ASA Responds to CRUSH RFI, Offers Perspectives on Use of Artificial Intelligence on Coding and Billing Practices

As the Centers for Medicare and Medicaid Services (CMS) looks to reduce fraud, waste, and abuse, a targeted approach focused on areas with demonstrated patterns of such inappropriate or fraudulent claims is the suggestion by the American Society of Anesthesiologists. This is the response ASA submitted to the Comprehensive Regulations To Uncover Suspicious Healthcare (CRUSH) Request for Information (RFI), which described potential pathways for CMS to reduce fraud, waste, and abuse, mostly with changes to coding and billing processes.  ASA's comments reflected the uniqueness of anesthesia billing and coding, including anesthesiology’s reliance on appropriate documentation from hospitals and other facilities, surgeons, and other stakeholders to submit Medicare claims.

Anesthesiologists, like other physicians, are committed to ensuring that claims submission and documentation are complete and accurate, enabling timely and appropriate payments by CMS and private payers. ASA agreed with the current CMS filing timeline that provides necessary flexibility for anesthesiologists and their groups to complete required coordination processes and resolve documentation discrepancies. CMS suggestions to reduce claim filing deadline to 90 to 180 calendar days would create significant challenges for hospital-based physician specialties such as anesthesiology. Unlike many office-based physicians, anesthesiologists rely on clinical and administrative information generated by multiple external parties before a claim can be accurately submitted. Required billing elements, including operative documentation, finalized anesthesia record times, modifier requirements, patient demographics, insurance verification, and coordination of benefits determinations, are frequently dependent on hospitals, surgical facilities, and ordering physicians.

Because anesthesiologists and their groups are dependent on submitting claims based upon information they do not control, anesthesiologists, ASA indicated anesthesiologists would likely face unnecessary burdens and challenges complying with shorter submission deadlines. Anesthesiologists would be particularly affected because anesthesiologists typically do not control patient registration, insurance verification, or other necessary documentation generated by the facility. ASA comments described clinical workflows, describing how anesthesiologists often do not include a face-to-face opportunity to confirm insurance or demographic information with the patient. Delays in obtaining complete documentation are common in complex perioperative cases involving teaching modifiers, medical direction requirements, reconciliation of electronic anesthesia records, or interfacility transfers requiring coordination across multiple health systems. When information is incomplete, anesthesiology groups must conduct follow-up actions with facilities, receiving hospitals, ordering physicians, patients, and health care surrogates, all of whom may not be able to respond within a shortened filing timeframe. 

ASA suggested CMS focus attention for a shorter claim filing deadline for certain high-risk items and services through a targeted, risk-based approach that focuses on categories with historically high fraud, waste, or abuse risk. Rather than applying a uniform deadline across all health care professionals, CMS may wish to identify high-risk services using existing program integrity data such as aberrant billing, improper payment audits, and fraud investigations.

ASA also commented on the use of Artificial Intelligence (AI) in coding and billing, expressing both our interest in having AI contribute to more streamlined processes and coding accuracy as well as caution against removing human accountability in the submission process. ASA acknowledged AI technologies have improved significantly in recent years but indicated that the final review and accountability for claim submissions should continue to rest with appropriately trained human reviewers. AI programs are in early stages of development and their use, efficacy, and accuracy are still under scrutiny. ASA suggested CMS move with caution before over-relying on AI for coding and billing processes. For now, use of a decision support system is likely the most efficient and realistic use for AI in coding rather than a complete solution. Because of this, ASA suggested AI should only be used to expedite workflows by automatically and intuitively assisting with coding and billing processes. AI solutions should be subject to human review and incorporate human coder feedback to continuously improve and minimize errors. ASA supported the training of AI solutions on the desired outcomes from the claims and CMS should take steps to increase pre-determined metrics such as first claim acceptance and net collection rates. 

For additional information, please contact ASA Department of Payment and Practice Management at [email protected].