Just when you think we’ve gotten over it, “fail first” policy pops up again. In this case, the North Carolina Department of Health and Human Services, Division of Medical Assistance (DMA) has announced a new “fail first” policy for antipsychotic medications (see North Carolina Medicaid Implements Fail-First Policy For Non-Preferred Antipsychotic Medications). If you’re not familiar with the concept, essentially a consumer of health care must “fail” on a “preferred” (meaning less costly drug) in order to get payment for a different medication. In this case, as of January 1, 2015, consumers in the North Carolina Medicaid program must “fail” on an antipsychotic medication that is on the “preferred” list before prior authorization will be granted for a non-preferred antipsychotic medication.
I think “fail first” policies are bad policy for a number of reasons – the health economics, the science, and the ethics of the policy are questionable. At the macro level, these policies are put in place purportedly to save money for the health plan. But the whole notion is a symptom of our “siloed” approach to health care resources. Is it really cost effective system design to force a “health catastrophe” in order for a consumer to get access to health resources (see What to Make of Current Policies About Access to Antipsychotic Medications: Short-Term Gain for Long-Term Pain)? Probably not if you’re looking at total consumer health resource use over time – rather than just considering spending on medication. And then there is the benefits administration of this policy. What is the criteria for a treatment “failure?” How is that documented and how does authorization for another medication actually happen?
Then there are the scientific issues. First, there is the issue of “bioequivalence” among antipsychotic medications. There is certainly controversy about whether these medications are similar enough to form a “class” and are really interchangeable (except for the now-generic versions of the same medications). And while the North Carolina list of “preferred” medications is long – the policy ignores that issue altogether. And, there is the new evidence from genetic research that “schizophrenia” is more than one disease (see Genetic Research Finds Schizophrenia Is Eight Distinct Disorders) – a situation not anticipated in the development and testing of the current crop of antipsychotic medications.
And then there are the ethical issues. There has been an emerging body of research that each “psychotic” episode (essentially the “failure” that is required) results in permanent damage to consumer brain functioning. (For more on this, see Does Drug Treatment Prevent Brain Damage In Early Psychosis Or Schizophrenia? and Schizophrenia in Translation: Is Active Psychosis Neurotoxic?) Is it ethical to require a “treatment default” that is so personally damaging to consumers? Can you imagine consumers for other types of health care services accepting care that operates under the premise that treatment failure with permanent long-term consequences is a requirement for access to some available treatments?
There is no doubt that policies like “fail first” policies have teeth. A recent review found that many health plans include “fail first” policies – policies that were cited a impeding consumer access to optimal treatment (see Medication Restrictions Significantly Affect Mental Health Treatment Outcomes; New Survey Cites Formulary Restrictions, Prior Authorization And “Fail-First” Rules As Obstacles). This is part of the mix of factors that contribute to a too-high proportion of people with mental illnesses not receiving treatment that reflect current thinking about best clinical practices (see For 39% Of Consumers With First-Episode Psychosis, Prescriptions Not Aligned With Best Practices).
When talking with my colleague, Jonathan Evans, the chief executive officer of Safe Harbor and a member of the OPEN MINDS advisory board, his perspective as a professional working directly with consumers brings an added perspective. “Fail first policies are a detriment to effective community treatment. I have many research papers demonstrating that “open” formulary designs improve treatment outcomes while saving health care dollars. But, in addition to the economics, I do wonder would we as a society permit “fail first” policies to apply to other major illnesses — heart disease, or kidney disease, or cancer? Could this be a parity issue? In addition, there is research indicating possible damage to brain tissue from the acute exacerbation of psychosis. Every time I see a health plan adopt this policy, I don’t see supporting data other than short-term financial gain.”
What is the solution to “managing” the cost of medications? I’m not saying that every consumer should have carte blanche selection of medications at the expense of the payer. First of all, I think the premise of managing medication costs is the wrong question. Payers shouldn’t be focused on the silos of medication costs – but rather should be focused on the total costs of a consumer disease state. But beyond that, consumers (and clinical professionals) need better tools for medication selection – about efficacy, side effects, and specificity which medications are truly clinically interchangeable. The selection guidelines shouldn’t be based only on the health plan’s price negotiations with pharmaceutical companies. Rather consumers should have access to selection guidelines that make use of the current state of genetic testing for medication matches, of the use of “big data” (practice-based evidence) on effectiveness in practice, and comparative effectiveness research.
I realize that every stakeholder in the U.S. health care system is struggling with improving their “value equation” and delivering more for the resources expended. But, I would argue that “fail first” policies are a crude tool in an era where other more effective approaches exist. In addition to some of the tools mentioned above, the notion of integrated care management (that includes an incentive for discretion in the use of total resources) may be part of the solution to this issue. What I do know is, increasing “fail first” policies is a step backward on many fronts.