Given the importance and complexity of these evaluations, it’s worth considering the effectiveness of the most common methods used today, the limitations of those methods and ways to make improvements or move towards more reliable results.

The three core components of estimating costs

The fundamental goal is to determine the expected claims cost for a population enrolled in a given provider network.

There are three core components to estimating network costs:

1. What mix of services will they receive?
2. From which healthcare providers will they receive those services?
3. What do those services cost when received from those providers?

Current cost-evaluation methods

The two most common methods today are discount analysis and claims repricing.

A discount analysis uses aggregated data to compare the percentage discount off billed charges between vendors. This method is practical and efficient, and it’s easy to apply under a variety of circumstances with little required detail or technical expertise.

However, the obvious drawback is that billed charges aren’t going to be the same between vendors, so a greater discount does not guarantee lower claims costs. In fact, billed charges themselves are somewhat arbitrary and can vary substantially by provider. There can also be some subtle inconsistencies in how vendors calculate discounts.

In claims repricing, each vendor “reprices” historical claims from the plan. This ensures the comparison is based on the same underlying data and is performed on the claims cost directly (rather than the discount from billed charges).

This approach is more technical and time-intensive than discount analysis and is prone to inconsistencies between vendors in how they perform the reprice. The results may also appear more rigorous than they actually are, as many vendors use discounts, aggregated at various levels, to arrive at the repriced claims amounts (not necessarily reflecting their actual contracts).

A Comparison of How the Two Methods Address the Three Core Components of Estimating Network Costs

With either of these methods, it‘s important to secure network-pricing guarantees. Without firm accountability, there’s an opportunity for exploitation or at best a reduced vendor incentive for due diligence.

Despite their limitations, when performed with skill and care, both methods can give meaningful and accurate relative results. Discounts correlating to costs is not an unreasonable assumption. For many plan sponsors, these methods are sufficient to make decisions with confidence.

That said, it’s worth considering where improvements can be made, not just in accuracy and applicability, but also in efficiency and value.

Reference pricing

Moving away from discounts (and billed charges in general) would be the most obvious place to start. This could be accomplished by leveraging an independent reference price, such as Medicare pricing, and comparing actual costs as a percentage of the reference price.

This method has the advantage of relying on actual claims costs (not highly variable billed charges) and should do a better job reflecting varying levels of acuity or treatment severity in the underlying experience. In general, a vendor paying a lower percentage of Medicare will mean lower claims costs, but there are some important nuances to that assertion. For instance, Medicare pricing includes facility-specific adjustments and other mechanisms or exemptions that may skew the results of a network comparison for a commercial population.

A process that avoids those adjustments and is only reflecting price differences on the basis of different weights or relative-value units would be ideal. A potential way to accomplish this would be to treat all facility care as though it occurred at the same facility (when calculating the Medicare price). Another option would be to compare the actual cost per relative-value units rather than as a percentage of Medicare.

In any case, along with the clear advantages there are some important limitations of this approach to consider:

• Vendors may not all have the required expertise or technical ability to perform reference pricing (particularly a customized version of Medicare).
• Calculations derived from Medicare experience may not translate well to a non-Medicare population for all types of care (e.g., maternity care).
• The process will likely be more time intensive than discount analysis or claims repricing for all parties.

In terms of the three core components of estimating costs, moving to reference pricing would significantly improve the cost of services received at identified providers, but with some tradeoff in terms of efficiency and complexity. This may be partially mitigated through an effectively managed process, including guides to assist the vendors and ensure consistent output.

There may also be scenarios where a mix of reference pricing and case rates, or even discounts, would strike the right balance between improved results and efficiency.

While reference pricing may alleviate the primary drawbacks of working with discounts, it does not fully address the challenge of different utilization mixes among competing vendors. The percentage of the reference price will vary by provider and by service. Moreover, the generalized historical result may not be indicative of the future results for a particular plan. This is an issue even when beginning with the plan’s historical utilization, as that utilization would not be static across vendors with material differences in their networks (members would shift to in-network options to some degree). Determining how exactly a given plan’s utilization would look on different networks, and forecasting those costs in a fair and objective way has largely been seen as being out of reach.

Prospective commitments

One novel approach is not to attempt to base the network comparison on historical results at all. Instead, medical vendors could be required to commit to plan-specific prospective rates or effective discounts, accompanied by strict (even dollar-for-dollar) guarantees for shortfalls. This is similar to what is currently common in contacts with PBMs.

This approach would allow each vendor to consider its own networks and management programs to arrive at proposed rates that reflect how they think plan utilization would occur in their networks. This could be done in a variety of ways to best align with plan objectives, budget and market nuances. Ideally, vendors would commit to some combination of costs relative to a reference price, cost per relative-value unit or case rates.

In all cases, the key is to align vendor incentives for network and cost management to those of the plan. Committing to prospective costs normalized to the underlying treatment severity, minimizing the reliance on billed charges and discounts, would be a big step forward in creating that alignment.

The obvious concern with anything prospective is that vendors could propose unachievable results to win business. To be able to challenge unreasonable results, it’s important to be aware of the variance between vendors’ historical results and what they are proposing.

It is also critical that the vendors are willing to stand behind their proposals by putting a substantial portion of their fees at risk. What is common today for the historical approaches, in terms of the share of fees being at risk for not meeting performance guarantees, is insufficient for a prospective approach. With reasonable safeguards in place, such as a risk corridor and certain exclusions, it should be possible to move towards greater alignment of incentives between plan sponsors and medical vendors, including a greater share of fees at risk for network cost performance specifically.

The "best" method depends on the plan

Although this discussion has focused on limitations in common current methods and moving in a direction that perhaps seems ambitious, it should be clear that methods like discount analysis or claims repricing may be the most appropriate method for many plan sponsors, depending on size, budget, timing and likelihood of achieving an accurate comparison.

The common thread in achieving success in selecting the right vendor partners is not any particular method, but the expertise and competence of the people applying that method, as well as having contracts that hold the vendor accountable for meeting their stated rate or discount promises.