Authorization and payment are one decision: linking bill review to utilization review

In most claims organizations, utilization review and bill review report to different leaders, run on different systems, use different vendors and measure themselves on different numbers. UR counts turnaround and guideline compliance. Bill review counts savings below billed charges. Neither is accountable for the thing that connects them: whether the care that was authorized is the care that was delivered and paid, and whether the care that was paid was ever authorized at all.

California law treats them as one decision. Operations rarely do, and the gap is where friction, liens and avoidable spend accumulate.

The statute draws the line clearly

Three provisions do most of the work.

Authorization is final. Labor Code section 4610.3 provides that an employer that authorizes medical treatment cannot rescind or modify that authorization after the treatment has been provided, for any reason. For a series of treatments, only the services not yet rendered can be modified.[1] In plain terms: once UR approves and the provider performs, the medical necessity question is closed. A retrospective medical-necessity denial at the bill review stage is not a savings opportunity; it is an invalid act that generates a dispute the payer will lose.

Payment runs on its own clock. Under section 4603.2, payment with an explanation of review is due within 45 days of receipt of an itemized bill with required reports and any written authorization. If the provider disputes the amount, they have 90 days to request second review, the employer must respond with a final written determination within 14 days, and undisputed balances are due within 21 days.[2]

Then the dispute forks by type. If the only issue is the amount, the provider has 30 days after second review to request independent bill review; missing that window deems the bill satisfied.[3] If the issue is medical necessity, it belongs to UR and IMR, not IBR. Which means the single most consequential act in the payment workflow is correctly classifying what kind of dispute you are in — and that classification depends on whether the authorization record is attached to the bill.

Four seams where the link breaks

1. Authorized care denied on the payment side. A provider obtains authorization, performs the service, bills, and receives an explanation of review reducing or denying payment on medical-necessity grounds. This is the clearest violation of section 4610.3, and it is common enough that provider-side billing firms publish standing advice about it.[4] Each instance costs the payer twice: the amount eventually owed, plus penalties, interest, lien exposure and the credibility of every legitimate review decision.

2. Delivered care that was never authorized. The mirror image. Services appear at bill review with no matching RFA or a mismatched service code. Without a link to the authorization record, the reviewer is left guessing whether the service was approved under a different description, was part of an authorized series, or was never requested.

3. Services that skip UR entirely and carry no price control. California's MTUS drug formulary exempts many first-line drugs from prospective UR, by design, to cut friction. But exemption from review is not exemption from cost. CWCI found NSAIDs' share of workers' comp drug spend rose from 14.2% to 23.5% after the formulary took effect, driven largely by two low-volume, high-priced drugs that are exempt from prospective UR and lack price controls.[5] When the utilization gate opens, the payment gate has to be the one paying attention.

4. Price control mistaken for utilization control. Fee schedules are effective at what they do, which is regulate prices, not volume. NCCI's analysis of workers' compensation fee schedules concludes that they control cost per service but have limited ability to move utilization toward group-health levels; utilization is the domain of guidelines, review and care coordination.[6] Bill review that squeezes unit price while volume drifts is optimizing half the equation, and UR that manages volume while pricing leaks is optimizing the other half.

What the research suggests the linked system is for

The empirical record on utilization management is a useful corrective to both sides. In a study of 9,319 workers' compensation cases, Wickizer, Lessler and Franklin found UM denial rates of only about 2% to 3%, many later reversed, with the measurable effect coming from length-of-stay management.[7] Review does not earn its keep by saying no. It earns it by shaping what is requested and how fast appropriate care is released.

What does move outcomes is coordination. Washington's Centers of Occupational Health and Education intervention, which paid providers for care coordination and timely reporting rather than for denials, produced meaningful reductions in disability among injured workers in a population-based evaluation.[8] And the clinical literature is blunt about what happens when the early window is mismanaged: early opioid exposure after a back injury is associated with roughly double the risk of long-term disability,[9] and non-indicated early MRI with substantially longer disability and $12,948 to $13,816 in additional medical cost per case.[10]

Those are treatment decisions with payment consequences that show up months later in the bill stream — which is exactly why the bill stream should feed back into review.

Five things to instrument

1. Give the authorization an identity that survives to the bill. Every approval should carry a durable identifier linked to claim, body part, service codes, date range and units, and that identifier should be matched at bill adjudication. Most "surprise" payment disputes are matching failures wearing a clinical costume.
2. Classify the dispute before you price it. Build the fork into the workflow: amount-only disputes route to second review and IBR; medical-necessity disputes belong to UR and IMR. Explanations of review should use reason codes that map one-to-one to the authorization state (authorized and matched, authorized and mismatched, not requested, exempt from UR, liability contested), because that mapping is what keeps a payer out of the wrong forum.
3. Reconcile authorized versus delivered, continuously. Compare units approved to units billed by episode, not by line. Physical therapy visits, injections and imaging series are where drift lives, and the pattern is visible weeks before it becomes a lien.
4. Watch the no-UR lane. For services and drugs exempt from prospective review, price and volume monitoring at the payment layer is the only control you have left. The NSAID pattern in California is the case study.
5. Close the loop back into review. Bill data is the only record of what actually happened. Feed it back: which approved services were never delivered, which guideline-concordant pathways finished on schedule, which providers generate repeat mismatches. That feedback is how a review program improves rather than merely repeats.

Governance is the point

Linking UR and bill review is not primarily a savings play. It is a governance play. When the authorization record, the clinical rationale, the service record and the payment decision sit in one traceable chain, four things become possible that are not possible otherwise: you can show a regulator how a decision was made, you can stop paying twice for the same argument, you can tell a provider network something true about its own performance, and you can measure whether faster decisions actually produced better recoveries.

California already writes the connection into statute. The work left is operational: one identity for the decision, one classification of the dispute, and one record that follows the claim from request to payment.

This article is general information about workers' compensation rules and practice, not legal advice. Check current statutes, regulations and case law for your jurisdiction.

Back to all articles