How to Thrive in the Wild West of New Targeting Techniques and Regulations

In the current regulatory landscape, pharmaceutical marketers may be looking back fondly on the days when HIPAA and HITECH were the only laws related to data privacy. But that landscape is changing fast, triggered largely by the California Consumer Privacy Act (CCPA).

“California started, and just about every state in the union now is looking at some sort of legislation at the state level because HIPAA doesn’t cover everything that consumers are concerned with when it comes to privacy,” explains Michael Weintraub, Founder and CEO at Medicx.

The big one to watch is Washington’s My Health, My Data Act (MHMDA), which makes it illegal to use personal identifiable information (PII) to infer or predict health conditions—unless, of course, that consumer has specifically opted in.

We can expect individual states to continue advancing their own privacy legislation, and cookies are officially disappearing after June 30, 2024, to boot.

Where does this leave life sciences marketers? And what are the best tools and techniques to optically target patients and healthcare providers without running afoul of new regulations?

Precision Medicine Needs Precision Targeting

Pharmaceutical brands and their advertising partners must be able to target patients with relevant ads—to get patients on the right medications faster, while also reducing ad waste. It is particularly important to accurately reach patients with rare diseases, those with cancers that must be identified quickly for optimal outcomes, and patients in underserved geographies.

This requires accurate profiling—determining what the ideal patient looks like—which helps brands determine who to target with their messaging. It also requires knowing how to reach these patients, including what channels they engage on, what sort of messaging and creative may be most relevant, what times of day they’re most receptive, etc.

The best targeting uses a combination of medical and prescription claims data plus consumer and lifestyle/attitudinal preferences data to predict things such as:

  • Treatment initiation
  • Prescribing proclivity
  • Disease progression
  • New brand adoption
  • Promotional response
  • Conditions not fully described by ICD-10 codes

For example, advertisers can use machine learning models to predict:

  • The likelihood of a patient switching brands based on patterns of similar comorbidities.
  • The receptiveness of a patient to switching brands to reduce number of doses required.
  • The propensity for illness or illness progression.
  • The likelihood of developing a comorbidity.

But according to MHMDA, advertisers can no longer use statistical modeling, machine learning, or artificial intelligence to determine which individuals are most likely to have specific conditions—which drastically limits these targeting technologies.

“Historically any of us smart analytically minded companies could have … taken a bunch of data and built models or algorithms to help you predict, at an individual level, what this person’s diseases are, what doctors they see, if they have been hospitalized,” says Weintraub. “The algorithms can do all that with the data—and sometimes they can be really accurate and you could be violating patient privacy as a result.”

More than ever, marketers need to balance this ultra-precision with privacy. One solution is to target by geography, at the nine-digit zip code level, instead of by the individual.

Use the Nine-digit Zip to Balance Precision with Privacy

Breaking the country down by nine-digit zip codes creates about 35 million hyper-local groups, each including somewhere between five to 1,500 households. For example, a single story of a 30-floor apartment building in Manhattan may be considered one of these “micro neighborhoods.”

Each one of these micro neighborhoods contains data signals that indicate whether or not individuals in the neighborhood have the targeted health condition. Brands and advertisers can prioritize these neighborhoods by the number of signals in each geography—without identifying who those specific patients are. The idea is to find a strong ratio of target patients to total population.

For example, let’s say that the data tells you two patients with fibromyalgia are in a micro neighborhood of 45 total people. Weintraub explains: “Run your ads to as many of those 45 people as you can with the hope those two are in that pool. It’s random with high precision. We’re looking for those fibromyalgia patients. We just don’t know which individuals they are or what homes they live in.”

This approach allows pharmaceutical marketers to target with precision—but not so much precision that you risk violating privacy laws. In this case, companies are not targeting any one individual or any single identifier out there in the digital landscape of media serving or ad serving. Instead, they would be targeting a neighborhood. To summarize:

  • Using only anonymous, de-identified data aggregated to identify target groups—and NOT using any individual or user-level information.
  • Executing privacy-by-design steps, such as removing hyperlocal groups where population or sensitive diagnosis thresholds are not met.
  • Conducting third-party HIPAA expert determinations every 12 months to ensure data use cases and associated processes cannot result in patient re-identification.

Connect the Dots with Multi-party Data

Increasing privacy regulations and cookie deprecation not only impact targeting—they also impact measurement. And measuring campaign effectiveness is already difficult. Measuring health outcomes—such as audience quality, office visits, and patient new starts—only works for about 8% to 15% of your campaign impressions as it stands today.

Without this kind of robust performance data, you’re ultimately making important campaign decisions in the dark. Using partial or inaccurate information can lead to making poor decisions, resulting in wasted ad spend and poor ROI—but most importantly, missed opportunities to improve the quality of a patient’s life earlier in their healthcare journey.

Some businesses see permission-based targeting as the solution to this measurement challenge. However, this approach alone is severely limiting as requiring someone to opt-in to brand communications adds barriers and reduces participation. First off, somebody would have to go to a website and provide their email address to opt in. Some of that’s passing privacy scrutiny, and some of it’s not. But more importantly, the adoption rate isn’t that high—it’s actually very, very low.

Identity resolution is a proven solution to cookieless marketing that connects identifiers across digital touchpoints into a single consumer identity. The process helps organizations create connections between what’s happening in the digital and offline world. Identity resolution can significantly improve your measurement capabilities, resolving 60%+ more impressions than cookies—ultimately doubling the pool of brand-eligible patients available for further analysis.

The process works by taking a multi-party data approach—combining a brand’s first-party data (e.g., opted-in contacts), second-party data (e.g., from publications, like X and Y), and third-party data (e.g., medical claims data, consumer preference data). By linking all of this data together, the brand can:

  1. Resolve advertising impressions to real (but de-identified) people
  2. Link impressions to medical claims data
  3. Create test and control groups to measure campaign performance for key metrics

Using identity resolution instead of cookie-based measurement or other cookieless approaches not only addresses privacy concerns, it also allows you to accurately measure more meaningful metrics—such as office visits, new patient starts, incremental lift, and patient lifetime value.

The Upside of Disruption

If we’re being honest, cookies were never a great solution. Industry pressures to move away from this technology and to adopt more privacy-centric techniques—such as micro-neighborhood targeting and identity resolution—presents an opportunity for brands to improve performance.

“We’re not getting a lot of accurate online information through cookies,” says Lauren McQuiston, Solutions Engineer at Medicx. “It’s exciting that we’re moving away from cookies. It’s forcing us to find new and better solutions that provide more confidence. We don’t have to be guessing at the best way to target or optimize a campaign—we can use real-world evidence and actually see a return on investment.”

So, while reaching patients may have never felt more daunting, the industry is, in fact, making meaningful strides towards safer, more effective, and more measurable methods—and it’s time for pharmaceutical brands and their partners to adopt them.

  • Frank Hicks

    Frank Hicks is EVP, Product Strategy & Partnerships at Medicx. Frank has spent more than 30 years designing, developing, and deploying best-in-class operations across a variety of industries. Upon completion of the Executive MBA Program at the University of Minnesota, Carlson School of Management, Frank co-founded Yatra Corporation, a development company for patented analytic applications. In 2005, Frank led Wolters Kluwer Health’s (formerly NDC Health) business and technology transformation program. Now at Medicix, which was recently acquired by OptimizeRx, where he will be focused on strategic technology and programmatic platform partnerships.


You May Also Like

An Authentic Approach to LGBTQ+ Marketing Could Save Lives

Last year, 56% of the LGBTQ+ population reported discrimination when seeking medical care,1 more ...

Healthcare Watch October 2013

Trend Setting: Crossix Solutions Identifies New Trend An analysis of 400 drug brands by ...