Big data

Innovation, data and healthcare – part 2

Louise Marston - 06.06.2012

Given Nesta's remit, I'm keen to make sure that I keep the focus of our Big Data work on innovation. But what does using data for innovation mean? What does it rule out?

I've come up with a spectrum of things you might use 'big data' for, from transparency and accountability at one end, through to new business models, products and services at the other end. In between are more incremental forms of improvement and innovation, including learning processes, experiments, and algorithmic decision making.

The definitions I've arrived at for each stage are:

·         Transparency - About openness and accountability, not analysis;

·         Learning - Understanding, generating insights, making connections; applying some analytic process to generate additional knowledge from the original data.

·         Experiments - Taking actions based on data, then reviewing changes; not just learning, but applying that learning to the system, and reviewing the impact it has; closing the loop.

·         Automation & customisation - Automating decisions or data access based on specific circumstance; customising what a user sees based on what you know about their likely needs.

·         New models - New business models, new products, new services; doing new things that are not possible without the data, rather than using data to improve things that would also happen without it.

 In the second of these posts about making sense of data and healthcare, I have tried to apply this model to healthcare, partly to test how useful it is, and partly to tease out some of the specific issues sitting within the vast 'data and healthcare'

	System perspective	Individual perspective
Transparency	Performance league tables All trials records Open access research	Patient access to their records
Learning	NICE NHS Evidence - understanding efficacy and treatment options Virtual trials - analysing data on to determine correlations	Peer learning from one group of patients to another
Experiments	Clinical trials	Patients Like Me - contributing to new trials
Customisation & automation	Error alerts - automatically identifying possible conflicting medications or contraindications at prescription or at the pharmacy	Decision-support systems - suggested diagnoses for doctors
New models	Insurance models - factoring in healthy behaviours to premiums	Health apps & devices for consumers

 

I've divided the table into changes at the system level, which affect large pieces of the puzzle, and the individual perspective, which makes a difference to an individual patient, and their experience of healthcare.

I think this structure could be usefully applied to a number of other areas, fractionating the idea of 'big data' into more manageable pieces, and allowing us to examine the routes to innovation in more detail.