Cyborgs, Ostriches, & Scrappers – Personas for Health Communication


Cyborgs, Ostriches, & Scrappers – Personas for Health Communication

  • Posted on

  • Written by Lisette Roman

Lisette Roman is a Senior Health Data Analyst at Decision Point that leads Decision Point’s latest health persona development initiative. We sat with her today to learn more about the project.

What are personas and how do they relate to work that Decision Point is doing on behalf of its clients?

The goal of our personas is to help our clients better tailor content to their members. Decision Point personas are models of real health plan members described in great detail and defined by their demographics, behaviors, goals, knowledge, beliefs, strengths and barriers. These personas help us understand the types of members in a health plan and how our clients can engage with them to encourage key behaviors: from healthy habits (such as going to the doctor to complete preventive screenings) to consumer behavior (such as staying happy and enrolled in a health plan). We then marry the insight provided by a persona with each member’s specific ability or propensity to engage in a desired behavior.

What we’re focused on in terms of personas is to achieve tangible, detailed segments of the population. We achieve this by taking a fresh approach to personas in two respects.

First, we have moved beyond simpler demographic segments of the population. Personas were initially popularized as a marketing tool to help companies communicate with their customers. These traditional personas were soft, conceptual descriptions of individuals which relied heavily on demographic information. Think of a segment like “suburban female, aged 50-55, 2 years on plan”. Is this enough information to engage this member in discussion about the importance of a mammography screening? We at Decision Point didn’t think so. Instead, we go to a greater level of detail to understand a member’s background and motivations related to how s/he chooses to live and take care of his or her health.

Second, we understand that the same member may have different personas, depending on the behavior we’re targeting. Consider the difference in the way that an individual with a fear of breast cancer diagnosis may behave in terms of scheduling a mammography versus taking her cholesterol medication. Might it be reasonable to expect that she may avoid the mammography but be on top of taking her cholesterol pills every day? It is this degree of nuance that enables our clients to engage an actual, tangible human personality, rather than simply a demographic statistic.

Tell us a little about the persona development process.

We have a multi-disciplinary team that includes linguists, physicians and data analysts. Together, we take a top-down approach to building the two main components of a persona: 1) the definition of the persona itself, and 2) the wellness profile associated with the persona.

To define the persona, we mine published literature examining the behavior of interest. This is the ‘top-down’ method, which allows the published literature to inform persona definitions, rather than a ‘bottom-up’ method, which would restrict our persona definitions to just the data available from our client. We fill in the gaps with other available data, such as neighborhood characteristics (what is the air quality of this member’s neighborhood? How many school-aged individuals are in school and how many working-aged individuals are working?) and consumer purchasing information (is this member a NASCAR fan? Does s/he read books about gardening?).

To define the wellness profile for each persona, we assign a value from 0-10 along eight key dimensions of wellness. This component of the persona, defined by a 0-10 scale, allows us to not just add detail to our definition of each persona but to also assign the “closest” persona to a member in the event that a member meets the definition of multiple personas for a single behavior. Our team leaves that clustering assignment to the data science team.

Walk us through an example of how this works.

Our persona development process starts with published academic literature and white papers. Together, we conduct literature searches to identify studies focused on a behavior of interest. Let’s consider mammography screening again. We take the body of breast cancer screening literature and extract statistically significant features related to completing the screening test and to avoiding the screening test. Next, our linguist takes the lead creating initial concepts of personas by iterating through combinations of these important features. Once we have some draft personas to test, we finally enter the testing phase: we build prototypes with actual client data, such as claims, laboratory data, consumer data, and census data. This process allows us to develop rich, tangible personas such as the Cyborg, the Ostrich, and the Scrapper. Let me describe these three personas in terms of their approach and attitudes to daily life and health-related habits which help us understand their mammography screening behavior.

  • Scrapper: “I got here through grit and determination”. This is an individual likely to get her mammography done. She faces major barriers in her daily life due to having a low income, living in an area without many grocery stores or access to fresh produce, and spends a disproportionate share of her income on housing. She also has a clinical condition that impacts her activities of daily living. Despite these barriers, she has an element of grit about her: she is a cancer survivor or an immigrant who has learned the English language
  • Cyborg: “I exercise, eat organic, and have a brand new knee. Leave me be.” This is an individual likely to pass on her mammography and other preventive screenings but likely to engage in elective procedures, such as knee replacement surgery. The Cyborg is a married middle-aged sportswoman with a healthy BMI, high income and high household net worth, who enjoys the outdoors and is an avid gym-goer or plays at least one sport seriously. She has a high level of education and views her primary care provider somewhat as a peer; in other words, she is likely to read up on and educate herself on health issues as they arise rather than visit her PCP. However, she is quick to engage in specialty care.
  • Ostrich: “I don’t want to know about any possible issues (Head in the sand).” This is an individual likely to forego her mammography screening. Although she tends to see her PCP most years, she has not completed a colonoscopy, mammography, or cervical cancer screening, despite recommendations to do so. She is very engaged with a behavioral health provider for her chronic anxiety; she also has a history of mild depression. She has a moderate-high income, holds at least a bachelor’s degree, and has a blue-collar job.

What are the next steps? How can this actually be used and for what purposes?

With Decision Point personas, we can identify actionable clusters of members, rather than more general population segments, that can be useful to our clients in several ways.

For example, we can use these detailed, tangible personas for targeted messaging in retention and satisfaction. Imagine how a health plan marketing team can tailor information about a member’s benefits for a Cyborg versus an Ostrich. With basic demographic segments, we might capture these two starkly different members in a single “suburban female, aged 50-55, 2 years on plan” segment. With “Cyborg” and “Ostrich”, the marketing team can highlight the benefits of interest to each, for compelling, personalized content.

Personas also allow us to engage members in clinical and wellness habits. If our client’s goal is to help members with diabetes to prevent retinal disease, we can identify members at high risk for not getting the needed retinal exam. Then, we can identify the personas of those high risk members, work with our client to develop targeted messaging that resonates with each persona, and provide those messages to the appropriate channel for that persona (be it a live call from a care manager, a general call center, or other). This approach can guide an outreach campaign of its own or support existing health plan resources, such as care management. Picture a care manager having at his or her hands not simply the clinical profile of a member but his or her detailed persona.

If we can better understand how an individual moves through the world and why, we can offer the right intervention to have the right conversations at the right time.