Please create an account to participate in the Slashdot moderation system


Forgot your password?
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. ×

Comment Re:CORRELATION != CAUSATION (Score 1) 216

Your experiment at the end is PRECISELY what I was recommending. Take people who are doing whatever they would have been doing anyway, and take another similar group and administer an intervention in the form of supplemental Vitamin D. The control group serves as the "low vitamin D" group not through a deliberate restriction, but by taking advantage of an already known tendency for people to be Vitamin D deficient. Not sure why you got all butt hurt about it when it was my idea.

The only difference is that I'm recommending that the researchers refrain from recommending vitamin D as a possible preventative for Autism until after the trial is conducted, whereas you have them making the recommendation before it is conducted. I don't even disagree with recommending people take a supplemented if they are Vitamin D deficient, only with researchers making a specific recommendation for Vitamin D as it relates to Autism until the trial is complete.

Comment Re:CORRELATION != CAUSATION (Score 1) 216

And these all show Correlation because they are picking through data not designed, but essentially found.

There are lots of environmental differences that are confounded with UV exposure and geographic latitude and while epidemiologist can try to balance for all of those factors, they cannot do so completely.

Comment Re:CORRELATION != CAUSATION (Score 1) 216

The researchers are not claiming that Vitamin D deficiency causes autism. They are claiming it increases the "Risk of Developing Autism", which is subtly different. Lots of chronic diseases and conditions are not so cut and dry as "Vitamin D deficiency causes Rickets".

For example, we know that smoking can increase your risk of developing lung cancer, but we also can point to specific examples of people who smoked for decades without ever developing Cancer. We know that processed meats appear to increase your risk of colon cancer, but the risk is fairly low and most people will never develop colon cancer despite a diet high in processed meats like bacon. That doesn't mean these risks are not real, but that the mechanism by which these things increase risk are probabilistic.

Even with the moral questions ignored, it may be impossible to show a 1:1 causation between any single environmental factor and autism development, but we can potentially learn which factors can causally influence risk. The trial presented above can show correlation, but not causality. The trail I proposed could address the potential for causality by showing support, or not, for a change in relative risk of autism development based on level of Vitamin D supplementation.

Comment Re:CORRELATION != CAUSATION (Score 1) 216

Maybe I'm not being clear.

1. enroll women early in pregnancy into a trial.
2. measure all enrollee's Vitamin D status on Day 1 of their enrollment.
3. Block mothers based on their preexisting Vitamin D status.
4. within block randomly assign half of the pregnant women to receive a Vitamin D supplement, and assign the other half a Vitamin D placebo pill
5. Take blood work throughout the pregnancy to test for their Vitamin D status throughout gestation, birth, lactation, until the child is say 4 to 6 years old
6. look at Autims prevalence in the two groups at the end of the study.

I'm obviously ignoring a lot of other factors that would need to be accounted for in the initial blocking step (economics, medical history, age of mother and father, etc.), but this is an experiment designed to change a single factor and to control for possible confounding factors BEFORE the treatment is applied. It can therefore speak to causality. The retrospective CoHort study relied on some bloodwork and a lot of questionnaires and cannot speak to cause and effect at all in the absence of corroborating evidence from actual experimentation.

Comment Re:CORRELATION != CAUSATION (Score 1) 216

I don't argue about that, but in my experience most parents are not exactly rational when it comes to fears surrounding autism. People give their kids bleach enema's from quack doctors they find on the internet. The authors didn't say "be sure you are not deficient" they said "take a supplement" and didn't specify how much. Even if the research did say "avoid deficiency for a host of good reasons, which may include autism prevention" in the interview, this is going to be turned into "take a supplement or your kids will get autism" as it is filtered through the press.

Most Americans, in my experience, are surprised that "More is Always Better" doesn't apply to vitamins. What happens when a mom who took vitamin D has a child with autism, because even if some cases are related to vitamin status, many are absolutely not? She may very well just up the dose for her next child, putting herself and the subsequent child at greater risk of Vitamin Toxicity.

Hell, Americans are not rational about vitamins either. Roughly 30 years ago they fought to prevent the FDA from regulating dietary supplements, including vitamin supplements due to a clever smear campaign by the big vitamin manufacturers. They even got big name actors (of the day anyway) to star in commercials where jackbooted FDA agents with M16's break into Mel Gibson's house to seize his vitamins from his kitchen cabinet. The FDA wasn't looking to ban anything, but to ensure that labelling claims were supported with evidence, but that didn't matter because people know that vitamins are GOOD FOR YOU (TM) and therefore more must be better.

Comment Re:CORRELATION != CAUSATION (Score 1) 216

As I pointed out to another commenter, testing this hypothesis more rigorously does not require putting anyone at elevated risk. You assume, incorrectly, that subjecting one group to a restriction would be necessary.

Having one group proceed as they would normal (change in risk of precisely ZERO), and have one group take a Vitamin D supplement. You can balance for pre-existing Vitamin D status so that the two groups are similar before starting. The Vitamin D supplement would constitute an altered risk (and one that may not be entirely quantifiable), but since the researchers feel the risk is small enough to justify a recommendation without additional data I fail to see why it would cause anyone pause in the context of a trial.

Comment Re:CORRELATION != CAUSATION (Score 1) 216

Vitamin D is not without it's risks. As a fat soluble vitamin, it can accumulate in fat stores. Particularly if people are taking higher doses. Vitamin D can be toxic at high enough concentrations, particularly in young children and presumably the developing fetus.

I consider this irresponsible because they have not confirmed that their recommendation does not constitute the lower risk option. They have just assumed that action is better than inaction, which is not a safe assumption.

A good example of this "unintended risks" can be found in numerous trials evaluating the effect of low fat diets for control of heart disease. Lot of these trials show small reductions in death due to heart disease, but increases in other causes of death for no net change in total death, or possibly a slight increase. At the end of the day, on average, there is no net change in your probability of dying by adopting a low fat diet. You just die of something else.

For this and many other reasons, medical policy should never be set based on a single trial, and DEFINITELY not on a trial that can only show "associations" that may turn out to be spurious.

Comment Re:CORRELATION != CAUSATION (Score 1) 216

No, I'm proposing that we limit our assertions to what can be supported by the available evidence.

There are various ways in which we could strengthen or disprove this association.

1. One could be to survey 1000 pregnant women for Vitamin D status. Then (accounting for status) provide half with a Vitamin D supplement for the duration of their pregnancy and look at the results. Since vitamin D supplementation is not a routine pre-natal practice no one would be at increased risk of Autism (assuming that the correlation is real, causal, and in the direction asserted by the authors).

2. We could look at how Vitamin D affects brain development in research models that already exist to study Autism. Thus building a mechanistic framework to justify considering the correlation detected above as causal.

We need to be clear that vitamins are not completely harmless, particularly the fat soluble vitamins, which includes Vitamin D. Unlike the B vitamins, which are excreted via the urine when consumed at high levels, fat soluble vitamins accumulate in body fat stores, and can reach toxic levels much easier. There can be too much of a good thing, particularly in young animals (including the developing fetus) who's storage capacity can be relatively quickly exceeded by small changes in supplementation.

Comment CORRELATION != CAUSATION (Score 1) 216

Ok, lets be clear this is a retrospective cohort study. That means that no experimental treatments were applied, so cause and effect cannot be determined on the basis of this trial. This is a clear example of what is wrong with most people's understanding of "BIG DATA". Just because the sample size is enormous, it does not immediately follow that differences detected are real, meaningful, or causal. In-fact, large sample sizes Guarantee that spurious differences will pop up in the data pretty regularly.

Even if the association is real, there is nothing in this data to say whether
A) low Vitamin D causes increased autism risk
B) Increased autism risk causes low Vitamin D
C) Or whether both are in turn caused by something else with improvements in one having no effect on the other..

The authors should never have made recommendations based on this data, other than that more trials are needed to determine whether or not this association is causal, or if it could be used as an indirect indicator of autism risk.

Comment Re: Missing information (Score 1) 166

Pure water isn't really an option at the municipal level. Even after the water leaves he plant it has to move through miles of pipes to get to your tap, and frequently the city that owns the plant and the city that owns the pipes serving your house are not the same cities. Never mind that the quality of the pipes IN your home can matter as well.

There is a practically infinite number of chemicals that can dissolve in water that would need to be tested forth confirm that the water is pure. And we need to deal with the fact that this is the real world and money/time/equipment/etc are finite resources. I'm not meaning to say that the EPA shouldn't find out what is there, but simply that the presence of a chemical is not enough to indicate anything nefarious. The next step is to determine if there is any reason for concern. They have a program right now screening chemicals for endocrine disruption potential. And not just at any concentration, but at the concentrations seen in the real world (the dose makes the poison after all). However, until we have more reason to be concerned than "look what we found" we should refrain from freakingn everyone out to no benefit.

Look at all of the scares around BPA. We were told it's everywhere, and collectively freaked out. Replacing it with other chemicals that are just as poorly understood. but when the tests were actually performed it was shown that while it *can* act as an endocrine disrupters in cell culture, it is so rapidly detoxified and excreted in the urine that the levels seen in people and water are well below any level of concern.

Comment Missing information (Score 1) 166

Exposure is only part of the toxicological equation. Is there causal (not just an associated) evidence of harm? Do we know what dose is necessary to cause this harm, and what the likelihood of harm is at a particular does? The EPA does an excellent job of surveillance. Unfortunately they do a less than stellar job of confirming there is a problem before someone turns exposure data into the basis for a panicked article about the latest "chemical".

Comment Re:Old price is the problem. (Score 1) 472

Valuable perspective, but it is only one perspective (the IT manager). And I'm not sure it is really relevant since the recommendations at are not for IT buyers but for personal use buyers. From the "Having to manage my own machine without support from a professional" perspective, a Mac has consistently been the best choice. Yes, they cost more, but I have fewer headaches (though by no means none) related to managing them in the handful of hours each week I have to use it when I'm not earning a living.

Furthermore, your experience is different from other IT managed scenarios in that you are managing for children, and the computers are shared resources. As I recall a couple of years ago there was a /. article about how IBM had moved to letting everyone select their own device and found that Mac's resulted in disproportionately fewer help desk tickets. The money spent on hardware resulted in savings from support. Your mileage may vary of course

I've personally helped dozens of people switch from PC's that they never really felt comfortable with, to Mac's which enabled them to get more from their machines. That kind of productivity gain and piece of mind can't be tabulated on a spreadsheet and so therefore has essentially zero value in an IT managed scenario, but for the day-to-day user of the machine it has great personal value. That is why Macs do so well with people buying their own machine, but not at large corporations.

Finally, the older MBAir that hasn't been updated will probably never be updated. The differences in portability between the Air and the MBPro are almost meaningless at this point. It exists in the line-up to hit a certain price point, and do so profitably. As long as it can fill that role, and still sell in meaningful volumes, Apple will keep it around. It's good business, regardless of what flamebait writers say.

Comment Re: Well... (Score 1) 609

No, I'm not slandering good folks based on a few bad apples. This is what eugenics was originally about. The Nazis didn't invent the concepts of racial superiority and racial purity. They've always existed, the Nazis just took it to the farthest extreme. Eugenics grew out of a combination of the racial ideals and a determine misinterpretation of the implications of Darwins works. Implementation of eugenics, even with only so-called "positive" selections is still dependent upon some external body to determine which groups deserve to be promoted, and is hopelessly biased and confounded with political and economic privledge.

Slashdot Top Deals

People are always available for work in the past tense.