Those Sleeping Pills May Be Killing You

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."

Comments

Did they adjust for crazy?

[elrous0]

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.

Re:Did they adjust for crazy?

[Aladrin]

I couldn't find anything in it to suggest they had actually done a double-blind trial, or even a half-assed blind trial, so their results are purely correlation, and not causation, despite the time they spent talking about causation. They do suggest that 'hangovers' from the drugs are a cause of traffic accidents and such, though, so they at least thought of that.

No mention about mental stability that I saw.

Re:Did they adjust for crazy?

[MozeeToby]

Most people taking prescription sleeping pills have been fighting sleep disorders for a long time, probably their entire adult lives. Getting terrible sleep for 30 or 40 years will probably increase your mortality regardless of what pills you're taking. Do the same study again only this time instead of looking at what drugs they're on, give them a sleep disorder questionnaire, drowsiness survey, and a sleep study. Then you'll have enough data that I actually care to look at your results.

Re:Did they adjust for crazy?

[ceoyoyo]

No. When you do an experiment, i.e. purposely manipulate one variable, you establish a causal connection. Identifying and explaining the mechanism is nice, and establishes the character and directness of your causal relationship. Trials are experiments.

Correlation comes from observational studies where you do not manipulate any variables yourself, you just look for natural or preexisting variation.

A simplified example - if I look at a bunch of people who take sleeping pills and a bunch who don't, and measure how likely they are to die, I get a correlation (maybe) - dying and taking sleeping pills are correlated, but I don't know if dying causes people to take sleeping pills, whether sleeping pills tend to cause you to die, or whether some other factor (being crazy maybe) causes you to both take sleeping pills and die.

If I take a bunch of random people and give some sleeping pills and others no sleeping pills, if the ones I give the pills die significantly more often then I can conclude that sleeping pills cause death (by some mechanism I don't yet know).

Re:Did they adjust for crazy?

[vlm]

None of these is satisfactory as they obviously don't adjust for things you can't measure

Or things you won't measure for whatever convoluted reason.

For example, back pain patients given powerful painkillers recover slower or not at all compared to no painkillers.

Example of false reasoning: I overstrained my back doing some overambitious carpentry alone. Intense pain when sitting or standing, laying on back not so bad. Went to doc, did not accept script for painkillers because I slept on my back just fine and everything I do sitting or standing is not allowed while on pain killers anyway (can't even drive to work if I'm high on painkillers). Also doc is all nervous that I'm dr shopping for abuse meds and really chilled out and got more helpful once he realized it was perfectly clear that I was only genuinely trying to fix my back. blah blah blah. The point is the diagnosis of "back pain" is the same for me and someone who's in agony even when lying down so they need painkillers just to sleep. No great stretch of imagination that the guy in more agony than myself is more F'd up and takes longer to recover (took me only about half a week, but I've heard if you really F up your back it can be semi-permanent, months maybe). Multiply this by 15K and you get a whopper like "taking painkillers means it takes months to recover from back pain diagnosis instead of days"

Correllation != Causation

[recoiledsnake]

The people taking the medications might be dying sooner because they have insomnia which is not fixed by sleeping pills easily. The study should not compare with the general populace since they are, by definition, better sleepers than the group that isn't able to get good sleep.

Re:Correllation != Causation

[mystikkman]

Bingo, this is like comparing the death rates of people taking heart medications versus people who don't and then claiming the medications are killing the folks when it could be that heart disease is what is killing them and the pills are not 100% effective at all times to deal with the problem.

Also, it has been proved that bad sleep is a killer by itself, so comparing the death rates of people with sleep issues who did and who did not take medications may actually show that not taking sleeping pills might kill you if you have insomnia, which is the exact opposite of what the headline is claiming.

Re:Correllation != Causation

[vlm]

Two "identical" guys one gets pill one doesn't is an anecdote.
12 thousand is not just a misdiagnosis. There must be something "different" about the 12K that did vs the 12K who did not, other than random chance. I don't think in the UK treatment plans are determined using dice or tarot cards, so there must be something special about the 12K who got the pills... or the 12K who didn't...

Re:Correllation != Causation

[mystikkman]

This was a not a controlled study but is a general experiment so that does raise the question of why a doctor would not prescribe sleeping pills to someone who is having a lot of trouble sleeping?

The only way to conclusively prove this would be to take patients who are going to be prescribed sleeping pills, split them into two while adjusting for other controls and have one group take the pills and other placebos and then monitor them.

Re:Correllation != Causation

[Anonymous Coward]

Those were not controls. They were simply a group matched on a very few parameters. Not everything. Not nearly everything relevant.

A control would have been to take people diagnosed exactly the same and giving some sleeping pills and the others placebos. That didn't happen. The sleeping pill group might have all had a family history of sleep apnea which was not diagnosed before giving the pills. We don't know because there was no randomization.

"Correlation != Causation" is EXACTLY the case we have here.

Re:Correllation != Causation

[recoiledsnake]

I'm tired of seeing these stupid comments every time an article on statistics is brought up. Clearly, a bunch of scientists doing studies along these lines know less about statistics and research design than some random Slashdot poster. Gee. Get over yourself.

Are you sure?

http://www.guardian.co.uk/commentisfree/2011/sep/09/bad-science-research-error

But in just this situation, academics in neuroscience papers routinely claim to have found a difference in response, in every field imaginable, with all kinds of stimuli and interventions: comparing younger versus older participants; in patients against normal volunteers; between different brain areas; and so on.

How often? Nieuwenhuis looked at 513 papers published in five prestigious neuroscience journals over two years. In half the 157 studies where this error could have been made, it was. They broadened their search to 120 cellular and molecular articles in Nature Neuroscience, during 2009 and 2010: they found 25 studies committing this fallacy, and not one single paper analysed differences in effect sizes correctly.

These errors are appearing throughout the most prestigious journals for the field of neuroscience. How can we explain that? Analysing data correctly, to identify a "difference in differences", is a little tricksy, so thinking generously, we might suggest that researchers worry it's too longwinded for a paper, or too difficult for readers. Alternatively, less generously, we might decide it's too tricky for the researchers themselves.

Why is it wrong for a Slashdot poster to have a conversation over the statistics involved when the headline is so sensationalist? What if someone reading stops taking sleeping pills that are helping them sleep and then get needlessly killed by insomnia because of bad statistics? Can't there atleast be a discussion on the statistics used?

I am tired of seeing stupid comments like yours that actually don't refute anything and instead attack the poster and call scientists infallible and above question.

Re:Correllation != Causation

[Rakishi]

Listen to your own advice and read the paper yourself before commenting.

They specifically compensated later on for difference in heart disease, asthma so on. The impact on the final result was minimal.

Re:Correllation != Causation

[Rakishi]

They tried to do that under the assumption that in general a lack of sleep probably doesn't kill you. The hypertension, or diabetes, or heart disease, or liver failure or obesity is that kills you. They did control for differences in those factors and found no change in their results.

Of course, the increased rate of crashing your car into to a wall at 90mph due to insomnia wouldn't be taken into account.

Re:Correllation != Causation

[FlavaFlavivirus]

From the paper: "A randomised clinical trial of sufficient duration and size could provide definitive evidence for or against the disturbing mortality hazards suggested by our study. Some American NIH reviewers have opined that a randomised trial of hypnotic lethality would be unethical. No such trial has ever been mounted, perhaps for reasons similar to the absence of randomised trials of cigarettes and of skydiving without parachutes." It is absolutely unethical to give persons with no history of sleeping problems a potentially-lethal drug. This is as good as it gets.

Most drugs are bunk

[Hatta]

Drug companies spend more on marketing than they spend on research. Is it any surprise that these stories keep coming up? SSRIs were going to cure everyone's depression. Now we find out that they're addictive, and only effective in the very worst cases of depression. Vioxx was going to usher in a new age of pain relief for arthritis, turns out it killed tens of thousands of people. Hormone replacement therapy was considered essential to prevent osteoporosis in women. Turns out it also causes bone remodeling that makes certain types of fractures even more common. Don't be surprised if we find out in the future that wonder drugs like statins carry risks we haven't been made aware of.

Pharmaceutical companies should not be allowed to market. Not to the general public, and not to doctors either.

Re:How often is this done?

[canajin56]

They're called "statistical studies" and they are used as evidence that a real study should be done, not that there's an actual effect in play. The problem with such studies is that they try their best to select an identical control group, but it's hard to do so. In this case it means matching the 15,000 people on the drug with 15,000 people who also have been diagnosed with insomnia (and for similar reasons), but all elected not to be medicated for it. Then you hope that that decision isn't in any way correlated with other behaviors that might increase or decrease the death rate.

Great Work

[trongey]

I'm glad they discovered that death risk increases with age.
Did you ever wonder how they come up with death rates that are less than 100%?

Re:Should have been triple-blind...

[icebike]

Sounds to me like data mining and meta analysis, which is all the rage today.

This study followed their subjects for an average of ONLY 2.5years. They clearly didn't follow them prior to the prescriptions.
Further the "controls" were selected based on superficial categories (age, gender, smoking, body mass index, ethnicity, marital status, alcohol use and prior cancer). Nowhere near a complete list of things that keep people awake at night.

And the causation argument still is the key here, since these drugs (several common hypnotics, including zolpidem, temazepam, eszopiclone, zaleplon, other benzodiazepines, barbiturates and sedative antihistamines) are not usually prescribed for people who have no problem sleeping.

Selection of controls was really the weak point here.

If you are under enough stress, or have some other problem keeping you awake, its as likely those issues are to blame as the use of these drugs. The headlines could just as well have been "Trouble Sleeping may be Killing you".