Consumer marketers are exceptionally good at segmenting consumers using a wide variety of personal information like demographics, browsing history and purchasing habits. Their goal is to provide the right information in the right format at the right time to stimulate action to improve measures like sales or customer satisfaction. They’re also relentless in calculating Return on Investment (ROI) to determine the effectiveness of a consumer campaign.
This approach has been consistent for decades. Marketers segment consumers into meaningful groups, gather increasingly personalized data on interests and behaviors, then use this knowledge to craft increasingly specific marketing communications/interventions. The current result is exponential growth in consumer data with an increased ability to track ROI down to specific individuals.
I also need to highlight that this capability is significantly enhanced by the efficient sharing and movement of data in the consumer space. This is a really important point. The more efficiently you share and gather information about individual preferences and actions the greater your ability to create marketing campaigns that work.
In this regard, the healthcare industry lags the consumer industry. Most healthcare systems didn’t start capturing electronic data in a cohesive way until 2010 or later – and most of that is clinical data. I discussed this in a previous post. Yet the growth in Population Health solutions now has Healthcare providers actively segmenting chronically ill patient populations (while gathering increasing amounts of patient data) to create interventions seeking to improve outcomes and lower costs.
What if we treated nonclinical interventions for chronically ill patient populations the same way marketers treat consumers? In other words, creating personalized nonclinical interventions for chronically ill patients AND relentlessly assessing each nonclinical element for ROI in the form of reduced claims activity?
I’ll explore this topic in a variety of posts. In this one, I’ll focus on three concepts – segmentation, data, and interventions – as they relate to the parallel approach by healthcare providers and consumer marketers. In subsequent posts I will explore some of the additional nonclinical data barriers, an alternative approach to evaluating nonclinical interventions and discuss how to approach non-compliant patients.
Where do we start?
To examine how we might approach healthcare interventions like consumer marketers, we need to examine a few elements. Namely, Clinical Risk Stratification (clinical segmentation), the growing use of non-clinical data, and the current approach to interventions.
Clinical Segmentation
Healthcare systems use risk stratification, a process using a variety of clinical information to segment chronically ill patient populations – and direct care where needed. Patients are grouped into four categories: low risk, rising risk, high risk and highly complex.
This summary from the National Association of Community Health Centers (PDF) does a wonderful job of describing risk stratification:
Risk stratification enables providers to identify the right level of care and services for distinct subgroups of patients. It is the process of assigning a risk status to patients, then using this information to direct care and improve overall health outcomes.
There is a heavy use of clinical information (claims, tests, EHR and pharmacy) to define segments by disease complexity and cost (low risk, rising risk, high risk and highly complex).
The highly complex segment represents 5% of the population but generates 50% of the healthcare cost. The combined high risk and highly complex segments represent 25% of the population but generate 80% of the healthcare cost.
This baseline sorting represents a way to design interventions and deliver appropriate care where needed, while targeting the highest cost segment of patient populations.
The growth in nonclinical data
Today healthcare systems are under growing pressure to reduce cost and improve outcomes, fueling interest to use nonclinical data as part of risk stratification to identify the chronically ill patients most likely to respond positively to an intervention. Let’s examine a few elements of nonclinical data.
Social determinants of health (SDOH:) the environmental conditions where people are born, live, learn, work, play, worship and age, affecting a wide range of health, functioning, and quality-of-life outcomes and risks.
Impactability: the assessment of the patient’s likelihood to respond positively to an intervention.
Patient Activation: the extent to which the patient has the knowledge, skills and confidence needed to manage his or her health and health care.
I want to highlight an important distinction. SDOH largely refers to external factors affecting health. Impactability and patient activation refer to patient attitudes, behaviors and knowledge.
This is a critical point. Consumer marketers utilize massive amounts of data on consumer behaviors in browsing and purchasing history to target them with ads. And we now have explosive growth in both clinical and nonclinical patient data documenting their behaviors in far greater detail. The growing power of Population Health to aggregate and analyze both clinical and nonclinical data opens the door to a novel approach in treating chronic illness. We can offer our standard clinical interventions and add highly personalized nonclinical interventions that capture data in ways that allow us to calculate their success (ROI) and improve future interventions.
Don’t we already have interventions?
We sure do. Every doctor visit for a chronically ill patient starts or maintains a clinical intervention. Information is stored for test results, doctor/patient interactions and more.
An incredible amount of research on crafting broader subgroup interventions already exists. Here’s a link to the CDC, describing interventions for a wide variety of conditions:
https://www.cdc.gov/chronicdisease/programs-impact/pop/index.htm
And you can find SDOH specific interventions here:
https://www.cdc.gov/socialdeterminants/cdcprograms/index.htm
These programs deliver value by advocating for early testing, care teams and more. Some even calculate Quality Adjusted Life Years, to measure improvements in quality and length of life. Unfortunately, many are generic in nature. They cover broad groups of patients with a heavy clinical focus, recommend the same generic intervention and are mostly government driven.
There are also state level programs that seek to target at risk patients with care management. In California, the CalAIM program seeks to provide our most at risk population with support that directly targets their nonclinical barriers to health. However, most of these programs are still in their infancy and require a considerable amount of coordinated work between the state and providers before they go live.
Where do we go?
Instead of crafting group-oriented interventions based on clinical information, we can move forward to target specific individuals with highly personalized interventions, based on their claims cost, geographic and nonclinical information. And more importantly, provide broader nonclinical interventions to patients who exhibit the highest propensity for impactability.
Why? Because any highly personalized nonclinical intervention is likely to be expensive. We can start by focusing on the patients with the highest potential ROI in the form of reduced claims cost, while targeting the remaining non-adherent patients with social engineering programs to stimulate changes in beliefs that prevent them from managing their health more effectively.
Next: Barriers to efficient use of nonclinical patient data
No Comments