Case Study Summary: During a proof-of-concept engagement, Penstock analyzed historical claims data for a commercial health plan and uncovered a systemic configuration error causing routine E&M services to reimburse at dramatically inflated allowed amounts. By mining paid claims data rather than relying on pre-pay edits or known error checks, Penstock identified approximately $1.8M in overpayments and enabled the plan to correct the issue at its source.

Background

During a proof-of-concept engagement with a commercial health plan, Penstock was provided historical claims data for analysis. As is often the case in early-stage evaluations, Penstock did not have access to the plan’s adjudication platform, configuration rules, or payment logic—only the data itself.

The plan had payment integrity vendors in place. The question was whether deeper data mining could surface additional value beyond standard audit programs.

It did.

The Discovery

While reviewing evaluation & management (E&M) services, routine office-visit codes that typically reimburse between $100 and $200—Penstock uncovered a major inconsistency.

In claim after claim, providers were billing modest amounts, often under $150, yet the plan was approving $3,698 for the exact same services.

That difference matters.

The billed amount reflects what a provider submits. The allowed amount reflects what the plan authorizes for payment based on its own configuration logic.

In this case, the plan didn’t just overpay, it approved an amount that was 18–30x higher than reasonable.

What the Data Revealed

This was not an isolated error or a single bad claim. The same pattern appeared repeatedly across the dataset, forming a clear signal that something was wrong at the system level. Over two years of historical claims, this single issue represented approximately $1.8 million in claim overpayments. Because the claims were adjudicating “correctly” according to system logic, the issue never triggered traditional payment integrity edits or alerts.

Root Cause: A Configuration Error

After Penstock presented the findings, the plan confirmed the issue was caused by a claim configuration error within its adjudication setup.

There was no provider misconduct.
There was no coding anomaly.
There was no documentation issue.

The system itself was allowing routine services at dramatically inflated amounts and had been doing so for years.

Why Traditional Programs Missed It

This issue persisted despite the presence of national payment integrity vendors already working with the plan.

Why?

• The claims paid as configured

• Overly simplistic anomaly detection methods deployed by other vendors could not detect the erroneous pattern

• No pre-pay rule would reasonably catch a misconfigured allowed amount

• The issue only became visible by mining paid claims data for patterns, not checking for known errors

This is the difference between running audits and deep post-pay data mining. This is post-pay done as intelligence, not just recovery.

From Post-Pay Insight to Prevention

Configuration errors like this are rarely detectable pre-pay. They must first pay incorrectly before they can be understood.

Once identified, however, they allow plans to:

• Correct configuration logic at the source

• Prevent future claim overpayments

• Improve financial forecasting accuracy

• Reduce downstream audit burden

Post-pay becomes the feedback loop that strengthens pre-pay.

‍

The Takeaway

This case study isn’t about a single recovery opportunity.
It’s about what happens when plans rely on static programs and legacy approaches.

$1.8 million sat undiscovered in the data—until it was mined.

‍