Big data at your bedside

July 2015

Woman asleep next to phoneMatt Pfeil, Chief Customer Officer and Co-Founder, DataStax, discusses how responsible use of big data can revolutionise healthcare.

Our lives are becoming more and more connected every day. According to the International Data Corporation (IDC), big data spending is expected to increase to an estimated £88 billion worldwide in 2015.

Private sector companies are connecting devices inside everything from cars to weighing scales and, believe it or not, plant pots. It is from these devices that sensor data is used to gather information about individuals.

As long as this gathered information is shown to improve customers’ lives, people seem happy to share their data.

For the NHS, however, connecting medical devices to gather data about patients is still an area to improve patient care. The same approaches to data used in the private sector could also be used to analyse data about people to support better clinical diagnosis.

Clinical use of data - improving decisions in real time

What clinicians need above all is visibility - the ability to see the early onset of conditions that would otherwise go undetected. By collecting data from medical devices, clinicians can capture information on the movement of conditions such as heart rate, blood pressure and oxygen saturation and then have real-time analytics carried out on this data.

By looking at current patient conditions against the overall set of clinical data, it is possible to spot potential patterns to aid in clinical diagnosis. Now, this involves creating a vast set of patient data to use as the basis for comparison.

One example of how this can work in practice is the treatment of sepsis, a blood poisoning disease that shows symptoms similar to other conditions until it has fully developed. Sepsis is one of the most expensive conditions that can affect patients. In the UK alone, the Sepsis Trust puts NHS spending linked to the condition at £2.5 billion.

The diagnosis of this condition is tricky as it looks like a range of other medical conditions in its early stages. For clinicians that have seen sepsis develop before, they may be able to diagnose it with successful treatment during the early stages. For clinicians without first-hand experience with sepsis, the diagnosis is harder.

This is where big data and real-time analytics can help. By calculating potential outcomes and making potential recommendations, big data can help provide the most effective treatment early and increase the chances of success.

Treating conditions like sepsis with big data involves looking for patterns in data sets based on data gathered in real time from the medical devices around the patient. US-based company Amara Health Analytics works with healthcare providers to capture all the patient data being created over time and then makes it available to them for analysis around sensitive conditions. For Amara, this approach allows them to look at more than 100 different data streams to track and analyse.

By analysing the individual’s data against other patients, it is possible to spot where sepsis might be developing. If a patient starts to exhibit the signs of sepsis, an alert can be sent over to the doctor or nurse and the information can be used as part of a wider clinical decision support service.

Thinking through the issues around data security and privacy

The use of big data around patient conditions is possible. However, there are security and moral points to consider. Patient data must be kept secure, private and available only to those that are authorised to use it. The Data Protection Act forbids unauthorised access to patient information as standard, while NHS professionals receive mandatory training on patient data confidentiality.

However, moving information about a patient’s conditions to a central data store where it is compared against other patients’ clinical details represents a new and potentially problematic course.

When first introduced, the NHS’ care.data initiative was a plan to digitise patient records and then share them across multiple NHS Trusts and other organisations across the healthcare sector. The goal for care.data was to help ensure that the quality of care received by patients is consistent across the NHS. Following public discussion around the role of care.data, this has recently been reintroduced to move all GP medical records to digital rather than paper-based systems.

As of March 2015, care.data will involve the digitisation of GP records for the use of the patient, so they can have access to their own medical data and understand more about their own health issues. Treatments and clinical information will be linked to this record.

The use of real-time data is different to this wider records project since it relies on having the scale of data to compare individual patient information against. However, the medical device data is only linked to the patient and is not designed to be part of care.data. Instead, the information is compared against the overall amount of anonymised patient data to spot patterns.

This data may be kept by individual NHS organisations in order for them to use, but more likely is that a central NHS body would support the secure storage of anonymised patient data, which would then be used for comparison.

Open source approaches to capturing data

In order to keep pace with the huge amounts of data generated from medical devices over time, it’s essential to capture the data as it is created. This time-series data has to be sent from the device to a database where it can be hosted and then collated for analysis. As so much data can be involved, traditional technologies like relational databases are not able to keep up with the speed at which new points of information are created and then need to be analysed.

There are several potential new and open source technologies available that can help NHS organisations capture data from medical devices. The challenge is dealing with time-series data at the scale required to help make analysis viable for clinical decision support. NoSQL databases can assist with the capture of data in real time, while tools like general engines can support analysis of data in real time. For long-term data analysis, an open source software framework can be used for batch data analysis.

Each of these technologies has their own advantages for handling data and can be used to meet specific requirements around writing and analysing information. Alongside this, the use of open source components allows organisations to see what they are using and keep the underlying infrastructure secure. The infrastructure components can either be hosted by the NHS organisations themselves, or run within a UK data centre that meets requirements around security of data.

Looking to the future

While big data has the potential to play a pivotal role in helping clinical staff with decision support and the development of clinical pathways, there needs to be respect for the individual’s right to privacy, anonymity and security around their data.

There should be no doubt that big data can contribute to greater clinical efficiency within the NHS. By looking at new technologies and innovative uses of data, IT can support clinical decision-making by offering recommendations faster and with greater knowledge. While, this has to be balanced against the wider citizen requirements around safety, security and anonymity of data, big data offers huge potential to improve the quality of care and change the way that healthcare is provided.

Image: iStock/51528039

There are no comments on this item

Leave Comment

Post a comment