Who Will Pay for AI?

In 2020, the largest U.S. health care payer, the Centers for Medicare & Medicaid Services (CMS), established payment for artificial intelligence (AI) through two different systems in the Medicare Physician Fee Schedule (MPFS) and the Inpatient Prospective Payment System (IPPS). Within the MPFS, a new Current Procedural Terminology code was valued for an AI tool for diagnosis of diabetic retinopathy, IDx-RX. In the IPPS, Medicare established a New Technology Add-on Payment for Viz.ai software, an AI algorithm that facilitates diagnosis and treatment of large-vessel occlusion strokes. This article describes reimbursement in these two payment systems and proposes future payment pathways for AI. Keywords: Computer Applications-General (Informatics), Technology Assessment © RSNA, 2021.

Evolving Radiologist Participation in Medicare Shared Savings Program Accountable Care Organizations.

PURPOSE: The aim of this study was to temporally characterize radiologist participation in Medicare Shared Savings Program (MSSP) accountable care organizations (ACOs). METHODS: Using CMS Physician and Other Supplier Public Use Files, ACO provider-level Research Identifiable Files, and Shared Savings Program ACO Public-Use Files for 2013 through 2018, characteristics of radiologist ACO participation were assessed over time. RESULTS: Between 2013 and 2018, the percentage of Medicare-participating radiologists affiliated with MSSP ACOs increased from 10.4% to 34.9%. During that time, the share of large ACOs (>20,000 beneficiaries) with participating radiologists averaged 87.0%, and the shares of medium ACOs (10,000-20,000) and small ACOs (<10,000) with participating radiologists rose from 62.5% to 66.0% and from 26.3% to 51.6%, respectively. The number of physicians in MSSP ACOs with radiologists was substantially larger than those without radiologists (mean range across years, 573-945 versus 107-179). Primary care physicians constituted a larger percentage of the physician population for ACOs without radiologists (average across years, 66.3% versus 38.5%), and ACOs with radiologists had a higher rate of specialist representation (56.0% versus 33.7%). Beneficiary age, race, and sex demographics were similar among radiologist-participating versus nonparticipating ACOs. CONCLUSIONS: In recent years, radiologist participation in MSSP ACOs has increased substantially. ACOs with radiologist participation are large and more diverse in their physician specialty composition. Nonparticipating radiologists should prepare accordingly.

Clinical Decision Support: The Law, the Future, and the Role for Radiologists.

Clinical Decision Support (CDS) was designed as an interactive, electronic tool for use by clinicians that communicates Appropriate Use Criteria (AUC) information to the user and assists them in making the most appropriate treatment decision for a patient's specific clinical condition. Policymakers recognized AUC as a potential solution to control inappropriate utilization of imaging and made CDS mandatory in the Protecting Access to Medicare Act of 2014. In the years since Protecting Access to Medicare Act, data on the potential impact of CDS has been mixed and much of the physician community has expressed concern about the logistics of the program. This article aims to review the legislation behind the AUC program, the events that have transpired since, and some of the challenges and opportunities facing radiologists in the current environment.

The Quality Measure Crunch: How CMS Topped Out Scoring and Removal Policies Disproportionately Disadvantage Radiologists.

PURPOSE: CMS implemented Merit-Based Incentive Payment System (MIPS) policies to cap points and remove "topped out" quality measures having extremely high national performance. We assess such policies' impact on quality measure reporting, focusing on diagnostic radiology. METHODS: Data regarding MIPS 2019 quality measures were extracted from the CMS Quality Benchmarks File and the Quality Payment Program Explore Measures search tool and summarized by collection type and specialty. RESULTS: Among 348 MIPS measure-and-collection-type combinations, 40.5% were topped out (56.6% of those with a benchmark) and 23.3% were capped. Among measures with a benchmark, the percent topped out varied (P < .001) by collection type: claims 82.7%, qualified registry 60.4%, electronic health record 11.6%. The percent capped was also greatest for claims measures (52.3%). Among 699 Qualified Clinical Data Registry (QCDR) measures, 63 had a benchmark, of which 44.4% were topped out. The percent of measures topped out also varied significantly (P < .001) by specialty, ranging from 0.0% (electrophysiology) to 95.0% (diagnostic radiology). Among 20 unique measure-and-collection-type combinations for diagnostic radiology, only one was not topped out, and 30.0% were capped. Among 20 radiology QCDR measures, 5 had a benchmark, of which 3 were topped out. CONCLUSION: CMS topped out measure scoring and removal policies disproportionately impact radiology, which has the highest topped out percentage among all specialties and only a single non-topped out measure. This asymmetry disproportionately impairs radiologists' MIPS flexibility and is anticipated to progress in ensuing years. Current CMS policies create a looming crisis for radiologists in MIPS. The high risk of an insufficient number of available quality measures creates an urgent need for new radiology measure development.

A Business Case for Artificial Intelligence Tools: The Currency of Improved Quality and Reduced Cost.

For data science tools to mature and become integrated into routine clinical practice, they must add value to patient care by improving quality without increasing cost, by reducing cost without changing quality, or by both reducing cost and improving quality. Artificial intelligence (AI) algorithms have potential to augment data-driven quality improvement for radiologists. If AI tools are adopted with population health goals in mind, the structure of value-based payment models will serve as a framework for reimbursement of AI that does not exist in the fee-for-service system.

Facility-Based Measurement in the Merit-Based Incentive Payment System: A Potential Safety Net for Which Most Radiologists Will Be Eligible.

OBJECTIVE. The purpose of this study was to assess the percentage and characteristics of radiologists who meet criteria for facility-based measurement in the Merit-Based Incentive Payment System (MIPS). MATERIALS AND METHODS. The Provider Utilization and Payment Data: Physician and Other Supplier Public Use File was used to identify radiologists who bill 75% or more of their Medicare Part B claims in the facility setting. RESULTS. Among 31,217 included radiologists nationwide, 71.0% met the eligibility criteria for facility-based measurement as individuals in MIPS. The percentage of predicted eligibility was slightly higher for male than female radiologists (72.9% vs 64.5%). The percentage decreased slightly with increasing years in practice (from 78.8% for radiologists with < 10 years in practice to 67.3% for radiologists with ≥ 25 years in practice). The eligibility percentage was also higher for radiologists in rural as opposed to urban practices (81.6% vs 71.3%) and in academic as opposed to nonacademic practices (77.2% vs 70.3%). However, the percentages were similar across practices of varying sizes. There was also a greater degree of heterogeneity by state, ranging from 50.9% in Minnesota to 94.0% in West Virginia. By overall geographic region, the percentage of predicted eligibility was lowest in the Northeast (64.7%) and highest in the Midwest (78.3%). A higher percentage of generalists met the 75% facility-based threshold than did subspecialists (77.3% vs 65.4%). When stratified by subspecialty, however, facility-based eligibility was lowest for musculoskeletal radiologists (38.1%) and breast imagers (45.1%) and highest for cardiothoracic radiologists (85.1%). For other subspecialties, predicted eligibility ranged from 66.0% to 77.8%. CONCLUSION. Most radiologists will be eligible for facility-based reporting for MIPS in 2019, with some variation by demographic and specialty characteristics. The facility-based option provides a safety net for radiologists who face challenges accessing hospital data for reporting quality measures. In general, radiologists should not alter their current MIPS strategy but should instead consider facility-based measurement as a contingency plan that could result in a higher final score.

Medicare Access and Children's Health Insurance Program Reauthorization Act: The Basics for the Breast Imaging Radiologist.

Breast imaging radiologists are considered by many to be leaders among diagnostic radiologists in the transition to value-based care. Many strategies for success in the changing healthcare landscape are exemplified by the day-to-day practice of breast imaging, including well-developed quality measures, standardized accepted best practices and terminology, and a prominent role in communicating with patients and coordinating care. Further development of these strategies will be important for continued success in both the Merit-Based Incentive Payment System and in alternative payment models.

MACRA 2.5: the legislation moves forward.

The Medicare and CHIP Reauthorization Act of 2015 remains the payment policy law of the land. 2017 was the first year in which performance reporting will tangibly impact future physician payments. The Centers for Medicare & Medicaid Services (CMS) considers 2017 and 2018 transitional years before full implementation in 2019. As such, 2018 increases the reporting requirements over 2017 in the form of a gradual phase-in while introducing several key changes and new elements. Indeed, it is the nature of the transition itself that led to the somewhat unique title of this manuscript, i.e., MACRA 2.5. Stakeholder feedback to the CMS regarding the program has ranged widely from the elimination of core components to expanding reporting to non-government payers. This article explores the potential impact on neurointerventional physicians.

Nephrogenic systemic fibrosis: possible association with a predisposing infection.

OBJECTIVE: Infection at time of MR contrast administration has been reported to predispose patients with renal failure to development of nephrogenic systemic fibrosis (NSF). We assessed the frequency of infection at the time of MR contrast administration in a group of NSF patients. MATERIALS AND METHODS: Eight patients developed NSF during 2002-2006, of whom seven received the MR contrast agent gadodiamide (Omniscan), with doses of 0.10-0.31 mmol/kg. Data for the following were available for only 2005 and 2006: numbers of infected and uninfected renal failure patients who received MR contrast material and number of contrast-enhanced MR scans in all patients. We extrapolated data to 2002-2006 to approximate rates of NSF in infected and uninfected renal failure patients using Fisher's exact test for association between variables and calculated odds ratios with 95% CIs. RESULTS: Five of seven NSF patients receiving MR contrast material had infections at the time of contrast administration. Three hundred thirty-four patients with renal failure received MR contrast material in 2005 and 2006 (29 infected). The rate of NSF was 6.7% in infected renal failure patients and 0.26% in uninfected patients. Data extrapolated to 2002-2006 yielded estimates of 75 infected and 750 uninfected renal failure patients. The association of NSF with infection was highly significant (p < 0.001) with an odds ratio of 25 and CI of 3.9-264.4. CONCLUSION: The association of NSF with infection was highly significant, supporting the hypothesis that infection at the time of MR contrast administration in renal failure patients predisposes to NSF.