To build and train high performing machine learning models, Lumiata transform's customer's raw data into longitudinal person records called Person360. Person360 provides a user with detailed information on a patient or member's records. The historical data includes the history of a person's claims, coverage, medication, labs, clinical encounters, social determinants of health, engagement, and more. Historical data allows end users to better evaluate a person's profile before taking action.
In machine learning, creating Person360 records allows Lumiata to train individual models that can be aggregated or leveraged for group models, disease models, and data segmentation. Segmenting populations based on certain characteristics affords an end user additional utility to train high performing models.
Our data model is highly customizable and gives customers the flexibility to add additional data elements that are needed for their ML/AI activities. Reach out to your Customer Success Manager to discuss adding the necessary data elements to your data model.
In the Person360 Summary page, a user can quickly see a snapshot of the data that is available to them in a data group. You can see information such as:
- Number of members or persons in the dataset
- Total allowed amount
- Top 10 ICD-10 codes by members and cost
- Top 10 Lumiata Disease codes by member and cost
- Population information
- Distribution by age and gender
In the detailed records section, users are able to explore a nested view of a person's history. This section allows users to explore the following:
- Person's claims history
- Demographic information
- Coverage period
- Claim types
- Cost information
- Codes: ICD-9, ICD-10, CPT, NDC, RxNorm, HCPCS, LOINC
The screenshot is synthetic data.
Updated 7 months ago