How AI Can Generate Ideal Audience Segments in Healthcare

Blog

The Holy Grail for marketers in every industry is to deliver the right message to the right audience at the right time. Healthcare marketers specifically have struggled to align all three elements of communications, particularly healthcare audience segmentation. The good news is that artificial intelligence (AI) technology can now help bring all the pieces together.

Generally, healthcare marketers have gotten pretty good at the “right message” part, crafting appeals that are concise and compelling.

As for the “right time,” other than time-sensitive communications (e.g., announcing a new physician hire or clinic opening), healthcare marketing programs should be always-on. After all, the need for annual wellness visits, mammograms, colonoscopies, and other preventive care is continuous.

But the “right audience” component has been challenging. How can you as a healthcare marketer be certain that you’re reaching the right patients with the most relevant message at the right time? That their information is up to date? That the patients you’re reaching out to haven’t already scheduled or recently had an appointment?

Here’s a look at why audience segmentation in healthcare is so critical, how current approaches fall short, and—most importantly—how AI-powered CRM intelligence technology can help generate better patient audience segments, automatically.

Why Accurate Healthcare Audience Segmentation is So Important

Ensuring that your messages reach the right audience is vital for several reasons. The first is to maximize relevance. Targeting the right audience ensures more relevant communications to each patient, resulting in higher response and activation rates. Conversely, irrelevant messaging is annoying. It will be ignored, and worse, if it happens repeatedly, may even spur patients to opt out of your communications.

The second reason segmentation matters is to provide clarity. Each patient should always know what the next best step is for them in maintaining good health. Clarity eliminates the risk of sending excessive, confusing communications.

A final reason for proper audience segmentation is to optimize business results for your healthcare system, driving patient volume to support service lines. Reaching the right patients—and just as importantly the right number of patients—on a continual basis keeps your utilization rates constantly high (but not too high), maximizing ROI and the value of your marketing efforts.

Provided is an example of how segmenting the right audience for outreach resulted in higher responses when compared to a general population segmentation. This example specifically measures Actium’s machine learning model effectiveness for cardiac health. Results concluded that the model population identified individuals who were most at need and more likely to take action for their health.

GroupCount% Filling out HRA% with Any Appointment% with PCP Appointment
Modeled Population17,6030.96%30.14%14.84%
Comparison Group3,5540.53%7.56%2.25%
Modeled Response / Comparison Outreach Responsen/a1.8x4x6.5x
Modeled Population: Uses Actium Health Machine Learning Clinical Models
Comparison Group: Typical demographic segmentation

How to Build Healthcare Audience Segments

Traditionally, there have been two use cases for building healthcare audiences, depending on the top-level objective.

New patient acquisition: Healthcare marketers purchase lists from third-party sources, based on geography and possibly filtered on demographic criteria. The appeal of this strategy is that it (potentially) enables marketers to reach new patients and bring them into the system.

healthcare audience segmentation

However, there are several downsides to this approach. The most significant problem is accuracy: research shows the average accuracy of third-party audience targeting is 59%, with even the best providers only reaching 72%. Put another way, in the best case, more than a quarter of your messages will either reach the wrong person or simply be undeliverable.

Among the other risks of buying email lists are that you’ll violate regulatory guidelines, harm your deliverability, annoy the recipients, enable poor patient experiences, and even be penalized by your email service provider. And, as low as the accuracy and high as the risks are for email outreach, both are even worse for SMS messaging.

Existing patient activation: Advantages of creating a strategy for existing patient activation include better predictive data; patient familiarity and established relationship with your organization; and the patient trust and loyalty you’ve earned over time.

While this is a more effective approach to driving patient volume and increasing capacity utilization than new patient acquisition, the practice of manually compiling lists for marketing outreach introduces its own set of shortcomings.

  1. First, the process of manually filtering patient data risks introducing bias. Even with the best of intentions, manual list building can run counter to equity objectives.
  1. Second is the issue of timeliness. Manual list building is time-consuming, meaning lists are usually outdated by the time they are complete and ready to use. Patients may have already scheduled an appointment or even seen a physician by the time they receive the outreach, making your message irrelevant and annoying. Even more importantly, do you have capacity to see the patients you are reaching out to?
  1. A final issue is predictability. Contacting everyone on the target list with an email blast runs the risk of overwhelming your call center, overbooking appointments in the short term, and leaving some patients frustrated that they are unable to schedule a visit in a timely manner. This leads to start-stop marketing campaigns, uneven and unpredictable volume, and a poor patient experience.

On the other hand, manually building and segmenting a list while throttling communications to optimize capacity utilization increases the risk that information will be out of date, leading to either mistimed outreach, unpleasant patient experiences, and tedious manual re-checking/re-working.

And of course, whenever you want to start a new marketing program, or restart an existing one, you’ll need to manually pull a new list, beginning this whole process over.

A Better Approach: Fully-Automated Healthcare Audience Generation

The AI-driven Intelligent Audience Generator™ (IAG) functionality built into the Actium CENTARI CRM intelligence platform eliminates manual list-building tasks and their associated drawbacks.

This technology updates your patient audience continuously, in real time, based on pre-established and prebuilt transparent qualifiers. Lists are always up-to-date and ready for use when you are ready to launch a new campaign, with no need to ever “build” lists. Real-time updating eliminates the risk of redundant or mistimed outreach (i.e., sending a message to a patient who’s just completed an appointment).

The AI modeling is transparent and designed to eliminate bias based on demographic, socioeconomic, or other factors. It automatically prioritizes list members to assure that patients who will benefit most from your services will be contacted first.

Built-in capacity sensing makes it ideal for always-on programs, constantly adjusting outreach based on response rates and available appointment capacity. This helps to smooth demand for services and consistently maximize capacity utilization.

In this example, the Intelligent Audience Generator displays the qualifiers for patients that are overdue for their Medicare Annual Wellness Visit, along with the audience counts from a healthcare system’s patient population data.

healthcare audience segmentation

How Automated Healthcare Audience Segmentation Works

The machine learning models in the IAG can leverage your EMR patient data to predict and prioritize your highest-risk patients for outreach, resulting in 4-8x times better outcomes than conventional approaches. They combine clinical and behavioral data, using machine learning to ingest and process billions of rows of data to model predictions and next best actions with speed and accuracy.

These models continuously learn from your patient data to gain insights and correlations into which patients are most likely to need services; the best time and method for reaching them; how to improve patient outcomes over time; and more.

Marketers can see the criteria used to generate the lists, both the qualifiers (e.g., age, gender, clinical history) and the disqualifiers/exclusion criteria (e.g., recent/scheduled appointments, available capacity, patient value). The technology and process are completely HIPAA-compliant as there is no PHI exposed to the marketing team.

Benefits of Automated Audience Generation

Generating audience segments for healthcare outreach benefits everyone involved: patients, the healthcare system, and you as a healthcare marketer.

Patients receive timely, relevant messages; they are not “over-communicated” with. Because patients are prioritized for outreach based on their risk factors, they have better health outcomes. This combination leads to a better overall patient experience.

Healthcare systems get better and more consistent and predictable capacity utilization. Patient outcomes are improved, resulting in higher ROI on value-based care reimbursement. Call centers aren’t overwhelmed, and instead experience steadier, more predictable call volume. Patient loyalty and lifetime value are increased.

AI-powered automated healthcare audience generation enables hyper-targeted marketing, which leads to:

  • Conversion rates up to 3X greater than the industry average, due to getting the right messages to the right patients.
  • A 50% increase in conversion through intelligent segmentation and personalized campaigns.
  • Consistent application of best practices based on specific patient conditions.
  • Highly relevant reminders to schedule preventive care visits.
  • Steady, high capacity utilization for appointments.

Finally, healthcare marketers get accurate, up-to-date audience targeting without manual administrative effort. More importantly, they can quantitatively demonstrate their value to the organization. Being able to show how their programs directly drive increased revenue via appointments booked makes a much more powerful business case for marketing investments than do typical marketing operational metrics like open and click rates.

Conclusion

Healthcare marketers are sometimes described as “data rich but insight poor.” They can have access to large amounts of patient information, but struggle to make the best use of it. Even without full realization of the “patient 360 view” that some healthcare CRM vendors promise, the amount of data is overwhelming.

That makes generating lists for healthcare audiences a frustrating and risky task. It’s impossible to manually build lists that are accurate, timely, and free from bias. The result is a less-than-ideal patient experience, unpredictable appointment volume, and continual start-stop marketing campaigns.

Fortunately, the Intelligent Audience Generator’s functionality can alleviate these frustrations. It provides continuously updated, highly accurate, bias-free lists ready for use at any time, and ideal for running always-on marketing programs. Marketers can trust the output because the process is transparent and uses predefined, expert-built qualification factors.

To learn more about connecting with the right audiences, across multiple service lines, and driving patient volume, watch the webinar recording: Maximize Patient Volume with the Right Audience.

Maximize Patient Volume with Targeted Audience Segmentation in Healthcare