Cutting Healthcare Costs With Price Transparency Data: A Field Guide For Self-Insured Employers

We just hosted a webinar on using price transparency data to find savings in self-funded plans. This post recaps the key ideas, shows the workflow we demoed on real data, and gives you a concrete playbook you can run in your market.

Identified claims data is useful but slow, sensitive, and often incomplete. Since mid-2023 the payer machine-readable files have become workable at scale. That unlocked a new path for self-insured employers, TPAs, and consultants to benchmark contracted rates directly across payers and providers, without touching PHI.

The opportunity is straightforward. Contracted rates for the same service at the same facility can vary dramatically by payer. If you can see those gaps before renewal, you can pick a better network partner, negotiate harder, and steer members to higher-value care.

The scenario we analyzed live

To keep things concrete, we used a fictional employer, Skynet, based in the Boise metro, with about 450 employees, mostly 25 to 35. High-utilization categories for this group include maternity and pediatrics. The local payer mix we considered included Blue Cross of Idaho, Regence, and Cigna.

We worked from a statewide Idaho slice of Gigasheet’s processed payer MRF data, enriched with:

• NPI and taxonomy data
• Medicare benchmarks
• Rate type flags and plan identifiers

We filtered to Boise area ZIPs to mirror where employees actually seek care.

‍

‍

Benchmarking the big spend: maternity and pediatrics

Childbirth episodes

We looked at DRGs common to deliveries, including MS-DRG 768 (vaginal delivery) and C-section DRGs. Even within Boise, the same delivery at the same hospital showed five-figure swings across payers. One plan that looked attractive in other categories became a liability on maternity once we drilled into delivery DRGs. This is exactly the kind of landmine a quick benchmark will surface.

Practical tips when you replicate this:

• Group by facility, payer, plan, and code so you see the rate ladders clearly
• Check rate types, especially per diem versus case rate
• Keep a separate view for VA facilities since those rates often price lower

Pediatric office visits

We pivoted to pediatrics and examined 99214 for established patient office visits. Using medians to dampen noise, we saw providers ranging from roughly ~190% of Medicare up to ~310% of Medicare for the same code in the same market. Sub-specialists predictably priced higher than general pediatrics, so filtering by taxonomy matters.

How to make this actionable:

• Build a provider shortlist by taxonomy, distance, and median allowed amount
• Compare each provider to Medicare percent, then to market median
• Keep both a citywide and ZIP-level cut to reflect true member travel patterns

Catching odd claims with quick spot checks

We showed how a plan sponsor or TPA can use the same dataset to sanity-check an expensive claim. Example: a MS-DRG 775 childbirth claim surfaced from a non-hospital place of service at $23,000. A targeted TiC query for DRG 775 across hospitals and home health agencies in Boise showed that in-home rates were not in that range for that payer. That does not prove an error, but it flags a claim for deeper review.

Checklist for outlier review:

• Match the place of service to where the service actually occurs
• Confirm rate type (per diem, case, fee schedule) and unit counts
• Compare against both plan-specific and market distributions
• If results look off, request medical records or re-price via your TPA

Steerage and network design with data, not guesswork

Once you know the spread, you can:

• Narrow networks around the high-value facilities for your top codes
• Offer member incentives to use better value sites of care
• Negotiate with TPAs and ASO partners using payer-published contracted rates
• Explore direct contracts with high-value primary care or pediatrics groups for predictable access and pricing

We highlighted a real pattern from outside Boise where a dialysis site across town was more than a thousand dollars cheaper per session. Simple incentives like rideshare credits covered the trip and saved hundreds per visit.

What to watch out for in the TiC files

Price transparency data is powerful, but you still need guardrails.

• Ensure apples-to-apples comparisons by place of service and rate type
• Remove nonsensical pairings like colonoscopies in ambulances
• Distinguish payer files from hospital-published shoppables, which are different artifacts with different refresh cycles
• Use medians, trims, and interquartile ranges to reduce the impact of zombie or stale rates
• Keep a clear data lineage so you can defend your numbers in a renewal meeting

Bringing AI into the workflow

We previewed Gigasheet’s Price Transparency Agent. In the webinar, we pointed it at the Boise dataset and asked for a focused analysis on 99214. The agent generated an executive summary, segmented rates by ZIP, and highlighted differentials versus market medians and Medicare benchmarks.

‍

‍
Two reasons this matters:

1. You get a digestible briefing for a benefit decision in minutes
2. You can click back to the exact rows the agent used, then spot-check assumptions inside Gigasheet

Think of the agent as an analyst that drafts the first 80 percent. Your benefit strategy and local market context still carry the final 20 percent.

How teams operationalize this with Gigasheet

• Load very large payer files once, then slice down for your markets and high-utilization codes
• Export filtered data to CSV, Excel, Tableau, or Power BI for handoff or visualization
• Use our API to automate monthly refreshes, generate recurring benchmarks, and feed downstream apps
• Join with claims extracts or eligibility files to weight by your true utilization

A common pattern is to keep the big lake in Gigasheet, export targeted slices for ad hoc work, and re-run the same saved views every renewal cycle.

A simple playbook you can run this week

1. Plan
   Identify your top 20 codes by spend and volume. For a younger population, start with delivery DRGs and pediatrics office visits and imaging.
2. Collect
   Pull the payer MRFs for your metro and the plans you actually use. Load into Gigasheet’s processed tables so you inherit NPI enrichment and Medicare benchmarks.
3. Clean
   Filter to appropriate places of service. Exclude impossible pairings and trim extreme outliers after confirming they are not legitimate rate types.
4. Enrich
   Add your utilization weights, geography, and provider taxonomy tags. Calculate medians and Medicare multiples.
5. Analyze
   Rank payers and networks for your top codes. Build shortlists of high-value facilities. Identify risky categories where one payer’s pricing undermines an otherwise good plan.
6. Act
   Bring benchmarks to your TPA and ASO discussions. Set member incentives for steerage. Pilot a direct contract where it clearly pencils out.

FAQs we covered live

Can I export the data?
Yes. Export slices to CSV for Excel or feed into Tableau and Power BI. Direct integrations are available.

Is there an API for automation?
Yes. Anything you can do in the UI, you can do via API with your key. Teams schedule monthly benchmarks and weekly anomaly sweeps this way.

Can I model total cost, not just unit prices?
Yes. Join payer rates to your utilization. If you lack claims, we can approximate utilization by taxonomy from deidentified studies, then refine with your data.

What about facility fees and bundled services?
Use the correct DRGs or bundled constructs for the episode, not just a single CPT line. Confirm rate type and place of service. Compare case rates to case rates.

Want help running this for your plan?

If you are a self-funded employer, a broker, a benefits consultant, or a TPA, we can stand up a market benchmark, a claims spot-check view, and a steerage shortlist for your top codes in a matter of days. We will also configure an AI analyst runbook you can rerun every month.

Send a note to sales@gigasheet.com and ask for the Self-Insured Benchmark Pack for your market. If you prefer to DIY, we will share a sample dataset and starter views for delivery DRGs and 99214.

Thanks for joining the session. If you missed it, watch the recording above and pause on the live drill-downs to see exactly how we set up the views. If you want a copy of those prebuilt filters and pivots, we are happy to share.

‍