Medical and Rx Claims Data

**Overview**
This dataset captures healthcare claims and service utilization data across multiple dimensions of patient care and provider activity. Each record represents a medical, dental, or procedural service delivered to a patient within a defined period and coverage plan.

Key features include:
- Time & Coverage Context: Service Period, Coverage, and Service Region fields establish when and where care was provided.
- Patient Demographics: Anonymized attributes such as Age, Gender, and Zip Code allow for demographic and geographic segmentation while maintaining privacy.
- Provider & Service Details: Doc. Type, Taxonomy, and Place of Service describe provider roles and settings of care.
Clinical Information: Primary and additional diagnostic codes (Prim._Diag, Additional_Diag 1–8) capture underlying conditions, while Procedure, Rev Code, Anes_Proc, and
NDC code
reflect medical, dental, or pharmaceutical interventions.
- Specialty Attributes: Dental-specific elements like Tooth Number and Tooth Surface enhance procedure-level granularity.
- Operational Metrics: Fields for Modifier(s) Procedure Units, Days/Visits/Treatments, and
Charge
enable analysis of treatment intensity, reimbursement complexity, and cost of care.
- Inc. From / Inc. Thru: Service inclusion dates define the precise billing or treatment interval.

**Use cases**

- Population Health Studies: Trends by age, gender, diagnosis, or region.
- AI/ML Modeling: Predictive modeling for risk scoring, claims fraud detection, and personalized care pathways.
- Healthcare Analytics: Utilization patterns, patient journeys, and cost analysis.
- Claims Optimization: Identifying billing inefficiencies, modifiers, and reimbursement opportunities.
- Clinical Research: Linking diagnoses, procedures, and service settings to outcomes.

**Sample SQL query**
Query to derive top diagnoses by frequency and cost:

SELECT
primary_diag,
COUNT(*) as frequency,
SUM(charge) as total_cost,
AVG(charge) as avg_cost,
ROUND(COUNT(*) * 100.0 / SUM(COUNT(*)) OVER (), 2) as pct_of_claims
FROM claimdata
WHERE primary_diag IS NOT NULL
GROUP BY primary_diag
ORDER BY frequency DESC
LIMIT 20;