How Feasibility Tools Can Accelerate Data Partnerships
Datavant powers the largest network of health data in the world. Our mission is to connect the world’s health data to improve patient outcomes, and we’ve made significant progress toward that goal: our partner ecosystem includes over 500 real-world data sources, covering all the major claims data sources, EHR data, SDOH data and novel data types such as genomics data and imaging.
With such a vast landscape, finding the right data partner is no simple task. Datavant today released Assess and Segment Builder, advanced feasibility tools within Switchboard that reduce the friction in health data exchange.
Navigating the health data landscape is necessary, and complicated
Fifteen years ago, the limit of a health data analysis might have been using prescription data to understand which providers were prescribing a certain therapeutic, or claims data to conduct an outcomes study. Today, the availability of additional data types like EHR, lab, and social determinants of health (SDOH) data, along with new providers, and traditional data types such as claims and prescription data, means more choice of data provider than ever. Additionally, the ability to link different data types using a common patient token means more powerful analyses and more context on the patient journey. Instead of just knowing what therapeutic a patient is on, you can understand why or why not patients are adhering to a treatment. This is only one example of how the healthcare data landscape has changed in recent years.
Two opposing forces that make navigating the health data landscape difficult for startups and industry giants alike are expansion and fragmentation. As novel data sources expand in availability, there is an increased demand for advanced analytics and data-driven insights. At the same time, this causes data fragmentation across institutions, burdening companies seeking to leverage the full breadth of available data. Preserving patient privacy is a final consideration in any health data exchange that has historically limited the utility of aggregated datasets.
The challenges are steep, but the insights gained have the potential to unlock significant improvements in patient outcomes. Such insights will be accelerated by having advanced tools to find and leverage the proliferating number of specialized data sources.
We convened industry thought leaders to brainstorm solutions
Navigating the expanding landscape of health data is one of the key challenges outlined by the Product Council. To solve this challenge, they highlighted the role technology can play in helping facilitate dataset exploration, overlap comparison, and data segmentation. Today we are thrilled to announce the launch of Assess and Segment Builder, two new tools within the Datavant Switchboard that will unlock these capabilities for our customers.
Datavant Switchboard now includes feasibility tools
Assessment of a potential partner’s dataset is a critical step in the health data exchange process. It’s common for companies to invest significant internal resources just to answer basic questions needed to vet a possible partner. Now with Assess and Segment Builder, these questions can be answered faster, in a neutral environment where multiple parties are involved, and without leaving the Datavant Switchboard which makes uploading and distributing datasets easier.
Assess is a toolkit that enables you to quickly understand patient counts within your own datasets or calculate the overlap across multiple datasets. With Assess, you can understand the data you have: quickly pull baseline feasibility counts for a customer’s request, or make sense of data across internal silos that may have formed due to mergers or acquisitions. Customers looking to form partnerships for academic or governmental research, clinical trials, employee health outcomes, or other research purposes can quickly evaluate the overlap or unique composition of multiple datasets to build the desired patient cohort. Assess contains four different types of reports that can be used independently or alongside each other for a variety of use cases.
Assess contains four different report types:
Overlap – Find the number of patients and records in common among up to 10 different datasets
Use this report for questions like “If I were to link this claims dataset, EHR dataset, and lab dataset, how many overlapping patients would I get?”
Overlap Comparison – Compare a base dataset against up to 10 other datasets simultaneously
Use this report for questions like “I’m trying to decide between many different claims data providers; which claims dataset has the highest overlap to my tokenized patient population?”
Stack – Calculate the unique patients each dataset contributes to a common pool.
Use this report for questions like “What’s the right combination of datasets to reach a total of 300 unique patients?”
Profile – Review the metadata of one or more datasets, including the number of individuals, records, and physicians and breakdowns of patient characteristics, such as demographics, disease states, and procedure codes.
Use this report for questions like “How many unique patients that meet Criteria X and Criteria Y show up in this dataset?”
Segment Builder is a tool that allows you to create a subset of a parent dataset using customizable join and filter criteria. With Segment Builder, you can create a segment that only contains the specific data you need (like breaking out patients from specific states, or with certain ICD codes). These subsets of your parent dataset can then be shared with a partner for distribution. You can also use your segments in Assess for feasibility analysis with other datasets, or to gather further insights on your own data.
With Segment Builder, you no longer need to re-upload individual cohorts or segments, and can instead upload your data once and create segments within Switchboard’s secure, private environment.
Before Segment Builder, I had to individually upload 12 different monthly datasets. Now I can just upload my entire parent dataset once, and in a fraction of the time, I can create 12 different segments by specifying my own filter criteria. I’m looking forward to finding other efficiencies with this tool that will help streamline our data workflows.”