Effective and nuanced patient stratification is the key to successful population health efforts. Whether a provider system, payer, employer, or another risk-bearing entity is managing a given population, leveraging data to effectively identify patients to target for an intervention is crucial. This involves leveraging as many data sources as one has accessible, aggregating the data, and then applying logic, groupers, algorithms, and machine learning to group patients into cohorts that will benefit most from an intervention. Stratifying patients in this way—by their impactability—is a prerequisite to a population health infrastructure yielding financial and clinical returns.
Listed below are four common population health challenges and solutions to support them. These approaches help support accurate and effective patient stratification to enable systems to overcome roadblocks and effectively carry out population health initiatives.
Understanding the chronic disease state of a patient population is critical to any population health strategy. Too often, population health teams only focus on the most common physical chronic conditions, such as hypertension, heart failure, and diabetes. While these conditions are key drivers of healthcare utilization, there are often comorbid mental health conditions that significantly contribute to a patient’s burden of disease. In fact, high-cost patients with mental health conditions tend to incur significantly more expenses than high-cost patients without mental health conditions.
Understanding how behavioral health conditions affect a patient’s overall health can help care teams manage current symptoms and prevent worsening of the condition, as well as manage the impacts on other health concerns.
Proving a return on investment (ROI) is a critical element of any population health program. ROI is necessary to make a business case for population health, to achieve clinician buy-in, and to scale adoption. Demonstrating the ROI of population health initiatives can be challenging, but several approaches will help.
From the get-go, a health system may choose to stratify patients based on evidence-based clinical and utilization variables. By leveraging the body of knowledge in the scientific literature and lay press, a system may cohort patients based on known impactability that has been proven elsewhere; the wheel need not be re-invented. By simultaneously measuring outcomes for an intervention cohort and a control cohort, PHM teams can calculate ROI in real time. The control cohort could include patients who are in a queue for a PHM intervention, who decline the intervention, or even the entire risk population (this is not a perfect comparison or control group but will provide directional ROI information). Finally, in the long-term, a more controlled and adjusted analysis conducted retrospectively can evaluate the precise clinical impact and financial savings for a given patient population, accounting for regression to the mean, and with statistical significance.
As healthcare continues to shift from volume to value and health systems work to overcome the COVID-19 pandemic, care delivery and access are also evolving. It is commonplace for patients to receive care from multiple providers inside and outside of the hospital, making it difficult for health systems to aggregate all patient data from numerous sources. Incomplete, fragmented data leads to care decisions that don’t consider the complete picture of a person’s health.
To make the most informed decision, care teams need access to comprehensive data sets. With a reliable enterprise data warehouse that aggregates and organizes data from multiple care sources (inside and outside of the walls of a system), patient cohorting can achieve a higher level of accuracy.
Population health strategies are heavily reliant on accurate algorithms to stratify patient groups. Providers and population health managers alike seek to understand the inputs and weighting underlying a given algorithm in order to buy in to the cohorts that it creates. A lack of transparency can result in a lack of faith in the output of algorithms and can also result in a complex, cumbersome process when an algorithm needs to be updated. As such, an algorithm that is both validated, flexible, and transparent is key. This can provide members of the care team with complete transparency into the population definitions as well as the ability to easily adjust those definitions as needed. The Population Builder™: Stratification Module allows for this agility and transparency.
Identifying the right patients is critical to any population health initiative. By leveraging a data-driven population health tool like the Population Builder: Stratification Module, care teams can be confident that they are identifying impactable patients in a comprehensive and transparent way and can feel empowered to invest in and measure interventions to target the identified cohorts.
Would you like to learn more about this topic? Here are some articles we suggest:
Would you like to use or share these concepts? Download the presentation highlighting the key main points.