Ribbon Health’s response to the CMS “Request for Information Regarding a National Directory of Healthcare Providers & Services”

Ribbon CEO Nate Maslak shares what it will take to build a National Directory of Healthcare Providers and Services

When CMS first issued a Request for Information regarding a National Directory of Healthcare Providers and Services, my first thought was: this is an incredibly challenging problem to solve. Given the complexity and importance of this problem, I commend CMS for spearheading the critical effort to establish a National Directory. There are many sources of truth for provider data (e.g., claims, credentialing, health plan directories, etc.), and if there was a quick way to spin up one unified database, the industry would have already done so. The best solution needs to rely on a strong set of partnerships between public and private sectors and include all stakeholders (e.g., health plans, providers, patient organizations), with CMS and other governmental organizations leveraging their power to set expectations and standards. Technology companies must then work together to build software, systems, and processes that solve the key challenges our healthcare system faces today: access, cost, and quality. 

I’ve personally felt the negative consequences of incomplete or inaccurate provider data. I experienced the challenges of our healthcare system when my mom was left in the dark as she tried to navigate the healthcare system, resulting in a series of medical bills that put her financial health at risk without improving her physical health. I was determined to fix this broken system and, along with my cofounder Nate Fox, launched the first iteration of Ribbon – a care navigation tool to help people find the right care at the right time for them. Users loved the solution but became quickly frustrated when the doctor they chose to see ended up being out-of-network, at a different address, or sometimes was no longer practicing. My cofounder Nate Fox and I recognized that this was not a navigation issue but an inaccurate provider data problem. And, that problem was a huge one.

Barriers to finding a provider can have a significant impact on health outcomes for patients, and drive up costs on both sides of the system. Billions of dollars of waste in our system is directly due to inaccurate and disparate provider data. If a patient finds a provider that has outdated insurance coverage information, and then receives a surprise bill for thousands of dollars, there’s a much higher likelihood that they’ll avoid seeking additional care. Beyond the obvious need for information that is accurate - we need to provide people with information on providers that will ensure a strong patient-provider match. Research shows that when someone goes to a provider that they identify with - (whether that be through language, gender, ethnicity, etc) they are more likely to stick with that provider and follow care recommendations. 

Why is the provider data problem so hard to tackle? What would it take to create a single-source-of-truth? I encourage you to read the full response from the Ribbon team on why this is such a difficult problem to solve across the industry, as we share our hands-on learnings and thoughts on how building a National Directory through public and private partnerships can strengthen access to high-quality and affordable healthcare for all. 

Ribbon Health’s response to the CMS “Request for Information Regarding a National Directory of Healthcare Providers & Services”

Ribbon Health (Ribbon) appreciates the opportunity to provide input on CMS-0058-NC and supports the effort to establish a National Directory of Healthcare Providers & Services (NDH).  Provider data is the connective tissue of the healthcare ecosystem, and maintaining accurate directory information is critical for patients, payers, and providers.  Below, Ribbon has outlined recommendations for how the Centers for Medicare & Medicaid Services (CMS) can address the challenges of building a directory at scale, which will require the collaboration of private and public sector stakeholders to develop the architecture and support the implementation of the NDH.  We provide these recommendations based on CMS’s specific questions as well as our expertise in data and directory management.

Ribbon is dedicated to building the infrastructure to improve healthcare access with actionable provider information, including insurance coverage, pricing, and performance data.  Since 2016, we have worked to build comprehensive and accurate directories for patients, providers, and healthcare stakeholders. Our data platform and provider data management tools create solutions to standardize data inputs, improve care navigation, and simplify referrals.  Given our vast experience working in this space, we look forward to assisting CMS and other key contributors to make healthcare information more accurate, accessible, and actionable.  

1. Questions on the establishment of an NDH

1a. What specific health information exchange or use cases would be important for an NDH to support?

Recommendation: CMS should articulate which specific use cases are to be supported by the NDH, and we recommend prioritizing uses that assist patients in finding care.

Existing provider directories often fail to provide accurate information to patients, leading to barriers in locating and identifying an appropriate care provider.  According to a Ribbon consumer survey conducted by OnePoll, one in three Americans has had a negative healthcare experience because provider information on their health plan’s website was incorrect.  ​​Moreover, without valid information, care navigators are unable to effectively route patients to appropriate providers and ensure healthcare access for patients.  

The problem extends beyond patients themselves being unable to find care.  Providers and referral coordinators face a similar issue in facilitating referrals, which are critical to managing a patient’s care journey and outcomes.  The NDH can further advance goals outlined in the CMS Hospital Readmissions Reduction Program (HRRP).  Traditional referral management workflows can be extremely time consuming and error-prone when coordinating care navigation, leading to higher costs and fewer patients treated.  These challenges can cause care continuity disruptions and lead to higher readmission rates in acute care settings.  By facilitating timely referrals, the NDH can reduce hospital readmission rates by ensuring patients receive better continuity of care post-discharge.

Without reliable provider data, payers also face certain risks, including (i) failing to to meet network adequacy standards and (ii) creating ghost networks where providers are listed as part of a network, but in practice, are either not in-network or are not accepting new patients.  According to a March 2022 report by the Government Accountability Office (GAO), which focused on provider directors in three U.S. cities, 22 percent of the phone numbers listed for psychiatrists who treat adult patients were wrong, and 21 percent of the psychiatrists were not accepting new patients.  

These inefficiencies ultimately harm patients by reducing their access to in-network providers and adequate networks.  Given these existing problems, CMS should focus first on ensuring that the NDH is formatted and built to improve the task of finding a relevant and available care provider.  While the CMS’s Request for Information (RFI) contemplates a host of use cases, it is important to start with the threshold issue of patient access.  Keeping access as the focal issue of the NDH will also ensure that the data needed, and the format used, can address the foundational issue of correctly and accurately locating available care.

1b. What provider or entity data elements would be helpful to include in an NDH for use cases relating to patient access and consumer choice (for example, finding providers or comparing networks)?

Recommendation: CMS should build intuitive data schemas to empower patients to improve access and consumer choice.  CMS should consider including provider data elements, such as gender, language(s) spoken, acceptance of new patients, telehealth designation, specialty, and conditions treated, to facilitate patient searches for available and appropriate providers. 

For the NDH to best support access to high-quality care, the data and underlying taxonomies must be built with the patient in mind.  Specifically, without more detailed data available on providers, such as in-network status, provider gender and language(s) spoken, acceptance of new patients, and telehealth designation, patients struggle to identify providers that meet them where they are and to select providers they can trust to continue care.  

Today, companies partner with Ribbon to match patients to the appropriate provider, specialist, or facility based on a patient’s request or need for care.  By deeply embedding itself into customer workflows, Ribbon quickly learned that baseline provider data (e.g., names, addresses, and phone numbers) is insufficient to enable proper care decisions, such as searching for a provider that treats a specific condition.  For this reason, Ribbon has taken a patient-centric approach to build a comprehensive provider dataset, most recently prioritizing the patient’s provider search experience by appending the provider directory with a bespoke specialty taxonomy.

While an important step in the right direction, the CMS provider specialty taxonomy was born out of academic departments and written in a way that someone without a medical or research degree may struggle to understand.  Billing and claims processing workflows have since been built atop the specialty infrastructure to enable institutions to have a common language.  Similarly, medical boards were also established to oversee specialty certifications, which leverage the same specialty jargon.  The American Board of Internal Medicine, for example, oversees hematology, nephrology, and rheumatology, among other specialities.

Unfortunately for the patient, diseases and conditions often span specialties, which means that the present-day terminology does not always align with a patient’s search for the right doctor based on her ailments and symptoms.  For example: a patient seeking care for “allergies” could receive care from a dermatologist, immunologist, otorhinolaryngologist (head and neck), or an allergy specialist.  Under the current CMS specialty taxonomy, a patient who searches online for an allergy specialist will be presented with a narrow, incomplete list of providers, especially if the patient does not know to search for key terms like “immunology” or “otorhinolaryngology.” 

To address this challenge, Ribbon prioritized provider searchability in its own directory and recognized the value of surfacing data on patient panels, conditions treated, and treatments performed.  Ribbon built its own specialty data field called “Focus Areas” to drive simple, intuitive search results for patients as a supplement to the CMS specialty taxonomy.  Rather than relying on care navigators to search conditions such as “atopic dermatitis,” “flexural eczema,” and “eczema,” Ribbon clusters each condition into an intuitive single search term (e.g., “eczema”) to facilitate patient searches for appropriate providers.  These examples persist across nearly every specialty and condition, resulting in inefficiency, waste, and delayed or inappropriate care.

2. Questions on the risks, challenges, and prerequisites associated with implementing an NDH

• What are the most promising efforts that exist to date in resolving healthcare directory challenges?  
• How could CMS best incorporate outputs from these efforts into the requirements for an NDH?  
• Which gaps remain that are not being addressed by existing efforts?

Recommendation: To address current challenges in identifying provider locations and facilities, CMS should build unique location identifiers. 

Presently, CMS holds payers responsible for keeping provider directories accurate, but there is little guidance available on required data fields and storage protocols.  The rollout of the National Provider Identifier (NPI) was a significant step forward for interoperability, introducing a standardized, unique key for individual providers and enabling organizations to aggregate multiple data sources together and improve health interoperability.  However, there is no convention like the NPI for locations or facilities data.  As a result, a single hospital could be tied to hundreds of variations of names, spellings, and acronyms, which further contributes to noise and fragmentation.  For example, “Stanford Children’s Hospital — Lucile Packard” (the women’s and children’s hospital based in California) may show up in any of the following forms: “Children’s Hospital Stanford,” “LPCH,” “Stanford Children’s Hospital,” and “Lucile Packard Children’s Hospital Stanford.”  

Ribbon has seen many customers in the industry attempt to use Type 2 NPIs as a proxy for unique location identifiers.  However, Type 2 NPIs do not specify the location type, such as care delivery sites or billing facilities, which creates challenges to understanding organizational relationships between locations, health systems, and payers.  As an example, when searching for “Cleveland Clinic” in Ohio in the NPPES NPI Registry, over forty unique Type 2 NPIs are returned for unlabeled, overlapping addresses.  Unfortunately, Type 2 NPIs are ineffective as comprehensive unique locations identifiers, especially for facilities like “Cleveland Clinic” which span multiple national locations.

The ripple effects of provider location errors are costly.  CMS released a third annual review of Medicare Advantage Organizations (MAOs) provider directories, which found that close to 49 percent of provider locations had at least one error, and “inaccuracies with the highest likelihood of preventing access to care were found in 41.75 percent of all locations.”  As a result of inaccurate location data, health plan teams are forced to manually validate rosters line-by-line, introducing a higher margin of error.  Provider referral volumes are also impacted due to inaccurate service locations listed.  Worst of all, patients could unknowingly show up at the wrong hospital or location, resulting in denied care or, receiving care and subsequently being sent a large medical bill.   

A single location can exist across hundreds of data sources, like directories, rosters, the United States Postal Service (USPS) address database, NPPES NPI Registry, and more.  Without a unique identifier or key for locations, it will be virtually impossible for CMS to facilitate connecting data sources together and to support interoperability.  Many Ribbon customers, in lieu of unique location identifiers, rely on TIN/EIN (Tax ID) as the location identifier since it is tied to billing relationships.  Unfortunately, this further proliferates data fragmentation because billing, tracked by Tax ID, is disconnected from care metrics, tracked by NPI.  The more disjointed the data, the harder it is to understand which locations are both high-quality and low-cost based on data inaccuracies.

In building our directory, Ribbon prioritized establishing unique identifiers for locations.  Similar to how Google Maps clusters data based on consolidated location data, local government data, and historical traffic patterns, Ribbon aggregates hundreds of sources using its own machine-learning methodology.  By using this approach, Ribbon has developed a national dataset of more than one million healthcare locations, with a universally unique identifier (UUID) for each distinct name and address combination.  These UUIDs enable Ribbon to remain flexible and nimble, supporting a wide range of use cases across customers, such as primary care provider selection and more.  As we look forward, Ribbon is exploring open-sourcing its provider location UUID data to drive deeper industry collaboration and facilitate interoperability.  Ribbon would be happy to share more insights with CMS on addressing this specific challenge when creating the NDH.

3. Questions on data verification and accuracy 

3a. What use cases would benefit from data being verified and what sort of assurances would be necessary for trust and reliance on those data?

Recommendation 1: To streamline data collection, CMS should aggregate existing, verified data sources, thus reducing the burden on providers for data entry.

Today, providers must provide data to multiple bodies, including payers, state licensure boards, credentialing services, and more.  These various data submission processes have inconsistent formats and upload methods.  This fragmented data collection is a contributing factor to provider disengagement, friction, and delays, which impacts care delivery and patient health outcomes at scale.  The Agency for Healthcare Research and Quality has noted the rising and concerning prevalence of burnout among clinicians, with some studies finding burnout rates of over 50 percent.  Provider burnout can lead to negative downstream impacts on access to care, patient safety, and care quality. A publication in the Journal of General Internal Medicine noted that roughly 9 percent of physicians who experience burnout are likely to have made at least one major medical error in the past three months. 

The NDH has the potential to alleviate this provider burden by automating data collection from existing and reliable sources.  Vast provider data already exists within payer portals, Independent Physicians Association (IPA) rosters, health system rosters, and medical boards databases.  If CMS can accurately aggregate existing data, the NDH can reduce the manual data entry burden on providers by building upon verified, highly confident data sources. 

Recommendation 2: To increase confidence in, and adoption of, the NDH, CMS should transparently report data accuracy, such as displaying confidence scores for provider data elements.

When CMS launched the NPPES in 2007, there was limited enforcement and accountability to ensure the data remained accurate.  This resulted in diminished public trust and lower adoption of the NPPES as a centralized directory.  To improve upon the NPPES implementation experience, the NDH could indicate how accurate each provider record is and could show improvements in accuracy over time.  Through transparent reporting of data accuracy, CMS could increase confidence in, and adoption of, the NDH.  

Given ongoing consolidation in the healthcare ecosystem, coupled with the inconsistent contracting schedules between health plans and providers, it is challenging to maintain and uphold provider data accuracy standards.  To foster trust with our customers, Ribbon reports on provider data accuracy through a confidence score model.  The model predicts the probability of data accuracy and assigns a score to provider names and locations.  The confidence score ranges from zero to five, where zero represents that the data is confirmed to be false, one represents less than 20 percent accuracy, four represents approximately 90 percent accuracy, and so on.  The scores also change dynamically to reflect the most current accuracy assessment.  This level of transparency reminds Ribbon customers that not all data can be expected to be accurate today and shows that our data is generally becoming more accurate over time.  Ultimately, confidence scoring enables our customers to monitor and manage provider data quality in real-time, mitigating risk and potential impacts on patient care.

4. Questions on alignment with CMS priorities 

• We want an NDH to support health equity goals throughout the healthcare system. What listed entities, data elements, or NDH functionalities would help underserved populations receive healthcare services? 
• What considerations would be relevant to address equity issues during the planning, development, or implementation of an NDH?

Recommendation: Ribbon believes the development of the NDH would directly accelerate the CMS Strategic Plan to protect and strengthen patient access to high-quality and affordable healthcare.  CMS should include relevant data elements that can expand equitable access to culturally competent healthcare.

The NDH has the potential to advance health equity initiatives by coalescing real-time provider data.  As you are aware, the Department of Health and Human Services (HHS) has established a national priority to increase choice, affordability, and enrollment in high-quality healthcare coverage by 2026.  According to the Office of the Assistant Secretary for Planning and Evaluation (ASPE), “[s]tudies show that people without health insurance coverage are less likely to receive necessary preventative care and screening services, have less access to care and experience worse health outcomes than those with health insurance coverage.”  To reduce uninsured rates and improve healthcare literacy, patients must have a frictionless and seamless process to evaluate and select health plans.  With a few clicks, patients should be able to compare health plan benefits and evaluate proximity to in-network facilities with a high degree of confidence.  The NDH is well-positioned to become the database that aggregates provider, location, and network data to establish a comprehensive source of information for patients about the providers and facilities that are in-network for any given plan.

The NDH can also directly support HHS Objective 1.3 by incorporating data elements that address culturally competent care.  In a 2017 HHS Survey, 32 percent of Black adults and 33 percent of Hispanic adults said it was very important to have a healthcare provider who shared or understood their culture, compared to only 13 percent for White adults.  By aggregating provider data – such as race and ethnicity, languages spoken, gender, LGBTQ+ affirming – the NDH can help reduce health disparities by ensuring that patients can access health services that are respectful of, and responsive to, their culture, language, and other needs.  We recommend these data fields be nationally-mandated in the NDH to ensure the consistent capture and standardization of data.

As noted in HHS Objective 1.3, it is vital to support underserved populations with an inclusive healthcare experience.  CMS is positioned to reduce health disparities by both aggregating and surfacing culturally competent provider data in the NDH, and by ensuring that the NDH data model is inclusive and specific.  As an example, listing the languages spoken by a provider and also capturing the provider’s language proficiency (novice, intermediate, fluent) would enable patients to find providers that meet their linguistic needs.  According to BMC Medical Education, if an individual can easily find a provider that speaks her language, her care outcomes will improve as the result of reduced medical errors, better understanding of and adherence to treatment plans, and higher patient satisfaction. 

Similarly, in constructing the NDH data model, CMS can promote health equity by ensuring the NDH captures the full breadth of ethnic populations and racial identities in the U.S. today.  By comprehensively capturing ethnic and racial identity data, payers can build targeted provider networks and health programs that serve the specialized needs of specific ethnic populations.  Through the NDH, CMS can advance its 2023 strategies to expand equitable access to comprehensive care.

5. Questions on the approach to NDH implementation 

5a. We are soliciting comments on the feasibility of a phased approach to implementation and potential opportunities to build stakeholder trust and adoption along the way: What entities or stakeholders should participate in the development of an NDH, and what involvement should they have?

Recommendation: CMS should build cross-sector public-private relationships to inform the development of, reinforce the public trust in, and improve the implementation of the NDH.

Currently, no single governmental body, organization, or company is positioned to solve the issue of provider data accuracy alone.  Cross-sector collaboration is critical to build and govern the NDH. Historically, partnerships between the private and public sectors have been successful in addressing complex, systemic issues across healthcare.  For example, the recent Da Vinci Project encourages industry leaders to open-source use cases and solutions built upon the Health Level Seven International (HL7) Fast Healthcare Interoperability Resources (FHIR) platform.  This collaboration has resulted in meaningful progress toward centralized, interoperable solutions on a national scale. 

Ribbon encourages CMS to engage with public and private stakeholders to leverage existing improvements and tap into institutional knowledge in order to continue developing solutions for provider data accuracy.  Ribbon has gained significant and useful experience in building and improving provider data platforms, and our industry-wide expertise can help inform the architecture, implementation, and maintenance of the NDH.  

We also encourage CMS to convene key stakeholders to continue sharing information, experiences, and ideas to build the NDH.  Discourse among government, technical, venture, and health leaders would spur a collective call to action and would assist in breaking down existing siloes that have, to date, served as barriers to building a better health information ecosystem.  Through industry-wide collaboration, CMS can establish the NDH that is best suited to drive positive healthcare transformation at scale. 

—

Download Ribbon’s full response to the RFI. 

‍