Let us examine the most common tool in use today by consultants and carriers to evaluate the competitiveness of payor networks: Uniform Data System, also known as Uniform Data Specifications (UDS). UDS is a compilation of book of business (BoB) discounts, developed by large health payors (i.e., health plans or payors) and benefit consultant organizations, and is considered the industry standard for measuring and reporting on provider networks across geographies where employees reside.
Plan sponsors care a great deal about the function of healthcare benefits within the realm of human resources, especially with respect to how their employees are engaged, the quality and methods by which their employees receive care, and how the total cost of care is well-managed. Plan sponsors know the health of their employees directly impacts productivity, loyalty, job satisfaction, and time away from work duties. To address these concerns adequately, no component within healthcare benefits carries a greater weight than the composition and characteristics of the provider network. Many attributes of the network must be examined and understood to comprehensively assess a payor’s provider network in addition to the use of UDS. For this reason, UDS is viewed by many as largely inadequate in its use as a foundational tool for evaluating and comparing provider networks across payors. Understanding the complexities of care coordination and high- performing providers, as described earlier, the reader can agree that a network discount is only one piece of numerous, complex elements that contribute significant differences to the true value of a network, and in fact, the methods by which network discount is calculated via UDS are viewed as inconsistent and oversimplified.
UDS began BoB compilation in the late 2000s for the purpose of helping self-funded plan sponsors decide, across the markets where their employees reside, the payors who have the most favorable healthcare pricing for inpatient, outpatient, and professional service types.1 As the reader can imagine, the answer to this research may significantly vary market to market. The most prominent determinant of “favorable healthcare pricing” has historically been the degree of unit cost discount from billed charges to allowed amounts achieved by the payors.
UDS data content (Note, not data usage) supplied by the participating payors is decided by the UDS workgroup.2 Content is compiled and shared twice annually over a rolling 12-month period, containing approximately six months of claims lag.¹ Data aggregation and layout per service type are dictated in detail.
Limitations of Using UDS Data
For some, particularly payors who are competing by virtue of submitting UDS data, the “U” in UDS may be more representative of “unstable” instead of “uniform.” Reasons for this sentiment are numerous and representative of the changing healthcare industry, growing complexities of the provider network, and the accuracy of the submissions themselves. The network is far from simply a roster of hospitals and clinicians who have agreed to participate within a marketplace, and the variety of contracting methods is too large to include in a concisely written document. In fact, as previously mentioned, the specific providers are not the drivers of the data alone; the membership geography is the starting framework. Let us examine the top reasons why UDS is usually viewed as only one of many considerations for the proper evaluation of a payor network, and when used as the foundation for network comparison, is in fact, a flawed mechanism.
The value of the UDS data can be augmented by a claims repricing exercise.1,3 With a repricing effort, plan sponsors and brokers have an opportunity to review and compare their historical claims data file to a claims file that has been reconfigured, or repriced, by payors using their current negotiated arrangements during the same, defined timeframe and with the same provider and service mix that was used by the plan sponsor’s incumbent payor(s).
An incumbent payor, who is competing to retain the plan sponsor as a client, may also request to reprice their own claims, using prospective, newly negotiated provider negotiated rates as compared to the retrospective, historical claims. Repricing brings additional detailed information for consideration, but also brings more complexity, time, and variation. Because each payor tends to have its own preferred repricing method, repricing conclusions cannot be sufficiently compared to other payors’ assertions. Because of these factors, although repricing can bring more insights, many believe the inconsistencies associated with the exercise outweigh the benefits. As with other elements discussed, specific consistent requirements for repricing, where the numbers are not presented variably at the discretion of payors, are essential in making this a more valuable supplement for network evaluation. For any variation amongst payor repricing methods, we need to understand the rationale for such divergence from a standardly recommended approach.
Let us revisit the mechanics of examining a discount for evaluation and how these leave gaps in the assessment of the overall network effectiveness.4 To reiterate, discount evaluation of a network using rolled up data submissions is oversimplified, vague, and allows for conjecture of the true value of the data as demonstrated by inconsistent results and rankings submitted by consultants, even within the same metropolitan statistical area. For example, when a payor, in one year’s time, is depicted as having a three to five percent increase in discounts, then the reader of this information knows a more reliable approach, where data usage, rules, and methods are consistently applied amongst consulting firms is definitely needed, as this depiction is simply not doable.
Regardless of the data layout “specifications” being named as such, even dollar fields such as “discount, allowed, or eligible” can be open to interpretation, affecting validity of the analysis.5 The data studied does not entirely, if at all, accommodate utilization patterns, the overall size and mix of the network, the actual charges billed, the population demographics other than to examine a three-digit level zip code for employee residence, clearly delineated out of network utilization, or other costs that are charged to the plan sponsor via bank accounts/ claim wire. Out of network utilization, in particular, can create misalignment of costs, because under certain circumstances, members may receive out of network services from independent/ third-party medical networks (e.g., Naviguard, etc.), where the payors may benefit financially from a total paid amount, while members face high, out of network payments. To leave dynamics such as these out of the picture limits the overall value of the assessment and comparison of payor networks, underscoring the need for payors and consultants to revisit and develop an improved approach. We need for consultants to consider the totality of data and information more judiciously that is relevant to a network’s total value and to not accept broad statements of value that are mentioned separately as adjustments or in an appendix and not a part of the core data under consideration. We also need to insist upon reaching agreement on the most appropriate methods of capturing those elements outside of claims activity that impact total costs. This should not be a concern over discovery of proprietary information, as the assessment is more broad-based than rate specific, even with said improvements.
While network breadth is oftentimes a compelling advantage, and many larger plan sponsors seek an expansive network, more of them are also looking for narrower, high-quality networks (or HPNs, as described previously), at least as another option for their employees, that may offer HPPs and/ or Value Based Contracting (VBC) providers. (In fact, questions about the presence of VBC providers have become routine in plan sponsor requests for proposals.) Yet, discount analysis can potentially offer more “credit” to the size and discount of the network over other important quality and cost-containing features. What we have seen in the industry is an ebb and flow of interest in narrow networks since they tend to mostly attract non-utilizers or those who already seek services from the designated providers contained within them – many of us are not willing to switch from network providers to whom we have grown accustomed even with a higher premium price point; meanwhile, providers tire of the narrow network configuration, where they render services to basically the same patients but at even lower reimbursements. All this to say that when a payor offers multiple network options for selection, each option offered is worthy of a series of questions and answers to discern best choices for plan sponsor employees.
With respect to value-based arrangements in place with providers, whether they sit within large or narrow networks, great variety in construct of these programs exists for providers in the commercial segment, making the granting of UDS “credit” to these programs less straightforward and misrepresentative of the value they may offer to the overall network. For example, certain accountable care organizations may only cover fully insured, self-funded, or both sets of members. The arrangement may be upside only with performance payments linked to quality, cost, or both. The glide path to risk is slow, but hopefully progressive, which means year-over-year, surplus or risk-sharing percentages may change. The value model may or mostly cover the total cost of care or may only apply to specific condition or procedure-based episodes of care such as total joint replacement, cancer care, or obstetrics. The VBC network may only apply to certain payor products and may be single or multi-tier, meaning non-VBC providers may also be providing services to the same members as the VBC providers, either as supplemental providers or as a second tier, for example. The timing of these payments is also highly variable. To illustrate, care coordination fees that are designed to incentivize quality or provider investments or that offer a vehicle for interim payments may occur monthly, quarterly, or vary with actual performance and outcomes. Sometimes, performance payments “simply” drive the FFS payment higher or lower based on specific outcomes. With total cost of care models, a reconciliation of quality and cost within the year under examination may occur several months after the end of the performance year, which may not be captured within the timeline assessed for the discount analysis. And by the way, especially with providers new to these programs, surplus and risk payment amounts are difficult to predict, and the amounts can vary depending on the parameters within the negotiated contract. Notably, contract measurement is highly negotiable with commercial VBC contracts (unlike Medicare VBC, which dictates measurements and performance standardly). Also, the methods by which the payments are made to providers differ such as via plan sponsor claim wire access, special fees with self-funded plan sponsors, or additional payments made directly by payors for fully insured client membership; or payors may decide to secure certain guarantees with respect to these models with their clients.
Additionally, when a payor declares a certain number of members/ employees are aligned with value-based providers, the sophistication of the arrangement greatly impacts the true value to which this translates for the member and plan sponsor. To be declared “attributed” to VBC providers does not mean a great deal if the care coordination, aligned incentives, and patient engagement are not excellently managed, and providers are not all-in with supporting operational investments and functions. A common element across most VBC arrangements is they tout their objectives to improve quality, manage costs, and enhance the member and provider experience (quadruple aim). Clear indicators of each of these focus areas should be apparent, and UDS data specifications certainly do not adequately address these. Not only this, with attribution based VBC models, patients/ employees may have no idea they are “attributed to” a VBC provider or what that means to them, so declaring member experience is enhanced requires more evidence of such. What all this means for discount calculations is that they may be more error-prone due to the degree of variability in payment amounts, methods, timing, and structure of VBC arrangements. For a more comprehensive view, then, taking the presence of these models into account and seeking validation of their impacts on total costs, quality, and member and provider experience are good practices.
Importantly, the overall composition of the provider network directly affects plan sponsor risk attraction. To illustrate, Blue Cross/ Shield (Blue) plans tend to have the largest and “friendliest” provider networks, which may translate into the worst risk for a plan sponsor in the form of attained discounts. This is because of the frequency and depth of claims exceeding outlier clauses (described in more detail below) due to sheer volume of claims that are a direct result of number of providers but also (frequent) Blue market dominance in addition to per-member-per-month (PMPM) costs. This, in turn, leads to the need for improved risk adjustment accuracy considering case, provider, and service mixes, which as mentioned, is up for considerable diversification in techniques.
Moreover, when a discount percentage is stated from billed charges, billed charges can meaningfully vary between provider organizations in the same geography, not to mention those outside the member’s zip code that are included in the analysis due to the need for medical travel especially in rural areas or for those seeking centers of excellence offered by payors such as many have noted with companies like Walmart. Not only this, but charges may be exemplified through several billing methodologies: straight fee-for-service (FFS) or they may be filed as a case rate, a global rate, or a per diem rate, and they may mirror Medicare billing methodologies or could be a hybrid of Medicare and custom negotiations. So, translating all these potential combinations into a simple discount is not sufficient, especially if plan sponsors are attempting to translate these discounts into per-member-per-month impacts. Furthermore, FFS contracts may also contain outlier provisions that basically cause a default to billed charges when the tallied charges meet a pre-defined threshold, and those charges may or may not be first dollar. Another occasional practice with payor-network partnerships is for the payor to “buy down” the network such that in exchange for providing an upfront payment, the provider, typically a large health system, agrees to provide more favorable FFS rates or not increase them. As the reader can deduce, great variation amongst payor-provider contracts and partnerships makes the discount calculation a limited, albeit standard, analysis.
Lastly, this document lists areas of skepticism within the payor community regarding the practices of the competition, how they are represented to plan sponsors and benefit consultants, and how these factors can significantly impact costs to plan sponsors. Many believe, even with specification standards, too much flexibility exists including using the data specification form appendices, where participants share additional information. This document is a call to action to all of us in the healthcare industry to expect more, not just more data, but more transparency in data that is meaningful for true network valuation and how this translates into consequences for patient care. The way to accomplish this is to include evidence of medical and pharmacy reimbursement structures that incentivize clinical outcomes and care coordination, and the data represented needs to reflect the efforts, programs with providers, and payment mechanisms from all sources (e.g., claim wire banking, upfront payments in exchange for FFS discounts, value-based incentives, percent of charge defaults and other outlier payments). Broad sweeping statements or adjustments cannot be taken at face value or narrative alone; those simply result in inconsistent assumptions and generalizations with “credit” given. Instead, value drivers within in the network must be more plainly, yet comprehensively, shared – and in the context of the market where members reside. See concerns below voiced by the payors and others in the health industry, who are concerned with fair, transparent, complete, and quality-oriented submissions for evaluation of provider networks, from which plan sponsors make important decisions regarding the health of their employees:
- Belief some payors do not abide by the data specifications, resulting in inflated discounts
- Insufficient auditing, questioning, and validating of the discounts asserted
- Suspected practice of including projected discounts for providers that will soon be in-network without consistent auditing of actuals
- Exploiting the inclusion or exclusion of custom networks to their advantage
- Applying out-of-network discount assumptions
- Suspicion some payors submit adjusted data as actual
- Need for better definition of claim types, which causes questions about claims classification accuracy and consistency amongst participants
- Provider mix differences between payors
- Three-digit zip code too high level
- Inclusion of claims where third-party vendor negotiations took place, making verification difficult
- Inclusion of VBC reimbursement that is wrought with assumptions, as previously described
- Exclusion of certain providers
- Exclusion of outlier claims
- Exclusion of pharmacy
- Exclusion of zero discount claims
- Exclusion of claims associated with providers under examination for fraud, waste, and/or abuse
In conclusion, for network valuation, as with any initiative of this size, scope, and importance, care must be taken to question, determine, and establish definitions of success, the goals to be accomplished, and a clear business plan with corresponding project plans that enable deployment and adoption on a sustainable basis. Furthermore, in-depth measures that examine quality, patient experience, and health equity in addition to costs require significant consideration. Traditional methods of valuing payor networks such as UDS have insufficient value as a standard vehicle for plan sponsors and benefit consultants to compare networks within geographies where employees reside. While the concept of “uniformity” is admirable, criticisms of UDS abound for a variety of reasons related to comprehensiveness, accuracy, consistency, and discretionary interpretation without the necessary auditing and validation of assertions made.
Watch for the next Risk Strategies Consulting white paper where our team of experts provides a thorough assessment of provider network valuations.
Want to learn more?
Connect with Stuart Piltch on LinkedIn, here.
Learn more about Risk Strategies Consulting, here.
 Myers, L. (2019, January). WakeDat: Frequently Asked Questions. WakelyBCS. https://www.wakely.com/sites/default/files/files/content/wakedatfaq.pdf
 Bidnet. (2020, March 9). Discount data specifications. Bidnet. https://www.bidnet.com/bneattachments?/664175957.pdf
 McCann, D. (2019, September 20). Beyond discounts: A new approach to Health Plan Financial Analysis. CFO. https://www.cfo.com/news/beyond-discounts-a-new-approach-to-health-plan-financial-analysis/657373/
 Hiles, A., & Reilly, P. Aetna get a sharper view of medical costs across all carriers. NEEBC. https://neebc.memberclicks.net/assets/White-Papers/aetna-get-a-sharper-view-of-medical-costs-across-all-carriers-october-2015.pdf
 Myers, L. (2012). Determining discounts. Milliman. https://us.milliman.com/-/media/milliman/importedfiles/uploadedfiles/insight/healthreform/pdfs/determining-discounts.ashx