Stuff that occurs to me

All of my 'how to' posts are tagged here. The most popular posts are about blocking and private accounts on Twitter, also the science communication jobs list. None of the science or medical information I might post to this blog should be taken as medical advice (I'm not medically trained).

Think of this blog as a sort of nursery for my half-baked ideas hence 'stuff that occurs to me'.

Contact: @JoBrodie Email: jo DOT brodie AT gmail DOT com

Science in London: The 2017 scientific society talks in London blog post

Monday, 26 April 2010

I'm still a bit rubbish at stats

I seem to suffer from a bit of a mental block with statistics, despite being a fan of Ben Goldacre, reading around the subject and attending classes on it. For some reason even though I can grasp it when I am in the middle of learning it, the knowledge seeps out at an alarming rate. I more or less have to start from scratch whenever I meet a new article.

Hopeless. But let's remember all the other things I'm good at :) Peculiarly I managed to get a first in statistics at university (though this was a modular exam at first year level and open book at that) but have really let the side down since then.

Abstracts don't help, writing things in a very paragraphy sort of way when what I really need is to see things written out line by line...

Effect of Valsartan on the Incidence of Diabetes and Cardiovascular Events
http://content.nejm.org/cgi/content/abstract/362/16/1477

Results The cumulative incidence of diabetes was 33.1% in the valsartan group, as compared with 36.8% in the placebo group (hazard ratio in the valsartan group, 0.86; 95% confidence interval [CI], 0.80 to 0.92; P (less than) 0.001).> placebo, did not significantly reduce the incidence of either the extended cardiovascular outcome (14.5% vs. 14.8%; hazard ratio, 0.96; 95% CI, 0.86 to 1.07; P=0.43) or the core cardiovascular outcome (8.1% vs. 8.1%; hazard ratio, 0.99; 95% CI, 0.86 to 1.14; P=0.85).

But what I need is something more like this:

Results
The cumulative incidence of diabetes was 33.1% in the valsartan group,
as compared with 36.8% in the placebo group
(hazard ratio in the valsartan group, 0.86;
95% confidence interval [CI], 0.80 to 0.92;
P<0.001). p="0.43)" p="0.85).

Conclusions Among patients with impaired glucose tolerance and cardiovascular disease or risk factors, the use of valsartan for 5 years, along with lifestyle modification, led to a relative reduction of 14% in the incidence of diabetes but did not reduce the rate of cardiovascular events. (ClinicalTrials.gov number, NCT00097786 [ClinicalTrials.gov].)

Actually that still doesn't help very much, I'm still a bit at sea. At this point I try applying mathematical functions to some of the available figures (basically dividing one number by another) to see if I get anywhere. I'm vaguely cheered to have spotted that 0.96 might be related to 14%...

If it's any help, I'm useless at reading maps too though I am very good at writing instructions - and delighted to hear that Mark Miodownik's lovely programme "How to write an instruction manual" has been uploaded to the @Speechification vault.

Is there an online journal club for remedial stats?

1 comment:

  1. I do of course mean that 0.86 might be related to 14%.

    But it might as well have been 0.96 for all the difference it will make to my competence in this particular area. Shame it's so bloomin' important...

    ReplyDelete

Comment policy: I enthusiastically welcome corrections and I entertain polite disagreement ;) Because of the nature of this blog it attracts a LOT - 5 a day at the moment - of spam comments (I write about spam practices,misleading marketing and unevidenced quackery) and so I'm more likely to post a pasted version of your comment, removing any hyperlinks.

Comments written in ALL CAPS LOCK will be deleted and I won't publish any pro-homeopathy comments, that ship has sailed I'm afraid (it's nonsense).