AI Integration and Systemic Reform Are Needed to Balance Resource Protection With Timely Access to Treatment

Emerging technologies and process reconsiderations may alleviate time and resource stressors oncology teams face when appealing patient insurance denials.

Although many multidisciplinary oncology teams are equipped to file high volumes of patient insurance claims and appeal coverage denials, the process is often less than seamless. Furthermore, denied insurance coverage can lead to interruptions and delays in patient care. Emerging artificial intelligence (AI)–based technologies may alleviate some of the time and resource stressors these teams face; however, wide-scale change is incumbent upon reevaluating the prior authorization process as a whole and creating an optimal balance between health care resource protection and ease of access to guideline-concordant care.

In an interview with OncLive®, Kate Baker, MD, MMHC, medical director of value-based care at Tennessee Oncology in Lebanon and Gallatin, explained that large practices, including hers, often have pharmacists, prior authorization coordinators, and other team members whose responsibilities include submitting claims to insurance companies, connecting patients to copay assistance programs, and appealing denials, thus lightening this day-to-day load for providers. However, findings from a 2022 survey conducted by the American Society of Clinical Oncology (ASCO) Association for Clinical Oncology showed that among 300 ASCO members practicing in the United States who responded to the survey, 52% reported having 2 or fewer staff in their practices who worked exclusively on prior authorizations.1 Regarding workload, 56% of respondents shared that their practices completed a maximum of 50 prior authorizations per week, and 53% of respondents reported spending a maximum of 40 hours per week on these tasks.

Additionally, circumstances can arise where providers are needed to step in to offer their expertise to strengthen an appeal. For instance, after a claim is denied following a previous denial and appeal, insurance companies may request peer-to-peer review calls, during which providers must provide additional information beyond the contents of the claim to explain why they have ordered a certain test or treatment for a given patient, Baker said. She estimated that these calls each take up to 20 minutes of clinic time and occur approximately twice a month in her practice, although those working at practices with fewer resources may face higher volumes. Results from the ASCO survey revealed that although completion of a single prior authorization that does not escalate beyond the initiating staff member was reported to take a maximum of 1 hour in 56% of respondents’ practices, completion of prior authorizations that do escalate was reported to take a maximum of 3 hours each among 63% of respondents.1

Many insurance roadblocks are related to defending the medical necessity of imaging tests, Samyukta Mullangi, MD, MBA, a medical oncologist at Tennessee Oncology in Dickson, added in a separate interview with OncLive. She noted that this aspect is largely based on trial and error and requires constant communication between medical oncologists, laboratory specialists, and front desk staff to create an appeal that will be approved.

Patient Harms Due to Prior Authorization Requirements (n = 300)

  • Treatment delay (96%)
  • Diagnostic imaging delay (94%)
  • Pivot to patient’s second choice of therapy (93%)
  • Increased out-of-pocket expenses (88%)
  • Therapy denial (87%)

How Do Insurance Denials Affect Outcomes of Patients With Cancer?

Mullangi explained that in many cases, delaying treatment initiation is inevitable regardless of insurance-related obstacles, due to patient needs to see additional physicians or receive further imaging. However, she and Baker stated that in other cases, waiting to administer treatment is not clinically defensible and is directly related to insurance denials, which can delay treatment at crucial points in the patient care journey.

Baker noted that the process of appealing denials can last between a few days up to a few weeks or even months, which is valuable time during which a patient with metastatic disease could be receiving therapy. For instance, Baker referenced a patient with metastatic colon cancer who experienced worsening symptoms and eventually died from their disease before receiving insurance approval for their third line of therapy. Data from the ASCO survey revealed that the most frequently reported patient harms due to prior authorization requirements included treatment delay (96%), diagnostic imaging delay (94%), pivots to the patient’s second choice of therapy (93%), increased out-of-pocket expenses (88%), and therapy denial (87%).

Mullangi emphasized the importance of planning in advance whenever possible to create a buffer period of 1 or 2 weeks between a patient’s diagnosis and the scheduled start of their treatment, to account for the time required for the prior authorization process, potential denials, and resulting appeals. However, she noted that this is not always possible, especially for patients who require immediate care.

In response to some of these patient-facing issues, the Tennessee Oncology Practice Society backed an amendment to a Tennessee bill, effective January 1, 2026, that prohibits health benefit plans from requiring step therapy protocols—requirements for patients to progress on or experience adverse effects with a less expensive drug before an insurance company will approve a more expensive treatment prior to providing coverage of approved prescription drugs—for patients with stage IV advanced metastatic cancer or a hematologic malignancy.2,3

How Might AI Help Oncology Teams Streamline the Prior Authorization Process?

“Submitting prior authorization letters for the same medicine for many patients…that’s a repetitive process,” Baker stated.

As a result, Baker explained that many practices, including Tennessee Oncology, are exploring ways to incorporate AI into the insurance submissions and appeals workflow to expedite parts of the process. She noted that AI’s strengths lie in quickly generating materials for repetitive tasks, which is directly applicable to insurance proceedings such as writing prior authorization and appeal letters. Additionally, she stated that AI scribe tools can take detailed notes about individual patient cases and import more diagnosis codes into the electronic health record, which may make prior authorization letters more informative for insurance companies and prevent many denials in the first place.

Mullangi concurred that AI may help streamline the more abstract aspects of the process. She highlighted the potential for natural language processing to compile relevant insurance coverage justifications from provider notes about a given patient, then compare those notes with established guidelines that the insurance company has for approving or denying claims. She also spotlighted the potential role for automated clinical decision support that could suggest materials or information that providers could add to strengthen the prior authorization requests.

“We’re still in a stage where a human would need to double-check the work being done by AI, but [AI] could take a lot off the plates of the humans and allow us to be a bit more efficient at getting those prior authorizations through,” Baker summarized. “I’m hopeful that all these tools will work together synergistically to help move this process along a little faster.”

What Large-Scale Changes Are Needed to Improve Prior Authorization Processes in Oncology?

Mullangi noted that the proliferation of prior authorization requirements has expanded beyond particular cases or uses for certain treatments and now encompasses almost every type of test and treatment, regardless of guideline compliance. She emphasized that the amount of payer strain related to fraud or waste that this process eliminates may not be enough to offset the magnitude of the logistical hurdles associated with appealing denials, particularly given the benefits that cancer therapies can provide when administered appropriately.

“Before creating a solution around AI, it would be good to revisit the whole concept and ask: Is this meaningfully deterring excessive waste or fraud in health care?” Mullangi underscored. “Are the risks that this system is imposing worth the benefits of cutting down this purported waste?”

She added that beyond imposing treatment delays, the logistical hurdles of insurance denials sometimes cause enough administrative burden to convince physicians and/or patients to abandon the effort to appeal and instead choose a different treatment plan.

“That is not how we should shave off spend,” Mullangi argued. “It should not be by creating so much friction in the system that folks give up out of exhaustion, rather than thinking through appropriate [vs] inappropriate care.”

References

  1. ASCO Prior Authorization Survey Summary. ASCO. Accessed October 6, 2025. https://cdn.bfldr.com/KOIHB2Q3/as/wt2thjbxbj24ncwnwjwzvk9c/2022-ASCO-Prior-Auth-Survey-Summary
  2. HB 0850. Tennessee General Assembly. Updated May 27, 2025. Accessed October 6, 2025. https://wapp.capitol.tn.gov/apps/Billinfo/default.aspx?BillNumber=HB0858&ga=114
  3. Fischer MA, Avorn J. Step therapy-clinical algorithms, legislation, and optimal prescribing. JAMA. 2017;317(8):801-802. doi:10.1001/jama.2016.20619

Related Content:

PublicationsOncology Business Management
VISION Data Show Promise of Tepotinib in Treatment-Naive NSCLC Harboring MET Exon 14 Skipping Mutations: With Catherine Shu, MD
Updated FIREFLY-1 Data Further Support Tovorafenib as a SOC in R/R Pediatric Low-Grade Glioma
ctDNA Assays Are Poised to Reshape Lymphoma Treatment Strategies: With Sarah Rutherford, MD