Lantern’s Savings Methodology Fixes the 7 Biggest Flaws in Traditional Healthcare Claims Analysis

Commercial health insurance provides coverage for more than 150 million Americans through employers, employer consortiums or labor unions, but all too often, it is impossible for those plan sponsors to understand what they have actually paid. At an expected average cost of $16,000 per employee in 2025 – that is a big bill to have questions about. The problems in traditional claims analysis run deep, but ultimately boil down to a lack of transparent, standardized or intellectually rigorous methods to quantify cost increases, price out episodes of care, and measure savings from interventions.

An accurate and granular understanding of spend is critical. In today’s economic environment, hard-dollar savings are a benefits leader’s north star. After years of amassing lots of individual point solutions in an effort to enrich benefits packages, improve health outcomes and reduce spend, those same leaders are now digging in to see which – if any – are actually living up to the promise of hard-dollar savings.

But comparing the value of third-party solutions is hard. In the words of a Fortune 500 benefits leader who attended a recent client summit, “It’s a nightmare because everyone calculates things differently.” She’s right, and to make the nightmare worse, there’s often very little disclosure about what’s included in episodes. Weak contract-to-contract savings are often masked with soft dollar savings and overly optimistic assumptions. Without a consistent approach, decision makers may believe they are making apples-to-apples comparisons when, in reality, the numbers tell very different stories.

Lantern is a Specialty Care Platform that connects self-funded organizations with access to its network of high-quality specialists in surgery, cancer and infusions care. By ensuring that members have a) convenient access to b) the right care in an appropriate setting from c) an excellent specialist and d) at a fair price, Lantern helps lower the cost of care and improve health outcomes for both plan sponsors and individuals. To date, we have saved our commercial clients more than $1 billion, and our ability to accurately measure and articulate those savings to each stakeholder we serve – employers, unions & trusts and consultants – is critical.

We believe that transparency is the key to meaningful cost comparisons. That is why we have made our entire savings calculation process publicly available – from what each episode includes, to how we quantify savings and the reasoning behind every step. By openly sharing our methodology, we highlight common barriers to effective savings analysis, address challenges that can lead to misaligned results and make healthcare costs and savings calculations easier to understand.

This approach should not be proprietary to Lantern. We want to set a new industry standard. Lantern‘s methodology has been reviewed and endorsed by leading academic and industry experts.  We thank those experts for their participation in this process, and we invite other industry stakeholders to join us in our push for greater transparency by providing feedback. Together, we can ensure that self-funded organizations and other at-risk groups have the tools they need to advocate for fair, understandable and more sustainable healthcare pricing.

Below, we summarize the seven steps we take to produce a more transparent savings methodology; the full document can be viewed here.

#1: Clean up missing or inaccurately transformed claims data after it passes through all the intermediaries.

Claims data passes through multiple intermediaries. During that process, procedure codes, provider identifiers and pricing details can go missing or be inaccurately transformed. For example, we often find examples of episodes without anesthesia claims; such claim codes can begin with a “0”, which can be eliminated in processing and then discarded as bad data.

#2: Allow six months to one year of claims lag, at minimum.

The delay in finalizing claims, known as claims lag, can significantly impact cost estimates. The costliest claims are typically the last to be finalized., If the analysis does not allow sufficient time for claims runout (six months to a year at a minimum), they will materially underestimate costs. For example, in our analyses, 36% of surgical claims dollars are still unaccounted for after 30 days, and 7% is missing after 180 days, which disproportionately affects higher-cost procedures. Consider an episode-specific example: If a total knee replacement with $46,000 in facility and implant spending linked to the DRG-code is missing due to claims lag, the total episode will look like ~$8,000 as opposed to $54,000. How can a benefits leader begin to decide where to focus cost savings efforts when their most costly episodes look to be very cheap? Even for small employers, claims lag can translate to millions in missing spend; not to mention the Fortune 500 employers who spend upwards of $2 billion annually on health care – 7% translates to hundreds of millions of dollars. Would you look the other way on a bill that was $140,000,000 off?

What claim lag means in practice

Average allowed costs per missing claim line:

• $6,627: the average claim cost missing after 365 days
• $11,624: the average claim cost missing after 455 days
• 819 days: longest duration of claim paid after the date of service

#3: Precisely define episodes of care

What is actually in an episode? Some definitions have wide guardrails and include follow-up care, readmissions from complications or pre-op imaging while others only include direct procedural costs. For a true apples-to-apples savings comparison, either definition can get you to a good answer, but the definition has to be specific and consistently applied. That just does not happen today. Lantern favors the more narrowly-scoped approach, analyzing only costs incurred between admit and discharge. Lantern’s procedure-specific savings methodology ensures precision by accounting for all universally essential cost components, including provider rates, site of service, anesthesia and, where applicable, implants, so that savings can be compared against a well-defined and universally applicable benchmark. Without a standardized approach that includes core cost drivers, employers and benefits leaders risk making misleading comparisons that do not reflect real cost savings.

#4: Account for differences in site of care & geography

Accurate savings analysis must account for differences based on site of care and geography, because both significantly influence pricing. Consider this: Medicare rates for the same procedure can vary by as much as three times across different markets, making a one-size-fits-all number ineffective. Additionally, reimbursement rates (and medical coding frameworks) vary dramatically based on whether a procedure occurs in a hospital inpatient setting, a hospital outpatient department or an ambulatory surgery center (ASC). Without precise adjustments for geography and site of care, savings comparisons risk being misleading and fail to reflect real-world pricing.

#5: Account for the actual procedure mix by individually reviewing claims outliers before dropping the extremes

Outliers should be accounted for, but decisions about those outliers should be based on an understanding of healthcare billing and clinical protocols, not just basic statistical trimming. Analysis should start with more of an Exploratory Data Analysis approach, and then logical rules based on data observations and industry expertise. For example, claims that fall materially below a Medicare floor are likely incomplete, while claims with costs many multiples in excess of Medicare may represent cases with complications or billing errors. As a result, simply cutting the top and bottom deciles may seem balanced, but that fails to account for the actual procedure mix or episode completeness. A disciplined approach to filtering outliers ensures savings calculations reflect true, comparable costs rather than distorted extremes.

#6: Don’t rely on machine-readable files (MRF) and “groupers”

It has become a more common practice to rely on machine-readable files (MRF) and “groupers” to calculate savings, but these tools in their current forms aren’t ideally suited for episodic procedural care pricing. While both can offer good insights in other contexts, they lead to incomplete, inconsistent and unverifiable results in the context of complex specialty care. MRFs introduce three shortcomings. First, they make it hard to adjust for patient acuity by focusing analysis at the individual claim-line level instead of the entire episode. Second, they artificially minimize the cost of implants as the variability in implant type is obscured by sharing a single price on an implant code. Third and finally, they often ignore contracting nuances like code limiting provisions. Existing groupers introduce additional inaccuracies by misclassifying procedures, inconsistently assigning site of service, applying a variety of faulty claims aggregation logical rules and failing to account for variations in provider billing practices. Grouper approaches can be quite effective, but tools need to be tailored to specific use cases instead of relying on a one-size-fits-all approach. Reliance on these tools in their current form leads to inaccurate comparisons and creates dependencies on proprietary methodologies that mask how costs are calculated. Both reasons make it impossible to verify whether cost comparisons and ultimately savings calculations are accurate.

#7: Implants must be included to ensure accuracy.

Implant costs are a significant driver of surgical expenses, yet they are often omitted or inconsistently represented in claims data. Inaccurate implant data skews savings calculations, particularly for procedures like joint replacements, where the median implant cost exceeds $10,000. Lantern’s methodology ensures that all relevant implant costs are captured, whether they appear as separate line items or are embedded in facility fees. Without this level of detail, cost comparisons fail to reflect the true financial impact of surgical procedures.

Conclusion

Without a rigorous, standardized approach to analyzing savings, plan sponsors are left without a clear picture of their spend despite ultimately bearing the financial burden. Lantern’s approach reflects a step toward greater transparency by providing open access to its savings calculation methodology.

We recognize that our effort will only have a true impact if it is part of a larger, collective push across influential plan sponsors, consultants and employer health organizations. The journey toward transparency in healthcare pricing and cost savings assessment requires industry-wide collaboration.

We humbly invite feedback, discussion and even challenges from others in the industry because we know that our perspective is just one piece of the puzzle. By working together and committing to accuracy in savings calculations, we can help drive smarter, more sustainable spending in healthcare—ensuring that cost reductions are not only real but also truly meaningful. One hundred and fifty million Americans are counting on us.

John Zutter is the CEO of Lantern, leading the company’s mission to expand access to high-quality, cost-effective specialty care through digital health solutions, specialty networks and bundled payments. Since joining in 2016 as a turnaround leader, he has transformed Lantern into one of the country’s fastest-growing employer-focused healthcare platforms, growing from 100,000 to over 6 million members and serving 1,000+ employers. Under Zutter’s leadership, Lantern expanded into oncology navigation in 2022 with the first complete cancer care solution for employers, and introduced the first independent and national infusion network in 2024. Prior to Lantern, he was a co-founding Partner at Dundon Capital Partners, held executive roles at Santander Consumer USA, and served as Vice President at J.P. Morgan in New York. A native New Yorker turned Texan, Zutter now lives in Dallas with his wife, four children and dog.