Healthcare organizations across the country are under enormous pressure to improve health outcomes, where the notion of “population health management” has suddenly evolved from a “nice-to-have” to a necessity. This prominent care initiative involves grouping patients together based on certain characteristics – age, gender, chronic conditions come to mind, just to name a few – and working to improve the health status of that group by monitoring, analyzing and improving the health of the individuals involved. Sounds simple, right?
The challenge with this approach is that there are an infinite number of ways to actually define and measure population health data, and not to mention a few data accessing hurdles potentially standing in your way (does the topic of interoperability ring any bells?). Thankfully, being aware of these common challenges as you embark on your population health journey will set you up for greater success – and less unseen hiccups – as you measure and analyze your results.
Below are a few factors to consider around the most common measurement obstacles standing in the way of population health success:
1. Defining a population: Did you know that 15 to 20 percent of the population represents more than 80 percent of healthcare costs? Effectively measuring, and providing care, to a patient population under a population health model requires not only understanding which individuals already require treatment, but also which individuals are the most likely to need it in the future. But how do we measure that?
The first step is to define who specifically is in that population, which isn’t usually an easy undertaking. Because of the frequent overlap in populations, the same person can often end up in a myriad of population cohorts (what if you’re 70 years old, a smoker and a vegetarian, for example?). To avoid over complicating things, it is important that segmenting those groups is done extremely carefully and with an understanding that there will be a lot of “double-counting.”
Most organizations don’t get this right the first time around, and end up having to revisit their approach at one time or another. Just remember that the more specific the definition of a population cohort, the more helpful it will be for data analysis. As your analysis helps you understand the patterns in the data – from identifying hotspots to targeting intervention strategies and assessing the impact – you can continually go back and refine your definition to make it more precise.
2. Understanding how to measure: A recommended approach for actively managing populations is to take subsets from the group and pinpoint opportunities for intervention. Let’s use diabetes as an example. The overall goal for this subset of patients is to move high-risk diabetics into a low-risk category and to move lower-risk diabetics out of the population completely.
One way to do this is through intensive insulin therapy intervention. After implementing this approach, the next logical step is to measure your results. Let’s say you achieved your goal, but the before and after results for the hemoglobin A1c tests haven’t changed. How can that be? Well, without getting into a detailed discussion of statistics, it has to do with the random variation inherent in populations.
Clearly, the end result is better than where you started, but that is why it is so important to be careful in how you measure, analyze and interpret results. There is almost always more to the picture than what initially meets the eye.
3. Accessing and connecting data: Perhaps the biggest challenge of all with population health is having access to all of the necessary health information. Your organization may have limited access to payer claims data, for example. In order for population health initiatives to be effective, it is critical for organizations to have the full patient picture available so that the best possible decisions can be made. Aside from data access challenges, another hurdle is around connecting that data to specific individuals.
Here in the United States, we still don’t have a universal patient identifier, which, much like a Social Security Number, are number codes that are used across all healthcare providers to identify individual patients. A number of other countries have started to make headway with this, and the superior state of their population health initiatives are a proof point to the value that this type of capability brings. Thankfully, as health information exchanges and interoperability initiatives continue to mature and payers are able to provide greater access to claims data, the opportunities to improve the measurement of population health programs will continue to grow.
Understanding time periods: When modeling population health data, time is another challenging factor that needs to be considered. Patient populations naturally change over time; people are born, people age and people pass away. This is another great example of why setting specific definitions for your population health initiative is so important.
Let’s say, for example, you are measuring the population of 40 year olds. What actually makes someone 40 years old? Did they just turn 40 today, or are they on the tail end of 40 and really almost 41? Setting a specific beginning and ending period for population measurement is just one of the many challenges of dealing with time, and is an all too common hurdle when it comes to measuring your results.
So with all of these major challenges standing in your way, where should you start to ensure that your population health initiative is successful? The good news is there’s a strong possibility that your organization already has the information needed to build the initial foundation for population analytics – such as your EHR and business intelligence solution, for example. With this foundation already in place, you can begin to move forward with attaining a better understanding of your patient population and implementing the necessary strategies to improve their health status both now and in the years ahead.
