IDBS Data Management Blog

The UK, as one of the world’s leading innovation centers, continues to be “good at generating great ideas in our universities but less good at turning them into products and businesses of the future.” This insight by David Willetts in his 2012 Policy Exchange speech highlights a global issue – how to translate research excellence into economic value.

Businesses with the potential to develop ideas need to be able to tap into relevant past research. Too much resource is wasted in the repetition of early stage research experiments and opportunities to stimulate innovation in response to research findings are often missed.

Industrial research and development (R&D) participants, who have a keen focus on this issue, believe that at least 10% of their research data is lost and must be reworked and at least 13% of people’s time is spent looking for data.

How much public money is spent on research? Here are some facts:

• The EU spends ca. $65bn (Euro 50bn)
• The US spends ca. $60bn
• The UK spends ca. $5bn (£3bn)

Even if the levels of inefficiency are the same as industry the amount of wasted time and effort represents an astonishing sinkhole for public money at a time of intense financial constraint.

For academia to capture and use knowledge assets there should be one place to go to share research, something akin to a national ‘Book of Research’. It would provide access, on a strictly controlled and secure basis, to those who make use of research data and are able and motivated to secure the IP generated by it. This could also accommodate the public need for open access to research information once IP is protected.

The good news is that open access to all academic data is on the radar. Governments are starting to recognize their pivotal role in creating the conditions and e-infrastructure to maximize the full economic value from research. The US Government recently pledged to increase access to federally-funded research findings and in the UK, the Research Councils UK (RCUK), is heading up free and open access to outputs from publicly-funded research which it believes offer significant social and economic benefits.

However, it is not good enough just to publish headline information to the public. The underlying research data should also be secured and made available to the academic community so that innovation is properly nurtured.

Any country seeking to drive an innovation agenda that fosters collaboration must encourage cohesive, effective data management that gets rid of silos across communities. This will help those countries that already excel in research to benefit commercially from those successes.

According to James Joyce, mistakes are the portals of discovery. Good science means making mistakes. I recently read a great article that encourages just that; researchers to be open and honest about their scientific mistakes. It recognizes that information borne from mistakes helps shape science.

The art is to not keep quiet about them because this actually risks leaving others to [unknowingly] repeat them. Tracking scientific mistakes for others to learn from, and looking at opportunities to share knowledge, is essential. This IS collaboration, a premise which beats at the very heart of science.

R&D generates high value data assets – the lifeblood of every laboratory. To avoid duplication and maximize scarce time and resources, quality data must flow through this knowledge ecosystem for it to thrive. Enterprise analytics require close collaboration to ensure metadata is captured alongside high context information, an ontology and its provenance. The intelligence of the community as they interpret and challenge the data must be captured alongside the experimental conclusion. Smart social tools, including tagging and commenting must be there ‘close-to-the-data’ to enable connectedness.

Only if there is high context – that includes the mistakes and false paths taken can we generate a useful Big Data asset. Competitive advantage is not going to be driven by choice of Big Data analytics alone but by the quality and provenance of data the analytics has to work on. We are all in a race to achieve the highest context, highest value distributed datasets to make data reusable. So when it comes to scientific mistakes, once is an understandable [and oft valuable] learning curve. Twice or more? Well, that risks looking distinctly careless and costly.

I’ve been talking recently about the idea that every R&D organization has an inner superpower waiting to be unleashed – but most of them are yet to discover it. In fact, many are squandering their most fearsome competitive advantage. What is our industry’s most constructive weapon? Data.

R&D centric companies across all sectors, from pharma to food, create, use and monetize information. Their raw asset – data – has value. It’s a capital asset. And when that data is added to, interpreted and shared, it becomes increasingly more complex and valuable. The creation of data assets requires a complex inter-dependent community of projects, supported by various teams, each providing skills and insight from discovery to delivery: an ecosystem of ideas and information.

Collaboration within Complex Iterative Processes

Too often today’s data ecosystems are suppressed by ineffective collaboration. Something as simple as efficiently moving data from one person to another, and effectively aligning data from internal or external collaborators, is continuously hampered. This is real life for the vast majority of researchers today and this status quo must be challenged and changed.

R&D scientists know that treating R&D as a linear process from basic research to product is flawed. And dangerously so. This heritage concept in no way reflects how teams really generate the information asset and, in practice, serves to entrench a siloed mentality.

In reality, data, information and knowledge are created and shared across complex, iterative processes that span research, development, patent filing, manufacture and post-market analysis. It’s a collaborative data ecosystem and an increasingly globalised, multiparty environment. The volume and complexity of the information has grown exponentially. Accepting this truth and working with it has profound meaning for how we use and further exploit both our internal, and global, communities to increase R&D productivity.

Exciting news from the US last week. The CommonWell Health Alliance, a not-for-profit trade association, has been formed with the promise of transforming the face of national healthcare. The headline act is to make health records easier to share and five major providers are coming together to create common standards.

In a bid to promote seamless interoperability and access to patient data across the healthcare system, these electronic medical record providers will promote common standards for sharing health data. It will mean that doctors can not only move patient records from one health system to another but also access relevant health records (pending patient agreement). A pilot program testing CommonWell Health Alliance’s plan will be conducted over the next 12-18 months, after which time the Alliance will be formally established.

Historically, a lack of interoperability among electronic health record systems and poor infrastructure for exchanging information have been the biggest barriers preventing the benefits of computerized health. The creation of the Alliance has been described as a matter of national importance by its founders and not a commercial effort but ‘an obligation’. They are to be commended for the herculean efforts that will be needed to digitize the content of an entire industry.

At IDBS, we are in a good position to witness how the siloing of health data in the US has been a major barrier to delivering connected healthcare. It is important that data interoperability standards are driven from within the industry. The Federal Government also has a key role to play in promoting this collaboration and incentivising interoperability, as it did with HIPAA.

There is more though. We need to aim for HiFi healthcare data sharing and collaboration. Aside from the debate around standards, which everyone agrees are necessary, we need to give focus to the quality of the patient data that is being stored and shared. High-quality, high-context data is a must. This will ensure records are not full of gaps or lacking in vital background details (context). This is a great and necessary first step along the road.

I urge you to read John Nosta’s recent article in Forbes.

He explains how digital technology has failed to keep pace with clinical innovations. I like to compare it to how the telecoms industry was 20 years ago – data is siloed, private and in no format to be shared. What changed? A shift change in connectivity of data (and then devices) that has changed the world; healthcare must do the same. In this digital age, where Electronic Health Records (EHRs) are everywhere, much of practical, medical decision-making continues to be paper and document driven. This makes data-needy questions even harder to answer. It makes insight very localized and limits information sharing and collaboration. There is certainly progress, through outcome access to EHRs driven by HIPAA (USA) or using national databases like CRPD (UK), and there is a growing rash of personal genomic informatics and ‘quantified self’ startups. However barriers to effective collaboration between the stakeholders of patient, clinician, pharma and payer still remain. As Dr Eric Topol espouses: patients – the consumers of healthcare, which now costs the USA one hundred times what it did in 1960 – deserve evidence-based medical care and that evidence is a complex mixture of genomic, outcome and research data.

The health sciences world is waking up to the reality that advances in healthcare will be enabled by advances in our ability to use data better – from the bench to the bedside and to the boardrooms of hospitals. It is important to stress that we are talking about high quality, high context data. As DrSalluzzo points out, automating a bad pharmacy process is hard to correct. The quality of data is intrinsic to its value. In truth, so much of the data patients are asking clinicians and researchers to rely upon is full of gaps. The insights gained from bringing together the best available clinical, research and biochemical data drive valuable changes in clinical practice, address data gaps and advance medical care.

Better data, better decisions, better life.