How AI Can Boost COVID-19 Vaccine Adoption
// By Betsy Keller //
As COVID-19 vaccine rollouts continue, it is clear that hope is on the horizon. But healthcare marketers know there are still challenges ahead in overcoming vaccine hesitancy before we inoculate enough people to achieve herd immunity and return to a more normal way of life.
Fortunately, new tools are available to help understand and address people’s concerns around the COVID-19 vaccines. Machine learning, enriched data, and advanced analytics can help predict who is likely to be vaccine hesitant and why, as well as what cultural, clinical, and environmental factors may influence their behaviors. Those insights can inform tailored communications to educate people about the vaccines and motivate them to get their shots.
Words Have Power
It is not enough to raise awareness about the availability of COVID-19 vaccines. It is equally important to use the right messages with the right people to drive action.
Our recent survey of 1,000 people across the U.S. found that nearly 7 in 10 adults say they plan to get the vaccine. On the surface, that is great news. But when you start to dig deeper into the data, you get a more complete picture of people’s intents and likely behaviors based on their backgrounds, ages, and other factors. For example, people who are younger, have limited income, and/or don’t have a college degree are less likely to get vaccinated. Gaps in intent could lead to future risk of COVID-19 infection.
Understanding people at a deeper level and leveraging insights from factors such as age, ethnicity, race, income level, transportation needs, and education can help healthcare marketers prioritize who needs to be reached out to, what is holding them back from getting vaccinated, which messages will resonate with them, who is likely to engage in outreach, and who will take action from it.
Overcoming Barriers – Lessons from Flu Campaigns
The survey[1] uncovered many concerns shared by Americans about the COVID-19 vaccines, including:
- Side effects
- The length of clinical trial times
- Distrust of government and pharmaceutical companies
- Disbelief in vaccines
It’s important to understand the barriers such as these that exist for people at an individual level to proactively address and overcome them.
Recently, we teamed up with a leading disease management company to drive flu vaccinations. The organization’s goal was to reach high-risk Medicaid members and remind them to get their flu shots. Since direct mail would take time to develop and have limited reach among this particular population, we created a call program to 23,000 people across eight states.
Each call was customized based on member data, as well as the various state Medicaid programs. The automated voice program featured a decision tree to guide the conversations to understand the reasons why people hadn’t gotten their flu shot. Some people didn’t have transportation to a vaccination site, some couldn’t afford the cost, others were worried about side effects or how it would interact with their current medications.
“Applying machine learning, which includes artificial intelligence, allows marketers to create data correlations and make predictions about people’s motivations and receptivity to specific messages.”
By uncovering these concerns, we were able to develop barrier-breaking messaging for the disease management company to use in targeted outreach campaigns to its members. The campaign resulted in increased flu vaccinations and improved member satisfaction: 65 percent of respondents got the flu shot and 50 percent said they found the call to be helpful.
There are two key lessons from flu campaigns that healthcare marketers can apply to COVID-19 vaccination efforts:
- Understand what hurdles you might have with your audiences and how you can overcome them.
- Develop barrier-breaking, personalized messaging to drive engagement and action.
Filling In Knowledge Gaps to Optimize Outreach
Most health marketers have access only to a fraction of data about their population based on clinical and claims information. By supplementing that data with broader lifestyle and demographic variables, it is possible to create a more comprehensive and actionable view.
Applying machine learning, which includes artificial intelligence, allows marketers to create data correlations and make predictions about people’s motivations and receptivity to specific messages.
This helps marketers focus outreach strategies and proactively address concerns for those more likely to be noncompliant versus those who may be first in line to get vaccinated. One area to look at is past behaviors, as they can be helpful in predicting future intent and action. For example, if someone got the flu shot in 2020, they may be more likely to get a COVID-19 vaccine.
Enriched, individual-level data can help marketers understand if someone is likely to be compliant with their medication regimen, if they’re a spender or a saver, if they have a childcare need, if they speak a language other than English, etc. All these data taken together can help inform analytics that yield engagement insights, including which types of communications they are likely to respond to.
“Demographics, health conditions, living environment, access to a car, purchasing habits, income level, prior channel receptivity, past program participation, and many other factors all play into a person’s motivations.”
For example, if a person is younger, they may be more responsive to text messages, and if a person speaks English as a second language, they may prefer to receive a phone call from someone who speaks their native language. Proactive outreach such as this based on individual-level data, advanced analytics, and predictive models can help marketers reach people before they become noncompliant.
One-Size-Fits-All Approach Is a Fit for No One
As seen with flu vaccine campaigns, it is possible to drive engagement by gaining a deeper understanding of the people you’re trying to reach and then communicating with them using the right messages at the right place and time.
People’s intents and behaviors are influenced by a combination of factors working together. Demographics, health conditions, living environment, access to a car, purchasing habits, income level, prior channel receptivity, past program participation, and many other factors all play into a person’s motivations.
Taken together, these data can help healthcare marketers understand which people may have certain needs or concerns, who is likely to have certain behaviors, which incentives can help prompt action, and what are the best ways to reach out to them and successfully motivate them to get vaccinated.
A well-thought-out approach and timely, personalized outreach to provide accurate information and proactively overcome potential barriers will go a long way in educating people that the single best way to protect against COVID-19 infection and help everyone return to a post-pandemic world is getting vaccinated.
Betsy Keller has more than 15 years of experience in healthcare data analytics, predictive and financial modeling, data warehousing, and program evaluation. At Welltok, Betsy manages a team of healthcare data analysts responsible for analytic services, including data acquisition, data warehousing and quality control, model development, and program evaluation across a suite of products and solutions. Betsy holds an MPH from Boston University, a BS from Trinity College, and is a licensed CPA.