Price Transparency in Medical Bills









Overview of Question

The demand for price transparency for medical services, particularly for “out-of-pocket” expenses, has increased to support the vision of consumer-driven healthcare. Price transparency aims to “allow patients to know exactly what they should expect to pay and lets them compare the reduced cash prices with the prices they would pay if they went through insurance.” The Centers for Medicare & Medicaid Services (CMS) set a deadline of January 2021 for price transparency for hospital services.

The price transparency regulations have two goals: Firstly to increase competition among hospitals and health systems and, ultimately, lower prices; and secondly to provide consumers with the information needed to make informed decisions. The initiative expects that transparency rules, along with new interoperability rules, could eventually lead to less price variation among healthcare providers, improvements in market differentiation, and a shift to value-based care. In quantitative terms, the federal government estimates that real-time transparency on individual “out of pocket” expenses can save insurers and members $154 million in medical costs. The three-year average annual burden and cost of implementing the rule will range between $5.7 billion and $7.9 billion for insurers.

Description of Current Status of Implementation

Despite the clear benefits, the recent implementation has been less successful. According to NBC News, The Journal of the American Medical Association found that hospitals are largely ignoring the law. The report analyzed 5,000 hospitals nationwide and found that just 300, or fewer than 6 percent, were fully compliant with the rule, meaning they had publicly published both machine-readable files and separate price estimators for shoppable items. Many hospitals were partly compliant, but 50 percent had neither of the two required components.”

It is surprising that there is no price transparency tool for healthcare services, specifically “shoppable services,” given the requirements from CMS over the past two years.

Description of Challenges and Causes for Implementation

So why has it been difficult to develop a price transparency tool?

•           Could it be due to technical complexity?

From a data science workflow perspective, implementation involves data collection, cleansing, access to data through a program like Python or R, and making the database machine-readable (which meets CMS hospital pricing transparency requirements). One can do exploratory data analysis and data visualization to present insights. This does not require any machine learning models as there is no hard mandate for prediction, classification, or pattern recognition. It is more of a data look-up project than a data science project, focusing on retrieving, reporting, and accurately reflecting the cost of all flavors of a service/procedure to patients. The challenge is not with the math or programming required for data collection; manipulation; or query/visualization as the CMS requirements do not ask for cross-entity data exchange. The CMS mandate is more on reporting than analysis, although some basic data visualization is always welcome.

•           Could it be a funding issue?

From the information above, implementation costs should be low. According to the NBC report, “The PatientRightsAdvocate.org report estimates the requirements would cost hospitals $12,000 to implement, which includes publishing price estimators and compiling machine-readable lists of procedure costs.” So clearly funding is not an issue either.

•           Could it be data availability and accessibility?

In this case, the availability of price data for common procedures. With almost all hospitals and clinics using some kind of electronic medical record (EMR) system and business management system to handle back-office functions, including billing and reporting, pricing data is accurately recorded in the system(s). However, this data may need to be queried from the EMR, practice management system, clearing house, or even banking systems. Today’s technical setup allows payments to be traced back to every encounter, with a CPT code for every procedure. Extracting this data may require a small data consulting project, but generally, the resources are readily available.

Also, although it is a valid concern to worry about sensitive healthcare data as the mandate is not asking for any HIPPA-sensitive information, it is asking for an average price tag for a standard procedure in one’s internal database (not asking for any integration with other entities’ databases) without traceability back to any patient information. This should not be an issue.

Then what are the REAL challenges?

Challenge 1: Resistance from data providers (Not “Can’t”, but “Won’t”).

After eliminating technology, process, and market factors (both government and patient demand for knowledge of expected fees before receiving service), the only remaining factor is people/stakeholders other than government and patients. Could it be that hospitals and other provider organizations do not want to provide price transparency?

The consequences of noncompliance are high: CMS encourages anyone to report noncompliance, and if confirmed, the hospital could be subject to a fine of $300 to $3,000 per day for each day of noncompliance.

So why do more than 50% of hospitals choose noncompliance despite this financial risk? It may be a business model challenge rather than a technical restriction. The seemingly costly noncompliance might be a deliberate decision after weighing the pros and cons of price opaqueness on the hospitals and other providers (who currently have no compliance issues, but this could change after CMS finishes with hospitals).

It is against the business objectives of payers and hospitals to enhance price transparency, as it promotes competition among healthcare providers and drives down prices. We should consider whether the concept of “healthcare consumerism” is a misnomer and an unattainable goal in today’s healthcare delivery mechanism, which does not align with price transparency.

According to a Harvard Medical School survey, over 75% of hospitals (a random sample of 100) failed to publicize their negotiated rates with payers for common services. Some cited that exposing their negotiated rates with different healthcare plans and employers would enable the other party to negotiate more aggressively on rates. However, the same survey found that over 80% of respondents had adopted “selective” compliance, with most implementing a cost estimation tool on their website. After filling out a dynamic questionnaire including personal choices of the location of service, insurance plan (full insurance information required unless “cash price” is chosen), and procedure name, users receive a highly “qualified” estimate (with a disclaimer that the hospital cannot be held responsible if the final bill amount differs).

Challenge 2: The raw data is messy and incomplete.

After reviewing some of the machine-readable files, a study found that data provided by health plans are inconsistent and often inaccurate or incomplete. For example, one health system’s prices submitted by one of its own health plans for one service line were up to 10 times the actual negotiated rate, and only a portion of the services it had negotiated was included in the file. These issues should be identified and resolved in the monthly reporting required of health plans under the Transparency in Coverage rule.

The challenges in reporting healthcare pricing in a machine-readable format were not unexpected. Pricing for ancillary services like laboratory and radiology is generally straightforward and maintained in fee-schedule or unit-cost-based pricing structures. More complex services, like surgical procedures and inpatient stays, are often priced based on acuity. The price may reflect the number of procedures performed, the length of stay, the type of implants, and/or the drugs required, and may also factor in the level of care required given the patient’s morbidity. This pricing is not conducive to a matrix-like reporting structure, but these issues will likely be resolved over time, resulting in a more accurate and fair representation of healthcare prices across the ecosystem.

Prospects for resolution of the question/use case

There is no doubt that price transparency in healthcare is no small task. The policymakers like CMS have initiated the push. From the data science community’s point of view, it might be pragmatic to tackle this requirement in smaller chunks starting with small prototypes to prove the feasibility of each technical solution point. Here are some examples of the “smaller” prototypes even small data science vendors/teams can create:

• Reduction of the data scope: select only the “most often billed” procedure codes
• Simplify the estimation tool inputs: Apply ML regression models on datasets from different providers and through “Feature Importance”, reduce the number of metrics in the tool estimation
• Regular “retraining” of the model (“price prediction”) with actual new billing information to improve the accuracy of the model.

Going forward, we can focus on providing transparency for patients’ total “out-of-pocket” costs, including both insurance-assisted costs and full “cash prices” for healthcare services beyond just common hospital services. However, pure price estimation for these services can be complex, as it involves many factors such as the selected procedure (which may or may not include professional fees from specialists), the location of the service, the provider (as each provider has different contracted reimbursements with in-network health insurance carriers), and the patient’s insurance or payer plan. It is difficult to calculate a uniform price, as evidenced by disclaimers on price transparency tools stating that estimates are “only valid for today” and are not guaranteed. To improve price transparency, we can consider aggregation of services and prices from different sources(providers), reduction of middlemen in data collection, and digitization of pricing information.

In the future, providers and health plans may use the new pricing information during contract negotiations, potentially changing the dynamics of these negotiations while the consumer/patients can finally make an informed healthcare decision based on the quality and price of the services.

1. The problem is the lack of price transparency for medical services, particularly for out-of-pocket expenses, which is hindering the consumer-driven healthcare vision.
2. The problem is significant because it makes it difficult for patients to know exactly what they should expect to pay and compare the reduced cash prices with the prices they would pay through insurance.
3. The Centers for Medicare & Medicaid Services (CMS) set a deadline of January 2021 for price transparency for hospital services with the goals of increasing competition among hospitals and health systems, lowering prices, and providing consumers with the information needed to make informed decisions.
4. Despite the clear benefits, the recent implementation has been less successful, with only 6% of 5,000 hospitals nationwide fully compliant with the rule.
5. The challenges to developing a price transparency tool include resistance from data providers who do not want to provide price transparency, which goes against their business objectives of enhancing price opaqueness and promoting competition among healthcare providers. The consequences of non-compliance are high fines, but more than 50% of hospitals still choose non-compliance despite this financial risk.