Dr Martin Knapp
What are you currently doing?
Writing, reading, relaxing….. and speaking when I get a chance!
When did you realise technology would be useful in healthcare?
It was all about the data, and how it could be used to make predictions and about when was the right time to put patients with chronic renal disease on dialysis. Also a need to detect change in clinical markers after renal transplantation, to tell us when a patient was starting to reject their transplanted kidney and to initiate investigation and increase anti-rejection therapy.
The nephrology discipline is more numerate than many others because of the volume of data that needs to be managed longitudinally when patients have kidney disease. Nephrology is rich in the needs for trend analysis and event detection of many kinds. Not only the quantity of data but the fact that it will only make sense if considered with what went before and, where it might be going.
Early in my role as a specialist nephrologist I started drawing graphs on specialised graph paper. Rapidly we found this not to be realistic by hand because of the volume of data. I also appreciated that by developing statistical algorithms we could quantify change and our prediction of trends.
How did your clinical career lead you to getting involved in clinical informatics?
An interest in chronobiology got me interested in informatics. Chronobiology examines periodic (cyclic) phenomena in living organisms, the cycles known as biological rhythms. Applying chronobiology principles to medicine Initially this was in looking at blood pressure variations and later at immune-suppression. At what time to give medications? Could we get a better result or less toxicity at different times?
What aspirations for the future of healthcare kept you going?
Facilitating the use of graphic systems to provide longitudinal trend data to treat individual patients, with the potential to build in algorithms to alert clinicians and patients to important change.
Any memorable experiences you are willing to share?
In 1983, after I left nephrology at Nottingham, I transferred myself to Obstetrics to create a Medical Information Technology Unit, at Queens Medical Centre. At Queens we directed computing at the process of having a baby, which is not dissimilar to having a transplant in many ways – making the system we had developed for nephrology very highly suited. Midwives were a bit different from renal nurses but certainly they too are very strong-minded people and very important. We had realised from talking to clinical staff that it is a good idea to please the people who were important. We set out to please the midwives! Their main gripe in life was the amount of paperwork they had to do. We focused the system on the activity at the end of birth, when midwives had to collate all the paperwork to send off in the various directions. It would take them 2 hours after a baby was born before that job was completed. By focusing use of the computer on the post-delivery process we got great enthusiasm and the project went very well.
At another site (also putting in a system for Obstetrics) the focus was on producing information for the Regional Authority, who had funded that project. There was very little gain for the midwives and much extra keyboard inputting. When Christmas and New Year came around all the computer development team went home. When they came back to the Hospital, with the Regional Authority system, the midwives just refused to use the system – they’d had enough – with extra work putting in data for managers without visible benefit at the clinical level. At Queen’s, using the system with the needs of the midwives taken into account, this produced data for management as a by-product, with very little non-clinical work needed from midwives.
When you look back, what is your sense about then and now?
In spite of the investment in IT health information systems they are not as good as they could be.
Presentation of information when it is shared between clinicians is still often not good and may not arrive before the patient arrives! Many hospitals send out discharge summaries to GPs only as a text summary. Selected graphic representation of data, correlated with text, e.g. medication type and dose, could convey a more insightful story of the patient’s illness episode in the context of the past and the future. There is a need when presenting information graphically to incorporate data that has been shared between both hospital and community healthcare providers.
Philosophically it would seem that clinical systems have often been designed to replicate the design of the paper record, rather than step back and explore what the computer can do to enhance the paper record, the workflow and help the clinician.
What advice would you give those currently working in and shaping the future of (IT-enabled) healthcare?
We need to be collecting more data (not less) so that we can better detect change e.g. in the patient’s response to treatment. Presenting to the clinician, data in a time series (the longer the period the better!) and presenting data graphically is one way we could achieve better integration of data and integration between health care providers…… using data as the link.
I have not been aware of enough meetings that bring together clinical and informatics specialities and which encourage fertilisation and cross-boundaries between medical sub-specialties. My example, of moving from nephrology to obstetrics, enabled us to apply lessons learned in nephrology to design a better system for obstetrics.
There has been lots of enthusiasm for computers, but often it has gone down a single track, usually as a natural instinct to stay with what we know and in our own area of interest. I would implore the future workforce collaborates more.