Monthly Archives: June 2012

Is stress a risk factor for dementia?

Gosh, this is a hard one to study. I have to hand it to the people at Southampton University who have just announced this project funded by the Alzheimer’s Society. It is clearly a massively important topic. To look retrospectively at stress is challenging because we all remember past events in a rather filtered way, and if the participants’ memories are impaired, doubly so. To look at it prospectively involves very long time periods and large cohorts (although as the Whitehall II cohort ages, this becomes a possible source of longitudinal data with a specific focus on both biological and experiential stress markers).

What interests me most of all is confounding – the close relationship between stress, lifestyle and socio-economic (dis)advantage. How can we really tease them apart to get the true effect that stress has, independent of the others? There are tools in the box, such as multiple regression and propensity score analysis, but they depend on having quantitative measures of stress and all the other factors that we believe really represent the underlying factor for each individual.

How will this measure against Austin Bradford Hill’s considerations for finding cause-and-effect?

  1. Strength of effect – it may not be terribly strong, especially if there are stronger determinants such as smoking mixed up with it
  2. Consistency across studies – this is the first study, others will be needed to try to confirm it.
  3. Specificity – this is a problem because we are far from lab conditions, and looking at the many varied sources of stress, how people deal with it, and what happens to their cognitive function, although at a molecular biology level it might be possible to be quite specific
  4. Temporality – the cause has to happen before the effect; this one at least should be OK
  5. Dose-repsonse – if we believe the markers of stress provide numbers at least in the right ballpark, then this could be one of the big arguments in favour of a link
  6. A plausible mechanism of action, and…
  7. …coherence between lab and field data – there have been animal studies so this one is a possibility.
  8. Experiment – I don’t think that one is getting past the ethics committee any time soon!
  9. Analogy with other risk factors – if the stress effect is hormonal then people with pathological hormone imbalances might be a useful population, if it is an effect of neuron-environment interaction (harmful plasticity) then it is going to be much harder to tease apart from the confounders.

So, a mixed bag, but certainly worth trying out. I wish them luck, but I’m glad I’m not trying to crunch the numbers for them.

Leave a comment

Filed under Uncategorized

Accessing national clinical audit & outcomes data

A process for researchers to access data from the National Clinical Audit and Patient Outcomes programme (NCAPOP) has just been published. This is a collection of 30 projects running in England (and sometimes other parts of the UK too) collecting data that is potentially useful in many areas of health care. You can see a PDF list of the projects here.

Leave a comment

Filed under Uncategorized

The spectrogrammer sleeps tonight

Nice visualisation of the world going to sleep and waking up again, courtesy of some clever people at Twitter. The look reminded me of spectrograms, which I’m sure would have some application somewhere in the world of multivariate stats or time series… Spotted (as so many good things are) at flowingdata.com. Well done to Nathan Yau for spotting Ramadan which nonplussed the Twitter guys. It also looks to me like there are late nights every weekend, and a spike in NYC for Thankgiving and one in both NYC & Sao Paulo for Easter. Or maybe I’m reading too much into it.

Image

Leave a comment

Filed under Uncategorized

Bristol event history analysis course

Bristol’s Centre for Multilevel Modelling are running this course in September 2012:

Discrete-time Event history analysis, 10-11 September 2012, University of Bristol

To apply, please go to Workshops on www.bristol.ac.uk/cmm and follow the link to the application form.

This workshop will introduce methods for the analysis of event history data (also known as survival or duration data) with a focus on discrete-time methods. Advanced topics such as modelling recurrent events, transitions between multiple states and correlated event histories will also be discussed. Examples of potential applications of these more advanced methods are studies of the timing of births and duration of unemployment spells (recurrent events), transitions between different labour market states, and the interrelationship between female labour force participation and childbearing (correlated histories).

Leave a comment

Filed under learning, noticeboard

ESDS events in June and July

The UK Economic and Social Data Service are running some useful workshops soon. Here’s a quote from the e-mail:

Family Resources Survey user meeting

Friday 22 June 2012

Royal Statistical Society, London

This meeting will provide a forum for the exchange of information and views between users and producers of the Family Resources Survey. It is aimed primarily at users and potential users of survey microdata.

 The meeting is free to attend and lunch is provided. To view the programme and book a place please go to  http://www.ccsr.ac.uk/esds/events/2012-06-22/

———————————————————————————

An Introduction to the International Passenger Survey

Friday 29 June 2012

Basement Computer Lab, Humanities Bridgeford Street Building, University of Manchester

 A one-day workshop to introduce the International Passenger Survey to those with no or little knowledge of the IPS.

 The workshop is free to attend and lunch will be provided. To view the programme and book a place please go to http://www.ccsr.ac.uk/esds/events/2012-06-29/

 ———————————————————————————

 Health surveys user meeting

 Tuesday 10 July 2012

Royal Statistical Society, London

 This meeting will provide a forum for the exchange of information and views between users and producers of the UK health surveys. It is aimed primarily at users and potential users of survey microdata. The programme contains a mixture of papers from data producers and researchers and a poster session at lunchtime.

 The meeting is free to attend and lunch is provided. To view the programme and book a place please go to  http://www.ccsr.ac.uk/esds/events/2012-07-10/

Leave a comment

Filed under Uncategorized

Football and domestic violence

An interesting article at the BBC website today claims that their own research has shown a link between England winning or losing football matches and the incidence of domestic violence. A closer look reveals the argument is really a lot shakier than that. We have numbers of incidents reported to police on the day of the match, compared to the same day a year ago, and alas! because England tend not to stay in these tournaments very long, there are only four matches to draw on. So that is four pairs of counts. The numbers are not given but one can work them out:

2510 (draw), 1890 (draw), 2427 (win), 3221 (lose) for the four match days and 2556, 1875, 1911, 2497 respectively for the previous year.

We could complain about the same date being a weekday one year and a weekend the other but the real problem is there’s not enough data. We need to know how much these incident counts fluctuate naturally from one day to another, and we can’t really tell from just four non-match days. You can see that there is an overlap between the match and non-match days, so common sense says this is not going to prove anything. Indeed if you wanted to prove that win/lose results were different to draws, then you’d only have two pairs in each camp.

Nevertheless you could go ahead and fit a generalized linear model (Poisson regression) to this, with year and result as predictors (if you try the interaction between year and result you will have run out of data, that’s how close to the edge this exercise is). This says p<0.0001 for an increased incidence with a win or lose result, which presumably is what Prof Brimicombe means when he says it is a “definitive and significant increase“. But Poisson regression requires a pretty specific formula for the natural fluctuation from one day to another (variance), and we can’t really test that with just four numbers. An alternative analysis is negative binomial regression, which does not have the stringent assumptions of the Poisson, but needs more data to be able to estimate the variance. Which of these alternatives best fits the data? We can’t tell because the negative binomial has too few data to converge to an answer for us.

Let common sense prevail. This is an important topic and an interesting analysis, but looking at four days is not going to reveal any causal link. It should be a springboard to applying for funding to do the research properly, not straight to the headlines.

R code:

library(MASS)
library(lmtest)
date<-factor(rep(1:4,2))
ipv<-c(2510,1890,2427,3221,2556,1875,1911,2497)
draw<-c(1,1,0,0,0,0,0,0)
result<-c(0,0,1,1,0,0,0,0)
preg<-glm(ipv~date+draw+result,family=poisson)
nbreg<-glm.nb(ipv~date+draw+result)
summary(preg)
summary(nbreg)
lrtest(preg,nbreg)

Leave a comment

Filed under Uncategorized

Massive Health

Wow, this is a very nice project. Collect data on what people are eating and how healthy they think it is through their smartphones, turn it into an animated visualisation. The green marvellous muesli of morning gradually gives way to the red naughty nachos of night time. I’ve just had a coconut macaroon and a cappuccino so I had better not sound too sanctimonious.

Image

Spotted via flowingdata.com

Leave a comment

Filed under Uncategorized