As the health care industry shifts from fee-for-service to value-based care, the ability to deploy effective population health initiatives has emerged as a critical factor for the success of organizations. With more risk shifting to providers, organizations need to go beyond merely analyzing data to stratify patients into condition-specific cohorts; they need to develop sophisticated approaches for weighing a wide range of variables to selectively deliver interventions and guidance to caregivers that help improve outcomes while lowering costs.
Population health has the promise to deliver big results, but many providers are facing roadblocks along the way. In order to successfully engage in population health, providers should focus on five critical areas.
Getting the data
Successful population health initiatives need access to relevant, accurate data. Thanks to electronic health records, providers have easy access to clinical data. Gathering data from partners and affiliates, as well as from payers, will provide the complete picture of the continuum of care and allow population health analytics to identify care gaps.
But providers also need access to data that goes beyond the clinical encounter. To effectively stratify their patients, providers need insights into their habits, personality and environment. Analytics are effective when organizations incorporate sources such as biometric telemetry, mobile and wearable app data, public data, and social data into their efforts. External data from nontraditional sources such as social media, government records or even weather patterns should also be incorporated into population health efforts to give further insights on a patient’s habits outside the doctor’s office.
In addition to data that provide insights about patient populations, providers need information to guide the treatment of their patients. Pulling unstructured data, such as insights from textbooks and journal articles, into a population health program will provide guidance on the best options of care for each patient cohort, helping providers deliver higher quality care.
Providers need to emphasize interoperability and data governance for population health initiatives to thrive. With disparate data sets stored in a variety of heterogeneous systems and formats, the extraction, transformation and loading process of accessing and normalizing data from various sources is difficult and time-consuming. Inaccurately translated data or importing processes that take too long can severely diminish the utility of population health strategies.
Organizations that participate in health information exchanges are at a considerable advantage over their peers as they are actively engaged in sharing and exchanging information to better understand their patients’ continuum of care. But, even with the connectivity and access to different records available from HIEs, there is often no good mechanism for sharing data at the volume required for population health management.
While interoperability is a hurdle for many organizations, growing pressure from the government, new standards outlined in legislation such as the Medicare Access & CHIP Reauthorization Act and new resources from industry groups such as Health Level Seven's Fast Healthcare Interoperability Resources will help drive gradual improvements on the interoperability front. Organizations that take a proactive approach to interoperability by following protocols outlined in FHIR and making efforts to standardize data are at a considerable advantage when it comes to population health.
Effective population health programs rely on extensive infrastructure to achieve the desired results. For health systems trying to determine the technical and human resource requirements to support their initiatives, it is important to remember that more is not always better. Many organizations have made or are considering big investments in building and maintaining large internal enterprise data warehouse capability, which may not always be the right choice.
In-house population health solutions can be a huge burden on internal resources, without providing the desired results. For many health care organizations, complicated and cutting-edge IT projects are simply not their strong suit. For others, they can set out to meet requirements that wind up changing radically by the time the project is nearing completion. Cloud-based or hybrid solutions that mix internal capabilities with outsourced software solutions can be much more flexible and much less capital-and human resource–intensive.
Patient and provider engagement
There are ultimately few opportunities for patients and their providers to engage directly in a clinical encounter, and smart use of technology, including natural language processing, cognitive computing and population health analytics, can not only ensure that each clinical encounter is focused on the most important care issues but also that all patient communications outside of those encounters are optimally focused. Successful population health programs need to not only identify the interventions —they need to give providers the right tools to work with their patients to make those interventions happen and achieve results.
All the data in the world will not improve outcomes for patients if they and their providers are not effectively engaged with each other, and with clinical goals and programs. One of the most important things population health programs can do is arm physicians with all of the data, but no more than is necessary, to understand and buy into the care plan. Doctors need to know not just that a particular test or treatment is recommended, but what data that recommendation is based on and what the specific implications are for their individual patients.
Providers also need tools to uncover the most-efficient and effective methods of engagement for each patient based on their demographics, conditions and preferences. The most successful approaches will employ a mix of digital and human communications depending on the age and stated preferences of the patient.
Scaling knowledge, expertise
To efficiently practice population health, organizations must be able to scale their operations to account for the large volumes of data and queries. With millions of data points to access, normalize, analyze and act upon, automation at every phase is key to manageable population health programs.
Even with automated processes, the sheer volume of information and potential correlations to scrutinize is too vast and complex for the human mind — or human-designed structured query language — to work with effectively. Not only do providers need to take into account all information available on their populations, but they need to combine that with all the research available for treatments and set standards of care to achieve desired results.
Organizations are realizing that to overcome information overload and scale population health projects to the next level, new innovative solutions, such as cognitive computing, are required. Cognitive computing promises the ability to analyze and interpret data to understand a population’s needs in a nuanced way, taking into account far more variables than a human could. Cognitive systems can also design tailored recommendations for care and generate useful insights for doctors that will improve outcomes for patients and let doctors focus more on care and less on literature reviews.
To elevate population health management strategies, organizations need to be able to identify patient cohorts with ever more subtle and sophisticated characteristics. Population health will be about much more than just identifying gaps in care or even the highest-risk populations; it will involve a whole a new level of care optimization based on intelligently assessing and balancing a vast number of variables and data points.
By pooling data, utilizing new engagement models, embracing cognitive computing and investing in the right mix of in-house and hosted technologies, providers will glean the insights that allow them to achieve the promises of population health.
Anil Jain, M.D., is a vice president of IBM Watson Health and former senior executive director of information technology at the Cleveland Clinic. He continues to practice and teach internal medicine at the clinic.