Euro-Sante :: Euro-Health

Internet Splat Map
Real-time data with hidden patterns

I once wrote about “on-demand, real-time, location-independent” healthcare. Today, we often think of this in the context of e-health or telehealth. The essential capability behind this string of terms was to capture the way that the ‘information value chain’ (digital information technology) could influence how healthcare is delivered. An interview I gave in Euractiv added that such technologies could be productively disruptive of cozy working practices in healthcare, by shifting the focus of healthcare decisively to the end-user, and away from provider interests.

For years we have seen investment in e-health technologies, but few services, and bold statements from the European Commission on the potential of e-health to bend the cost curve down. Now everyone wants to bend the curve down; most approaches, regretfully, rest on reducing activity within existing ways of working rather than adopting new ways. One difficulty here is obviously that with disruptive innovation in healthcare, healthcare work will change and that will have an impact on professional practice — when was the last time a health profession was made obsolete by technology. For some types of surgery, radiologists may become obsolete through interventional radiology which integrates real-time radiological technology into surgery. Something to think about, perhaps.

These thoughts bring up one important aspect of the use of information technology in healthcare, the ability to understand the present better in order to use resources better. Hospitals are notoriously bad at forecasting real-time demand for healthcare. They make various predictions that demand for emergency services may follow the cycle of the full-moon (true), and may correlate with large public events, but run-of-the-mill day to day capacity and resource management treats on-demand healthcare as an emergency

There is a technique used in some areas called ‘now casting’, an effort to correlate real-time information with short-term modelling of resource use, and to anticipate short-term demand. The European Centre for Disease Prevention and Control, in a June 2009 report on surveillance during a pandemic, includes amongst it various methods the use of now-casting. Weather forecasts are a weather model which is continuously updated with real-time data. We may also be familiar with public health surveillance and digital technology has improved the quality of our models. But real-time techniques (such as mapping) takes us into a new realm where demand can be better understood, to avoid inappropriate queueing and service rationing. For Europe, such capabilities to track information would be important, as information, like viruses, doesn’t respect borders; and real-time information healthcare capabilities for one member state would be capabilities for all.

The problem we have is that we are generally good at looking for things we know about, but not for things we don’t know about, and therefore are unable to anticipate. So our public health pandemic systems are surveillance systems, which focus on things we already know about; they cannot identify short term changes with emergent problems — unless we are looking, we won’t see.

A robust real-time healthcare information system would be agnostic to specific issues, as it would work to identify emergent patterns, and provide a picture of the situation to enable shorter-term or immediate responses. We aren’t yet that smart to know what to look for — people have predicted the end of infectious diseases, the end of history, the end of poverty. The advantage of digital information technology is its ability to search for emergent patterns, to correlate perturbations as potential precursors of something to come (like monitoring a heart patient at home to detect early tell-tales, so the ambulance can be sent the day before the heart attack).

The advantages would be immediate, from better hospital resource management, to tracking emerging infections, to improving the ability of front-line services to be in the right place when they are needed.

Do you have suggestions?

Comment on this post with your suggestions on areas where improved data management in real-time could both drive down costs and improve healthcare. Also comment if you are already doing this.

Want to know more?

Mapping of influenza using real-time data feeds has featured on Google; a Canadian firm also does health mapping: Infonauts.

Use of now-casting in weather is used by meteorologists; the UK’s Met Office explains how it works here.

Nowcasting consumption using Google data is here.

Here’s one from 2004! using nowcasting of air quality to issue health alerts; linking this data with hospital admissions data for, say people with COPD, might offer improved real-time responses, perhaps even an anticipatory capacity to tell people to stay indoors. A UK project does this (this link is a Wikipedia entry on the project).

Author :