Big Data & Analytics Heroes
Award-winning inventor and businessman Shahid Shah’s perspective on big data in healthcare draws on more than a quarter-century of technological experience. Keep reading to catch his vision for how data analytics and the Internet of Things (IoT) will change the face of healthcare.
How are big data and analytics changing business strategy for healthcare?
The US healthcare system has traditionally compensated independent providers with fees for diagnosing and treating patients (the fee-for-service model, or FFS). That system created the world’s best care, but the costs of providing such healthcare have risen at unsustainable rates, an unintended consequence of the system’s emphasis on volume: The more care they have provided, the more reimbursement providers have received. Thus FFS creates an incentive for healthcare providers to provide expensive services, services that bring in more money than other options do.
Even worse, FFS sometimes encourages healthcare providers to abstain from “doing right” by their patients, because sometimes healthcare providers are not reimbursed for doing right. For example, patient education, telemedicine and wellness advice are all known to improve health, but providers aren’t paid to provide such services, so they don’t encourage them.
To help arrest cost growth, the healthcare industry is gradually abandoning FFS in favor of a value-based collaborative reimbursement model in which healthcare providers work together to provide care. In this model, providers are compensated for the results and outcomes they achieve rather than simply collecting fees for the services they perform.
Encouraging though it is, the move from FFS to value-driven care is a paradigm shift in healthcare, and it will require unprecedented access to data and analytics. Personalized medicine, patient engagement, care coordination, treatment engines, predictive diagnostics and bundled payments are just some of the benefits of value-driven care, but none of these is possible without effective big data tools and next-generation analytics approaches.
How are the Internet of Things and wearables reshaping healthcare?
Almost all healthcare treatments require top-notch diagnostics—after all, the more you know about a person’s illness, the more able you are to treat that illness. Better still, the more you know about an entire population’s illness, the more effectively you can create drugs and care plans for specific groups of patients within that population. Thus the Internet of Things and healthcare wearables are revolutionary if for no other reason than that they create opportunities for self-service in healthcare.
What’s more, the healthcare IoT is making possible remote data capture and diagnostics, obviating the need for an expert caregiver to be in the same physical location as a patient. As healthcare wearables and IoT devices begin to reshape medicine, the types and amounts of data that become available will create an environment in which machine learning drives both prescriptive and predictive analytics, and all in the service of patients.
What are the biggest challenges healthcare providers face when getting started with big data and analytics?
The biggest challenges are probably the regulatory and structural design of the healthcare industry itself. Unlike retail, or other industries that compete directly for consumers’ dollars, the healthcare industry pays for services through a reimbursement, or intermediated payment, model. In short, patients don’t pay directly for services—rather, insurance companies pay, or the government pays.
Such intermediation hinders industry attempts to get data, then analyze it and use it to create products and services. Regulatory and payment impediments abound for those trying to make use of healthcare data, and running afoul of them is discouragingly easy.
Companies such as Amazon rely on real-time data to inform their recommendations of products and services, and credit reporting companies can let an automobile dealer know whether a prospective buyer can afford a loan. But such innovation is difficult to achieve in healthcare because of payment and regulatory hurdles. For example, even though automated data-driven remote diagnostics and telemedical systems can provide superior care for many common ailments, healthcare providers do not receive reimbursement for offering such innovative methods of treatment.
In terms of big data and analytics really delivering a return on investment, what is the market still missing?
Computing return on investment (ROI) requires understanding costs, prices and revenue associated with products and service—not just at a surface level, but in painstaking detail. Unfortunately, very few healthcare insurers or providers have a clear and detailed view of the costs they incur when they perform complex services for patients. Providers who seek a clear picture are often unable to tie insurance payments to specific patients’ costs, never before having been required to do so—rather, they are accustomed to simply applying fees to broad categories of services. But without metrics in place to understand costs, healthcare providers are unable to realize the full benefits offered by data analytics.
Another significant obstacle to proper delivery of ROI for big data analytics is the dearth of data-driven processes and procedures in patient care. Most standards of care, or care plans, have been determined not through rigorous use of population-size data analytics, but rather by local trial and error over long periods. Moreover, they are sometimes the result of what some call “eminence-driven medicine,” in which institutional or personal reputation drives care delivery decisions.
Though physicians are certainly taught about data and how they can use data to conduct rigorous clinical trials, the amounts of data once available to physicians are dwarfed by the quantities of data available today. For clinicians trained in the data use techniques of yesteryear, next-generation gene sequencing, modern electronic health records, modern diagnostics devices and patient-generated healthcare data can be unmanageably difficult—not through any fault of the physician, but because the physician’s training in big data and analytics tools is inadequate to deliver an acceptable return on investment.
How do you think big data and analytics will handle the data growth over the next 10 to 15 years? Do you think the market requires another shift in technology?
Patient care specifically, as well as the business of medicine in general, has been in the hands of humans—particularly clinicians—for most of its existence. Caregivers have operated in an isolated insight economy for many years, deriving insights from clinical trials, literature and other slow-moving data. Accordingly, although they have practiced cognitive pattern matching since the beginning, they have worked without access to large, fast-moving data sets, leaving them to learn by trial and error. Indeed, many modern treatments are the result of the tried-and-true techniques we lovingly call “clinical trials.”
To inaugurate medicine as a full participant of the insight economy, big data and analytics must bring machines and software into the diagnostics and treatment mix—perhaps to a greater degree than some care providers will find comfortable. But we must become comfortable with machine-assisted diagnostics—just as we have become comfortable taking directions from a smartphone GPS app, and just as we have become comfortable riding in aircraft that can take off and land themselves using the power of geographical and meteorological data.
Over the next decade, we must learn how to make clinical diagnostics and analytics tools transparent, allowing us to converse with, say, our smartphones as we do with physicians today. In the coming cognitive era, such devices will offer in-depth knowledge about the internal state of our body, the quality of our diet and the nature of our exposure to disease vectors, including during travel. Using a wealth of information, big data and analytics tools will be able to give patients and their caregivers treatment suggestions that would not otherwise be available. In the same way that Siri and Cortana tell us about the weather using data gathered by millions of sensors around the world, next-generation analytical tools will bring us significant clinical tools hidden inside our wearable and handheld devices.
If we are to make the most of the next 10 to 15 years, tools such as IBM Watson will need to “go to medical school” to learn about patient care alongside their human counterparts, allowing coming generations of students to accustom themselves to taking diagnostic data and treatment suggestions from such tools. Moreover, in view of the complexity of the human body, another shift in both technology and human behavior must also occur, creating interfaces between humans and machines at whose nature we can only guess today. Just as modern pilots learn to treat complex aircraft machinery as an extension of their own body, so the clinicians of the future will understand not only their own abilities, but also the abilities of the computers they use—learning when to rely on the tools they use and when to rely on their own strength and, in doing so, achieving the full potential offered by the union of technology and person.