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

Sunday, 31 January 2016

Communicating Risk in Health Information (pt 2) - Patient Information Forum event: side effect risks & bowel cancer screening sessions

Last Tuesday I went to PiF (Patient Information Forum)'s event on communicating risk in health information (you can download the presentation slides). It was very good and inspiring and I thought I'd gradually type up my notes.

If you spot any mistakes please let me know, @JoBrodie or my obvious gmail address is somewhere at the top of this blog.

My notes from sessions 1 and 2 are in my previous blog post
Communicating Risk (pt 1) - Patient Information Forum event: health literacy and data visualisation sessions
Session 1: Health risks of low health literacy
Janet Solla, Director, Community Health and Leaning Foundation
Session 2: Using data visualisation to explain risks
Eluned Hughes, Breast Cancer Now

My notes from sessions 3 and 4 below are in this blog post, below
Session 3: Factors influencing the perception of side-effect risk information
Peter Gardner, Head of School of Psychology, University of Leeds

Session 4: Developing a risk communication for bowel cancer screening
Sam Smith, Cancer Research UK Postdoctoral Fellow, Centre for Cancer Prevention, Queen Mary University of London, Wolfson Institute of Preventive Medicine

Tuesday 26th January 2016, 10.30-4pm (with very nice food) at NCVO offices, King's Cross

Session Three - Factors influencing the perception of side-effect risk information
Peter Gardner, Head of School of Psychology, University of Leeds
[slides, PDF]

Peter showed us the patient information leaflet (PIL) for a blood pressure drug which had a long list of fairly unpleasant possible side effects (adverse events) - basically a litany of potential doom and gloom. You could easily imagine that reading this would put people off taking the tablets.

One particular problem is the language used in terms of the relative frequency of these side effects, and what people understand by them.

In testing the official EU terms 'very common', 'common' etc they found that people understood that 'very common' means over 50 per cent whereas in fact it actually refers to 'more than 10 per cent' of people taking the medication. In fact I tested this out later that evening in a group of human-computer interaction experts with whom I was having dinner (on our project which is about making medical devices safer) and they agreed that 'very common' would be more than 50%, as do I (even though I'm fairly familiar with PILs and know what they're referring to).

People actually greatly inflate the risks thanks to the use of these unclear words.

Peter and his team have done a bit of testing via a pop-up on a cancer charity page, which invited website visitors to take part. It was slow going as only around 15 people completed the survey each month so it took ages to get enough to start doing some number crunching. Note that participants were self-selected (visiting that particular website, presumably wanting information and likely to be quite motivated... a large proportion of them also turned out to be educated to degree level)

They tested the effects of stating the risk of side effects as verbal descriptoers, percentages, frequency statements and combinations of these.

For example, "If 100 people take this medicine then 3 would get constipation" (they also looked at 'would' and words like 'will' vs 'may' but there weren't significant differences on that).

Verbal descriptions such as 'common' or 'rare' produced markedly less accurate risk estimations on their own though percentages performed quite well, though less good for things that were lower risk.

People preferred combination statements such as "affects 1 in 500 people (0.2%)" ['affects more than 1 in 10 patients is known as a 'frequency band'].

I wondered, but didn't think to ask, how this information affects people's behaviour if they experience these side effects - do they take them less seriously because they're 'just' a consequence of the medication, so ignore them, or do they monitor themselves for the onset of anything a bit worrying.

Fortunately both NICE (National Institute for Health and Care Excellence) and EMA (European Medicines Agency) are fans of some sort of combination - words and numbers. However this is based more on consensus rather than evidence, because everyone thinks it's a good idea (it probably is but doesn't have a particularly strong evidence base as yet). What is the best combination?

There's certainly a possibility of verbal descriptors having a framing effect on people's understanding but what's particularly clear is that people don't even share understanding of the same risk descriptors. They also want to test with other medications (after all there's also an emotional response to certain types of treatments - you might be pretty relaxed about ibuprofens for muscular pain but much more tense about something that's meant to be keeping your blood pressure normal) and also, partly in light of the presentation earlier on the day about health literacy (session one), they want to test numeracy levels and look at other individual differences. Also to investigate the effect on actual behaviour

See also
Berry DC, Knapp P and Raynor DK (2002) Provision of information about drug side-effects to patients Lancet Mar 9;359 (9309): 853-4

Session Four - Developing a risk communication for bowel cancer screening
Sam Smith @SGSmith_87, Cancer Research UK Postdoctoral Fellow, Centre for Cancer Prevention, Queen Mary University of London, Wolfson Institute of Preventive Medicine
[slides, PDF]

The NHS' Bowel Cancer Screening programme got going in 2008 and involves sending people between 60 and 74 a thing to poo into and presumably send it back. This is so a faecal occult blood test can be done (blood within the poo structure, rather than bleeding). I did wonder when Sam mentioned that this involves putting a small poo sample on card whether people were actually carding their poo at home, but I think not!

This collection is accompanied by a booklet that explains what the test aims to do and what's involved and they wanted to test how well it was understood. One measure of readability is the Flesch-Kincaid measure which looks at sentence and word length and for comparison the Harvard Law Review is about 30, The Sun newspaper is 76 and the NHS Bowel info is 62. 

Interestingly (depressingly) there's a strong correlation between affluence and the returning of a kit - this can mean a danger of introducing inequalities in screening for bowel cancer because only those who return are being screened. Obviously they don't want this to be the 'fault' of the booklet, where people can't understand what's being asked of them and what it all means.

They also did a study where they got people to read the booklet and at various points asked them to 'think aloud' so that they could record people's impressions, and comprehension. In one example someone misunderstood the phrase 'about one in 20 people in the UK will develop bowel cancer in their lifetime' to mean 'that's one in four of the population, isn't it?'. People struggled to get the gist of the facts-based booklet - there were too many facts, some not really that relevant and some text had too much scientific detail.

They also tested the development of an overview booklet (to give people a 'gist' of what's going on). In this document they used more sign posting to help people find their way through it, and also used vernacular language (bowel instead of colorectal). They also called it a 'two minute' guide but someone suggested that that might imply the recommended time to read and might put people off who wanted more info, or felt they'd take longer to read it. They also rejected 'a simple guide' as that might be a bit patronising.

Although their 'gist' leaflet was well understood it arrived with the 'facts' document and the results indicated that the gist leaflet didn't appear to change people's intentions about screening, even if they had more knowledge about it, so no real difference in uptake. One thing they'd like to do is test the gist leaflet as a standalone thing.

No comments:

Post a Comment

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).