There is a famous ad for Maxell cassette tapes showing a man reclining in a chair in front of a speaker — hair windswept, tie flapping, lampshade tipped back like a sail, martini glass sliding off the table. The voiceover promised that even after 500 plays the high-fidelity tape would still deliver high fidelity. That was the literal message. The visual was the figurative one: The signal was so faithful it felt like a physical force that could move objects.
When there is a signal but not yet proof that a drug may cause harm, do payers change access rules? And if so, how? A safety signal, in the usual industry sense, is information about a potential causal association between a drug and an adverse event, drawn from adverse-event reports, clinical-trial data, and real-world studies. The question is whether that kind of signal is enough to affect formulary status or medical policy. The answer is sometimes — but less often and less visibly than one might expect.
Payers do not usually act like safety monitors
Published work on payer decision making, along with Currant’s recent interviews with pharmacy and medical directors, points to a useful starting point: Payers are generally reluctant to monitor adverse events themselves. They prefer not to be the party tracking side effects in real time. They tend to be more comfortable making access decisions than running safety surveillance.
Formal safety programs such as REMS shift monitoring and risk-management obligations to manufacturers, prescribers, pharmacies, or other parts of the health care system, not to payers. Payers may require compliance with those programs, but absent that kind of outside framework, they usually do not build their own safety-monitoring programs.
Payer reluctance to act as safety monitors narrows the likely forms of payer action. When payers do respond to safety signals, they typically shape access rather than monitor harm — and the response is rarely dramatic. It usually looks like utilization management, not safety monitoring: nonpreferred formulary placement, prior authorization, step therapy, or coverage with evidence development (CED).
Payers move when others translate safety signals
Safety concerns generally affect access when a potential causal association becomes visible to outsiders — translated into a form a payer can defend and use. That happens when trial data show adverse events of concern or imply a need for ongoing safety monitoring, when real-world evidence suggests problems in broader use, when a guideline or consensus statement identifies use that carries an elevated safety risk, or when a regulator or public payer turns the concern into a policy signal. Two examples illustrate the point.
Anti-amyloid drugs for Alzheimer’s disease. These drugs are a clear example of payers conditioning access on externally translated safety and benefit concerns. Medicare did not provide unrestricted coverage. Instead, CMS used CED — a Medicare policy that covers a treatment only when it is used within a CMS-approved clinical study or registry. The April 7, 2022, national coverage determination (NCD) covers FDA-approved amyloid-directed monoclonal antibodies under CED when the NCD criteria are met.
The CED decision was not based on safety alone; questions about clinical benefit were central. But safety was part of the picture. CMS’s decision memo discussed amyloid-related imaging abnormalities, or ARIA — potential harms such as brain swelling or bleeding. Those concerns reinforced the case for caution.
Anti-amyloid coverage has since evolved. After traditionally approved drugs entered the market, CMS allowed broader Medicare coverage through registry-based CED. That is wider access than the Aduhelm-era path, but still not unrestricted coverage. The key point holds: Safety-related uncertainty did not lead Medicare to monitor adverse events as an ordinary payer function. It led Medicare to condition coverage on structured evidence generation.
Rosiglitazone. Rosiglitazone shows that commercial plans and Medicaid programs can respond differently to the same safety signal. A 2007 FDA warning about cardiovascular safety concerns led some commercial plans to implement prior authorization, including edits intended to block rosiglitazone for patients who take medications that compound its cardiovascular risk. But the Medicaid response was much weaker. Ross and colleagues found that nearly all state Medicaid programs continued to cover rosiglitazone as a preferred drug after the 2007 FDA warning, with minimal change in prior authorization.
The Medicaid–commercial split matters. The same safety signal moved some payers and not others — a reminder that safety concerns affect access unevenly across the system.
Integrated delivery networks are a notable exception
The pattern of payers acting only after outside parties translate a safety signal has one notable exception: integrated delivery networks. When the payer and the provider are part of one organization — such as Kaiser Permanente — adverse events can be observed clinically, not just through claims data. The connection between clinical observation and coverage decisions inside an IDN is different from what exists at most payers, even if its practical effect on access decisions is hard to measure from outside.
Safety is one factor among many
Boxed warnings, the FDA’s most prominent drug-safety alert, do not consistently change payer behavior either. One analysis of Medicare Part D formularies examined 101 new boxed warnings and major updates affecting 68 oral drugs between 2008 and 2015. The share of formularies providing unrestricted coverage fell only modestly — from 65.4% before the warning, to 62.6% one year later, to 61.9% two years later. If a boxed warning produces this kind of muted response, then a preliminary safety signal — by definition less settled — is even less likely to move payers on its own.
Step-therapy research suggests that payers do not always prefer safer options. Recent work has found that protocols sometimes require patients to try agents that are not aligned with clinical guidelines, and that protocols are frequently more stringent than the guidelines themselves. That matters because guidelines often incorporate benefit–risk judgments, including safety and tolerability, even when they are not framed as “safety” documents.
Step-therapy practice does not prove that payers ignore safety. It shows that safety does not always control the decision. Economic considerations — including cost-effectiveness, which folds in efficacy — and administrative simplicity often weigh more heavily.
What the evidence tells us
The literature on safety signals and payer behavior is thinner than one might expect, and there are two explanations. Both are probably true.
One explanation is that safety signals are often less decisive than other factors, especially when the drug offers a meaningful clinical benefit or fills an unmet need.
Another explanation is that safety’s role is hard to observe from the outside. Public formularies and policy documents can show that a drug is nonpreferred or subject to prior authorization, but they often do not reveal why. Internal committee discussions are usually private. Net prices and rebate arrangements are confidential. Even when safety matters, it is bundled with cost, efficacy, and general uncertainty in ways outsiders cannot disentangle.
Wind you cannot see
The Maxell ad worked because it made an invisible signal visible. With payer decisions, the visibility runs the other way. While the access rule is public, the economic and clinical forces that shaped it — safety only a component of the clinical ones — usually are not. From the outside, it can be hard to tell whether safety was a breeze in the background or a real force pushing the decision.
No Comments