The graveyard shift? You're not kidding!

A BBC report of a new study by sleep researchers suggests that night shift workers have higher risk of various health problems than do we daytime doodlers; heart attacks, cancers and type 2 diabetes.  This is because the expression patterns of many genes are based on the day-night cycle, and the 'chrono-chaos' of night work upsets lots of body functions, the story says.

The study, published in the Proceedings of the National Academy of Sciences, found that mistimed sleep caused gene expression to fall significantly.  Genes affected included those having to do with circadian rhythms, or the maintenance of our sleep/wake cycles.

One can't be totally surprised, although one might expect that the graveyarders would get used to their diurnal cycle and do just fine.  One has to wonder if there are other things about who chooses to do night work, or doesn't have options, so that nightshifting is a consequence rather than cause.  In that case, nightshifting would be a confounder relative to the health implications rather than their cause.

The point here is rather just a brief one, that we and many others have repeatedly made.  If these types of variables are not known or taken into account, or there isn't enough of this risk factor detectable in the study sample, then attributions of causation of what is measured will be inaccurate of misleading. This is one of the challenges of epidemiological research, including the search for reliable risk factors in the genome.

Then there is the question, related to an earlier point above, whether any genetic risk factors lead the bearer to look for nightwork and hence appear to be associated with some health result only indirectly.  What about variants in the chrono-genes?  Many such questions come to mind.

Inferential chaos?
Maybe, therefore, the chrono-chaos is a different form of informational and inferential disorder.  A disorder of incorrectly done studies.  As we know, many results of association studies, genetic or otherwise, are not confirmed by attempts to replicate them (and here we're not referring to the notorious failure to report negative results, which exacerbates the problem).  We don't know if the 'fault' is in the study design, the claimed finding of the first study, other biases, or just bad statistical luck.

A piece in Monday's New York Times laments the high fraction of scientific results that are not replicable.  This topic has not gone unnoticed; we've written about different reasons for nonreplicability over the years ourselves.  The degree of confidence in each report as it comes out is thus surprising, unless one thinks in terms of careerism, a 24/7 news cycle and so on.

Not to defend fast food, but does it really cause asthma?

Here's a case for confounding variables if there ever was one: a new study suggests that 'fast foods' cause asthma, rhinoconjunctivitis and eczema.  Indeed, the BBC tells us that teens who eat fast foods three or more times a week are 39% more likely to have severe asthma than kids who eat three or more servings of fruit each week. Sounds pretty clearly causal, doesn't it?  Well, maybe not.

The results are from a long-standing international study of asthma, begun in 1991 because of concern over increasing asthma rates.  Called ISAAC, the International Study of Asthma and Allergies in Childhood, the study is based in New Zealand but is associated with centers all over the world.  Asthma incidence began to rise in the mid 1980's and the trend has yet to be satisfactorily explained by environmental epidemiology.  And of course, many tens of millions of dollars have been spent looking for a genetic cause, without success.   

An ISAAC study published in 2004 reported that eating hamburgers was associated with increased risk.  Other studies have found some association with maternal diet during pregnancy, as well as childhood diet. 

Results from the latest paper relied on questionnaires translated into 53 languages, and included more than 500,000 children from 54 countries. Along with lifestyle questions, teenagers, and parents of younger children, were asked about symptoms of asthma, watery itchy eyes and eczema.  And, quoting from the paper, they were asked the following about their diet: ‘In the past 12 months, how often, on average, did you (did your child) eat or drink the following: meat; seafood; fruit; vegetables (green and root); pulses (peas, beans, lentils); cereal; pasta (including bread); rice; butter; margarine; nuts; potatoes; milk; eggs and fast food/burgers?’.  They could choose 'never or occasionally,' 'once or twice per week,' or ' ≥3 times per week.'

Covariates considered included exercise, television watching, maternal education, maternal smoking in the first year of life and current maternal smoking.

While the BBC story makes a case for fast food being the cause of asthma, rhinoconjunctivitis and eczema, the story is much less clear if you look at the details actually presented in the paper. In teenagers, fruits are 'protective' if eaten once or twice, or 3 or more times a week,

[m]ilk was inversely associated with current wheeze once or twice per week, severe asthma ≥3 times per week, VQ [current wheeze], current and severe rhinoconjunctivitis once or twice per week and current and severe eczema once or twice per week as well as ≥3 times per week for current eczema. Vegetable consumption was also inversely associated with current wheeze ≥3 times per week and VQ once or twice per week and ≥3 times per week.

So, milk was protective if consumed once or twice per week, but not if consumed three or more times per week.

'Risk factor foods' were butter, fast food, margarine and pasta ≥3 times per week.  And seafood was positively associated with 'severe asthma, current and severe rhinoconjunctivitis and current and severe eczema. Butter, margarine, nuts, pasta and pulses [legumes] were positively associated with one or more conditions.'  If there is some consistent causal factor -- dairy fat, say, why butter but not milk?  Why not eggs of meat, if it's just fat that's the risk factor?

In younger children (6-7),

For all centres combined, eggs, fruit, meat and milk ≥3 times per week were inversely associated with all three conditions, current and severe. Cereal ≥3 times per week was inversely associated with severe asthma. Vegetables once or twice per week and ≥3 times per week were inversely associated with current and severe wheeze as well as current and severe rhinoconjunctivitis and ≥3 times per week with current eczema.

So, eggs and meat are protective, whereas they aren't for older children (aged 13-14).  Indeed, risk factor food for younger children is fast food only.  And, the only food category both age groups have in common is fast food.  Which, of course, could include a large variety of things. 

To add to the difficulty making sense of these data, "there is considerable variation for some foods between centres within countries."  Indeed, take a look at the odds ratios (ORs), the measure of the strength of association between variables and outcome, for fruits and fast foods by age and country.  They are, by and large, all over the place.  So, the odds of having severe eczema if you eat fruit in Mexico, say, vs. never eating fruit are 0.15, or 2.0.  That is, risk ranges from much lower than never eating fruit, to twice as high. The odds of having current wheeze if you eat fruit in Spain are 0.50 or 2.0.  The overall data may well suggest a trend, but it's not one that is reliably evident center by center, or even within each country. 

So, apart from the question of why this story has been picked up by the media (well, we know why -- anything that indicts fast food has the potential to be sensational), how is it that these results can be all over the place like this?

For one thing, and we've said this numerous times in numerous posts, food questionnaires are notoriously unreliable.  Do you know how many times per week you consumed pasta in the past 12 months?  Or fruit? 

And, who is likely to eat more fast food?  Kids in cities.  Higher risk of asthma in cities has been documented over and over again.  It isn't clear why that is, though allergy to cockroaches has been suggested, or proximity to asthmagenic industry, although this contradicts the other widely held hypothesis that too much cleanliness is causal (this is the Hygiene Hypothesis, which we blogged about here), and kids on farms have frequently been found to be at lower risk of asthma, for whatever reason.  Ok, it's possible that fast food is the reason, but it's also possible (and we'd say likely) that it's a confounding variable, a factor that stands in for an actual but unknown and unmeasured risk factor that kids in cities are more exposed to than kids in rural areas.  We aren't saying that's so, just suggesting it's a possibility; whether this study shows this rural/urban disparity is not clear. 

The asthma epidemic is relatively recent.  Having fresh fruits and vegetables year-round is also relatively recent, so most grand and great-grandparents were disease-free for some other reason.  Unless, of course, fruit and veg are cancelling out the effect of fast foods in kids who eat both, which we can't tell from the report, but that does seem to be a rather contorted explanation, if so.

It is not our purpose to defend fast food here, but really, how unhealthy is it, relatively speaking?  Yes, it packs a lot of calories, and for a lot of people that's not a good thing, but what else might be wrong with fast food?  And we are talking about fast food as a whole -- Big Macs, fries, fish sandwiches, milk shakes; everything lumped into one.  The authors write, 'Biologically plausible mechanisms for the relationship between fast food consumption and asthma and allergic disease could be related to higher saturated fatty acids, trans fatty acids, sodium, carbohydrates and sugar levels of fast food and possibly preservatives.'  Maybe, but if so, what have fast foods got that bacon, beef, any desserts, and so on don't?

And, the authors suggest that fruits and vegetables are loaded with antioxidants, which are said to be protective -- against almost everything.  Except that multiple large double blind studies have found that they aren't.  So, again, it's possible that fruit and vegetable consumption is a proxy for some other unidentified protective factor.  Or not. 

Given that the data used in this study are recall data, the diseases self-reported, the causal reasoning purely speculative, and that there's a possible problem with multiple confounding variables, we're going to go out on a limb here and suggest that, based on this study, the diet/asthma-eczema-rhinoconjunctivitis connection is a tenuous one at best.

If a bird calls in the forest and no one hears it, is it there?

An interesting discussion has been going on in the British birding world for a few months now.  The topic was mentioned on the 8/23 airing of "Material World," which is how we came to hear about it.  The show included mention of a discussion of the issue in "British Birds" starting here, in March, with follow-up here. The author of the original piece, Richard Porter, is one of the UK's top birders.

Willow warbler

Fifty years ago, in the spring of 1962, I took part in the first year of the BTO [British Trust for Ornithology] Common Birds Census, my patch being in the grounds and adjacent woodland of the National College of Food Technology at Weybridge, in Surrey. One of the commonest birds was the Willow Warbler Phylloscopus trochilus and most were located by ear. If I repeated the census now, and assuming there had been no population change, my counts would be much lower, as I would be failing to pick up many singing birds. The ageing process!

Bird counts, done by numerous expert volunteers, are crucial to knowledge of which birds are where when, and to the construction of bird atlases and so on.  Good birders rely as much on their hearing as they do on their eyesight to locate and identify birds, if not more, so hearing loss is a common topic of conversation among aging birders -- "I can no longer hear the black and white warbler, though I could last spring" -- and always mourned.  High pitches are often the first to go with age-related hearing loss, and of course bird song is generally high-pitched.  But then again, most birders know when they aren't hearing a bird song -- they see the beak moving but hear nothing, so how much of a problem this is isn't clear.

The stereotypical birder is an eccentric old lady traipsing about in the brush with binoculars around her  neck.  How representative she is of all birders is not known: birding organizations, at least British bird organizations, currently don't collect data on volunteers.  But what if birders are by and large an aging population, and bird counts rely predominantly on them for data?  Bird counts could look like they were shrinking when they were in fact going in any direction; up, down, or nowhere.  The BTO acknowledges that this could be a problem, but they don't believe it's a major one, saying there are plenty of young birders coming along.  There surely are, but we still don't know the age distribution of bird counters.

Whatever the demographics of this particular group, this story interests us for a few reasons.  First because it is curious to wonder how accurate bird counts actually are.  Already there's the sampling nature of the counts -- birders can't be everywhere and count every bird and yet it's not possible to know what proportion of the population of any given species is missed.  So, if some are missed in areas that were thought to have been well-covered, this can produce an inaccurate count even of areas thought to be accurately counted, as Richard Porter suggests.

But it also represents an issue that is true of probably every survey ever done: there are variables, known and unknown, that influence the outcome but that can't be measured, or that aren't measured because no one thought to measure them. These are known as confounders in epidemiology, and the field is plagued by them, and they probably explain more non-replicable results, more false negatives and more false positives than any other factor.  But because of their very nature, we most often can't know which. 

Tasmania Tiger, 1906, Washington DC

And what about extinctions?  Is the Tasmanian Tiger extinct or isn't it?  Thousands of "unconfirmed" sightings have been reported since it was declared extinct in 1936, for example.  Are they credible?  A good succinct piece about what biased reporting means for extinction statistics over at Smithsonian.com suggests that what we know about extinctions depends pretty heavily on reporting rates.  A lot more extinctions are being documented in the US, Australia, and other rich nations and many fewer in many parts of Africa, and, say, Paraguay.  It turns out that the difference is pretty clearly due to discrepancies in reporting -- everyone's doing an equally poor job of protecting endangered species.  And then, of course, there's all the politics of conservation biology -- it's not all bad if your favorite species looks more endangered than perhaps it is because that gives funders motivation to send money your way.  How common that is, of course, is unknown.

Are we getting at least a good sense of the direction population counts are heading in?  We hope that reporting bias isn't so bad that even this is unknown.

"Oh, don't be such a crybaby! Walk it off!" (says the harsh coach. And he's right!)

If this story doesn't make the point about the elusive nature of genetic causation, what will?

About 10,000 cases of breast and bowel cancer could be prevented each year in the UK if people did more brisk walking, claim experts.

Cancer is a disease of misbehaving cells.  But 'walking' can reduce the risk.  How can that be?

A tumor and its metatastic dissemination is a clone of cells, often if not usually descended from a single misbehaving cell whose misbehavior is inherited by the daughter cells it produces when it divides.  The initiating event may be a mutation in the usual sense (change in DNA sequence in some body cell, such as in the lung or intestine), or it may be due to some other trigger. But it is generally a clonal trigger.

So how can an organism-level exposure, such as to walking or other exercise, have anything to do with what happens to one your billions of cells?  The answer is unclear, but what is clear is that it works through confounding.  That is, it's not exercise itself, but something correlated with it.  In this particular case, if the story is correct, the confounder is obesity.  That leads to chemical agents circulating (related to normal body steroids, perhaps) that can modify cell behavior.  The less obese you are, goes the argument, the less of these mutagens are flying around your bod and bumping into DNA here and there, raising the odds the wrong gene in the wrong cell will be damaged.  Or something like that.

If this explanation is right, obesity is itself a correlate or confounder of the circulating molecules.  Thus walking is correlated with obesity which is correlated with the actual causative agent.  If so, it is easy to see why it is so damned difficult to find a 'cause' of cancer.  Genetic susceptibility variants that are measured are confounded by these other factors.  If undetected, the variant itself becomes an additional confounder.

The confounders are not perfectly associated with each other.  Not all who exercise are slim, not all slim people exercise, not all of either need have any given genetic variant, countless variants across the genome could individually contribute, and what about smoking, diet, and other factors, measured and unmeasured?

If there are too many such factors, even if all are measured, sample sizes for detecting their individual or combinational effects may be prohibitively large.

Welcome to the world of complexity!

We are gathered together to . . . .

There's an interesting paper (or, actually, it's not a paper, it's a description of an ongoing study; Lee et al., Cohort Profile: The biopsychosocial religion and health study (BRHS)) in the December International Journal of Epidemiology, (2009 38(6):1470-1478) but it probably interests us for all the wrong reasons. It's a teaser of a piece, a profile of a sample that's been chosen for a study of the effects of religion on mortality, and a call for collaborators.

The piece describes the aims of the study, and the characteristics of the cohort being studied, thousands of Seventh Day Adventists living in Loma Linda, CA, both black and white. The idea is that Seventh Day Adventists live longer than non-Adventists, even when their healthy lifestyle is controlled for (the relevant lifestyle factors mentioned in the piece are vegetarian diet, not smoking, social support and eating nuts -- the Adventists' position statement on diet doesn't emphasize nuts, we notice, but maybe in Loma Linda, CA, nuts are a big part of the diet?). The investigators are collecting piles of data on religious observance, social and economic variables, as well as health indicators, to try to figure out what it is about religion that contributes to longer life.

This is curious, really, and not a little tantalizing. Because the Loma Linda study is still underway, there are no results yet to report, so we turned to the literature to see what's already been written on the correlation of religion and mortality. It turns out there's quite a lot, including, for example, this meta-analysis of 42 different studies of this question, which concludes that "[r]eligious involvement was significantly associated with lower mortality (odds ratio = 1.29; 95% confidence interval: 1.20-1.39), indicating that people high in religious involvement were more likely to be alive at follow-up than people lower in religious involvement." Indeed, the IJE paper notes that "In fact, Hall concluded religious attendance was more cost-effective in increasing longevity than statin-type medications." Powerful stuff.

Obviously, the question is Why? What is it about religion that protects people from death? (This is a little paradoxical, since religion is ultimately supposed to protect people after death, or maybe even increase their desire to hasten death so as to get to Heaven quicker -- but, ok, we'll stick with the premise.) Do religious people live healthier lives?  Is it healthier people who answer 20 page questionnaires on their religious views?  Could it be simply the fact of belonging to a group (which, admittedly isn't really so 'simple', since what is it about group membership that would be protective, and how would you figure it out?)? In which case, do people in book groups or Elks Clubs or stamp collecting clubs live longer, too? That is, is religion a proxy variable for group cohesion? If so, why all these studies considering religion as though there were something specific and unique about its practice that explains longer lifespans?

Which there is. It's prayer. But the possibility that the power of prayer is the explanation doesn't seem to get mentioned in this field, where instead it's all talk of social and psychological variables. (The way to test the power of prayer, of course, is to look at whether Unitarians live longer, too. ) We're not suggesting that it is prayer, just pointing out that it's odd that epidemiologists are looking at the association of religion with health when religion has everything in common with any other group behavior, except this one thing, which isn't being considered. Why privilege religion, then?  It just throws confounding variables into the mix, already a potential problem in even the best of studies.

Though, at least one recent study did look at the power of prayer to heal the sick. What the investigators found was that if you knew people were praying for you, it improved your recovery. But if you didn't know that, you didn't recover more quickly. So this suggests that it's the power of knowledge of prayer rather than divine intervention because of prayer. That's consistent with the psychological effects of believing you have something going for you -- the placebo effect, if you will.

And, anyway, if the vast majority of the American people consider themselves to be religious, as shown in poll after poll, who are all these religious people living longer than? Why is our life expectancy embarrassingly lower than that of much more secular societies such as in Japan or Europe? Would it be enough for us to move to Europe, take advantage of their civilized national health care systems (where society really is trying to make you better), and forget Mass?