Two way search

Dave Winer, the father of RSS, writes about getting the data flowing both ways.  Or, at least the potential for this, by having search engines use his blog to make results more relevant. And less random. He figures it won’t be sinister as he only reveals general stuff about himself on his blog-what city he calls home, and many other details that will help disambiguate his search query (his example: differentiating firmware driver from car driver).

He likes this idea so much that he bought the domain www.twowaysearch.com!

By the way, the tidbit above was found by going to TwitterFox (which makes Twitter much more accessible in an optional popup,  in my opinion, as it is never the main event), looking at a Guardian Tech tweet, going to a link in their article, and then finding the two way search in Jeff Jarvis’s Twitter feed on his blog here. Jeff’s blog is righteous as well-he has written a post linking to the MistakeBank, a ning site where people share and try to learn from their mistakes.

NHS Evidence: 3 months on

Went to update workshop held by London Links (no not the bespoke jewellers ahem) and had update on NHS Evidence:

• See librarians as advocates of the service
• NHS Evidence is still at an early stage so it is about managing expectations and it will take time for people to use and trust it
• Aim is to be akin to NHS Search, a single source of information for the NHS
• Working on integration into 3^rd party systems such as hand held devices
• Highlighted new areas: drugs and horizon scanning, commissioning, public health and e-learning modules
• More resources that feed into the FAST search (front page) are being ‘ingested’
• First determinations of accreditation scheme will be published soon
• Eyes on Evidence bulletin you may have noticed uses the Specialist Collections to promote new evidence
• They are reviewing many areas of the service including journal provision and the Specialist Collections
• User testing is continuing and they said they would be happy for librarians to volunteer for this
• Release 2 will be in October where they hope to roll out personalisation and improved search functionality (user ranking)
• Release 3 will possibly include the ability of third parties to upload content and/or local information, this is where the look and feel of the site may change more significantly than merely building on NLH
• Aim is not to duplicate the work of NHS Information Centre or DoH but wants to group all relevant resources in one place. [In answer to a question about attracting commissioners]

There was an interesting report from a test between NHS Evidence, TRIP and Pubmed using a series of clinical questions which (albeit using the surrogate outcome of number of systematic reviews for quality as opposed to relevance of results) found that NHS Evidence is not doing badly. Reinhard Wentz, the ex-medical librarian who carried out the tests said us info pros could learn something from clinicians about single line searching. I’d like to see a more thorough test of this (and my colleagues are talking about testing the utility of Emtree headings so hey which is more positively riveting).

In my modest opinion: I tend to use the primary sources although that’s because of local protocol in the main. However if they could 1. make it more explicit what resources they are ‘ingesting’ (their words) and 2. refine the results to take into account of both currency and relevancy then it may be useful. Look forward to seeing how it is promoted (they are planning to liaise with medical schools to get it on curricula amongst I assume other things).

Why we must ration health care

Peter Singer, the controversial ethicist based at Princeton University, writing in the New York Times argues for a rational debate about rationing in the face of decisions such as those around cancer drugs such as sutent:

You have advanced kidney cancer. It will kill you, probably in the next year or two. A drug called Sutent slows the spread of the cancer and may give you an extra six months, but at a cost of $54,000. Is a few more months worth that much?

 
If you can afford it, you probably would pay that much, or more, to live longer, even if your quality of life wasn’t going to be good. But suppose it’s not you with the cancer but a stranger covered by your health-insurance fund. If the insurer provides this man — and everyone else like him — with Sutent, your premiums will increase. Do you still think the drug is a good value? Suppose the treatment cost a million dollars. Would it be worth it then? Ten million? Is there any limit to how much you would want your insurer to pay for a drug that adds six months to someone’s life? If there is any point at which you say, “No, an extra six months isn’t worth that much,” then you think that health care should be rationed.

This debate will run and run…

Clinical Reader: Malicious or just stupid?

I’d never heard of Hanlon’s razor before. Apparently it is an adage that reads “Never attribute to malice that which can be adequately explained by stupidity”. I rather like that; I’ll have to remember it. It is a nice way to look at the world, as we can all be stupid sometimes. Take for example the new company Clinical Reader. The company has an online product that basically seems to be an RSS reader but that they decide which RSS feeds they’re going to track, not you. I came across it a few days ago as some clinicians were discussing what a good service it was on some mailing list. Well, I thought, that’s clinicians for you. The product didn’t seem bad enough to comment on, but likewise certainly didn’t seem interesting enough to comment on either – at least not from my point of view. So I thought to myself “well, our handsome and fragrant readership will probably come across it soon enough, and they can make their own perfectly balanced minds up about it…”, and left it there.

But today I see another twist on the Clinical Reader story, and what happens when you get involved in social networking tools without really knowing what you’re doing. The thing was that Nicole Dettmar (evidently a fan of The Prisoner) had pointed out in her blog that Clinical Reader were implying that they had been awarded ‘five stars’ by institutions such as the British Library, the NLM, Imperial College, The Lancet etc. She pointed out to them on Twitter than the NLM does not endorse anything, and that they ought to do something about it, and promptly received a reply threatening legal action (I love the use of ‘kindly’):

I mean to say, what a stupid (or malicious) thing to do. Of course everyone picked up on it and they received a barrage of tweets and blog commentary. As of writing they have since backed down, which they should do because they are plainly in the wrong, saying “We are keen to engage the twitter community the tweet made by a junior member of the team was poor judgment”. However the “five stars according to…” graphic remains throughout their site once you get past the first page.

Malicious or stupid? You decide. Either way it doesn’t reflect well on Clinical Reader.

Human slaves…in a robot world

I was reading a very interesting article in the New Scientist about, er, robots, that can detect our emotions. Or ‘machines’, if you prefer (I noticed that the title had been changed to ‘robots’ from ‘machines’, perhaps in a bid to raise eyebrows).

Yes, robots are fun and so are machines, but where is the health connection, you ask. Machines can distinguish between the 6 ‘basic emotions’, fear, disgust, surprise, anger, happiness and sadness, but only if an exaggerated expression is presented to them. (Digression: reading about basic emotions or universal expressions always makes me a little skeptical as different people have different ways of expressing themselves-you’ll find lots of exaggerated smiles in North America, but comparatively few in the UK).

My favourite part: machines have gotten quite skilled at differentiating between real and fake pain in humans.  One study by Littlewort et al saw computers correctly classify pain 88% of the time. Human volunteers got it right 50% of the time-perhaps they were guessing. Will this technology have ramifications for reducing numbers of benefits fraudsters and pill-poppers as well as people trying to separate pain into organic or psychosomatic causes? Oh, and teaching autistic children to correctly identify facial expressions?

I hope so. Although it is easy to see the scarier side of technology like this. Who knows, maybe we’ll all be forced to adopt poker faces in the future to avoid ‘mood profiling’ and targeted advertising.

Using Twitter to track disease outbreaks

A few months ago I wrote a short post on Phil Baumann’s 140 healthcare uses for Twitter. Well, Chris Thorman over on Software Advice has written a longer piece on the potential of Twitter for identifying and tracking disease outbreaks in real-time. To get some vaguely reliable data from Twitter, rather than the mess of misinformation with the occasional piece of truth thrown in which is Twitter as of today (oh, cynic that I am), it would  be necessary to have a uniform set of diagnostic codes, “hash tags” and a proper authentication system, e.g. as Chris writes:

“…This adoption by doctors would need include a verification system that only allows trusted or authenticated users to tweet about information contained in the EMRs. What we’re trying to avoid is aggregating a whole mess of data related to a particular disease. Authenticating users to make sure they are who they say they are avoids this problem. With a uniform set of diagnosis codes and a proper authentication system, suddenly the trending data sent out by these verified doctors’ tweets goes from speculative to extremely reliable.”

I do actually think that Twitter or a similar technology could be very useful in tracking the early signs of a condition, or any other “rare event” pattern. In fact I’d be amazed if we weren’t using it this way very soon. What is required however, as always, in some kind of central, trustworthy institution to organise, analyse, study and disseminate the data that comes in, as well as ensuring the diagnoses are correct and not just false positives. That’s the hard part, not some Dr in a clinic in Great Yarmouth tweeting that one of their patients has a sniffle. Anyway, take a look and see what you think. Twitter fans will like it at least. Another step towards world domination.

NHS health records to private sector?

Personally, I’d be delighted if the Tories got Google or Microsoft or someone to build the NHS records system. Might actually work. Read the news story on the HSJ here. That nice (?) Mr Cameron also says he will review all quangos should they come into government, including the bloated (did I say bloated?) NICE and Care Quality Commission. Keep the public sector on their toes eh David? Mind you, the opposition always says they’ll get rid of quangos, until they get into office then they just create a whole lot more. Jobs for the boys (and girls).

fefoo – not another search engine

I’m not sure why it’s called fefoo, but it’s quite a nice little tool. It’s not another search engine. Rather, it’s a tool through which you can search a whole range of different search engines – Google, Bing, Yahoo, of course, but also many of those funny little ones you’ve never heard of, like Viewzi (fun graphical interface), Spezify (yet more fun graphical interfaces) and LexxeAlpha (no fun graphical interfaces, but rather powered by “advanced natural language technology”, though still returns Wikipedia first).  The search pages give a little tool bar at the top that allows you quickly to try your search in another search engine. You can also look for blogs, images, torrents, people, movies etc. It’s all quite useful, to tell the truth. And finally, if you’re truly hardcore, rather than specifying in the drop down menus that you want to search Yahoo, say, you can use command line searches, in this case, for a search on ‘Tom Baker’, “:yahoo Tom Baker”, though unfortunately it does not seem that you can combine command lines searches, e.g. “:yahoo :images Tom Baker” for, you guessed it, images of Tom Baker. Oh well. It’s a nice little tool nonetheless, and helps ensure you venture beyond just Google from time to time.

What will US healthcare reform look like?

Fascinating article  in Economist about the United States healthcare system that proves that throwing money at a system doesn’t necessarily relate to either efficiency or quality. They may end adopting compulsory insurance and some form of the QOF to remedy this. Will they set up a US version of NICE? It’s too early to tell although if they do I’d be happy to help…

Promotion envy

I regularly suffer from promotion envy when other folk have much slicker promotional materials / concepts than me.  The latest dose has been caused by this lovely campaign in the states promoting the public library service – Geek the Library (alerted by this post at It’s all good). 

One of the key things for me is the combination of linking peoples passions to the ability to learn more about them through the library.  The use of web 2.0 type sites gives a way to be involved and hopefully generate some discussion and improve awareness amongst a younger audience.

As an avid public library user I would love to see something like this in the UK.  I have geeked all manner of things via my public library. 

On the health informatics side I noticed that NHS Evidence recruited a whole stack of marketing / promotion / brand management types.  They will almost certainly have more capacity in this area than NHS Libraries have ever even contemplated (more promotion envy).   Should there be a lesson for all parts of the library community?