Evaluating the Impact of Restricting AI-Driven Medical Necessity Decisions to Physicians on Health Insurance Utilization Reviews

Utilization Review (UR) and Prior Authorization (PA) processes are meant to make sure healthcare services are used appropriately and costs are managed. Usually, healthcare providers send clinical information to insurance companies. Then, insurers decide if a treatment or service is needed based on clinical rules and the patient’s condition.

Recently, health plans and insurance companies have begun using AI tools to help with these reviews. AI can look at patient records, clinical notes, and medical rules to flag cases for human check or suggest if treatments should be approved, denied, or need more checks. This can make decisions faster, reduce paperwork, and improve accuracy in routine cases.

Even with these benefits, there are worries about relying too much on AI. AI might miss some details about each patient or show biases from its training data. Because of this, rules now say that clinicians must oversee AI decisions to keep things fair and correct.

Regulatory Framework Governing AI in Medical Necessity Decisions

In the United States, both the federal and state governments have started making rules about how AI can be used in healthcare management. California is especially active in making clear laws.

Federal Actions:

• In October 2023, President Joe Biden’s Executive Order told the U.S. Department of Health and Human Services (HHS) to create policies for using AI in healthcare, focusing on protecting patient rights.
• The Centers for Medicare & Medicaid Services (CMS) made rules affecting Medicare Advantage plans:
  • The 2023 CMS Medicare Advantage final rule says that medical necessity decisions must consider each patient’s unique situation and cannot rely only on AI.
  • The 2024 CMS Interoperability and Prior Authorization Final Rule requires faster replies to PA requests. It allows AI to help, but only clinicians can make the final decisions.

These federal rules make sure AI supports human clinical judgment instead of replacing it.

State Laws – Focus on California:

California passed new laws about AI in healthcare. Senate Bill 1120 (SB 1120), starting January 1, 2025, controls how AI can be used by health care plans and disability insurers during utilization reviews. Some key points are:

• AI must use detailed clinical data from patients’ histories and provider notes, not just general datasets.
• Only licensed physicians or qualified healthcare workers can make final medical necessity decisions. AI cannot deny, delay, or change care on its own.
• Health plans have to explain their AI policies to regulators, providers, patients, and the public.
• AI systems must follow anti-discrimination laws to treat all patients fairly.
• The California Department of Managed Health Care (DMHC) and Department of Insurance will audit AI use to check if the rules are followed.
• Penalties for breaking the rules can include fines and other actions.

These laws stress human control and openness while letting AI help.

Implications for Healthcare Organizations and Health Plan Operations

Hospital leaders, physician group owners, and healthcare IT managers need to make many changes to follow these rules. The main effects are:

1. Necessity of Physician Oversight

Healthcare groups must make sure AI-supported decisions are reviewed and finalized by physicians or qualified professionals. Doctors use their judgment to decide medical necessity. Workflows must be designed so AI suggestions go through a human check without slowing down patient care.

2. Transparency in AI Use

Organizations have to clearly tell patients when AI is used in decisions. Patients should know how to talk to a human if needed. This is important for those managing patient communication.

3. Data Privacy and Security

In California, AI-generated data is treated as personal information under the California Consumer Privacy Act (CCPA). This means healthcare groups must protect AI data like other health information, using strong safeguards to stop unauthorized access.

4. Non-Discrimination Assurance

Health plans need to check that AI does not cause bias or unfair treatment. This means ongoing monitoring and updates to AI systems to follow federal and state anti-discrimination rules, especially for diverse patient groups.

5. Preparation for Audits and Reporting

Health plans and providers in California must keep detailed records about AI tools, training data, policies, and compliance efforts. Regulators will review this information. Teams must have good data systems ready for audits.

AI and Workflow Automation: Supporting Compliance While Enhancing Efficiency

These laws limit AI from making final medical decisions but do not remove AI’s help in utilization reviews. Healthcare groups that use AI automation with strong clinical oversight can improve efficiency and still follow rules.

AI as a Decision Support Tool

In utilization management, AI can quickly find patterns, pull important clinical information, prioritize urgent cases, and make first checks to help human reviewers. This lowers doctors’ workload so they can focus on tougher cases.

Streamlined Prior Authorization Processing

The CMS prior authorization API rule starting in 2027 requires payers to give PA decisions fast: 72 hours for urgent cases, 7 days for normal ones. AI can help meet these deadlines by automating routine checks, verifying documents, and making initial recommendations for humans.

Integration with Electronic Health Records (EHRs)

AI tools working with EHR systems can pull out patient data during clinical work. This cuts delays from manual chart reviews and uses AI’s ability to sort large data quickly.

Communication Automation with Compliance

AI automation tools can send patient notices about utilization decisions and AI use. These systems add required disclaimers about AI and tell patients how to reach a human provider. This follows transparency rules like California’s AB 3030.

AI Training and Continuous Quality Monitoring

Using AI needs regular checks. AI models should be updated with current clinical data to avoid old or biased results. Organizations should have ongoing quality checks and feedback from clinicians to keep AI accurate and fair.

Challenges and Considerations for Medical Practice Administrators and IT Managers

Following rules like SB 1120 brings challenges for healthcare administrators and tech leaders.

Balancing Efficiency and Compliance

Automating reviews promises faster results and cost savings. But systems must clearly separate AI support from final human decisions to meet the rules.

Ensuring Physician Engagement

Doctors have heavy workloads. To get quick reviews of AI advice, systems need easy interfaces, alerts, and maybe more staff to help manage utilization reviews well.

Data Management and Privacy Compliance

Because AI data counts as personal info, groups must strengthen policies for data security, encryption, access control, and breach response. Compliance teams must keep up with changing privacy laws.

Transparency with Patients and Providers

Administrators should plan how to explain AI’s role simply to patients and providers to avoid confusion or distrust. Training staff on these explanations helps keep patient trust.

Preparing for Regulatory Audits

Records on AI use, clinical decisions, and policies must be clear and routine. IT systems should create audit trails automatically to lessen manual paperwork.

Final Thoughts

Limiting AI-driven medical necessity decisions to licensed physicians affects how utilization reviews are done in health insurance. Federal and state rules, led by California’s SB 1120, set clear standards. Human clinical judgment stays at the center of care decisions while AI helps as a support tool. Medical practice administrators, owners, and IT managers across the U.S. must balance following rules, working efficiently, protecting patient privacy, and being clear with patients. Using AI as a support tool in utilization reviews can improve administrative work without breaking laws or ethics. As AI changes, watching new rules and keeping strong human oversight will stay important to manage health insurance reviews well and responsibly.