Executive Snapshot

Figure 1

Source: IDC, 2020

December 2020, IDC #US47023720 SITUATION OVERVIEW

The COVID-19 pandemic gave urgency to a legacy challenge for healthcare organizations: the need for actionable insights to drive tailored care programs for high-risk individuals. This IDC Perspective showcases a health plan that successfully identified members at risk of poor outcomes from COVID 19, which drove member outreach. Effective programs for high-risk individuals must identify the drivers of risk and align them with the appropriate services to mitigate risk. While there is constant discussion in the healthcare industry of the need for actionable insights, the progress has been slow, even with the use of machine learning (ML) and artificial intelligence (AI).

The COVID-19 surge gave rise to the development of machine learning algorithms to identify at-risk populations. Many methodologies that define risk are rules based, using Centers for Disease Control and Prevention (CDC) or World Health Organization (WHO) guidance. Such approaches identify more individuals than any organization could effectively manage. Blue Shield of California and Cogitativo (www.cogitativo.com) took a different approach. Cogitativo used its data science platform for healthcare to build ML algorithms to identify high-risk individuals based on a variety of factors in order to help Blue Shield prioritize members for outreach and to develop programs to reduce risk factors.

The "risk-scoring engine," which was developed as a result of the collaboration, identifies the individuals at risk clinically and structurally because of COVID-19. Using the Blue Shield data on its 4 million members, geolocation data, and early results from research on the progression and the outcome of individuals with COVID-19 and other SARS viruses, Cogitativo was able to identify at-risk populations based on a variety of factors. By parsing out individuals into homogenous populations (in terms of COVID-19 risk), tailored outreach programs were developed by Blue Shield.

The outreach was part of Blue Shield's focus on using technology to deliver personalized care, in this case, by identifying and contacting members with heightened needs during the pandemic. The approach included members who tested positive as well as those most likely to suffer serious negative outcomes if they got COVID-19. Combining AI technology with old-fashioned personal connection, Blue Shield was able to reach tens of thousands of members to offer personalized options and advice. The vignette described in the following section demonstrates the value of balancing technology with personal touch. It illustrates that the path to health requires more than clinical care and that, when a health plan demonstrates that it cares, it drives much higher member engagement and loyalty.

Marissa's Story

After struggling through weeks of intense body aches and fatigue with COVID-19, Marissa Duenas found herself feeling paradoxically grateful that she had contracted the disease.

"I lost my husband last year," Duenas said. "When I had COVID, it was so stressful without him; still, [it] helped me to realize, 'Marissa, you have your insurance, you have your job, you have your family and your friends. You have to keep going!'"

One thing that helped Duenas persevere was getting a call from a woman she refers to as her "angel" — a Blue Shield of California case manager who checked in to find out what she might need to get through the disease safely.

Duenas, 68, lives alone. Her adult daughters had taken in Duenas' dog, so she did not have to care for it during her illness. Duenas' biggest challenge was buying and preparing food. The case manager was able to set her up with two weeks of meal delivery, a service that Blue Shield has been providing to members who test positive for COVID-19 during the pandemic.

"Just to feel that somebody's calling you to know how you're doing, it's probably the best experience I ever had," Duenas said. "Knowing that other people care for you, and your insurance is there when you need it."

The Blue Shield of California and Cogitativo collaboration checks all the boxes of a successful strategy to manage high-risk patients. The available data includes a broad range of sources, the algorithms delivered insights that were actionable, and the interventions went well beyond the clinical needs of patients. The balance of technology and personal touch, and its impact on member loyalty, was a key component of success.

The Process

Cogitativo combines its knowledge of healthcare with the academic experience of its data scientists to assist clients in understanding data and how it can be applied to complex healthcare challenges. The company's methodology consists of:

• Defining the challenge/problem
• Ingesting noisy structured and unstructured data
• Uncovering the insights
• Making actionable recommendations on process changes that allows clients to get solutions that are based on the intellectual property frameworks brought by Cogitativo, combined with the client's unique capabilities

Cogitativo collaborated with Blue Shield of California to identify a "reasonable" number of members who, if they got COVID-19, would be more likely to have a poor outcome. With a membership of 4 million, it was critical for Blue Shield to prioritize its outreach to a manageable number of individuals. Like most health plans, Blue Shield sponsored community outreach and education around COVID-19 but recognized that, for the very highest risk patients, a more personalized outreach was required.

Cogitativo augmented the Blue Shield claims data with research results from studies on outcomes from COVID-19 and other SARS viruses. It built a "predictive engine" and provided an interactive visualization that Blue Shield could use to access the data tables that identified the various factors that might result in increased risk to its members. For example, in identifying individuals who had not seen a physician in over a year, Blue Shield could predict the number of individuals in the risk cohort and the recommended action. Blue Shield could then customize its approach to outreach. With support from Blue Shield, Cogitativo is now discussing use cases within various U.S. government agencies.

The Next Steps

Blue Shield's initial success with a broadened understanding of initial risk factors of COVID-19 has set the organization on a path to identify and address other factors that represent high risk. The work has begun to identify and reach out to individuals of races and ethnicities known to be at a higher risk of exposure to COVID-19. Those populations include African Americans, Latinx, Native Americans, and Pacific Islanders.

Blue Shield and Cogitativo are also working on algorithms for prioritizing the distribution of the COVID 19 vaccine, once available. The challenges of supply chain management and distribution logistics that lie ahead are unprecedented. The ability to prioritize the most vulnerable individuals based on factors beyond "clinical risk" and to proactively reach out to individuals to ensure that they receive the needed care will go a long way toward managing the outbreak of COVID-19.

ADVICE FOR THE TECHNOLOGY BUYER

There are many lessons learned that healthcare organizations should take into consideration as a technology buyer, including the following:

• Be prepared to operationalize the insights before you invest in AI predictive analytics. It's important to understand the questions you want answered from the analytics to get accurate results. The insights have limited value if they do not drive better decisions.
• Look for vendors that understand the iterative process that AI algorithms require, as the process of tuning and validating the algorithms takes time and ongoing learning. Even after the algorithms are in production, they continue to need maintenance and updates.
• Test any vendor's understanding of healthcare, as it is a complex industry that is heavily regulated and currently under tremendous financial and operational pressure because of COVID-19. Healthcare organizations do not have time to teach vendors about their industry. An often cited reason for AI failure is that the insights could not be operationalized, knowing the industry will help to mitigate this risk.
• Don't underestimate the value of collaboration, as successful AI requires both technical and domain expertise and as algorithms need ongoing attention. Use reference checks to evaluate the ability of a vendor to deliver on the partnership promises.

Learn More

Related Research

• IDC FutureScape: Worldwide Health Industry 2021 Predictions (IDC #US45834920, October 2020)
• IDC's Worldwide Digital Transformation Use Case Taxonomy, 2020 Value-Based Health (IDC #US44983520, July 2020)
• IDC TechBrief: Integrated Clinical and Social Care (IDC #US46142420, April 2020)
• COVID-19 Industry Impact: Health and Life Science (IDC #US46180320, April 2020)
• AI-Based Automation Framework for Healthcare (IDC #US45069019, May 2019)

SYNOPSIS

This IDC Perspective discusses how COVID-19 gave urgency to the need to identify individuals at the highest risk of a poor outcome if infected. Blue Shield of California, in collaboration with the data science-as-a-service provider Cogitativo, developed a machine learning risk-scoring engine that delivers actionable insights, enabling Blue Shield to proactively reach out to its most vulnerable members with critical services, including food.

"The balance of technology and personal touch often delivers the highest value experience that can lead to improved loyalty." — Cynthia Burghard, research director, IDC Health Insights.