Datavant and Kythera recently hosted a webinar that covered several topics related to patient privacy and identity resolution, ranging from the important difference between identifying and knowing to covering real-world applications using patient-level information, such as patient engagement campaigns, patient hub operations, and rebates auditing. Here are some key highlights from the event.
Patient event information is critical to organizations using patient-level data. It is well recognized this data is fragmented, coming from many different sources and in many different formats. Being able to effectively link and access this data is a key step in creating a more complete picture of a patient. In order to preserve the privacy of patient-level data, the processes of expert determination and identity resolution must be applied to data at scale. We will explore expert determination and its related HIPAA regulations in another discussion, however, the basic concept of identity resolution is to replace protected health information (PHI) and personally Identifiable information (PII) with an encrypted token so patient privacy is preserved while and the ability to link and analyze health events across various data sets is maintained.
When using data at scale, ensure that the data you are using is of the highest quality so that you receive the highest possible match rate. This is a challenge when data is coming from different sources and in different formats, whether it is from EHR, claims, patient hubs, wearables, or from your own, in-house data. Data that is standardized will result in a much higher match rate, so make sure you take the step to standardize your data, first. Starting with data aligned to a single standard and format allows encrypted patient tokens to be applied across all your data sources and elements, replacing PHI/PII and ensuring privacy while maximizing analytical options. Much like an old song that has been remastered to make it sound clearer, Kythera’s Wayfinder platform remasters healthcare data to an industry-leading standard resulting in a match rate above 99.5%.
Deidentified, transaction-level, patient data is used for a variety of applications, such as understanding trends in patient behaviors. These patterns can be observed through the use of patient cohorts, which require an in-depth examination of patients and their healthcare events which fit specific criteria or requirements. In one example used in the webinar, we discussed how to build more effective patient cohorts by starting at the individual patient level instead of building a cohort using aggregated data. Essentially, you go from viewing the world as an aggregated group to seeing a “knowable, anonymized individual”, leading to a more informed, longitudinal view. Assembling cohorts from the individual-level, instead of the group level, you can consider questions such as…
• “How does this person behave?”
• “Are they compliant with their medications?”
• “Do they regularly visit their doctor?”
With cohorts built from this level of information and understanding, you can customize your outreach, campaigns, or offers at an individual level, resulting in much higher levels of effectiveness. If you would like many more examples of how to apply patient-level data to improve your connection with your market, please take a few minutes to browse through the webinar.
To learn more about the identity resolution process, how to use patient-level data to improve clinical or commercial outcomes, how Kythera’s Wayfinder platform establishes a standard that enables data to be accurately combined across sources at scale, or to talk with us about a real world situation you are facing, please reach out. We are happy to help.
Data is an asset which businesses and industries can not ignore. If used properly, data provides its users a competitive advantage, however, without a sound approach, data may fall short of its full potential. To help ensure success, organizations should start with the critical first step of identifying what questions need to be answered and what problems need to be solved, then quickly mobilize to identify what data is needed to answer those questions. Data needed for strategic purposes almost always comes from different sources, in different formats, and in varying quality that can complicate utilization, especially for machine learning applications. As data and analytics are evolving, so to are data technology platforms - and keeping up is critical.
Information is expensive. The costs of acquiring, handling, and understanding information to achieve actionable, business value directly affects the bottom line. Not investing in information can be even more expensive.