How Snowflake’s Data Marketplace Helps Personalize Healthcare

The approach to personalized healthcare and personalization has many strengths and weaknesses, and there are a variety of approaches to combat challenges. We’ve asked Milissa Campbell, Health Plan & Life Sciences Consulting Lead, to help define and understand personalization strategies.

What is personalized healthcare?

The first thing most people want to know is what competitors are doing. In today’s world, that competition is every individual’s digital and consumer experience — not just that of healthcare. Retailers like Amazon have set such a high bar for the digital experience and personalization that the expectations in every industry have quickly risen to incredible heights.

The right way

So how exactly do you provide the Amazon-like experience when it comes to healthcare? According to Milissa, personalized healthcare requires sending the right message and/or care to the right people in whatever way is best received by those individuals.

• The right message: This is individually optimized for that person to include outcomes or quality of life. It anticipates a need and/or answers a question before it’s asked.
• The right persons: The health ecosystem has a lot of roles, patients, members, providers, caregivers, spouses and guardians. All have within them personas that are defined by the behavior and the way that they interact with healthcare. The right person or persons has an unmet need, whether they know it or don’t, that the messaging or care will address or assist with addressing.
• The right channel: This is where data becomes the real power tool in your toolbox — using analytics-driven identification of the channels and messaging that’s most likely to drive the desired behavior.
• The right time: The right message is anticipating a need or behavior change and driving action toward changing that or mitigating an undesired behavior or change.

Building a framework

The maturity of your personalization strategy and program is dependent on two primary things: your analytics maturity (how you curate, ingest, and use your data) and your operational maturity (how you act on the output from those analytics). We can assess a healthcare organization’s personalization maturity based on the definitions that exist within the framework we use.

• Passive personalization — This is the same as no personalization. It means some people get the same message in the same way. Think about a commercial, website or poster. An example in healthcare might be a poster in a doctor’s office reminding you about the importance of getting a mammogram. It’s a great message and reminder, but not everyone who needs to see it will see it, and it might not be compelling enough to drive any behavior.
• Generic Personalization — All people get the same message delivered to them at the same time in the same way, or it could be certain people getting the same message at the same time in the same way. This could be a mass email that goes to all clients, or even all female clients, between the ages of 40 to 64, reminding them to schedule their annual mammogram. While we’re narrowing down the audience for the intended message to a targeted list, the timing would also be important. How many patients are scheduling an appointment six or seven months in advance? It’s important to deliver this message at the right time, such as 30 to 60 days out from when the patient is due for their annual mammogram, to ensure the patient’s appointment aligns with their timeline.
• Targeted Personalization — Certain people get the same message delivered at different times in the same way. Targeted personalization is a little more specific than generic personalization. In the mammogram example, it could be the same mass email that is delivered to individuals on their birthday every year.
• Customized Personalization — Certain people will get a customized message delivered to them at a determined time in the same way. What this means is that women who are aged 40 to 64 (who have not had a double mastectomy), determined to be at risk of not having their mammogram, and have not had a screening within the last two years will receive an email reminding them of the importance of completing their annual screening.
• Personalized Care — Certain people get a customized message delivered to them through the channels that are most likely to be received and acted upon by the individuals. In our example, women aged 40 to 64 and men at risk for breast tissue cancer, who have not had a mammogram in a time period appropriate for their risk factors will receive outreach through means that are appropriate for the individual and are provided easy-to-use scheduling capabilities.

After you have that appointment scheduled, you just rinse and repeat. Go back to your analytics, figure out what behaviors you need to be driving to keep patients booking appointments, and proceed with follow-up care.

Milissa and her team use this framework when they go into a healthcare organization — they talk to stakeholders, look at what they’re doing, and help them figure out where they are today, where they want to be tomorrow – and then put together a map of how to get there.

What ingredient is key to success?

That’s pretty simple — data is the key (which shouldn’t be too surprising). But what data should healthcare organizations be using? The answer is: any data you can get your hands on. The way that all of us interact with our healthcare is not only dependent on the CPT codes in our medical records and the inpatient stay that we’ve had. It’s a function of our socioeconomic status, our education, our support system, and our experiences with the healthcare system.

Our healthcare behavior and experience are multifaceted — therefore, your data should be, too. Instead of asking limited questions of a narrow data set, use vast, broad datasets and let the data tell you what is important, predictive, and actionable.

So what do I do once I have these datasets?

It’s not only about having access to all that data, it’s about the ability to get to it quickly, have it be usable, and have your analysts and business users start delivering insights and designing interventions.

This is where the Snowflake Data Marketplace enters the picture, filling a need that many organizations have. More than 200 commercial and open datasets across 16 domains are seamlessly integrated into your data.

We’re all trying to tackle the social determinant of health problems, and we all know that data is the key to making that work. Data helps us understand it and address social determinants of health. Within the data marketplace, you have access to data covering all the domains of social determinants of health and beyond.

Embrace change for what lies ahead

As you go on our journey and leverage the Snowflake Data Marketplace, let your analytics and your business teams start looking in the marketplace, start looking at your data samples, and work collectively and creatively to come up with the analytic strategies to address the questions that you’re trying to get answered.

Start with your analytics and business owners and ask if personalization is a key part of your strategy. Obviously, the answer should be a resounding “yes.” If it is, then the right stakeholders are at the table looking at the data and putting together strategies to solve this.

Remember, it’s not only your IT people, it’s not only your analytics people. Bringing creative groups together to design experiences and creatively think about what data to throw at a problem is the way that you are going to make a change in personalized healthcare.