Email
The centre has released the main lessons generated by a seminar it held in December 2009, which involved NHS staff, suppliers to the health service, analytic software firm SAS and delegates from the private sector.
Paul Goodwin, professor of analytics at the University of Bath, told the event that he saw four major trends in this area: text analysis, use of social networks, visualisation of data and automatic large-scale forecasting.
Brian Derry, executive director of information services for the centre, told SmartHealthcare.com that each of these has potential for health and social care informatics. "This is all about 'predicting and preventing, not failing and fixing'," he said, quoting one of the seminar's speakers from publishing. "Historically, the public sector model would be to report annually, rather than follow things day by day."
But he added that it is now possible to accelerate the analyse of data, allowing organisations to spot patterns, identify and analyse risks and make connections far quicker than in the past. This can allow managers to nip problems in the bud, rather than react to them after they have happened.
Derry said that text analysis, looking at data held in unstructured written formats, is helpful in linking social care records to health information. This is particularly helpful when treating elderly patients who are also helped by social services. "If you want to understand them, you have got to find some way of joining this information up," he said.
Social networks can provide both a way to reach, and a way to hear from, groups of people with a particular condition, who often congregate electronically through such systems. "One of the big challenges for any public service is getting feedback," Derry said, and social networking can be employed for this, although its use is variable. "But where it works, it can be incredibly powerful and incredibly quick."
Visualisation of data – graphing, mapping or animating information – is useful in communicating information to non-specialists, but also helps specialists spot trends that are not obvious on a spreadsheet, Derry said. This could be particularly powerful in planning: "You can start seeing that certain groups are heavy users of services, and that these aren't very well connected," he said. This could allow health organisations to identify those running higher risks, and attempt to help them at an early stage: "In insurance terms, it's a risk management model."
Finally, automatic large-scale forecasting can help organisations spot trends much earlier than before. Derry said that the seminar heard that banks look for "little flurries" of activity from a few users, which indicate how many people will behave in future, while retailers make use of weather forecasts to predict demand for product lines. "The commercial world does this routinely, but the public sector is just getting its head around it," he said. "We shouldn't be afraid to do this."
;
Comments