Hospitals continue to publish machine-readable files (MRFs) containing pricing and negotiated rate information in accordance with CMS’ price transparency rule that became effective January 1, 2021. As of March, our audit of 50 hospitals found that 14% of hospitals are complying, which aligns with other studies that reported compliance rates of 17% and 20% at the end of Q1. While CMS contemplates more aggressive enforcement, many are eager to start using MRF data to inform strategic decisions, especially once payers publish their MRFs in 2022. However, our audit of the current MRF landscape found that even when posted the files are inconsistent and incomplete. Early conversations with vendors and plans suggest that payer files will have more expansive health data, but the market is likely to remain more opaque (than transparent) in the near-term.
What makes hospital MRFs misleading right now?
Retrieving and normalizing MRF data requires an enormous analytical effort from anyone hoping to make meaningful sense of the information that’s been published. From a technical perspective, no two MRFs look the same – file types vary and information in columns from different files aren’t standardized to match one another. However, even when datasets are clean enough to work with, some key features of the current MRF landscape make them challenging to use when making meaningful comparisons or evaluating strategic decisions.
- The data lacks context.
While hospital MRFs must report strategically sensitive information like payer-identified negotiated rates, they aren’t required to include relevant metrics that explain how to interpret rates. MRFs don’t include indicators such as quality scores, outcomes data, experience scores, care coordination support and patient reviews that demonstrate the value of services. They also don’t incorporate financial or market factors that impact rate setting such as patient risk mix, utilization volumes, market share, alternative payment models and site of care payment differentials. Of course, intangible characteristics like purchaser preference, brand name, strategic partnership and community relationships aren’t captured by MRFs at all, even though they play a key role in contracting.
- The data is an incomplete snapshot of any given market’s pricing.
The availability of files in any given market depends on hospital compliance, which is not yet the norm. The completeness of MRF data for a market will vary based on factors such as pressure from peers or the media. Providers in different markets publish at different paces, and some may continue to hold out entirely until CMS adopts harsher enforcement.
Hospitals also have strategic incentives to publish or not publish and may choose to post rates based on the type of organization they are. For example, high-cost organizations such as Academic Medical Centers may choose not to comply because they are the highest-cost provider in their market. Conversely, organizations that are lower-priced may publish to promote themselves as a more affordable alternative to their peers.
- The files may exclude data, and it’s hard to confirm if something is missing.
We can never be certain that all required information is included in an MRF at any given time. Files are not updated in real time, and elements such as items and services, codes, descriptions, or rates may be inaccurate or missing altogether. Further, even if a hospital discloses all the payers it contracts with (e.g. UnitedHealthcare), it may not list every product that payer offers (e.g. United Healthcare Options PPO or UnitedHealthcare Signature Value HMO). These exclusions and the rationales behind them differ by provider and will change over time as hospitals continue working towards full compliance. Some hospitals may take the approach of posting as many services as possible, while others may exclude lower-volume or contractually complex services while they get their bearings.
- The data is not meaningfully standardized.
Most challenging of all, hospitals and plans have widely variable definitions of services and base rates, making apples-to-apples comparisons difficult. Some hospitals may include professional or technical fees in their base rates because they employ physicians, while others without contracted physicians may only include technical fees. Hospitals may also use different codes for services, such as the charge code, billing code, revenue code, CPT/HCPS, or NDC, complicating service-level price comparisons.
While hospitals are supposed to provide the underlying rates for services in alternative contract structures such as bundles, capitations, per diems, or risked-based payments, CMS’ vague policy language leaves significant flexibility for hospitals to decide what information is ultimately published. Complex payment processes force organizations to choose a base value to report as their rate, and hospitals do not disclose what these rates represent or how they were calculated. Confronted with significant logistical hurdles and short timelines, hospitals may have taken shortcuts such as using historical payment amounts as proxies for more complex contractual data that is difficult to distill down to a single rate.
So, what can I use this information for right now?
Given these dynamics, we’ve examined how confidently you can use files now versus what requires better data to do in the future. After evaluating each of these actions, we awarded them a confidence grade which roughly translates to how usable insights from each analysis would be to inform strategic decisions.