Earlier this month was Delivering Better Health Services, the 2009 joint annual conference by the Health Services Research Network and the National Institute for Health Research’s Service Delivery and Organisation Programme. As an SDO-funded researcher, I was asked to present and my talk is now available here. I discussed our meta-narrative review of electronic patient records and work on the anti-coagulant service based at the Whittington Hospital. This work suggests that we need a greater focus on communication across boundaries, and an ongoing praxis involving clinicians, informaticians and researchers.

For a conference on delivering better health services, it was remarkable how several talks described the delivery of worse health services, a result of the distorting effect of targets on clinical care or problems with financial conflicts. Prof. Steve Shortell, from Berkeley’s School of Public Health, opened the first day with a focus on integrated care, how it demonstrably improves health outcomes and the importance of appropriate funding models. Much that followed that day considered how the NHS’s recent focus on the choice agenda and an internal market directly works against achieving integrated care. Notable here was Matthias Beck’s analysis of the Local Improvement Finance Trust scheme, a miniature public-private partnership scheme for the NHS, which delivered their immediate goal of procuring modern buildings, but at much greater cost than expected.

Dr Henry Potts

Amy McKeown, a graduate of our MSc in Health Informatics course and soon to have a honorary appointment at CHIME, has had three conference presentations accepted recently. Amy’s Master’s thesis was a realist review on the use of computerised cognitive-behavioural therapy (CCBT) in the prevention of mental health problems. This included a case study on Xanthis, the CCBT product developed by her own company, Xanthis Ltd.

The research showed there to be a strong evidence base for CCBT as being effective in preventing and treating the early stages of stress, anxiety and depression. Amy’s work was recognised with a £1000 prize for the best health informatics dissertation at UCL, awarded by Dr Foster Intelligence. Amy and I are now disseminating the work further, with two presentations accepted at the First International E-Mental Health Summit 2009 in Amsterdam in October (“Realist literature review of computerised cognitive behavioural therapy (cCBT) for prevention and early intervention in anxiety and depression” and “Real world case studies of a preventative computerised cognitive behavioural (cCBT) package being used to support employees within large UK organisations”), and a third accepted at Medicine 2.0: Social Networking and Web 2.0 Applications in Medicine and Health in Toronto in September (“Computerised cognitive-behavioural therapy for prevention and early intervention in anxiety and depression: A case study of Xanthis”).

Dr Henry Potts

I love this site. You see some great signs in hospitals too. I’m not thinking so much of the official ones but more the ones that are just printed off and tacked up. Like the sign above a computer used by radiology registrars which is headed “How to copy a blank image to an empty examination”.

It gives instructions for two ways to do this. An easy way, and a hard way. The easy way doesn’t look that easy, so I’m guessing that the system designers didn’t anticipate that users would routinely want to copy blank images into empty investigations. Well, why would you?

The answer is interesting, since it tells something about what happens when carefully designed systems are released into the wild. The system in question is a Picture Archiving and Communication System. These systems are the essence of a modern radiology department and deal with the acquisition, storage and display of digital images. I guess a lot of radiologists spend a lot of their working day at PACS workstations.

But if you look at radiology as a component of a healthcare organisation, its function isn’t to produce and store images, its function is to generate reports. The report tells the referring physician what the radiologist can conclude about the patient’s condition. It is the report that the referring physician wants, not the image. And actually, you can have a report without an image. I’m guessing that in scans that are reported in real time, such as ultrasound, the radiologist sometimes forgets to save an image. I imagine too that sometimes patients are entered onto the system in advance of their appointment, so that if they don’t turn up, there is an entry that doesn’t have an image.

So you have a tool, PACS, which is so useful that it becomes the primary tool around which a radiologist organises his or her daily work, but which is performing a function (storing and organising images) which isn’t quite the one the radiologist wants (keeping track of patients, creating and transmitting reports). So when you have a patient who doesn’t have any images (and therefore isn’t on PACS) you get a difficulty. And the simplest way for the busy radiologist to work around this is to copy blank images into empty examinations on PACS.

What I don’t understand, however, is why anyone would want to do it the hard way?

Peter Knight and Prof. Dennis Protti spoke last night at City University, the event being the annual Open Evening of City’s Centre for Health Informatics. Dennis Protti, who lives and works in Canada, is a big name internationally in health informatics, and especially known for international comparisons of national strategies (see this for example).

Protti’s talk focussed on the idea of the single patient record. One patient, one record. So when you go to your GP, he or she has access to information about any stay in hospital, A&E visit or outpatient consultation, and when you go to hospital, the doctor there can see everything previously recorded about you by your GP.

This kind of integration has been focus of a lot of work health informatics for a long, long time. Decades. And there has been an awful lot of talk about how to achieve it: is the data to be pushed or pulled, collected together instantly or available if required, what kind of architecture would support this, what standards are required, can we make some magic Babelfish middleware to allow all the systems to understand each other?

In the NHS we’ve had at least two iterations through national strategies incorporating this idea. The first (circa 1998) had limited ambitions in this direction, with the focus being on integration within local communities. The final verdict on that strategy as a whole was that it didn’t offer much to help chief executives meet their more pressing priorities. So the money the centre had intended would be spent on IT somehow seemed to get spent on reducing waiting times, meeting quality targets, improving star ratings etc. The next national strategy (NPfIT) did have major ambitions in this direction: an explicit commitment to create a Summary Care Record for everyone. And that hasn’t happened either, but that’s a much longer story than we have time for here, and wasn’t really part of last night’s talk.

What was in the lecture was a review of six or so jurisdictions where this integration has been achieved: Hong Kong, Jordan, a Swedish county, a Spanish region, the U.S. Veteran’s Administration and, again in the U.S., Kaiser Permanente. They are all pretty small scale (max. 8 million patients). They all seem to have been achieved on the back of political ambition and organisational support, often as part of a wider drive to improve healthcare rather than through an initiative focused on IT. Technology doesn’t seem to be the problem.

Some in the audience wondered whether achieving this integration is really what we should be worrying about. I wonder that too. Interesting, then, that Peter Knight’s talk on the Research Capability Programme was all about record linkage. Record linkage? That’s connecting data items, and it is being done here to allow routinely collected patient data to be used e.g. by pharmaceutical companies in drug trials. I think that this is a great project. But I’m conscious that my sense of the risks and benefits is conditioned by a professional self-interest not my views as a citizen or a patient. I was intrigued though, by the juxtaposition of the two talks. We’re pushing ahead with ‘Record Linkage’ but stalling on the ‘Integrated Care Record’. As though, as a society, we’re happy to allow government agencies to associate data about treatment we receive from X with data about outcomes measured later by Y, but we don’t want the A&E doctor to know that our GP thinks we’re allergic to penicillin.