Big Data’s Big Ask

Words by Saskia Pronk

Data is relentless; just one human body contains nearly 150 trillion gigabytes of information.

The power to access and analyse these data sets is predicted to bring a utopia of improved patient care and clinical productivity, so it’s no surprise that a deal-making frenzy totalling nearly $4.7bn took place in the healthcare and pharmaceutical industries last year. Despite the real opportunities offered by patient data, there are many issues to resolve before the ‘data utopia’ becomes a working reality and planning for the sheer amount of data will prove crucial.

One industry-wide problem lies with the reality that patient data is fragmented. At the 2018 FT Global Pharmaceutical and Biotechnology Conference, London, a panel addresses the increasing issue. “It becomes hard to get one view of a given patient”, explains Pam Cyrus, VP & Global Head of Medical Governance at Bayer. Ruminating on the US system’s lack of communication, Cyrus explains: “An individual patient may go from one healthcare insurance company to another, so when you try to follow someone longitudinally – where they might see several physicians, it becomes fragmented.” However, Rajni Aneja, Global Head Digital Strategy & Transformation at Novartis, believes data interoperability is happening: “In the US, there is no shortage of data, we have the adoption of electronic medical records – but in the UK, the fragmentation doesn’t exist as much because of the single-payer system [NHS]. We’re making progress towards this, but are we perfect or ideal? Probably not.” This leads Aneja to question: “How do we get to this next level of sharing?”

The challenge is to accept what it is and let the data speak for itself

The second problem is determining which data should be used and how best to achieve it. “We often think of data’s opportunity as making medicine more precise, which is really important”, says Peter Donnelly, Founder and CEO, Genomics plc, “But there’s another piece which is less obvious, and that’s using these data resources to fundamentally understand human biology better.” This allows the identification of individuals who have genetic variations, where for pharma “It’s as if they’ve been on a weak version of your proposed inhibitor all their life”, the so-called ‘nature’s clinical trial’ explains Donnelly. Some companies have already made big bets in this area; in July 2018, GSK invested $300m in 23andMe to analyse this marriage of drug discovery and genomic data.

“All of human history has been living a giant trial”, supports William Mayo, Chief Information Officer at The Broad Institute-MIT and Harvard, “Finding variations which work or don’t work, that’s pretty much called evolution.” That’s why the question of ‘are we getting the right data or are we missing the signal amid the noise?’ concerns Mayo, “This obsession with ‘clean data’ implies that there is an absolute thing which is clean data and then there is a tier which is not clean.” According to Mayo, it’s inaccuracies like these which lead to a bias, saying: “For all we know, we’re cleaning out the signal, so the challenge is to accept what it is and let the data speak for itself.”

But is this the same for patient-empowering, real-time-data-producing wearables, apps, and other digital therapeutics? If an intervention isn’t regulatory-approved, does it still bring value to pharma and healthcare? Mayo informs: “There is massive signal in that data, and if you start from the position that it’s not clinical grade, you will give yourself an excuse to not look at it as useful.” Aneja agrees, stating: “It’s a form of real-world evidence. You have a consumer who is engaged and even motivated – and you’re getting real biometrics knowing about them from their lifestyle. Something you wouldn’t get from a clinical trial.” From this, Aneja believes it opens a new area of opportunity to be capitalising on data as an asset to say: ‘Can I make insightful decision making here and form intelligent correlations?’. What’s more, bringing these multiple data sources and inputs together helps alleviate that issue of data fragmentation.

Conversely, Cyrus believes it’s important to separate these technologies into ‘medical software’ and ‘what can supplement’ to determine how best to extract their value. For example, “If you’re talking about changing a label or getting approval from a health authority, you’re going to have to demonstrate and validate it as medical software.”

But that doesn’t mean it doesn’t have value if it isn’t validated as medical software, she explains: “You just use it as one other decision point in managing a patient, it’s not the only thing you use to make a treatment decision.”

It’s not straightforward to get the right data – but the power there is extraordinary

With big data and advanced analytics driving this paradigm shift in the healthcare and pharma industries, planning for the sheer amount of data is crucial, which includes decreasing fragmentation and ensuring accessibility. Pharma’s business value lies in capitalising on genomic data to aid drug discovery, adopting the so-called ‘nature’s clinical trial’. With investments expected to grow to more than >$7 billion by the end of 2021, the opportunity is immense, as Donnelly states: “It’s not straightforward to get the right data – but the power there is extraordinary.”