All Collections
Custom Patient Targeting
Custom Patient Targeting

FAQ for Custom Patient Targeting

Updated over a week ago

What is Custom Patient Targeting?

Custom Patient Targeting is a new privacy-safe patient targeting solution designed for healthcare brands. It offers the precision and customization you need to drive your best patient outcomes. It doesn’t rely on user-based targeting, ensuring compliance with all laws, policies, and guidelines from HIPAA, NAI and DSPs. Using AI-powered predictive modeling, we build a just-for-your-condition targeting model that bids on individual impressions based on aggregated patient behavior.

What seed data can I use to define my patient?

Many healthcare providers don’t have access to first party data or are reluctant to share it due to privacy concerns. Customers can use a combination of search terms and ICD-10 codes (derived from RWD vendors providing claims data) to seed models.

What are ICD-10 Codes

ICD (or International Classification of Diseases) and 10 (which is the latest version) is a standardized method to precisely classify over 70,000 diseases and conditions with a specific diagnosis (DX) code. For example, the code: E10.9 indicates "Type 1 Diabetes" while E11.9 indicates "Type 2 Diabetes". When Pharma brands market their drugs, they often will have a list of DX codes that their drugs are marketed to help with.

What is RWD?

RWD (or Real-world data) represent healthcare data that is collected in a privacy-safe way during the delivery of care. This can include a wide range of data sources ranging from things like payer and claims data (databases of media, pharma, dental claims from insurance companies), patient surveys, electronic health records, and more.

How do you use search terms to seed models?

Dstillery's large opted-in data panel allows us to see when certain searches are made for healthcare or drug-related product names, symptons, and/or co-morbidities. If we are able to find enough devices that search for a specific keyword, we can use those devices as seeds for our model. Initial research as shown that these search keyword-based models can perform on par with first-party pixel based models.

Why do we recommend both Search terms and ICD-10 codes to seed a model?

We’ve found that our search data from our opted-in panel has been really predictive for patient definition. Being able to use anonymous patients searching for specific drug names, symptoms, and co-morbidities in the largest search engines as seeds has been proven to provide strong performance. Combining this data with ICD-10 codes provides the most accurate definition of a patient for a majority of our customers.

How are you able to understand the patient journey without user tracking?

Custom Patient Targeting is powered by AI that identifies and learns the patterns of how, when, and from where anonymous opt-in devices from our panel browse the web. These behavioral patterns help the AI predict how anonymous patients browse the web. On the targeting side, rather than scoring users, we score impressions on their likelihood of targeting your desired patient. This allows us to get a level granular targeting unheard of in healthcare all without using any user based data

Did this answer your question?