Sunday, February 17, 2013

In Which Tam Does My Work For Me

Tam is a grad student, which means she has a lot more time than I do.  I mean, I've got a book of several hundred sudoku puzzles to do and a slew of computer games I am working on and this presses on me.  (This leisure thing takes up a lot of time if you are really committed to it!)  So this week I've outsourced finding blog content to Tam (without her knowledge, but she totally came through).  So I present:

1) An awesome advice column "How to tighten up your game at work when you're depressed" from a site that looks worth spending a lot more time on. 

I think that a lot of people (esp. young ones, new to office work life) don't get it about how much appearances matter and how observing professionalism can be beneficial, even if you're not suffering from a serious mood disorder. 

I also found that this really resonated with me: "Jerkbrain Lie: Feeling depressed (lazy, horrible, avoidant) means that you can’t (get to work on time, complete work tasks, do writing that you need or want to do, do housework)."  For me, it's usually feeling anxious, stressed, sick, or reactant (just hating/resenting/etc. whatever idiotic thing my boss/advisor/professor/society/the angel of my better nature insists that I do) that presents itself as the big barrier, but it can function the same way.  I have found that controlling my behavior is a lot easier than controlling my feelings, so I make use of the "I don't have to like it, I only have to do it" mantra to pretty good effect.  I think it's important that people give themselves permission to not feel pumped and enthusiastic about something they are doing, and to not really enjoy the process at all.  And then afterwards, to feel like, "I did that and I didn't even want to" and maybe even "Thus I am a rockstar."  Because sometimes, just doing basic life tasks is rockstar material.  Today I am going to be a complete fucking rockstar and put some miles in on the treadmill even though I am feeling physically blah (because I know that moderate exercise is very helpful for improving my symptoms even though all I want to do is lounge about with my heating pad and a lot of dark chocolate).  I might not like doing it, but I don't have to.  

2) A blog post describing some newly published research supporting the “butter is better than high omega-6 veg/seed oils” argument.  The kicker: patients who were put on the safflower oil diet did have lower total and LDL cholesterol than those who kept their same diet, but they were more likely to die!  Oops!  

From the paper:  “As expected, increasing n-6 LA from safflower oil in the SDHS [Sydney Diet Heart Study] significantly reduced total cholesterol; however, these reductions were not associated with mortality outcomes (results not shown). Moreover, the increased risk of death in the intervention group presented fairly rapidly and persisted throughout the trial. These observations, combined with recent progress in the field of fatty acid metabolism, point to a mechanism of cardiovascular disease pathogenesis independent of our traditional understanding of cholesterol lowering.” 

(Plus: "All authors have completed the ICMJE uniform disclosure form ... and declare: support from the Life Insurance Medical Research Fund of Australia and New Zealand and the Intramural Program of the National Institute on Alcohol Abuse and Alcoholism for the submitted work; no financial relationships with any organizations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.")

So yeah, chasing a target cholesterol level with the assumption that it will keep you healthy/alive is not the smartest play here, which I think my readers already know even if their doctors do not.  Perhaps eventually medical practice and public policy will catch up with the science.

9 comments:

Tam said...

This division of labor actually works really well for me. I have time and energy to mindlessly surf the web and forward interesting links, but summarizing them in such a way as to inform an audience and possibly sell them on the ideas is really beyond my capabilities, energy-level-wise.

Actually, grad school or no, I think I enjoy reading things and forming opinions about them way more than I enjoy trying to convey those ideas or opinions to others.

So if this were a radio show, I guess I'd be kind of like your producer.

Tam said...

On a more substantive note, the thing about cholesterol reminds me of the weight issue, where I think being obese (if not moderately overweight) is associated with increased risk for various health problems, but where I'm not sure there's good evidence that trying to lose weight actually improves health or longevity long-term.

It reminds me a little bit of the general political question (which often breaks down along right/left lines) of whether, given a problem, trying to solve it is a good idea, given that attempted solutions often have perverse effects.

I don't want to advocate a purely fatalistic world-view, but I think sometimes a bit of cheerful resignation can come in handy.

Sally said...

And I like writing about ideas more than doing my own research (even if that research only involves surfing the web, apparently) so it works for me too.

I love the idea of "cheerful resignation." That deserves more thought and development.

Sally said...

Tam, I was reminded of the weight issue as well when reading the cholesterol stuff, too.

One thing I didn't glean from the article was whether people started with high cholesterol levels and where the people in the two treatment conditions ended up. Given the sample, I assume they mostly had elevated levels. I think it's interesting to consider these cholesterol results in a general way -- it's as if this article were saying: Even though there is an intervention that reduces your metric of interest (in this case, cholesterol), that intervention is STILL more likely to kill you ... probably because your metric of interest isn't really a very reliable indicator of anything. Will we eventually find that the same is true of body weight? I don't know. But it's kind of a non-starter if we cannot find an intervention that effectively reduces body weight in the long term.

Sally said...

PS I really liked how the paper presented some nice empirical results (from an experiment! with random assignment!) and also discussed a mechanism by which the perhaps perplexing results make sense. I have not read a thousand nutrition science (etc.) articles to make any good generalizations, but this article has a better integration of behavioral findings and more fundamental biochemical theory and lab findings than I am used to reading. A lot of the behavioral work (e.g., epidemiological studies that follow a big sample of people and look for correlations among food intake and health outcomes) fail to find significant results that support their hypotheses and don't really get into the why. Of course, there are a lot of reasons to find null results (esp. in correlational studies!) but it's nice to see an actual counter theory presented rather than what amounts to "we're sure that the medical establishment is right about this but we just need to work harder to get the results."

Debbie said...

A friend of mine diagnosed with depression, who nevertheless gets to work on time, gets things done, etc., once wrote a very effective rant against people whining about not feeling motivated or not being in the mood to do things.

We are lucky that we actually sometimes are in the mood to do things, but when I find myself putting things off too long for this sort of reason, I can remind myself of him.

Strings of imperfect correlations are problematic: Death rates are correlated to cholesterol level (but not perfectly), cholesterol level is correlated to diet (but not perfectly), so death rates might or might not be correlated to diet in the same way. It's tempting to attack things from this angle because it's so much easier to test cholesterol (and weight) than death rates, but sometimes that strategy backfires.

Mmm, butter.

On your PS, I didn't read the article, but YES. Yes to experiments, yes to random assignments, and yes to accepting the results you get. I even had a zoologist I did typing for where I had to say, "Sorry, I'm not typing the phrase 'not statistically significant, but...'" Actually, I think he made me type it anyway (he wasn't willing to give up on the idea that differences too small to be seen with the given sample sizes might exist), but he did have to waste several minutes dealing with me.

Sally said...

Debbie -- Yeah, null hypothesis significance testing can kind of be a bitch, can't it? I personally agree that "surely God loves the .06 nearly as much as the .05" so I can understand the zoologist's frustration with the fixation on .05 in many of the sciences (though currently in the social sciences, you do sometimes see .10 reported as "marginally significant," though usually only in papers that have p<.05 results too). And if you work in a field that routinely underpowers studies (psych is BAD about this, for example), you're kind of screwing yourself over on your p values. At the same time, it can be overly tempting to use the arbitrariness of .05 and the somewhat shaky foundations of the whole NHST approach to justify to yourself making a big deal out of non-significant results, like your zoologist might have done.

Debbie said...

I've done very little science, but mostly I decided that if the differences were so small they didn't show up given my sample size, then they weren't that exciting anyway.

But the whole point (in my opinion) is that you pick a question where you don't already know the answer (which also seems sadly rare), and then you design the experiment so that you can accept the answer either way.

Why yes, it did take me a long time to come up research topics.

Sally said...

Debbie -- Obviously there isn't time for that kind of thing! You must be always collecting data, even if it's crap! If you run 20 studies, one of them should turn out with a statistically significant result....