Those Sleeping Pills May Be Killing You 237
dstates writes "A recent article in in BMJ Open reports a strong association between the use of prescription sleeping pills and mortality. The study used electronic health records for 2.5 million people covered by the Geisinger Health System to find 12 thousand who had been prescribed sleeping pills and a matched set of controls. Death rates were much higher in the patients taking sleeping pills and the risk increases with age. Kudos to the authors for publishing this in an open access journal."
Re:Neat but not surprising (Score:5, Interesting)
I quit smoking because I don't want to get cancer, and I don't want to smell bad all the time, and I don't want to be out of breath walking up the stairs. That said, I loved smoking. I still miss it every day, but the risks are greater than the rewards.
Stop taking my sleeping pills? Hah. Have you ever been so tired that you get a sore throat? Or that you argue with yourself at a stop light, "no, don't close your eyes, I know it would feel really really good but if you do that you'll miss the green and might not wake up until somebody knocks on your window"? Your legs shake, you feel sick to your stomach, your palms sweat constantly, your eyes try to close with all their might until you can hear the muscles straining in your ears.
Now try feeling like that for months on end. Stop taking my sleeping pills? Fuck that shit, I'd rather die early.
Re:Did they adjust for crazy? (Score:3, Interesting)
Re:Did they adjust for crazy? (Score:4, Interesting)
"The population is mostly of low socio-economic status, having less than high school education and less than one-third are insured under the Geisinger Health Plan."
"We were unable to control for depression, anxiety and other emotional factors because of Pennsylvania laws protecting the confidentiality of these diagnoses."
The results of this study ought to be interpreted in light both of the socio-economics, demographics, and regional characteristics of the population studied, and of the potentially crucial categories of comorbidity that were excluded.
My own use of zolpidem (Ambien) was during a time of an extraordinary convergence of situational stress factors. Once the stress conditions resolved, I was able to discontinue the drug.
Re:Did they adjust for crazy? (Score:2, Interesting)
Re:Those Hospitals May Be Killing You (Score:5, Interesting)
Re:If you have a hard time sleepin (Score:3, Interesting)
I "trick" myself into exercising by just picking up a sport and sticking with it. Treadmills are boring, but sport can be fun with the right crowd.
Nailed it on the head.
Back when I embarked on getting into better shape.. I struggled to force myself to do the recommended weekly exercise. It was 30 minutes I really would rather spend doing something else. Then I got into a floor hockey thing some guys at work had going and it literally changed everything. I saw the light. Not only was I getting way more exercise than I was doing jumping jacks in my basement.. but I actually _looked forward_ to it.
From there I got into badminton, then squash, then skiing.
I feel great now (I didn't buy it when others told me, but doing more exercise really does give you more energy) and exercise has gone from something I have to do if I don't wanna have a heart attack in 20 years, to something I do purely for enjoyment, with the health benefits a side benefit.
Also a side benefit I look much more attractive. I never thought I cared about my appearance, but I guess deep down I kinda did. I feel a certain confidence that I never new was missing.. if that makes sense (and I get if it doesn't, because I barely understand it).
Re:Correllation != Causation (Score:4, Interesting)
Another bogus study recently was on the correlation of Diet Soda to Heart desease.
http://www.reuters.com/article/2012/02/20/us-soda-diet-idUSTRE81J03920120220 [reuters.com]
The problem is the factors that bring people to diet soda are overweight or diabetes which are both factors that would tend to increase risk of heart desease. Here too the people that are prescribed sleeping pills are a self selected group on some problem associated with sleep. They can say they have factored in all the external variables but that may not be that case or there may be other factors related to sleep problems that are more to blame and this study would not be complete or accurate unless it followed an equally large group of people that would have been given sleeping pills but were not and/or given a placebo, You don't know if giving a pill is the problem say vs not giving a pill as one factor. If you give a placebo you don't test that variable.
The study said it matched "matched controls with no hypnotic prescriptions" but they did not match with those who would have been prescribed meds but weren't. You may only be seeing the effect of the problems they had. I could be that the meds actually lengthened life vs the group not given them. This is the problem with vacines that have side effects, but the cure in general is much better for society than the side effects.
But that kind of study is hard to come by so you may just have to go on the data you have but as in religion and politics attribution can be a grand evil.
Re:Should have been triple-blind... (Score:5, Interesting)
This is not really true. The purpose of a double-blind experiment is to set up a study with a controlled variable and observe the outcome. This is a meta-analysis, which looks at previously gathered data and tries to see if there are interesting patterns. The problem with such analysis is that although "blind" in the sense that it does not influence results, it is not "blind" in the choice of data. Whether intentionally or not, by cherry-picking data it is easy to create associations where none exist. This is further biased by the fact that only positive results are reported - no one writes of all the "no correlation" results they may have found through different choices of matched sets.
For example, I am sure that I could take a piece of data such as daily temperature and pick a subset of the stock market that happened to correlate with it - something that is likely entirely a figment of the data sets. This is the danger in such studies and it explains why they are NOT in any way the same as a double-blind trial.
Re:Should have been triple-blind... (Score:4, Interesting)
"This is further biased by the fact that only positive results are reported - no one writes of all the "no correlation" results they may have found through different choices of matched sets."
If you're honest with your stats you multiply your p-value by the number of comparisons you did. Yes, some of us do this. There's nothing wrong with retrospective analyses, it's just that so many people do the stats incorrectly.
Re:Did they adjust for crazy? (Score:5, Interesting)
Most of the people I know who take sleeping pills are not necessarily the most stable people in the world to begin with. Sorry to all you Ambien fans.
Theoretically, yes. In practice I don't think so.
This 'confounding by indication' is one of the biggest problems in pharmacoepidemiology. We know that people take meds because there is something wrong with them. We also suspect that taking certain meds over a long period of time is bad for you, particularly if you are already at high risk. So how can you separate those effects? A lot of statisticians spend a lot of time thinking about this, and 'adjusting for everything you can think of', propensity scoring and very tight matching of cases and controls seem to be the most often used solutions. None of these is satisfactory as they obviously don't adjust for things you can't measure. Use of instrumental variables is another possibility but there is rarely a good instrument to use.
Ideally you would run a randomised trial of a med to check whether death rates or adverse drug reactions are higher in the group taking them, but this is impractical because often the required trial would be enormous (massively expensive and time consuming), would have to recruit many of the 'high risk' people that are the groups most at risk of excess mortality but are usually not recruited into trials, and could only really examine one compound at a time. Also trials exclude people taking many other medications, or with comorbid medical conditions, because these may be unsafe and would again dilute the true effects - however it is likely that unknown drug-drug interactions are the cause of a lot of the problems we think we are seeing.
It's easy to snipe at this kind of research since its 'correlation not causation' but this really is the best that is possible at the moment when trying to answer these extremely important questions regarding drug safety. If anybody has any better ideas we'd be glad to hear them.