Milan

The increasing availability of big data presents major opportunities for biopharma, so we need to question how we as an industry use it to deliver localized value to the NHS.

But first, what exactly is big data, and what specific role does it play in healthcare?

Clearly, there is almost limitless data in healthcare, covering practically every procedure, condition and drug, and coming from a vast number of providers. But to simply say that big data is just a collective term for all the information flowing between organizations is to miss the point. If we are to realize the potential of big data, we need a stronger definition.

At Quintiles we begin by looking at it in terms of volume, velocity and variety. Thanks to an explosion of sources, from the proliferation of electronic health records, national aggregated data sets, linked data sets – and increasingly the digitization of years of research and development – we have more data than ever before.

And one of the reasons why is due to velocity. If we go back 20 years, it would be inconceivable to have access to real-time data. But now, we almost take this speed of delivery for granted. Social media has radically changed the landscape, while wearable devices mean that information can be collected without any effort at all. But that instant delivery comes with more demanding expectations. If we can now receive data in near- or actual real-time, then we are also expected to act immediately to interpret and implement the findings. The other element is variety, because data now comes in so many formats, whether it's structured, such as numeric data in traditional databases, or unstructured sources such as email, video and audio.

We recognize that what makes big data in healthcare even more of a challenge is down to two further factors: variability and complexity. Data flows can be highly inconsistent in healthcare and working with so many sources means it is a huge undertaking to link, match, cleanse and transform data across systems. Standardizing data into single records has been fraught with difficulty, while the unstructured sources like social media are very difficult to standardize too.

Governing big data

Having so much data at our disposal does not necessarily mean we can do anything we want with it. As Dr Junaid Bajwa discussed during our most recent webinar, the governance in the UK requires biopharma to comply to appropriate use, including whether data should be anonymized, how long it can be kept, and how securely it is stored. He also highlighted a disconnect concerning the bodies involved. Therefore, we as an industry must navigate the governance challenge carefully and ensure we put patient needs first and use the data appropriately, a major challenge in itself.

The current impact of real-world data

It is clear that real-world data is already being used to improve patient outcomes. At Quintiles we use a combination of primary and secondary care data and large aggregated sets to create value-added services, but the key is to understand what level of data is appropriate for the insights you are trying to discover. For example, how does big data influence outcomes audits? In this instance, we drill down from national level data and NICE guidelines to specific CCG or individual hospital data, where we can see how local patient experience, practice and interpretation of guidelines are impacting on the use and effectiveness of a product. Through this, we can understand why there are variations in uptake and adherence compared to other geographies.

Similarly, through simulation modeling, we help create local, geographically specific CCG plans that are based not on the general assumptions of national datasets, but on the variables of their own unique population. Understanding how typical – or more likely atypical – a CCG is compared to national data is invaluable. It is this use of data that allows us to build simulations based on real-word information, revealing a great deal about how certain pathway modifications can impact on budgets, adherence and patient outcomes.

The future of big data

We are clearly only at the beginnings of understanding what we can do with big data, and we will, in the future, see more about how we can link health data with social data, so we don't just understand impact on patient outcomes and the healthcare economy, but also on the productivity of the population. And of course this will inevitably push us to grasp the challenge of how we embrace social media. At the moment, we aren't able to be methodical about analyzing blogs, online communities, tweet conversations and hashtag trends, because it is extremely difficult to even know where to start with such a huge but unorganized resource. The future means discovering how to structure these outputs and learn how they can add value to the more traditional data sets.

At the same time, there will be a move to more prescriptive analytics, using data not just to capture behavior patterns, but to actually promote changes. In a sense, it will mean turning the agenda around, so it's less about understanding what is happening, and more about how we use the data to drive behavior change.

   

This post is the fourth in an ongoing series on UK market access:

Topics in this blog post: Evidence, Commercialization, EU, Market Access, Value, NHS