
<!--quoteo(post=1594782:date=Jan 4 2007, 01:01 PM:name=Depot)--><div class='quotetop'>QUOTE(Depot @ Jan 4 2007, 01:01 PM) [snapback]1594782[/snapback]</div><div class='quotemain'><!--quotec--> Since medical records fail to contain dietary or vitamin D intake, and would also not include the amount of exposure to sunshine, I'll agree with Black Mage. Totally. <!--QuoteEnd--></div><!--QuoteEEnd--> And as pointed out by Patch, the size of the sample, which is bloody huge btw, removes those factors. The study will include people who never see sunshine, it will include people who get more then they should. It will include people who get enough sunshine, but not enough calcium etc etc, you get the idea. This means results will average out in both groups, meaning that the drug <i>should</i> be the only factor.
<i>should be</i> | <i>is</i> the two are not the same.
also, next time your math professor tries to cover statistics, try not to fall asleep. it'll do you some good. a random sample of 100k is "huge" for exit polls, not for linking a drug to a health risk. there's a reason we have this thing called a "double-blind test"
<!--quoteo(post=1594798:date=Jan 4 2007, 09:58 AM:name=Black_Mage)--><div class='quotetop'>QUOTE(Black_Mage @ Jan 4 2007, 09:58 AM) [snapback]1594798[/snapback]</div><div class='quotemain'><!--quotec--> <i>should be</i> | <i>is</i> the two are not the same.
also, next time your math professor tries to cover statistics, try not to fall asleep. it'll do you some good. a random sample of 100k is "huge" for exit polls, not for linking a drug to a health risk. there's a reason we have this thing called a "double-blind test" <!--QuoteEnd--></div><!--QuoteEEnd-->
<!--quoteo(post=1594798:date=Jan 4 2007, 02:58 PM:name=Black_Mage)--><div class='quotetop'>QUOTE(Black_Mage @ Jan 4 2007, 02:58 PM) [snapback]1594798[/snapback]</div><div class='quotemain'><!--quotec--> <i>should be</i> | <i>is</i> the two are not the same.
also, next time your math professor tries to cover statistics, try not to fall asleep. it'll do you some good. a random sample of 100k is "huge" for exit polls, not for linking a drug to a health risk. there's a reason we have this thing called a "double-blind test" <!--QuoteEnd--></div><!--QuoteEEnd--> I know they're not the same nitwit. There's always possiblity they screwed up, which is why I used "should be". 2nd, I've studied clinical trials. You only need a few thousand. 3rd, this isn't even a clinical trial. No drug has actually been given out by the researchers. All they're doing is looking at medical records of people who have already taken the drug. What does a double-blind test have to do with this research?
<!--quoteo(post=1594775:date=Jan 4 2007, 06:36 AM:name=Black_Mage)--><div class='quotetop'>QUOTE(Black_Mage @ Jan 4 2007, 06:36 AM) [snapback]1594775[/snapback]</div><div class='quotemain'><!--quotec--> but they're not reporting on anything else. the study is *at most* an interesting statistic that warrants investigation. <!--QuoteEnd--></div><!--QuoteEEnd--> oh hey, look at what that black mage guy said
<!--quoteo(post=1594816:date=Jan 4 2007, 11:12 AM:name=Black_Mage)--><div class='quotetop'>QUOTE(Black_Mage @ Jan 4 2007, 11:12 AM) [snapback]1594816[/snapback]</div><div class='quotemain'><!--quotec--> oh hey, look at what that black mage guy said <!--QuoteEnd--></div><!--QuoteEEnd-->
What are you basing this all off? Who has access to the study itself through JAMA?
Please don't tell me you're mounting these counter-arguments based on the NEWS article, written by some random journalist who probably specialises in dentistry.
<!--quoteo(post=1594879:date=Jan 4 2007, 04:43 PM:name=emperor_awesome)--><div class='quotetop'>QUOTE(emperor_awesome @ Jan 4 2007, 04:43 PM) [snapback]1594879[/snapback]</div><div class='quotemain'><!--quotec--> What are you basing this all off? Who has access to the study itself through JAMA?
Please don't tell me you're mounting these counter-arguments based on the NEWS article, written by some random journalist who probably specialises in dentistry. <!--QuoteEnd--></div><!--QuoteEEnd-->
Right now all I can get to is the abstract
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec--> Context: Proton pump inhibitors (PPIs) may interfere with calcium absorption through induction of hypochlorhydria but they also may reduce bone resorption through inhibition of osteoclastic vacuolar proton pumps.
Objective: To determine the association between PPI therapy and risk of hip fracture.
Design, Setting, and Patients: A nested case-control study was conducted using the General Practice Research Database (1987-2003), which contains information on patients in the United Kingdom. The study cohort consisted of users of PPI therapy and nonusers of acid suppression drugs who were older than 50 years. Cases included all patients with an incident hip fracture. Controls were selected using incidence density sampling, matched for sex, index date, year of birth, and both calendar period and duration of up-to-standard follow-up before the index date. For comparison purposes, a similar nested case-control analysis for histamine 2 receptor antagonists was performed.
Main Outcome Measure The risk of hip fractures associated with PPI use.
Results: There were 13 556 hip fracture cases and 135 386 controls. The adjusted odds ratio (AOR) for hip fracture associated with more than 1 year of PPI therapy was 1.44 (95% confidence interval [CI], 1.30-1.59). The risk of hip fracture was significantly increased among patients prescribed long-term high-dose PPIs (AOR, 2.65; 95% CI, 1.80-3.90; P<.001). The strength of the association increased with increasing duration of PPI therapy (AOR for 1 year, 1.22 [95% CI, 1.15-1.30]; 2 years, 1.41 [95% CI, 1.28-1.56]; 3 years, 1.54 [95% CI, 1.37-1.73]; and 4 years, 1.59 [95% CI, 1.39-1.80]; P<.001 for all comparisons).
Conclusion: Long-term PPI therapy, particularly at high doses, is associated with an increased risk of hip fracture. <!--QuoteEnd--></div><!--QuoteEEnd-->
And since how certain diets affects your health is still pretty indefinite/contested, it's probably a good thing that they increased the scope of the case studies to balance out the limitations on individual health data.
<!--quoteo(post=1594899:date=Jan 4 2007, 06:17 PM:name=emperor_awesome)--><div class='quotetop'>QUOTE(emperor_awesome @ Jan 4 2007, 06:17 PM) [snapback]1594899[/snapback]</div><div class='quotemain'><!--quotec--> And since how certain diets affects your health is still pretty indefinite/contested, it's probably a good thing that they increased the scope of the case studies to balance out the limitations on individual health data. <!--QuoteEnd--></div><!--QuoteEEnd--> Nothing is contested regarding the effects of vitamin D deficiency or how the lack of sunshine can directly affect calcium absorption.
Unless you've lived in a cave eating nothing but carefully selected Vit. D-deficient foods for 8 years, I doubt that it's much of an issue. Even in England, sunlight gets through the cloud cover.
<!--quoteo(post=1594806:date=Jan 4 2007, 03:40 PM:name=briktal)--><div class='quotetop'>QUOTE(briktal @ Jan 4 2007, 03:40 PM) [snapback]1594806[/snapback]</div><div class='quotemain'><!--quotec--> You really only need a few thousand people. <!--QuoteEnd--></div><!--QuoteEEnd--> For what? To get a statistically accurate medical study, you need tens of thousands at the very minimum, and even then, the uncontrollable variables will corroupt the data. A medical study based purely on doctor's records, of only, say, 5000 people? <b>Hideously</b> inaccurate, because of the vast uncontrollables. If you want a study of only 5000 people, at 77, then you have to control everything else by outside means, with would actually corrupt the data more...
No, to conduct a study of <i>anything</i> scientifically, you need as large a sampe as possible, to minimize corruption of the results due to uncontrollable variables...
<!--quoteo(post=1594937:date=Jan 5 2007, 02:19 AM:name=Lt_Patch)--><div class='quotetop'>QUOTE(Lt_Patch @ Jan 5 2007, 02:19 AM) [snapback]1594937[/snapback]</div><div class='quotemain'><!--quotec--> For what? To get a statistically accurate medical study, you need tens of thousands at the very minimum.
No, to conduct a study of <i>anything</i> scientifically, you need as large a sampe as possible, to minimize corruption of the results due to uncontrollable variables... <!--QuoteEnd--></div><!--QuoteEEnd--> Off-topic I know, but a clinical trial of a new drug does only need a few thousand as pointed out here <a href="http://en.wikipedia.org/wiki/Clinical_trial" target="_blank">http://en.wikipedia.org/wiki/Clinical_trial</a>, which also matches what I've been taught at college. 2nd, large as sample as possible is rarely the best idea, simply because past a certain number of samples, accuracy ceases to rise very far, and then becomes simple pointless logistics, and a waste of time.
Gallup conducts its polls with 1000 respondents. According to statistical theory, polling 2000 people would narrow the margin of error from 3% to 2%, but according to the real world, that will double the cost of conducting the poll. (<a href="http://www.galluppoll.com/help/faq.aspx?ID=249" target="_blank">essay</a>). It's very surprising how accurate such a small sample can be if it's randomized properly.
In medical surveys, the large number of biological random variables might blur the results more than a political survey, but these researchers had an enormous sample size to work from and they were hopefully looking through those records carefully.
<!--quoteo(post=1594937:date=Jan 4 2007, 06:19 PM:name=Lt_Patch)--><div class='quotetop'>QUOTE(Lt_Patch @ Jan 4 2007, 06:19 PM) [snapback]1594937[/snapback]</div><div class='quotemain'><!--quotec--> For what? To get a statistically accurate medical study, you need tens of thousands at the very minimum, and even then, the uncontrollable variables will corroupt the data. A medical study based purely on doctor's records, of only, say, 5000 people? <b>Hideously</b> inaccurate, because of the vast uncontrollables. If you want a study of only 5000 people, at 77, then you have to control everything else by outside means, with would actually corrupt the data more...
No, to conduct a study of <i>anything</i> scientifically, you need as large a sampe as possible, to minimize corruption of the results due to uncontrollable variables... <!--QuoteEnd--></div><!--QuoteEEnd-->
I am going to find your Statistics teacher and bonk them hard on the nose. Hard.
beyond the empirical evidence, it just makes sense, because you need ACID to digest minerals, e.g. calcium. this is basic biochemistry. so if you take antacids, you don't have enough acid in your stomach to digest enough calcium to prevent bone loss. it doesn't take a nuclear engineer.
<!--quoteo(post=1594956:date=Jan 5 2007, 03:25 AM:name=TychoCelchuuu)--><div class='quotetop'>QUOTE(TychoCelchuuu @ Jan 5 2007, 03:25 AM) [snapback]1594956[/snapback]</div><div class='quotemain'><!--quotec--> I am going to find your Statistics teacher and bonk them hard on the nose. Hard. <!--QuoteEnd--></div><!--QuoteEEnd--> It's not about statistics. The results that a staticitian would be happy with, and a scientist would be happy with, are two completely different things. Especially when it comes to medical results. Also, your love of statistics doesn't taken into account the political mess this study could cause. 3% margin of error isn't good enough when you're talking about a drug that a large proportion of the elderly community would appreciate, seeing as they seem to be more at risk from heartburn.
No, a statistically "accurate" result is not what you'd want here, what if that 3% were enough to flip the evidence the other way? All of a sudden, Nexium is safe, and has no link, or a dubious link to hip fractures and osteoporosis, and people keep taking it. Then all of a sudden, another peer group comes along, and goes "hang on, they did a medical records study of only 5000 people, and cleared Nexium because of it"? You'd be thrown out faster than a speeding bullet from a minigun...
You can't just apply a statistical approach to anything, you need to take into consideration any possible effects that your study would have. Clearing, or linking, Nexium to any possible secondary effects would have a distinct lashback, if those results weren't of a large enough group to effectively eliminate the margin of error.
you know the members of your off-topic forum are getting old when a thread about hip fractures is a sweeping success <img src="style_emoticons/<#EMO_DIR#>/nerd-fix.gif" style="vertical-align:middle" emoid="::nerdy::" border="0" alt="nerd-fix.gif" />
<!--quoteo(post=1595031:date=Jan 5 2007, 03:06 AM:name=Lt_Patch)--><div class='quotetop'>QUOTE(Lt_Patch @ Jan 5 2007, 03:06 AM) [snapback]1595031[/snapback]</div><div class='quotemain'><!--quotec--> It's not about statistics. The results that a staticitian would be happy with, and a scientist would be happy with, are two completely different things. Especially when it comes to medical results. Also, your love of statistics doesn't taken into account the political mess this study could cause. 3% margin of error isn't good enough when you're talking about a drug that a large proportion of the elderly community would appreciate, seeing as they seem to be more at risk from heartburn.
No, a statistically "accurate" result is not what you'd want here, what if that 3% were enough to flip the evidence the other way? All of a sudden, Nexium is safe, and has no link, or a dubious link to hip fractures and osteoporosis, and people keep taking it. Then all of a sudden, another peer group comes along, and goes "hang on, they did a medical records study of only 5000 people, and cleared Nexium because of it"? You'd be thrown out faster than a speeding bullet from a minigun...
You can't just apply a statistical approach to anything, you need to take into consideration any possible effects that your study would have. Clearing, or linking, Nexium to any possible secondary effects would have a distinct lashback, if those results weren't of a large enough group to effectively eliminate the margin of error. <!--QuoteEnd--></div><!--QuoteEEnd-->
This would make a lot more sense if the approval process for drugs did not involve numerous studies for which a margin of error at or around 3% is perfectly acceptable.
<!--quoteo(post=1595031:date=Jan 5 2007, 06:06 AM:name=Lt_Patch)--><div class='quotetop'>QUOTE(Lt_Patch @ Jan 5 2007, 06:06 AM) [snapback]1595031[/snapback]</div><div class='quotemain'><!--quotec--> It's not about statistics. The results that a staticitian would be happy with, and a scientist would be happy with, are two completely different things. Especially when it comes to medical results. Also, your love of statistics doesn't taken into account the political mess this study could cause. 3% margin of error isn't good enough when you're talking about a drug that a large proportion of the elderly community would appreciate, seeing as they seem to be more at risk from heartburn.
No, a statistically "accurate" result is not what you'd want here, what if that 3% were enough to flip the evidence the other way? All of a sudden, Nexium is safe, and has no link, or a dubious link to hip fractures and osteoporosis, and people keep taking it. Then all of a sudden, another peer group comes along, and goes "hang on, they did a medical records study of only 5000 people, and cleared Nexium because of it"? You'd be thrown out faster than a speeding bullet from a minigun...
You can't just apply a statistical approach to anything, you need to take into consideration any possible effects that your study would have. Clearing, or linking, Nexium to any possible secondary effects would have a distinct lashback, if those results weren't of a large enough group to effectively eliminate the margin of error. <!--QuoteEnd--></div><!--QuoteEEnd-->
Protip: if the 3% margin of error was enough to possibly flip the evidence, then there wouldn't be a statistically significant difference and the conclusions of the study would be different.
Also the study said people taking the medicine had a 44% higher incidence of hip fractures, so that's kinda safe even +/- 3%.
Comments
This amazing photograph shows the all time record lows recently reported by our field reporter Black Mage.
<img src="http://www.blizzard.com/diablo2/wallpapers/hellfrozenover/HellFrozenOver-notext-640x.jpg" border="0" alt="IPB Image" />
As you can see, the Lord of Terror himself, Diablo, has been forced to resort to setting fire to the corpses of all the heroes who have tried over the 8 (OMG) years of his reign. More on this news as it happens. Now back to our scheduled program, Nexium Linked To Hip Fractures: a Natural Selection Discussion, with your host, Depot...
<img src="style_emoticons/<#EMO_DIR#>/biggrin-fix.gif" style="vertical-align:middle" emoid=":D" border="0" alt="biggrin-fix.gif" />
Since medical records fail to contain dietary or vitamin D intake, and would also not include the amount of exposure to sunshine, I'll agree with Black Mage. Totally.
<!--QuoteEnd--></div><!--QuoteEEnd-->
And as pointed out by Patch, the size of the sample, which is bloody huge btw, removes those factors. The study will include people who never see sunshine, it will include people who get more then they should. It will include people who get enough sunshine, but not enough calcium etc etc, you get the idea. This means results will average out in both groups, meaning that the drug <i>should</i> be the only factor.
the two are not the same.
also, next time your math professor tries to cover statistics, try not to fall asleep. it'll do you some good.
a random sample of 100k is "huge" for exit polls, not for linking a drug to a health risk. there's a reason we have this thing called a "double-blind test"
<i>should be</i> | <i>is</i>
the two are not the same.
also, next time your math professor tries to cover statistics, try not to fall asleep. it'll do you some good.
a random sample of 100k is "huge" for exit polls, not for linking a drug to a health risk. there's a reason we have this thing called a "double-blind test"
<!--QuoteEnd--></div><!--QuoteEEnd-->
You really only need a few thousand people.
<i>should be</i> | <i>is</i>
the two are not the same.
also, next time your math professor tries to cover statistics, try not to fall asleep. it'll do you some good.
a random sample of 100k is "huge" for exit polls, not for linking a drug to a health risk. there's a reason we have this thing called a "double-blind test"
<!--QuoteEnd--></div><!--QuoteEEnd-->
I know they're not the same nitwit. There's always possiblity they screwed up, which is why I used "should be". 2nd, I've studied clinical trials. You only need a few thousand. 3rd, this isn't even a clinical trial. No drug has actually been given out by the researchers. All they're doing is looking at medical records of people who have already taken the drug. What does a double-blind test have to do with this research?
but they're not reporting on anything else. the study is *at most* an interesting statistic that warrants investigation.
<!--QuoteEnd--></div><!--QuoteEEnd-->
oh hey, look at what that black mage guy said
oh hey, look at what that black mage guy said
<!--QuoteEnd--></div><!--QuoteEEnd-->
Because "anything else" should be the same.
Please don't tell me you're mounting these counter-arguments based on the NEWS article, written by some random journalist who probably specialises in dentistry.
What are you basing this all off? Who has access to the study itself through JAMA?
Please don't tell me you're mounting these counter-arguments based on the NEWS article, written by some random journalist who probably specialises in dentistry.
<!--QuoteEnd--></div><!--QuoteEEnd-->
Right now all I can get to is the abstract
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->
Context: Proton pump inhibitors (PPIs) may interfere with calcium absorption through induction of hypochlorhydria but they also may reduce bone resorption through inhibition of osteoclastic vacuolar proton pumps.
Objective: To determine the association between PPI therapy and risk of hip fracture.
Design, Setting, and Patients: A nested case-control study was conducted using the General Practice Research Database (1987-2003), which contains information on patients in the United Kingdom. The study cohort consisted of users of PPI therapy and nonusers of acid suppression drugs who were older than 50 years. Cases included all patients with an incident hip fracture. Controls were selected using incidence density sampling, matched for sex, index date, year of birth, and both calendar period and duration of up-to-standard follow-up before the index date. For comparison purposes, a similar nested case-control analysis for histamine 2 receptor antagonists was performed.
Main Outcome Measure The risk of hip fractures associated with PPI use.
Results: There were 13 556 hip fracture cases and 135 386 controls. The adjusted odds ratio (AOR) for hip fracture associated with more than 1 year of PPI therapy was 1.44 (95% confidence interval [CI], 1.30-1.59). The risk of hip fracture was significantly increased among patients prescribed long-term high-dose PPIs (AOR, 2.65; 95% CI, 1.80-3.90; P<.001). The strength of the association increased with increasing duration of PPI therapy (AOR for 1 year, 1.22 [95% CI, 1.15-1.30]; 2 years, 1.41 [95% CI, 1.28-1.56]; 3 years, 1.54 [95% CI, 1.37-1.73]; and 4 years, 1.59 [95% CI, 1.39-1.80]; P<.001 for all comparisons).
Conclusion: Long-term PPI therapy, particularly at high doses, is associated with an increased risk of hip fracture.
<!--QuoteEnd--></div><!--QuoteEEnd-->
And since how certain diets affects your health is still pretty indefinite/contested, it's probably a good thing that they increased the scope of the case studies to balance out the limitations on individual health data.
<!--QuoteEnd--></div><!--QuoteEEnd-->
Nothing is contested regarding the effects of vitamin D deficiency or how the lack of sunshine can directly affect calcium absorption.
You really only need a few thousand people.
<!--QuoteEnd--></div><!--QuoteEEnd-->
For what? To get a statistically accurate medical study, you need tens of thousands at the very minimum, and even then, the uncontrollable variables will corroupt the data. A medical study based purely on doctor's records, of only, say, 5000 people? <b>Hideously</b> inaccurate, because of the vast uncontrollables. If you want a study of only 5000 people, at 77, then you have to control everything else by outside means, with would actually corrupt the data more...
No, to conduct a study of <i>anything</i> scientifically, you need as large a sampe as possible, to minimize corruption of the results due to uncontrollable variables...
For what? To get a statistically accurate medical study, you need tens of thousands at the very minimum.
No, to conduct a study of <i>anything</i> scientifically, you need as large a sampe as possible, to minimize corruption of the results due to uncontrollable variables...
<!--QuoteEnd--></div><!--QuoteEEnd-->
Off-topic I know, but a clinical trial of a new drug does only need a few thousand as pointed out here <a href="http://en.wikipedia.org/wiki/Clinical_trial" target="_blank">http://en.wikipedia.org/wiki/Clinical_trial</a>, which also matches what I've been taught at college. 2nd, large as sample as possible is rarely the best idea, simply because past a certain number of samples, accuracy ceases to rise very far, and then becomes simple pointless logistics, and a waste of time.
In medical surveys, the large number of biological random variables might blur the results more than a political survey, but these researchers had an enormous sample size to work from and they were hopefully looking through those records carefully.
For what? To get a statistically accurate medical study, you need tens of thousands at the very minimum, and even then, the uncontrollable variables will corroupt the data. A medical study based purely on doctor's records, of only, say, 5000 people? <b>Hideously</b> inaccurate, because of the vast uncontrollables. If you want a study of only 5000 people, at 77, then you have to control everything else by outside means, with would actually corrupt the data more...
No, to conduct a study of <i>anything</i> scientifically, you need as large a sampe as possible, to minimize corruption of the results due to uncontrollable variables...
<!--QuoteEnd--></div><!--QuoteEEnd-->
I am going to find your Statistics teacher and bonk them hard on the nose. Hard.
it doesn't take a nuclear engineer.
<!--QuoteEnd--></div><!--QuoteEEnd-->
Yeah, but unions are unions.
Yeah, but unions are unions.
<!--QuoteEnd--></div><!--QuoteEEnd-->
What about intersections?
I am going to find your Statistics teacher and bonk them hard on the nose. Hard.
<!--QuoteEnd--></div><!--QuoteEEnd-->
It's not about statistics. The results that a staticitian would be happy with, and a scientist would be happy with, are two completely different things. Especially when it comes to medical results.
Also, your love of statistics doesn't taken into account the political mess this study could cause. 3% margin of error isn't good enough when you're talking about a drug that a large proportion of the elderly community would appreciate, seeing as they seem to be more at risk from heartburn.
No, a statistically "accurate" result is not what you'd want here, what if that 3% were enough to flip the evidence the other way? All of a sudden, Nexium is safe, and has no link, or a dubious link to hip fractures and osteoporosis, and people keep taking it. Then all of a sudden, another peer group comes along, and goes "hang on, they did a medical records study of only 5000 people, and cleared Nexium because of it"?
You'd be thrown out faster than a speeding bullet from a minigun...
You can't just apply a statistical approach to anything, you need to take into consideration any possible effects that your study would have. Clearing, or linking, Nexium to any possible secondary effects would have a distinct lashback, if those results weren't of a large enough group to effectively eliminate the margin of error.
It's not about statistics. The results that a staticitian would be happy with, and a scientist would be happy with, are two completely different things. Especially when it comes to medical results.
Also, your love of statistics doesn't taken into account the political mess this study could cause. 3% margin of error isn't good enough when you're talking about a drug that a large proportion of the elderly community would appreciate, seeing as they seem to be more at risk from heartburn.
No, a statistically "accurate" result is not what you'd want here, what if that 3% were enough to flip the evidence the other way? All of a sudden, Nexium is safe, and has no link, or a dubious link to hip fractures and osteoporosis, and people keep taking it. Then all of a sudden, another peer group comes along, and goes "hang on, they did a medical records study of only 5000 people, and cleared Nexium because of it"?
You'd be thrown out faster than a speeding bullet from a minigun...
You can't just apply a statistical approach to anything, you need to take into consideration any possible effects that your study would have. Clearing, or linking, Nexium to any possible secondary effects would have a distinct lashback, if those results weren't of a large enough group to effectively eliminate the margin of error.
<!--QuoteEnd--></div><!--QuoteEEnd-->
This would make a lot more sense if the approval process for drugs did not involve numerous studies for which a margin of error at or around 3% is perfectly acceptable.
It's not about statistics. The results that a staticitian would be happy with, and a scientist would be happy with, are two completely different things. Especially when it comes to medical results.
Also, your love of statistics doesn't taken into account the political mess this study could cause. 3% margin of error isn't good enough when you're talking about a drug that a large proportion of the elderly community would appreciate, seeing as they seem to be more at risk from heartburn.
No, a statistically "accurate" result is not what you'd want here, what if that 3% were enough to flip the evidence the other way? All of a sudden, Nexium is safe, and has no link, or a dubious link to hip fractures and osteoporosis, and people keep taking it. Then all of a sudden, another peer group comes along, and goes "hang on, they did a medical records study of only 5000 people, and cleared Nexium because of it"?
You'd be thrown out faster than a speeding bullet from a minigun...
You can't just apply a statistical approach to anything, you need to take into consideration any possible effects that your study would have. Clearing, or linking, Nexium to any possible secondary effects would have a distinct lashback, if those results weren't of a large enough group to effectively eliminate the margin of error.
<!--QuoteEnd--></div><!--QuoteEEnd-->
Protip: if the 3% margin of error was enough to possibly flip the evidence, then there wouldn't be a statistically significant difference and the conclusions of the study would be different.
Also the study said people taking the medicine had a 44% higher incidence of hip fractures, so that's kinda safe even +/- 3%.