Innovaccer And Value-Based Healthcare In The Age Of COVID-19

by Thomas H. Davenport and Randy Bean -

Over the past decade or more, we’ve been hearing the mantra of “value, not volume” when we speak to healthcare executives. One doesn’t have to look terribly hard for the reasons why payers, providers, and consumers of healthcare in the U.S. would seek more value. It is widely reported that as a nation we spend far more per capita than any other country in the world on healthcare, but our outcomes are nowhere near the top. It seems likely that the US will also not come out on top when the cost per capita of treating the coronavirus is compared to the effectiveness of treatment in the country.

Thus the various stakeholders in the industry have been seeking greater value in care, although relatively little progress has been made in the U.S. healthcare system. There are many reasons for slow progress, including the complex structure of the industry and vested interests that want to preserve their roles and profits. Those factors, however, will be difficult to change and will require considerable time and political energy. Our focus is on how information and technology can help to address the problem.

Information is not a new perspective on the value-based care issue; the idea of a “medical home” was first conceived by the American Academy of Pediatrics in 1967. It argued that the best decisions about child health could be made only if all information about the child’s medical history could be brought together in one place—a physical or virtual home for a patient’s care. That is still a key assumption behind the “patient centered medical home” (PCMH) as it is defined today. Other assumptions, according to the U.S. Agency for Healthcare Research and Quality, include that the large majority of care be provided by one team of care providers, that care is for the whole patient, that care is coordinated across multiple provider venues, that services are easily accessed, and that care is evidence-based and of high quality.

One description from Harvard Medical School described how the PCMH could work in practice:

Your primary care physician will be one member of a team who will offer comprehensive care all under one “roof.” Moving forward, people will pay for their health insurance, and their primary care doctor will receive one flat payment from insurance to cover most of the care provided. The amount will be based on the patient’s health issues and complexity. There will be far less “fee-for-service,” that is, billing for each visit. Services such as behavioral health and nutrition will be located in the office. Physicians and patients will determine specific health goals, which can then result in bonus incentives. Practices will be rewarded for things like helping a patient lose weight and get blood sugar under control — that is, for keeping them healthy and out of the emergency room and hospital.

Unfortunately this type of care is still more a vision than a reality for most patients in the U.S. healthcare system.

What Are the Technical Obstacles to Value-Based Care?

Lack of information integration in the healthcare system is one primary reason why value-based care through a PCMH is not often provided. It is difficult or impossible to coordinate care in a medical home if the provider doesn’t have all the information about a patient’s medical situation and past care. In provider organizations, the electronic medical record (EMR) is the primary vehicle for storing past medical history and care episodes. However, it is rare that EMRs have complete records of care for a patient across all provider organizations. Important healthcare information is also contained in health insurance claims, and can also be found in physician practice management systems, health information exchanges, and provider revenue cycle systems.

In addition to data integration, technology to support value-based care would need to be able to:

• Accurately predict the cost of care management for a particular population;
• Manage the workflow of interacting with patients in a value-based approach;
• Show physicians the cost implications of different care protocols;
• Make recommendations for how to improve health and lower costs; and
• Offer patients targeted behavioral nudges to improve their own health.

Innovaccer and Value-Based Care

Until recently, it was difficult or impossible to integrate information from all of these sources on a patient and employ it in a PCMH context. But that is the focus of Innovaccer, a “healthcare data activation” company founded in 2014 and based in San Francisco. Innovaccer offers data integration across all of the sources I have mentioned, applies the data in a care management workflow context, and supplies analytics on care and population health. Value-based care is possible because the provider knows what care costs, and can apply preventative care and early interventions into expensive health problems. Paul Grundy, a physician who is often described as the “godfather” of the PCMH movement, is the company’s Chief Transformation Officer.

When we spoke with Grundy, he said that three things have to happen in order to move to the PCMH model:

• Health plans have to pay for managed care rather than fees for services;
• There needs to be a cultural shift in medical training to focus on the whole patient and establishing a relationship of trust; this is most important for primary care and internal medicine
• Technology has to allow clinicians to get at all the data on a patient, including the “social determinants of health” that have a strong influence on wellness but are not currently captured in healthcare systems.

Grundy said that progress is being made on all three factors, but he is working with Innovaccer on the technology. He elaborated that the software can take in data from 70 different sources—pulling it in, processing it, cleaning it, and managing it over time—and then put it in front of the physician. For a particular patient it might identify three gaps in care, and four medical or health interventions that would have the most impact on the patient.

Grundy argued in our interview and in an article on managed care that we need AI-based prediction in our healthcare systems. He believes that EMRs are not very helpful to the PCMH because they are based on backward-looking billing processes and information. He feels that the healthcare industry needs to move beyond simply capturing what was done to a patient, and move toward predictions of what might happen and recommendations for how to prevent medical problems before they occur.

While Innovaccer does not yet have the full capabilities to predict what medical problems a patient will encounter in the future or make recommendations for how to prevent them, it is using a broad collection of patient data to predict an individual patient’s risk score—the future cost of care. Such predictions could be very useful to providers or payers in contractual negotiations over the cost of care management for a patient.

These capabilities are also useful in treating the coronavirus and COVID-19. Innovaccer has made available (at no cost) several new capabilities allowing healthcare providers to offer telehealth-based advice to their customers on whether or not they are infected with the virus. It has also created COVID-19 dashboards for providers to monitor their patients with the disease. 

Physicians of Southwest Washington (PSW), based in a state that was an early “hot spot” for COVID-19, quickly adopted the Innovaccer’s Covid Management System for assisting patients who were considered at high risk for serious illness in their ACO. Tamra Ruymann, Chief of Digital Health at PSW stated:

“Utilizing Innovaccer’s Covid dashboards we were able to identify high-risk patients and our care management team implemented the Innovaccer telehealth solution for video visits to assess if the social needs were met, such as food and medication needs to reduce the likelihood of being exposed.”

Innovaccer at a Large Healthcare System

Another Innovaccer customer is a Midwestern system of healthcare facilities and services with multiple clinics, medical centers, hospitals and care locations. The head of population health analytics leads the use of Innovaccer there. They’ve been using the company’s software for about five years. They primarily use it for data integration that enables population health use cases and managed care contracts. The system has many large care management contracts with Medicare, Medicaid, and commercial payers.

The population health director says the system has a variety of EMRs, and Innovaccer allows the integration of patient data from all of them. They integrate payer claims data into their data warehouse using the software. In addition to the data integration capabilities of Innovaccer, the chain also uses the company’s care management platform as its primary workflow tool. It manages the entire patient care process, from admission to discharge to ongoing monitoring and patient contact. The healthcare system also uses Innovaccer’s analytics dashboard, but does not do prediction or recommendation using the software. The head of population health analytics argues that predictive and prescriptive analytics still need substantial human interaction to tune the models, so he is not focused on automating that type of analysis or training a machine to do it.

Overall, however, the customer is quite positive about the capabilities Innovaccer offers the system. He noted:

“If you want to expand the complexity and number of managed care contracts, you have to have a partner like Innovaccer. It’s just not possible otherwise.”

Value-based healthcare, despite being discussed for decades, is still in the early stages in the United States. Political discussions about who will pay for care in the future will certainly influence the discussion. However, there is little doubt that information integration will be necessary to deliver care at higher quality and lower cost, and that predictive and prescriptive analytics will ultimately be necessary as well. The healthcare system will remain at least somewhat fragmented, and the software Innovaccer has can help to knit it together.