MasterClass Session 2 – Transforming Raw Claims Data Into a Strategic Lever
Most self-funded employers maintain a basic level of TPA summaries or automated portals. This leaves a massive gap between simply having raw data and actively using it to execute a strategy. Advisors must drive a shift away from traditional carrier-controlled data silos toward independent data warehouses that aggregate clean, unmasked streams of medical and pharmacy claims.
The Four-Part Data Inputs for Plan Design and Decisions
Advisors cannot engineer a meaningful renewal or point-solution strategy without continuously verifying a complete data set across four core inputs:
- Claims Data: Medical and Rx claims at the line-item level — paid, allowed, service dates, diagnosis, procedure, place of service, and member identifiers. Enables cost, utilization, and driver analysis.
- Plan Design Data: SPD, SBC, deductibles, copays, coinsurance, OOP max, networks, exclusions, and tier structures by plan option. Contextualizes utilization and member behavior.
- Premiums and Funding: Equivalent rates, employer/employee contributions, enrollment by tier, fixed costs, admin fees, and funding arrangement. Anchors true cost and projected savings.
- Stop Loss Contract: Specific and aggregate deductibles, contract basis (12/12, 12/15, paid), lasers, reimbursement terms, and disclosure requirements. Defines catastrophic risk exposure.
The Runout Distortion and Incurred-With-IBNR Logic
Relying on raw, immature paid claims reporting across a 3-to-6 month period will completely distort renewal projections. This is because clinical providers submit claims through complicated, multi-layered clearance and negotiation channels and paid data lags. These can lag the actual clinical events by a long gap.
- 30 Days of Runout: Captures an average of only 20% of true incurred costs.
- 90 Days of Runout: Captures an average of only 60% of true incurred costs.
- 365 Days of Runout: Only reaches approximately 85% of total incurred costs.
Advisors who evaluate plans strictly on a short-term paid basis are looking at understated trends. High-dollar inpatient stays and complex specialty pharmacy claims carry the longest processing lag, meaning short-term views paint a false picture of plan savings. To eliminate soft, inaccurate renewal advice, all evaluation models must calculate projections utilizing formal actuarial Incurred But Not Reported (IBNR) adjustments based on at least 12 months of history.
Categorizing Drivers: High Prevalence vs. Acute Severity
To prevent the mispricing of self-funded health plans, advisors must strictly separate their population analytics into two isolated tracks, avoiding the trap of tracking them interchangeably:
| High Prevalence | Acute/Catastrophic | |
|---|---|---|
| Clinical Makeup | Chronic and lifestyle-driven conditions, including type-2 diabetes, hypertension, moderate musculoskeletal issues (MSK), obesity, and foundational behavioral health struggles. | Major clinical traumas, late-stage cancer treatments, complex organ transplants, extended NICU stays, and active cell/gene therapies. |
| Actuarial Nature | Highly stable, predictable, and remarkably consistent year-over-year. | Extreme volatility; can easily swing a mid-sized group’s loss ratio by more than 20 percentage points overnight. |
| Advisor Levers | Scaled vendor management, direct point solutions (e.g., intensive clinical condition management), and targeted plan-design adjustments like steerage, value-based benefit incentives, and preventive care checkups. | Specialized case management, center-of-excellence (COE) facility steerage, proactive specialty drug carve-outs, and robust stop-loss placement. |
The Flaw of Basic ICD-10 Code Level Analytics
Relying on basic ICD-10 diagnostic reporting to craft a modern renewal strategy is dangerous. An ICD code operates strictly like a baseline smoke detector—it tells an advisor that a diagnosis exists, but it provides zero insight into whether they are dealing with a minor or major case.
For instance, identifying a claimant with code C50.911 (Malignant neoplasm of the right breast) fails to answer the critical financial and clinical questions needed for plan management:
- Staging and Grade: Stage 1 vs. Stage 4 cost profiles differ drastically.
- Trajectory: Is the treatment course active, ongoing, or resolving?
- Pharmacy Integration: Speciality drug infusions often exceed the medical claim itself.
- Site of Care: Is care delivered at a hyper-expensive health system?
Advisors must transition to episode-level roll-ups that integrate predictive risk scoring with pharmacy-medical claims tracking. Once clinical visibility is established, advisors can execute a clean, data-driven predictive loop:
- Identify: Look for members with rising risk, have documented gaps in preventative screenings, or have an upcoming speciality Rx that might hit prior to the stop-loss report.
- Prioritize: Rank interventions by a strict formula of: [Probability x Clinical Impact], rather than focusing on historical raw claim dollars. Filter by clear member reachability while separating preventable risks from unavoidable claims.
- Act: Execute targeted clinical outreach, re-tune plan designs to leverage point solution contracts, and disclose emerging claimants to stop-loss carriers proactively to lock down optimal renewal terms.
