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 2018/19 scientific society talks in London blog post

Friday, 25 April 2014

I'm almost embarrassed by how much this made me giggle: Bus stopping 'ping'

A couple of years ago while on a bus I noticed that the bus's 'Stopping' indicator sign, which lights up and pings whenever someone presses the 'let me off' button, had managed to lose its first four letters. This meant that the indicator just said '' while actually making a pinging sound. I have to admit that I was overcome with childish giggling every time it was pressed and had to escort myself off the bus, before I finally managed a journey without laughing for just long enough to record a brief video. Here it is, just under three seconds long.

I wonder if this is a case of determinative nominism ;)

Wednesday, 23 April 2014

April 2014, finding hashtagged tweets from November 2013 - a case study

Earlier today I spotted (I have a saved search on Twitter for my own blog) that someone (@PH_AdvocateEU) had been trying to uncover old tweets from a conference, using information in an even older blog post of mine which is now ridiculously out of date.

Since I like trying to find old tweets and haven't done it for a while I got in touch and asked if I could use it as a case study, to see what could be found. They agreed and I had a look - not much through usual methods but a 'handsearch' (combing through participants' tweets) looks the most promising, but it is the most labour-intensive unfortunately.

The hashtag in question is #EUCOPDBrx13

All the tweets that I've collected (there are more out there, but fiddly to uncover) are in this Storify:
#EUCOPDBrx13 - a collection of conference tweets

1. Has anyone already done the work for you - not in this case
It's possible that someone else has collected the tweets together and embedded them into a blog post or Storify etc, so search for the tag there first (you can use Google's Blog Search tool - I didn't find anything on either site in this particular case though.

Sometimes people add a widget to their website which feeds in hashtagged tweets and stuff can be picked up from them (not in this case though it seems).

Note that there may be paid-for services that do this, but I'm afraid I don't know about them. If you have a server and know what an API is you may be able to call them up from Twitter's servers.

2. - a few
Twitter's own search box is a really good place to start, I've recently been surprised and delighted to find really old tweets on it - note though that I'm using the web browser version, not a smartphone or tablet app.

When you first search anything on Twitter you're shown the 'top tweets', you need to click on 'All' to see what else is available... and then you need to scroll down to the end (and keep scrolling until you're met with a note saying 'Back to top' which is Twitter's way of telling you that all available tweets are loaded).

You can see them here (although as time passes fewer of them will appear presumably)

Doing this brought up 21 tweets - one was @PH_AdvocateEU telling me which hashtag to look for and seven of them were from 'bots' (automated accounts) reporting that the tag was (at the time of the tweet being sent) trending, which certainly suggests that there should be quite a lot of tweets.

Below was the earliest tweet I could find (it's a picture of the tweet with its address beneath it because embedding tweets doesn't seem to work very well on this blog. Subsequent tweets are embedded and show up as plain text).

If you are using Storify or you don't need to use the Embed Code to add a tweet, just its URL. Here are the URLs of the tweets above. If you hover over the datestamp (or timestamp) of the tweet you can then right click / copy link location to get the URL. (this is the one shown as a picture above)

I may have missed one or two!

3. Search Google - a couple
An ordinary search on Google can also bring up tweets. Don't forget to investigate cached copies of things - if you see a tiny green arrow to the right of a site's address click on that and choose cache (this is an archived copy of the page).

I found a couple that didn't show up when searching directly on Twitter

4. Targeted search restricted to names - not much, not tried all!
We know who was tweeting about the conference

We know from the content of their tweets who else was there, or participating
@imi_Ju, @karinkadenbach, @IPCRG, @yankeeu, @EU_H2020, @mikegalsworthy, @mikakosinska, @NABedlington, @humedsci

Although it doesn't bring up much in this case it's worth adding these names to the hashtag (one by one alas, this is fairly laborious I'm afraid) to see if you can find any other tweets that they sent or were mentioned in.

To find tweets they sent
from:imi_Ju #EUCOPDBrx13

To find tweets mentioning them
@imi_Ju #EUCOPDBrx13

I'd have to concede that this hasn't been particularly successful in this case though!

4. Manually scroll back to the relevant point of people's timelines - very promising, but hard work
Twitter will let you scroll back 3,200 tweets' worth on anyone's page. If they're very chatty this might not help if you're looking five or six months later of course.

I tried this for @EU_COPD, and found this - they've published fewer than 500 tweets so scrolling through wasn't too onerous. Each of these could be captured in a storify by collecting the URL (where it says 28 Nov, in grey - that's the timestamp that has the unique address for the tweet).

I'm not going to collect all of the URLs but a couple are (the last two in this pic)

5. Instagram and other tools - haven't checked
People take pictures, they tag them - have a look on Instagram for the same tag and see if there's any info there (you can include these in Storify stories I think).

Wednesday, 16 April 2014

Is there a "metadata protection act"? How do we think about metadata

by @JoBrodie

Recently I have spotted some tweets about an alternative health magazine (What Doctors Don't Tell You) which is offering its readers a taped recording of discussions with various people about cancer treatments. The implication in the advert is that the magazine can't say too much about the contents of the tape in the text of the offer because doing so would bring the Cancer Act 1939 upon them.

They're probably right, in two senses. Offering advice, to the public, about cancer treatment falls within the Act but the text as seen in this bit of marketing does not itself offer any advice about treatment, so is fine. I've no idea if the recording itself would breach the Act (regardless of how it's acquired by members of the public) or if it's problematic because receiving it is contingent on taking out a subscription... or in fact not at all (I'm not a lawyer).

This made me wonder: what's the deal with not saying something but just linking to it or inferring it? Does that count? No idea.

It reminded me of a small collection I've been making that relates to jigsaw ID and how all sorts of things can be inferred from metadata. Efforts can be made to conceal your data but the metadata about your can be leaky.

The jigsaw ID and metadata stuff is quite separate though from an advert about cancer information - this is about me thinking through some thoughts about the way in which we consider indirect information. But I suppose both would be covered by my notion of a 'metadata protection act' in which everyone had to be super careful about how they point to things.

Convinced that others might have thought this exact same thought I searched on Twitter for the jokey phrase "metadata protection act" but found nothing on Twitter and only two hits on Google. So I've sort of stumbled upon a sort of Googlewhack if nothing else ;)

Newspapers have guidelines on reporting abuse cases, to avoid indirectly implying the victim without naming them directly. Do we have anything else for metadata? If we did, I wonder what difference it would make to this advert. 

Friday, 11 April 2014

I'm seeing a lot of people asking this: "Do you know about the Cancer Act of 1939?"

This question seems to have cropped up rather a lot in recent weeks*. If you type "cancer act" 1939 into Twitter you'll see a stream of tweets about it - most of them seem to be pointing to only one or two forum posts, and they all seem to say the same thing. There seem to be a few convergent conspiracy theories about it.

I'm intrigued as to why there's a sudden (apparent) interest in the Act.

Briefly, it's an Act of Parliament that makes it an offence to offer to treat someone for cancer or give advice about treatment...

"4 Prohibition of certain advertisements.

(1)No person shall take any part in the publication of any advertisement—

(a) containing an offer to treat any person for cancer, or to prescribe any remedy therefor, or to give any advice in connection with the treatment thereof;"
...unless you are making the information available for healthcare practitioners.

There haven't been very many prosecutions under the Act, though things probably don't get that far as most people will remove misleading claims after discussions with Trading Standards. A few websites have closed down, events have tried to move venues (doesn't really work, still illegal) or speakers have been removed from the programme - it all seems to be quite low level stuff really.

Incidentally I've known of the Act since at least 2010, it's fairly well-known among skeptic bloggers I think.

*to be fair I've not really been looking for it before now so I don't know if it's always been this much talked about, or if this is a real and recent increase.

Additional comments policy - note that any approved comments are very likely to be published as plain text with no website hyperlinks. This is both an anti-spam and anti-conspiracy-theory strategy.

Tuesday, 8 April 2014

Complaining via the ASA (Advertising Standards Authority) or European ASA about a non-UK advert

by @JoBrodie

You can complain about adverts appearing in other countries (eg on non-UK websites).

The UK's Advertising Standards Authority is a member of the European Advertising Standards Alliance (EASA) along with 23 other European countries. If an advert appears in another country you can report it to the ASA and they'll pass it on to the EASA while liaising with you, thanks to their cross-border complaints arrangement, or you can also complain directly through the EASA.

I discovered this by accident when complaining about an advert for a diabetes clinic that appeared on Facebook - even though it was based in another country I reported it as it was clearly targeting a UK audience (and making misleading claims). I wasn't sure if there was much that could be done, but there was. The ASA told me they were passing it on to the regulator in that country and they kept me informed throughout - and my complaint was upheld.

The 23 European countries other than the UK are:

Austria, Belgium, Bulgaria, Czech Republic, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Lithuania, Luxembourg, The Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, Switzerland and Turkey.

There are also 7 non-European countries that have regulatory links:

Australia, Brazil, Canada, Chile, India, New Zealand and South Africa.

Some examples of adjudications appearing on the ASA's website where the trading address isn't in the UK.

South Africa
Consumer code | Complaint form
If advertisers don't play fair then the ASASA will take out an ad-alert.

Sunday, 6 April 2014

Small post for Sophia about genetic testing for diabetes

For Sophia - I realised this was going to take up more tweets than those following both of us might have patience for.

For everyone else - I was a diabetes science information officer for eight years until June 2012 and as such there may be newer information that I'm not aware of, if you spot any errors in my reasoning below do let me know, ta :)

General 'stuff' about genetic testing for diabetes, saving money for health service

1. What is the benefit to the individual or to the state for sequencing genome with respect to diabetes risk?
I'm not aware of researchers into diabetes genetics having called for widespread genetic testing post-birth (or at any other time), even in people at risk.

I might be wrong but I don't think genetic testing (we're really talking about Type 2 diabetes here) tells you much that you don't already know. Age, family history, weight and activity levels are possibly a better predictor of T2D. Family history obviously implies genetics, but it seems to be a pretty multifactorial sort of thing.

As the information below (which comes from Diabetes in the UK 2012) suggests, most cases of Type 1 diabetes don't seem to have an obvious family link, though there are genes that increase risk of developing the condition.
Type 1 diabetes
Although more than 85% of Type 1 diabetes occurs in individuals with no previous first degree family history, the risk among first degree relatives is about 15 times higher than in the general population.

On average: if a mother has the condition, the risk of developing it is about 2–4 per cent
if a father has the condition, the risk of developing it is about 6–9per cent
if both parents have the condition, the risk of developing it is up to 30 per cent
if a brother or sister develops the condition, the risk of developing it is 10 per cent
(rising to 10–19per cent for a non-identical twin and 30–70 per cent for an identical twin).

Type 2 diabetes
There is a complex interplay of genetic and environmental factors in Type 2 diabetes. It tends to cluster in families. People with diabetes in the family are two to six times more likely to have diabetes than people without diabetes in the family.
There doesn't seem to be very much that someone can do to prevent getting Type 1 diabetes, whereas health interventions can help reduce the risk of Type 2 (and prevent progression from "prediabetes"to full-blown Type 2 diabetes). However the lifestyle advice given to someone at risk of developing T2 is pretty much the same as would be given to anyone: eat a variety of foods, not too much, maintain a healthy weight, do a bit of exercise but doesn't need to be marathon-running to help.

Perhaps giving someone information in black and white from a genetic test might make them more likely to follow this healthy advice (but is this coercion or compliance!) however the tests themselves, even in high-risk individuals, don't appear to be all that sensitive or of use in clinical practice.

Genetic Screening for the Risk of Type 2 Diabetes: Worthless or valuable? Diabetes Care 2013
"Genetic testing for the prediction of type 2 diabetes in high risk individuals is currently of little value in clinical practice.
The limitations of genetic risk models are small effect size of genetic loci, low discriminative ability of the genetic test, small added value of genetic information compared with the clinical risk factors, questionable clinical relevance of some genetic variants in disease prediction, and the lack of appropriate models for studies of gene-gene and gene-environment interactions in the risk prediction. 

For improvement of the genetic risk models in the future, the definition of type 2 diabetes and classification of subtypes of diabetes should be more precise, new sequencing techniques should be applied to identify low-frequency and rare variants having a large effect size, non–European ancestry populations should be investigated to identify new variants relevant to type 2 diabetes prediction, studies of structural variation and epigenetics should be performed to identify new variants relevant to type 2 diabetes prediction, and modern statistical methods should be developed and applied in studies of gene-gene and gene-environment interaction in large populations." - emphasis added.
I don't think there's much benefit in genetic testing either to a person at risk of diabetes, or the state at this stage.

2. What is the benefit to the individual or to the state in diagnosing diabetes early?
With undiagnosed Type 1 diabetes things can go wrong very quickly as with no insulin available and rising glucose levels the person can go into diabetic ketoacidosis (DKA), a metabolic emergency that can and does result in death. So clearly a benefit to the individual there! Once T1D is diagnosed if the person doesn't take sufficient insulin this problem can re-occur.

Diabetes UK had a campaign to get parents more aware of the symptoms of Type 1 diabetes - the 4Ts: toilet, tired, thirsty, thinner (going to the loo more, drinking more water to compensate and losing weight along with being tired as the body's not getting the glucose fuel it needs).

With undiagnosed Type 2 diabetes the person may creak on fairly happily, and fairly unaware that there's a problem. Here insulin is still produced by the pancreas and the main problem is that the body (organs, muscles) become less sensitive to it). Any symptoms are generally put down to 'getting older' and that's why symptoms are pretty useless for Type 2 (good for Type 1), and 'risk factors' are much more important - as in T1 a blood test is diagnostic.

About half of people who are diagnosed with Type 2 diabetes will already have some signs of diabetic complications, so early diagnosis can help to prevent these from worsening (it's controversial but not unfeasible that some complications may be reversible if caught early). Plus people tend to feel a lot better once their glucose levels are brought back to healthier levels (also it's not just glucose, diabetes is a cardiovascular condition and blood pressure and blood fats / lipids are also problematic and need to be monitored).

There are other rarer forms of diabetes and some of these can be linked strongly to a particular gene - monogenic forms of diabetes can be probed with genetic testing however the person is already likely to have been diagnosed with 'diabetes' (probably Type 2*) and this testing really just refines the diagnosis rather than spots the existence of the diabetes. The advantage of a correct diagnosis is that the person gets the right treatment (in some cases this can be changing from insulin to tablets).

From the state / money point of view - diabetes complications cost a lot of money and diabetes medicine costs a lot too. Hospital admissions and cardiovascular treatments are probably the big costs and likely to increase with an aging population. Delaying or preventing complications saves money, though offset by more people taking medication. Insulin's main side-effect arises because it's incredibly difficult to get the dose right - if too much is taken then blood glucose levels can plummet and the person may need medical treatment. If too little is taken then the person may experience the metabolic emergency DKA mentioned above, which requires hospital treatment.

See also 
Article: Health Economics
Estimating the current and future costs of Type 1 and Type 2 diabetes in the UK, including direct health costs and indirect societal and productivity costs Diabetic Medicine 2012

I think a more feasible thing is perhaps to forgo predictive testing at this stage (maybe it'll be more useful in the future) but improve prompt rather than early (pre) diagnosis of any kind of diabetes so that the person can maintain good health for longer. 

*After I tweeted this post and asked for corrections @sparklyredshoes pointed out that people ultimately diagnosed with a monogenic form of diabetes are possibly more likely to have been initially misdiagnosed with Type 1 (not Type 2 as I've said above) as they're often 'young and skinny'. However there aren't really enough stats on this so she suggested that it was probably safer for me not to commit to which type of diabetes people are commonly misdiagnosed with :)

Saturday, 5 April 2014

Optical effects: my "rolling shutter" camera half-captured a flash going off at science event, delighting me

To be honest I'd have thought I'd have seen more photos like this. I understand that the chances of taking a shot like this again are low but given the gazillions of photos taken it seems like someone else would snap something similar, but searching on Google for 'mid-flash' or 'midflash' doesn't bring anything up - maybe I need better search terms.

This photo was taken by accident by me on 16 July 2010. I was at the Royal Institution to celebrate the SciCast awards and generally have a lovely science communication-y time, which I did. SciCast was celebrating the short films, made by children and teenagers in schools, that explained something scientific to a non-specialist audience. It was great fun and inspiring.

I'm not sure what I was trying to take a photograph of as so little of it is in focus, but I was delighted with the unintended outcome anyway.

Apart from the 'ooh that's a bit unusual'-ness of it, what I like about this photo is that it caused me to learn a little bit more about how iPhone cameras and other smartphone cameras work. The image is created by a 'rolling shutter' which scans from one side to the other. This means that one side of the image is recorded before the other one is, and if you're recording something moving fairly quickly then you can get some fantastically odd distortions.

Here's a relatively slow-moving wind turbine on Blackheath which I took on 29 August 2009 during the fab Climate Camp. There was a session on how to make wind turbines and we lowered this one to the ground to have a good look at it before re-erecting it, at which point it started spinning again. I tried to record a video but it looked rather odd.

My iPhone 3GS camera has a CMOS (Complementary Metal Oxide Silicon) sensor, which brings a certain distortion on images - here's a helpful video explaining the difference between CCD (charge coupled device) which doesn't scan.

Thanks to this article (Everything you wanted to know about rolling shutter) on the phenomenon I also found this great video taken from inside a guitar, using an iPhone 4 which has captured oscillations from the strings - note that this is just an artifact, the strings are actually vibrating much more quickly but this is what the sensor has picked up.

Anyway, back to SciCast - the winning video was Gravity, Mass and Weight by Oliver Madgwick who was 17 when he made it, with lego. It's brilliant :)


Incidentally, when you upload a video file to Blogger it shows you the image below while the video is 'rendering' or doing whatever it does for Blogger to be able to display it, thought I'd capture the image for posterity.