The Clay Feet of Research

Looking back on it now, I guess what I really learned in graduate school was that you can’t trust research. You really can’t. There’s so much variation in methods and methodology — and so little true objectivity — that essentially we can and do interpret data to mean anything we’d like it to mean.

This has lots to do with relying on correlations, i.e. the patients who had the most X factor also developed the most Y condition. But this does not prove that X causes Y. No way, baby. Remember that historical study showing that the regions of the world with the most storks also had the most births? Yeah, right.

Being an inquisitive person with diabetes following the latest research, I try not to get too disillusioned. I know the scientists are making progress. They are. With all the efforts, and all that funding, they have to be!! But for every study stating milk products cause diabetes, there’s another illustrating the health benefits of dairy consumption for diabetics. Wha?!

And now this. Researchers at UCLA recently discovered that tight BG control isn’t necessarilyFrustrated beneficial! Diabetics they studied with poor glucose control actually had a 35 percent lower mortality rate than patients with A1C levels no greater than 7. Wha?!

OK, maybe it’s a fluke, but awfully disconcerting following the latest DCCT results showing that tight glucose control supposedly makes a huge difference in long-term health.

I guess you just can’t trust research. Geez. And as if it weren’t dicey enough, some people have to go and just outright cheat. They’re saying Dr. Hwang’s indiscretions in faking stem cell research will “set the already embattled field back indefinitely.” I’ll say.


5 Responses

  1. B.A. Smith
    B.A. Smith January 18, 2006 at 10:03 am | | Reply

    Your Feet of Clay post is dead on. The med-media is so desperate for copy that strange “research results” from UCLA that you mentioned get all-too-prominent placement. Correlation doesn’t equal cause, as you pointed out. And, with a sample of 123 diabetics and no statement about the actual experimental design of the study – nor how strong (accurate) the results were – we faithful followers of the DCCT imprimatur shouldn’t loose any sleep. And certainly not change our control strategy. Nevertheless, I get upset whenever I see “research results” peddled to inform us common folk when all it’s really doing is confusing us with potentially biased, partial and wrong information. After all, I’m left-handed and, we now know from supposed “research results”, that my lifetime is going to be shorter than you right-handers, BECAUSE I’m left-handed. Say what? So, when we diabetics happen to see such “research results” played up in the press, we should remember to take these results with a VERY large “grain of salt.” This is because simple, bi-variate correlations, that purport to link one variable value (in this case, A1c numbers – which themselves are subject to error and variation) to something as hugely complex as an individual’s death rate, cannot possibly provide meaningful insight about causal relationships. The UCLA “researchers” should be sent back to beginning bio-statistics.
    So, when I see “research results” like the UCLA study, I chalk them up to academia’s too strong publish-or-perish bias, rather than anything consequential.

  2. Kassie
    Kassie January 18, 2006 at 10:24 am | | Reply

    The UCLA headline caught my eye earlier today. Of course, if you actually read the blurb, you learn that they are talking about heart failure patients. My issue (on top of those you raise) is that the catchy headlines are so misleading.

  3. Bill the diabetesdoc
    Bill the diabetesdoc January 18, 2006 at 5:38 pm | | Reply

    The article about inverse relationship between A1c and outcomes is by cardiologists who don’t know a darned thing about diabetes — they never looked at glycemic control, and only looked at a single A1c value in their SAS (sick as sh**) patients.

    You can relax, the DCCT (1400 patients followed over 10 years by multiple endocrinologists) is a better study than 123 patients with nasty heart disease followed for one A1c by cardiologists.

    BTW, the entire study is on-line at AHJ’s website:
    An unexpected inverse relationship between HbA1c levels and mortality in patients with diabetes and advanced systolic heart failure

  4. Anonymous
    Anonymous January 19, 2006 at 12:56 pm | | Reply

  5. AmyT
    AmyT January 19, 2006 at 1:24 pm | | Reply

    Thanks for the input, guys. And yes, I’m aware of the latest “cell transplant” news — taken again with a grain of salt.

Leave a Reply