Health IT and Data: Don’t Forget the Patient

Anil Jain
Anil Jain

Guest post by Anil Jain, MD, FACP, chief medical officer, Explorys, and staff, Department of Internal Medicine, Cleveland Clinic.

Nearly every aspect of our lives has been touched by advances in information technology, from searching to shopping and from calling to computing. Given the significant economic implications of spending 18 percent of our GDP, and the lack of a proportional impact on quality, there has been a concerted effort to promote the use of health information technology to drive better care at a lower cost. As part of the 2009 American Reinvestment and Recovery Act (ARRA), the Health Information Technology for Economic and Clinical Health (HITECH) Act incentivized the acquisition and adoption of the “meaningful use” of health IT.

Even prior to the HITECH Act, patient care had been profoundly impacted by the use of health information technology. Over the last decade we had seen significant adoption of electronic health records (EHRs), use of patient portals, creation of clinical data repositories and deployment of population health management (PHM) platforms — this has been accelerated even more over the last several years. These health IT tools have given rise to an environment in which providers, researchers, patients and policy experts are empowered for the first time to make clinically enabled data-driven decisions that not only at the population level but also at the individual person level. Not only did the 2010 Affordable Care Act (ACA) reform insurance, but it also has created incentive structures for payment reform models for participating health systems. The ability to assume risk on reimbursement requires leveraging clinical and claims data to understand the characteristics and needs of the contracted population. With this gradual shift of risk moving from health plans and payers to the provider, the need to empower providers with health IT tools is even more critical.

Many companies such as Explorys, a big data health analytics company spun-out from the Cleveland Clinic in 2009, experienced significant growth because of the need to be able to integrate, aggregate and analyze large amounts of information to make the right decision for the right patient at the right time. While EHRs are the workflow tool of choice at the point-of-care, an organization assuming both the clinical and financial risk for their patients/members needs a platform that can aggregate data from disparate sources.  The growth of value-based care arrangements is increasing at a staggering rate – many organizations estimate that by 2017, approximately 15 percent to 20 percent of their patients will be in some form of risk-sharing arrangement, such as an Accountable Care Organization (ACO). Already today, there are currently several hundred commercial and Medicare-based ACOs across the U.S.

There is no doubt that there are operational efficiencies gained in a data-driven health system, such as better documentation, streamlined coding, less manual charting, scheduling and billing, etc. But the advantages of having data exhaust from health IT systems when done with the patient in mind extend to clinical improvements with care as well.  We know that data-focused health IT is a necessary component of the “triple-aim.” Coined by Dr. Donald Berwick, former administrator of the Centers for Medicare and Medicaid Services (CMS), the “triple-aim” consists of the following goals:  1) improving health and wellness of the individual; 2) improving the health and wellness of the population and 3) reducing the per-capita health care cost. To achieve these noble objectives providers need to use evidence-based guidelines to do the right thing for the right patient and the right time; provide transparency to reduce unnecessary or wasteful care across patients; provide predictive analytics to prospectively identify patients from the population that need additional resources and finally, use the big data to inform and enhance net new knowledge discovery.

Provider adherence to guidelines remains low causing omissions of critical care processes or introducing harm. The appropriate use of computable guidelines into health IT tools have led to best practice alerts, order sets and facilitated documentation. Despite real concerns over the context and volume of alerts, many studies have shown that this form of decision support has led to improved adherence of clinical guidelines resulting in improved outcomes. These studies have been shown to enhance not only wellness by recommending immunizations or screening tests when overdue, but also in managing chronic disease such as in the case of recommending the initiation of insulin therapy in a diabetic not well controlled with oral pills.

However, most EHRs cannot balance the nuances of the complex patient with the constraints of its decision support tool. The big data analytic solutions can further augment the EHRs native ability to deliver alerts at the point-of-care with advanced predictive decision support based on data aggregated from multiple sources, analyzed in complex algorithms and then delivered in visualizations that simplify prioritized actions. The patient truly benefits when all available data is used to arrive at the best personalized course of action. These alerts can also be delivered to patients through patient portals allowing them to make informed decisions away from the point-of-care.

The impact of measuring unnecessary or wasteful care cannot be overstated. An Institute of Medicine report in 2012 suggested that more than $200 billion is spent on care that does not benefit the patient. Certainly avoiding that care when known would benefit the patient. In addition, providing this data back to providers not only allows them to see their performance across their attributable population, but if done a risk-adjusted manner, they also see how they compare to their colleagues – powerful motivation for behavior change. Most providers, without access to this data, really do not have a sense of their pattern of resource utilization for their patients and how much of it may not be necessary.  Providers typically focus on one patient at a time. However, armed with this data, especially when involved in a value-based care contract, the provider has every incentive and opportunity to optimize care for his or her patient.

It is impossible to talk about the value of big data without also speaking to the analytics that allow that data to be converted to useful information and knowledge. For example, at Explorys, data scientists use more than 200 billion data elements to produce predictive models encompassing more than a 100 variables to risk stratify patients into risk bins. These risk bins represent the patient’s individual risk of being readmitted to the hospital after a recent discharge for congestive heart failure.

Patients who may be in the highest risk bin can be offered more resources during the hospitalization and immediate post-discharge period to reduce the likelihood of re-admission. The benefits to the patients are obvious; moreover, in a value-based reimbursement model, the economic benefits to the health system are great motivators for either developing and building a data strategy or partnering with those that offer a solution. Organizations do realize that size matters when it comes to predicting events and a very large set of data is necessary when attempting to build the best models of care and predict clinical events or costs.

Finally, big data has a direct impact on knowledge consumption and creation. In the traditional knowledge transfer paradigm, bench to bedside (or research to bedside), it is often said that patients reap the benefits of new discoveries approximately seven to 10 years later. Perhaps because of the rigors of the scientific method, various communication silos, or delayed adoption because of our education and training process, this delay is addressable with health IT. Whatever the reason, health IT does simplify the dissemination of knowledge at the point-of-care by allowing systems to deliver information relevant to a specific patient; e.g., many EHRs have linked their systems within online textbooks and references.

Dissemination of information that a patient may be a candidate for a novel therapy introduced in the market or that the combination of several biomarkers examined in the background should prompt further diagnostic testing all benefit the individual patient. When we couple this delivery of knowledge to what is most cost-effective and efficacious, perhaps we can call it “bench to bedside to bottom-line.” This is the direction that benefits patients most especially when looking at patients on high-deductible plans.

Moreover, in addition to driving knowledge to the point-of-care, big data also drives net new knowledge discovery through real-world data. The ability to discover patterns and correlations in the information can lead to new research questions or even provide new insight into widely accepted treatments. This type of data mining is not a replacement for traditional randomized clinical trials, but in the appropriate setting can be a cost-effective precision guidance system for researchers. Moreover, this real-world data gives us the ability of assessing what actually happens to patients when cared for outside of the artificial environment of a clinical trial. This type of surveillance on a very large clinical data set can detect harm and benefit long before that signal would be detected by a single health system. It is no wonder than groups, public and private, are working on ways to safely and securely accumulate large volumes of patient data to provide this needed service. For example, Explorys has accumulated years of de-identified aggregated data from nearly 40 million patients for analysis to improve care for everyone.

Health IT and the associated tools that generate, aggregate and analyze data are a key component of improving the health and wellness of our population while reducing cost and whether done one patient at a time or a population at a time, providers need the health IT tools to address the data demands that are being placed on them to address the challenges and opportunities of the healthcare transformation occurring around them.

Write a Comment

Your email address will not be published. Required fields are marked *