By Sheila Talton | December 18, 2017
A few years ago, we had landed a few consulting engagements, including one with a large health system. We built a data management strategy for the system and then asked its technology leaders about their data-sharing capabilities and how they managed data integrity.
The client assured us that they used one of the larger electronic health record systems, and they felt their data quality problems were solvable with a few process changes.
But as we started digging into their infrastructure, my team and I quickly came to see that in fact the processes were not their only problems. Different parts of the organization used different point solutions, purchased from different vendors. Each solution created and stored data in a different way, and the sharing of data with other functions inside the organization was limited. Data on clinical outcomes, finances, staffing and patient experience resided in separate places – despite the fact that each of those functions depend on one another to improve operations. There was no existing enterprise solution that could take data from all the different departments, then standardize and normalize it to provide a holistic picture of performance, quality and cost of care.
Those findings led me to a revelation: Healthcare providers had all the data they needed to generate predictive, actionable insights for their patients and their businesses, but aggregating and integrating the data was another story. They also lacked the resources to prepare the data in a way that was timely, cost-effective and would produce insights to drive quality-of-care decisions and, in turn, significantly affect outcomes.
That insight, combined with two decades of helping other industries build effective business solutions that used data effectively, led me to start a company, Gray Matter Analytics. I wanted to see if data analytics technologies that have been applied in other industries – housed in an efficient cloud-based solution that would automate the data-integration process – could be applied in healthcare, both to improve the quality of care and to better manage the cost of care.
Demonstrating Outcomes Requires Better Analytics
I knew that creating a cloud-based data-analytics company would present a way of delivering solutions unlike anything healthcare providers had experienced in the past, and I believed the timing was right. Providers are under immense pressure to improve quality of care and reduce the cost of care. Even before the Affordable Care Act, the Centers for Medicare and Medicaid Services (CMS) had created significant incentives for providers around quality-of-care metrics – and significant penalties for failing to show improvement on those measures and on readmissions.
Those penalties have become a source of tremendous frustration for providers who believe that they are making progress, but can’t demonstrate it. That’s not because they don’t have the data – indeed, they possess data on everything that happens to every patient who comes through their doors. They can of course draw some conclusions from looking at the data generated by point solutions scattered around the organization, but the delivery of medicine is a team effort and the patient journey touches many departments.
Truly painting a picture of quality and performance requires capturing the full patient journey, from the admissions process to clinical procedures across practice and functional areas to outcomes. And more importantly, we don’t want to just report on what happened, we want the ability to predict what will happen if certain procedures are not performed in acute and post-acute situations. Then we can empower care mangers to intervene and prevent adverse outcomes and readmissions.
Predicting the Future
Take for example a system that seeks to better manage its cost of caring for patients with diabetes. Today the system is primarily, if not exclusively, focused on the population that has been diagnosed as diabetic. But its healthcare practitioners need to look at those patients who are likely to contract diabetes, based on weight, blood sugar, hereditary and environmental factors. Somewhere in the health system, all that data is available, but neither the physician nor the care manager likely has access to the data – or, more importantly, to the insights it could provide.
But what if the Chief Medical Officer could review the patient population, segment it by risk factors and create preventive care coordination, outreach and patient engagement to reduce the chances of contracting the disease? At the same time, the system’s CFO could model the cost of caring for the patient population that might contract diabetes and begin to plan on how to manage the cost of care. That information would guide her in negotiating contracts with payers. Most critically, the system’s leadership would have the ability to monitor – daily, weekly, monthly – trends among this patient population and be proactive in care and cost management. These types of interventions are possible only if the system implements data analytic solutions that creatively access the required data using predictive analytics.
But the technology is a small part of the equation. Integrating technology into daily workflow is critical, as is the evaluation and maintenance of algorithms. Implementing predictive analytic tools and solutions in healthcare is a means to an end — improved outcomes, including lower costs. Fully realizing the benefits from a specific solution requires a structured and thoughtful approach, involving the right people, with the right skill sets, at the right time. The key to successful implementation has little to do with the solutions or models, for the most part. Success depends on the time, effort and resources set aside for communication, change management and making analytic solutions a seamless part of user workflow.
It is not possible for the average health system to embark on this journey without partners. The implementation requires skills that are beyond the expertise usually found inside a health system. Just as health systems use professional real estate managers to manage their buildings, they need solution architects, data scientists, advisory experts in change management and process/workflow to help manage the application of data analytics to the delivery of care.
A New Kind of Solution
I believe these partnerships should be structured around successful outcomes – much like value-based care itself. The nature of the partnership should not be episodic but continuous, because quality-of-care improvement is continuous, as is the pace of change. The solutions must be cost effective, flexible and quickly implementable. They must be able to manage data quality and accessibility issues with tools that transform the data to make it useable. Further, that transformation should not take years, but a few months. Healthcare organizations must react quickly to the changes that are happening to payment models and to their patient populations.
We need clinicians at the head of these analytics solution partnerships. Payers and CMS are creating outcome and quality metrics, but health systems must incorporate quality metrics from their clinicians. Clinicians understand and have innovative ideas on quality improvements that affect the cost of care, but too often their metrics are not part of the solution. That must change.
Finally, internal IT organizations need to be incented to simplify their technology environments and to stop building solutions themselves; instead they should think about partnering with solution providers. Building solutions internally is expensive and time-consuming. Often by the time the solution is delivered, the problem has changed. Healthcare systems don’t have the time, money or resources to keep running on this hamster wheel. CIOs must embrace the complexity of their data, recognize the lack of resources and time to implement, and be open to partnerships that will deliver timely, cost-effective value.
How Success Happens
During my career in technology and my service as a director on public company boards in industries like manufacturing, utilities, banking and consumer products, I have reflected on how successful change happens in business. I believe the most important component is the commitment of the organization’s professionals, including senior leadership, to embrace, not loath, change. Creating a culture of embracing and valuing partnerships is also essential. This means recognizing that the business’s core competencies are manufacturing, delivering energy, etc. – not being a software developer or a technology integrator – and partnering with specialists who provide the competencies needed for change.
For example, when I supported Cisco in expanding its presence in China, India, Mexico and Brazil, I saw a clear recipe for success in the governments’ use of private-public partnerships. It is important to recognize the core competencies you have, and the core competencies you need to acquire from a partner to be successful. In the case of healthcare, providers’ core competencies are diagnosing complex health conditions and applying the right care pathways to heal patients. Healthcare does not have, nor should it have, all the core competencies in technology, data management, data analytics application development, and change management. It needs a group of people who can support efforts to use these technologies and be the interface to manage its partners.
I believe we are on the cusp of a revolution. That revolution will unleash the power of data to improve quality of care and the efficiency of care delivery. Partnerships with clinicians, IT, healthcare administrators, solution and advisory services providers, together with alignment of shared goals and financial values, will be the essential ingredients in healthcare’s success when it comes to using data to improve the quality of care. Using technology and predictive analytics to enhance the clinicians’ effectiveness will be the keys to unlocking innovation in care delivery.