Changing the data paradigm in healthcare


Our choices on interoperability will determine the digital transformation of entire industries

“No man is an island

Entire of itself,

Every man is a piece of

the continent

A part of the main.”

— John Donne

As the world becomes more connected and megabytes of data are shared every second, no company can afford to remain an island today. Interoperability is the future of a technology-led world where machine learning is expected to solve deep-seated problems of poverty and healthcare.

The recent stand-off between Flipkart owned e-wallet Phone Pe and ICICI Bank demonstrates how critical the issue of interoperability is. As India hurtles towards a cashless economy, an interoperable financial system is a prerequisite. A MobiKwik user should be able to transact at a store where the merchant uses another wallet, say Paytm.

The United Payments Interface set up by the National Payments Corporation of India does lay down the infrastructure for interoperability. However, there is a lack of consensus on the governance of this new interoperable ecosystem. Until that is fixed, we won’t see the full impact of mobile money.

Hopefully, it will only be a matter of time before it gets sorted out as the financial sector has always been a leader in interoperability. In the past, banks have done it with ATMs and cards, connecting systems and standards at a global scale so that a customer can access money from anywhere, anytime.

We have also seen successful examples of interoperability in communication-oriented industries such as telecom, where networks went from walled gardens to open systems.

What ails healthcare?

Surprisingly, the healthcare sector, where interoperability is of critical importance, still remains a walled garden.

If we look at the healthcare industry today, it is composed of a series of closed ecosystems within which information remains trapped. Every healthcare provider has its own diagnostic labs and attendant partner organisations that together work to solve a patient’s problems but there is not much sharing of data outside that loop. Patients are not allowed to move from one ecosystem to the other with portability of data. Worse still, the data is almost exclusively owned by providers rather than patients, with every healthcare provider trying to gain advantage by owning the patient’s data and locking them into its system. This has prevented the creation of a patient-centric healthcare system.

In an ideal world, a patient could move his or her data from one healthcare provider to another completely seamlessly and in real time leading to better outcomes and cost savings. Efforts along these lines are already underway in the EU but still face significant regulatory lobbying.

Interoperability also has a big impact on machine learning. The less interoperable an ecosystem is the less effective is machine learning. Take, for example, an area such as cancer diagnostics.

If you have enough data inputs indicative of cancer, you don’t really need specialists to make the diagnosis — an algorithm can do it. The problem here is that without interoperability, you are not going to see significant learning effect in machines, as it won’t accumulate adequate knowledge to make the correct call.

Fragmented strategies

A good example of interoperability that has helped in public good is the way airlines decided early on to share data to make flying safer. This collaborative approach where airlines and airports shared exact location data of crafts not only has resulted in safer skies, but reduced delays and helped cut costs.

Unless we see that level of sharing, it will be difficult for technology to solve problems. Take the example of self-driving cars where every company is following a fragmented strategy.

Tesla has a lot of data but not enough to create a fully self-driving environment. Google has a lot of non driving data. All car companies have islands of data. But none of these are talking to each other. As a result, the overall pace at which self-driving technology is progressing — while rapid — is still much slower than if these systems were interoperable.

The reason there is so much reluctance to share data is that with interoperability, companies face a loss of control. Perhaps governments may need to build incentives for interoperability to open the gates for collaboration.

As for companies, they will need to tweak their business models to take advantage of an interoperable environment, and may even need to create alternate sources of monetisation. Too many companies have built business models that rely on closed data paradigms.

Open data initiatives

Globally, there are a host of initiatives to break through these silos of communication. A few years ago, Gates Foundation and Markets for Good announced a Grand Challenge of data interoperability. The idea was that if data from public health statistics, private clinics, poverty tracking, etc. could all be viewed together, it might lead to breakthrough solutions and create a financially inclusive society. Several companies have committed themselves to the Gates Foundation initiative. Some healthcare IT start ups have emerged that are bravely trying to “unbreak” the silos. But the roadblocks persist. In the future, not just healthcare, but every industry that is transitioning to digital ecosystems will face the challenge of interoperability. All industries will need learning systems, some of which will achieve truly effective knowledge only through open, interoperable data. Understanding data as public vs. private good will gain importance. For that, governments will need to step in with unifying rules, standards and procedures. Companies, on the other hand, will need to innovate towards new business models that rely on open data paradigms to ensure faster value creation at the industry level without loss of value capture at the company level.

Paul Choudary is the author of Platform Scale and Platform Revolution and the founder of Platformation Labs. Narayanan is Editorial Consultant

Published on February 07, 2017