Large insurers of all stripes have already transformed big data into a source of competitive advantage; it is used to more accurately price and tailor products to customer needs, understand risk exposure and lower costs.
Given the torrent of new patient data from electronic medical records (EMRs), financial claims and operational transactional data from internal organizational IT systems offer important business and clinical intelligence that can be used in the near term. In addition, external or remote data can be added to this core “small data” warehouse to add additional context and value as the technology and systems mature.
But value-based healthcare organizations should not wait for complex big data solutions to get immediate value from their existing internal small data sources. These small data are immediately available in provider transactional healthcare systems and can be used to target patients who would most benefit from customized wellness programs, cost incentives and other interventions meant to change behavior and improve health.
Take University Hospitals Birmingham in the United Kingdom, for example, which switched from paper to electronic recordkeeping 12 years ago. Daniel Ray, the hospital’s head of informatics, explains that they have been extracting data from the back end of clinical systems and using it to evaluate the wholesale quality of patient care. In so doing, Birmingham’s executives recognized that prescribed drugs—particularly antibiotics—were not being administered to patients as much as 15% of the time. By retraining nurses, Birmingham has driven that down to 2–3%.
Richard Bakalar, MD, KPMG International’s Global Healthcare Center of Excellence and managing director with KPMG in the US, points out that more and more hospital administrators recognize that a complete overhaul of internal processes and culture is required to use data to improve patient care. He notes that the data need to be aggregated and put in the proper context, so that they can be used for decision-making, not only for individual patients, but also for population health decisions for patient cohorts. Having near real- time access to this information through automated data feeds is critical, as is pushing that data to the point-of-care decision-makers, whether it’s the primary or backup provider, the engaged patient or the health administrator. In Dr Bakalar’s estimation, hospital governance is indeed moving in this direction.
When re-using existing clinical data clearly shows that certain interventions produce tangible results, those interventions can be rapidly adopted throughout a healthcare system. That is what happened at Providence Health & Services, a chain of hospitals on the west coast in the US. In 2011, Providence’s St. Vincent Hospital in Portland, Oregon, launched the Modified Early Warning System (MEWS), a computer program that identifies patients most in need of immediate medical care based on changes in vital signs such as blood pressure and respiratory rate.
By identifying those patients and establishing new intervention processes, St. Vincent saw a 14% reduction in mortality rates over three years—about four lives saved per month. Now MEWS is in 27 Providence hospitals, all of which continually trade notes with each other to further improve the process. Improvement in one hospital’s performance drives the other 26 hospitals to work harder to incorporate any new processes into their workflow, according to Dr Shelley Sanders, a faculty instructor at St. Vincent Hospital.
As healthcare providers perfect their data processes, they’re learning to drill down even deeper into patient records to solve problems. MedStar Washington Hospital Center in Washington, DC, for example, worked with the University of Maryland to investigate the problem of “contrast nephropathy”— serious kidney damage suffered by some patients who undergo imaging procedures. To determine how many cases existed and who suffered the damage, the researchers looked for patterns over time, according to Dr Mark Smith, director of the MedStar Institute for Innovation and one of the co- developers of Amalga, a data-processing system for hospitals that was acquired by Microsoft.
Although the data have not yet pointed to specific processes hospitals can change to prevent contrast nephropathy, Dr Smith believes the ability to track the side effect over time will ultimately help physicians recognize the patients most at risk. What’s more, he says, technologies like Amalga are evolving to the point where anyone in the hospital can use data to solve tough problems. Rather than outsourcing them to some data science group, medical staff can pose their own simple questions as well as build more sophisticated queries.
One such initiative that highlights how to connect data to wellness can be found in a program from South Africa–based insurer Discovery Health. In 1998, the company implemented a new incentive-based wellness program called Vitality. Members of Vitality give Discovery Health access to information about their everyday habits—from the food they buy to how often they are visiting the gym—data that allow the company to quickly improve the program. Emile Stipp, group health actuary at Discovery, says that, over time, these data have revealed important links between wellness behavior and healthcare outcomes, from hospital admissions to the prevalence of chronic diseases. When Vitality sees, for instance, a member struggling with weight and not exercising, it can target that member with communications to make them aware of the plan incentives.
In one example, Vitality offered 50% gym discounts to members to encourage them to exercise. Over time, Vitality’s managers noticed that some members weren’t using the gym enough to derive real health benefits—a trend they wanted to change. So in 2010 they informed their members that the discount would drop to as low as 10% for anyone who joined a gym but rarely used it.
The strategy worked: total gym visits by Vitality’s 1.5m members increased by more than 6% to 25.7m in 2014 from 24.1m in 2013. Vitality also introduced a program in 2009 called HealthyFood, which provides discounts of up to 25% on fruits, vegetables and other nutritious choices. These programs have helped to dramatically improve health outcomes. Discovery’s data show a 10% reduction in hospital visits among Vitality’s most engaged members between 2011 and 2013, and that the cost per patient is 14% lower among Vitality members compared with non-participants.
Vitality exemplifies the increasingly popular trend among healthcare providers to carve up big data and use the pieces to improve the health of defined patient populations. “I use the term ‘small data,’” says Dr Bakalar. Insurers and hospitals alike are collecting and analyzing reams of patient data and using them to identify individuals or small groups of patients who stand to benefit the most from targeted interventions.
Now healthcare providers can answer specific questions using data generated from a range of sources—from insurance claims to electronic medical records to smartphone apps. And they can apply what they learn to empower patients to take better control over their own health outcomes. Mr Stipp says Vitality has developed an app for patients with diabetes that allows them to record their blood-sugar measurements and sends out-of-control results directly to their doctors. The app also prompts patients to attend regular checkups, as well as to exercise and watch their diets. Dr Bakalar expects to see more such partnerships between physicians, insurers and other healthcare providers— alliances centered on using data to target specific patient populations. Says Dr Bakalar, “That kind of interactive empowerment of patients will have huge implications in helping them improve their own health with fewer medical errors, especially when they have more financial accountability.” The patient, physician and manager of the future will likely be making far more of their decisions informed by data, and data about what’s happening right now, not days ago.