How Health Insurers Use Big Data to Reduce Hospitalization Rates: The Independence Blue Cross Case

Health insurance systems are a widely adopted model for managing medical expenses worldwide and constitute a crucial component of social insurance. The adequacy of these systems directly impacts whether citizens’ health can be effectively safeguarded. Originating in the West, health insurance systems in Europe and the United States have become increasingly sophisticated after more than a century of development. Particularly in the current era of booming internet-based healthcare, insurers are adapting to the times by leveraging new technologies and big data to predict customers’ health status. Recently, National Public Radio (NPR) in the United States interviewed a couple who had experienced severe illness, offering insights into how U.S. insurance companies utilize personal customer data to help them access necessary medical services.Carol and John Iovine are an ordinary elderly American couple. After John suffered a serious illness in 2014, their insurance company assigned him a health coach. The health coach not only helped Carol and John obtain the medical care they needed but also prevented John from requiring further hospitalization.

The first words out of John Iovine’s mouth were an apology. “Please forgive me; my memory of this isn’t very clear,” he said. “I had a stroke.”Signs of that stroke were evident everywhere in their home: a bed set up in the dining room, a shower installed in the kitchen, and more. Slender and dressed in blue pajama pants, John sat in a wheelchair. Yet he may have been overstating his memory problems.“We were classmates at Harding University… just not far from here,” he said. It was 1952, the year he met his future wife. “She was wearing a long red sweater and had fiery red hair. I told myself, ‘She is the one.’” John spoke animatedly about his first encounter with his wife, Carol.“Heaven rewards the diligent, and I finally got my wish,” he said with a smile.Sitting beside her husband, Carol recounted to the reporter the full story of the stroke. At that time, John’s health was in very poor condition. It started with an ulcer, followed by bowel resection surgery. Then came the stroke—and much more after that.“He developed pneumonia, jaundice, and sepsis; there was also a blood clot in his right lung,” she added. All of this occurred between October 2013 and January 2014.John had worked as a painter, and the severe illness kept him hospitalized for a total of 79 days. During that period, he was frequently unconscious and rarely left his hospital bed.“Ah, it was simply hell,” he said.“Sink or Swim”Last April, after several months of treatment at a rehabilitation facility, John Iovine was finally discharged and returned home.The period when patients are discharged and left to “fend for themselves” is a focal point of concern for the current healthcare system. For a long time, many patients like John Iovine have intentionally lost contact with their providers during this phase, ultimately resulting in readmission.Industry insiders state that such readmissions are avoidable. This single issue costs Medicare $15 billion annually, representing a significant financial drain. This is why Medicare launched an initiative several years ago to penalize hospitals that readmit patients too quickly, prompting many hospitals to pay greater attention to this problem.Now, insurance companies are also working to propose their own solutions.Independence Blue Cross, an insurer headquartered in Philadelphia, has made it its goal to identify which patients are likely to be hospitalized within the next three months, said Somesh Nigam, its Chief Intelligence Officer, in an interview with reporters. He stated that Independence Blue Cross is working to identify sick or elderly and frail individuals among its customers—those at potential risk of hospitalization.To achieve this, the company processes vast amounts of medical data in its possession, including claims bills, laboratory results, medications, height, weight, and family medical history. It also incorporates information about the communities where customers reside, such as local poverty rates. “The medical data we use to build these algorithms is roughly equivalent to five Wikipedias,” said Nigam.Computer algorithms filter all this information and calculate a score for each patient, ranking them by risk level based on their scores.Subsequently, Independence Blue Cross assigns a staff member, known as a “health coach,” to each high-scoring customer. These health coaches provide free health advice to customers and recommend beneficial additional services.“This collaborative effort is very useful for patients,” said Nigam. Health coaches can tailor health information to patients’ needs, help them schedule medical appointments, address medication-related issues, or assist in arranging transportation to outpatient clinics. In some cases, health coaches can even arrange for home care nurses. “All these measures have begun to show results,” said Nigam. “Hospitalization rates in our region have shown a quite significant downward trend.”The first group of customers requiring extra attention on Independence Blue Cross’s list numbered 18,000. One marker of its initial success was a 40–50% reduction in the expected hospitalization rate for patients with congestive heart failure.The Iovine family was among the beneficiaries of this early success. Carol Iovine’s life changed significantly after her husband suffered a stroke: she needed to manage his new medications and assist him with showering and using the toilet. The couple had to attend numerous outpatient visits and therapy sessions, for which they had to rent a wheelchair-accessible van.Carol said that things began to change considerably with the help of a health coach, who assisted her in managing her husband’s various needs. “Once, John needed a blood test, and the hospital asked me to take him to the emergency room alone for the blood draw,” Carol recalled. “I immediately replied, ‘Uh, I can’t do that.’” She promptly called their health coach, Donna Crockett, and explained the problem. “After that, a nurse came to our home to draw his blood,” Carol said.The key point is,The amount Medicare spends on home visits by nurses for health guidance or on streamlining outpatient processes is far less than the cost of readmission.Establishing RulesData forecasting is expected to save insurers on premium costs, leading a growing number of healthcare experts to closely examine both the potential and possible flaws of these predictive algorithms.“There is significant interest in this field right now,” said Glenn Cohen, a professor at Harvard Law School who has written about the legal and ethical issues arising from the intersection of healthcare and big data. “There is a major convergence underway among the medical field, computer science, and patient experience sectors.”Despite his confidence in the prospects, Cohen remains concerned. “Do customers whose personal data are used to build algorithms have the right to opt out of predictive programs?” he asked. “Are they forced to participate? Furthermore, can they even know that their data are being used?” Cohen noted that these questions remain in a gray area, as this emerging field has yet to establish standards for handling such information.Independence Blue Cross states that it strictly adheres to federal medical privacy guidelines regarding anonymization and uses this information solely to better serve its members. However, the company does not ask individuals enrolled in its health plans whether they choose to opt in to data forecasting. “These data are used only to improve or coordinate medical care,” said Nigam. “This is our responsibility, and it is widely recognized as such.”From the perspective of customer health, care coordination has profoundly improved John Iovine’s life. Since Independence Blue Cross assigned him a health coach, he has not been rehospitalized.Insurers report that due to the strong initial results of data forecasting, they plan to expand its scope. The company is collaborating with NYU Langone Health on its next target: type 2 diabetes. The goal is to identify individuals at highest risk of developing diabetes before symptoms appear, and then actively intervene to help prevent the onset of the disease.Translated by Chen Xin; Edited by Mo Renying