Key Takeaways
- Four-pillar therapy in reduced ejection fraction heart failure returns roughly $3,600 to $6,900 per patient per year at current Medicare Part D pricing.
- Price, not evidence, sets the sequence. Patients receive spironolactone before the better-evidenced finerenone because one costs $17 to $169 a year and the other about $9,264.
- Prior authorization on drugs with no therapeutic equivalent cannot redirect utilization. It delays or eliminates it, making patients 6.75 times slower to fill an SGLT2 inhibitor.
- The Pharmacy Quality Alliance measure now being written could make prior authorization and tier placement visible as quality inputs, if two specification choices are made correctly this year.
A patient with preserved ejection fraction heart failure will usually receive spironolactone before finerenone. Price decided that, and the evidence points the other way. Benefit design has stopped being the financing layer beneath clinical decisions and started making them.
A sequence nobody chose
Within the mineralocorticoid receptor antagonist slot in heart failure, two agents compete. Spironolactone runs $17 to $169 per year. Finerenone runs roughly $9,264. Prescribers overwhelmingly start with the steroidal agent and reserve finerenone for patients who fail it or cannot tolerate it, which is a rational response to what patients can afford.
Look at what that sequence does to the evidence. Finerenone earned its indication for ejection fraction of 40 percent or above through FINEARTS-HF, a dedicated randomized trial in 6,001 patients that met its primary composite of total worsening heart failure events and cardiovascular death, with a rate ratio of 0.84. Spironolactone in this same population rests on TOPCAT, which missed its primary endpoint overall and showed benefit only in a post hoc analysis of the Americas enrollment, against a roughly fourfold difference in event rates between regions.
So the standard of practice puts patients on the less firmly evidenced agent first and holds the better evidenced one in reserve. That is the reverse of how step therapy is supposed to work, where the inexpensive first step is the well-established one. No committee voted for this arrangement. No guideline recommends it. Price produced it, and it now functions as clinical policy across a population of millions.
Benefit design used to determine what care cost. It now determines what care happens.
The MRA slot is the cleanest illustration because the two agents sit in the same therapeutic position with a five-hundred-fold price difference between them. The pattern is not confined there. Once you look for benefit design acting as a clinical instrument, it appears at every stage of the heart failure treatment pathway, and the evidence for each instance has become quite specific.
Two heart failures, two regimens
Any argument about heart failure therapy has to start by separating the phenotypes, because most economic analysis quietly does not. Heart failure with reduced ejection fraction, roughly 40 percent or below, is the phenotype that qualifies for all four therapeutic pillars. Heart failure with preserved or mildly reduced ejection fraction, which accounts for roughly half of all cases, is treated with an SGLT2 inhibitor backbone plus an MRA. A beta blocker or an ARNI enters that regimen only when a separate condition calls for it.
Building each regimen at current Medicare Part D pricing produces very different pictures, and the patient-facing column is where the argument lives.
| Regimen component | Payer cost | Patient out of pocket |
|---|---|---|
| Reduced ejection fraction, four pillars | ||
| SGLT2 inhibitor, negotiated price | $2,142 to $2,364 | $536 to $591 |
| ARNI, generic to brand at negotiated price | $730 to $3,540 | $60 to $885 |
| Steroidal MRA | $17 to $169 | $6 to $24 |
| Beta blocker | $8 to $79 | $3 to $6 |
| Full quadruple therapy | $2,900 to $6,200 | $605 to $1,506 |
| Preserved or mildly reduced ejection fraction, two to three agents | ||
| Base case, SGLT2 inhibitor plus steroidal MRA | $2,160 to $2,530 | $542 to $615 |
| Escalation, SGLT2 inhibitor plus finerenone | $11,400 to $11,630 | $2,100, at the cap |
Two things fall out of that table immediately. In the preserved ejection fraction base case, the MRA costs a patient $6 to $24 per year and there is no beta blocker or ARNI in the regimen at all, which means the SGLT2 inhibitor accounts for nearly the entire out-of-pocket burden. Its tier placement is therefore the highest-leverage benefit design decision available for the larger of the two phenotypes, and it is a single decision rather than a program.
The escalation line carries the second point. Adding finerenone roughly doubles the cost of the most expensive four-pillar regimen and consumes a patient's full annual out-of-pocket ceiling by itself. That is the price of following the better trial.
The economics argue the opposite way
None of this suppression is even rational for the plan, which is the part most likely to change a conversation. A 2026 economic evaluation using Medicare-linked data from the Get With The Guidelines registry followed 50,598 older adults hospitalized with reduced ejection fraction heart failure. Mean one-year cost ran $41,802 per patient, with $25,172 attributable to hospitalizations. Modeled full quadruple therapy was projected to reduce hospitalization-associated expenditures by $9,780 per patient annually.
That study left its bottom line open, reporting a range from $8,556 in savings to $6,347 in net cost, because the result turns entirely on what the drugs cost. Current pricing closes that question. Negotiated Medicare prices took effect in January 2026 at $295 per month for Entresto, $197 for Jardiance and $178.50 for Farxiga, against 2023 list prices of $628, $573 and $556. Generic sacubitril valsartan reached the market in July 2025. Against $2,900 to $6,200 in annual drug cost, the net result lands at roughly $3,600 to $6,900 in savings per patient per year.
One honest boundary applies. The offset figure comes from a model rather than an observation, and it combines trial effect estimates multiplicatively to produce an 87 percent relative reduction in heart failure hospitalizations that no real-world population will reproduce. Treat $9,780 as a ceiling. The conclusion survives a generous haircut, and the study enrolled reduced ejection fraction patients only, so none of its economics transfer to the preserved ejection fraction population without separate evidence that does not yet exist.
Within that boundary the finding is unambiguous. In the phenotype where the evidence is strongest, plans are spending administrative dollars to slow down a therapy that pays them back.
What prior authorization actually does here
Prior authorization exists to steer prescribing toward a lower cost therapeutic equivalent. That is its design purpose and, applied to drug classes where equivalents exist, a defensible one. Across most of the heart failure regimen no equivalent exists.
An SGLT2 inhibitor has no generic substitute that produces the same outcome, and neither does an ARNI in reduced ejection fraction. The control cannot redirect utilization in those cases. It can only delay or eliminate it, and it moves cost from the pharmacy line to the medical line while doing so.
The measured effect is large. A 2026 analysis linked prescription orders to actual pharmacy fills across 2,183 heart failure patients newly prescribed an ARNI or an SGLT2 inhibitor, which allowed the authors to observe prescriptions that were ordered and never filled. Patients facing a prior authorization requirement took 3.03 times as long to fill an ARNI and 6.75 times as long to fill an SGLT2 inhibitor. Median time to first fill for the SGLT2 inhibitor ran 31 days with the requirement against 4.6 days without it, and roughly one in five patients facing it never filled at all. The requirement fell more often on patients identifying as Black or Hispanic and on those living in lower-income neighborhoods.
Finerenone looks at first like the legitimate exception, since a genuinely cheaper alternative does exist in the same slot. That exception is weaker than it appears. Prescribers already sequence the steroidal agent first on their own, so an authorization edit on finerenone duplicates a step that has usually already occurred. The requests it does stop belong disproportionately to patients with a documented reason to skip spironolactone, whether intolerance, gynecomastia or prior hyperkalemia. A plan defending its heart failure utilization management should be able to say what each edit actually changes, and across much of this regimen the honest answer is delay rather than substitution.
Where the patient feels it
Prior authorization governs whether therapy begins. Cost sharing governs whether it continues, and the two act at different points, so addressing one leaves a predictable share of patients failing at the other.
A cohort of 94,610 commercially and Medicare-insured adults with diabetes, heart failure or both showed twelve-month adherence to GLP-1 receptor agonists falling from 72 percent at copayments under $10 to 60 percent at copayments of $50 or more, with SGLT2 inhibitor adherence moving from 77 percent to 72 percent across the same range. Every person in that cohort was insured, which is the point of the finding. Coverage did not produce use.
Map the cost table onto those bands and the pressure point is obvious. An SGLT2 inhibitor copayment of $45 to $49 per month falls in the middle band. Brand Entresto at $74 falls in the high band. Generic sacubitril valsartan at roughly $5 falls in the lowest. Generic conversion moves one of the four pillars into the most adherent tier and leaves the SGLT2 inhibitor as the binding constraint on persistence in both phenotypes.
A common assumption about the Part D redesign needs correcting here. Four-pillar therapy generates $605 to $1,506 in annual out-of-pocket cost, which sits below the $2,100 catastrophic cap for 2026. The cap does not protect a standard reduced ejection fraction regimen. It engages only when a high-cost agent enters, and finerenone alone consumes the entire ceiling.
What these mechanics produce downstream is documented. In a 2026 cohort of 6,111 patients discharged after a heart failure hospitalization, only 54 percent of new guideline-directed prescriptions were filled within seven days, and by six months 42 percent of patients remained adherent and 51 percent persistent to their discharge regimen. Since the modeled offset assumes patients take the medication, a plan capturing roughly half of that persistence realizes closer to $2,000 or $3,500 per patient than the $3,600 to $6,900 available. For a plan carrying 10,000 post-discharge members with reduced ejection fraction, the arithmetic points to tens of millions of dollars in offset left uncollected each year.
The measure being written right now
The Pharmacy Quality Alliance convened a technical expert panel in April 2026 to develop a health plan measure titled Evidence of Guideline-Directed Medical Therapy in Persons with Heart Failure, spanning ARNI, ACE inhibitors and ARBs, beta blockers, MRAs and SGLT2 inhibitors. PQA measures are the pipeline into Part D quality programs, which makes the specification written this year the thing that determines what plans are held to later. The panel is mid-work, which is the useful moment to be specific about what the measure has to solve.
The phenotype problem is the hard part. A four-pillar standard applied to a preserved ejection fraction patient scores appropriate care as a failure, since that patient should not be on a beta blocker or an ARNI absent another indication. Avoiding that requires ejection fraction, and ejection fraction lives in the echocardiogram report rather than in a claim. Diagnosis coding is a weak substitute. ICD-10 does separate systolic, diastolic and combined heart failure, but the unspecified codes remain billable and heavily used, the systolic and diastolic descriptors are physiologic labels rather than ejection fraction thresholds, and mildly reduced ejection fraction has no distinct code at all. This is the same coding limitation I have described upstream in Heart Failure’s Stage A and B Blind Spot, arriving now at the treatment end of the disease.
A second gap sits inside the drug classes. The announced concept counts mineralocorticoid receptor antagonists at the class level, and a class-level numerator is satisfied by spironolactone at $17 per year. A plan could post strong measure performance in preserved ejection fraction heart failure while every finerenone request in its book sat behind an authorization queue, and the measure as scoped would not register it. Either the specification accepts class-level equivalence and says so plainly, or the preserved ejection fraction population needs agent-level logic.
One path around the phenotype problem already appears in the evidence. The 2026 discharge cohort measured adherence and persistence against the regimen each patient was actually discharged on rather than against a fixed standard. The prescribing decision encodes the phenotype, which keeps the measure claims-native and removes the ejection fraction requirement entirely. PQA already operates proportion-of-days-covered measures in Part D, so the machinery exists. The blind spot is undertreatment, since a reduced ejection fraction patient started on two pillars would score well. Pairing a persistence measure with a separate completeness measure that does require clinical data would cover both failure modes, and it would place the burden of clinical data sourcing only on the measure that genuinely needs it.
The reason any of this matters to a plan is what such a measure would connect. Prior authorization criteria and formulary tier placement sit today with utilization management and pharmacy benefit design, insulated from quality accountability. A dispensing-based measure makes both of them direct inputs to a reported rate. I have written separately about PQA's position between guideline authority and payer operations in PQA at the Inflection Point, and this is that leverage applied to a specific disease.
One more exposure, and a correction
The disparity finding in the prior authorization analysis deserves accurate framing rather than the framing it usually gets. It is a fair-access exposure with no Star Ratings mechanism currently behind it, because CMS declined to implement the Excellent Health Outcomes for All reward, formerly the Health Equity Index, for 2027 and returned to the historical reward factor while removing eleven administrative process measures. Anyone presenting this to a payer audience as a Stars revenue risk will be corrected in the room. The exposure is real and it surfaces through nondiscrimination review, accreditation standards and Medicaid contracting instead.
Finerenone is the access question to watch next. It sits outside negotiated pricing, carries a monthly patient cost near four times the level at which adherence measurably degrades, and now holds an approved indication in the larger of the two phenotypes. The conditions that produced the sequencing inversion described at the top of this piece are about to apply to a much bigger population.
Heart failure has an unusually complete evidence base and, in the reduced ejection fraction phenotype, an unusually favorable economic case. Four-pillar therapy at current Medicare pricing returns roughly $3,600 to $6,900 per patient per year. What stands between that evidence and the patient is not a missing trial, an absent guideline or an unaffordable drug. It is a set of administrative controls that were built to manage substitutable spending and are now applied to therapies with no substitute.
This is the opportunity in the white space. The measure now being specified at PQA is the first mechanism that would make prior authorization criteria and tier placement visible as quality inputs rather than as private financial decisions. Whether it does that depends on two specification choices being made this year, namely how the measure handles ejection fraction phenotype and whether it counts MRAs by class or by agent. Those choices are answerable now and expensive to revisit later.
A companion argument sits upstream of this one. Where this piece concerns therapy the system can see and impedes anyway, The Heart Failure We Don’t Count concerns the four fifths of the heart failure population the measurement apparatus never sees at all. Read together, they describe a measurement failure before diagnosis and a payment failure after it.
Frequently asked questions
Does guideline-directed heart failure therapy pay for itself?
In reduced ejection fraction heart failure, yes. A 2026 economic evaluation of 50,598 Medicare beneficiaries found mean one-year costs of $41,802 per patient, with $25,172 attributable to hospitalizations, and projected that full quadruple therapy would reduce hospitalization-associated expenditures by $9,780 per patient annually. At current Medicare Part D pricing the regimen costs roughly $2,900 to $6,200 per year, leaving net savings of about $3,600 to $6,900. No comparable offset figure exists for preserved ejection fraction heart failure.
Do all heart failure patients need all four drug classes?
No. Reduced ejection fraction heart failure, roughly 40 percent or below, is the phenotype treated with all four pillars. Preserved and mildly reduced ejection fraction heart failure, roughly half of all cases, is treated with an SGLT2 inhibitor backbone plus an MRA. A beta blocker or ARNI appears in that group only when a separate condition warrants it. Economic analyses built on the four-pillar regimen apply to the reduced ejection fraction population alone.
Why is prior authorization on an SGLT2 inhibitor different from ordinary utilization management?
Prior authorization is designed to steer prescribing toward a lower cost therapeutic equivalent. For SGLT2 inhibitors in heart failure no equivalent exists, so the control cannot redirect utilization and can only delay or eliminate it. A 2026 analysis found patients facing the requirement took 6.75 times as long to fill an SGLT2 inhibitor and were 2.23 times more likely never to fill it, with median time to first fill of 31 days against 4.6 days without.
How does price affect which mineralocorticoid receptor antagonist a patient receives?
Prescribers generally start with a steroidal agent such as spironolactone, costing $17 to $169 per year, and reserve finerenone, at roughly $9,264 per year, for patients who fail or cannot tolerate it. That sequence follows affordability rather than evidence. Finerenone earned its indication for ejection fraction of 40 percent or above through FINEARTS-HF, a dedicated trial in 6,001 patients that met its primary composite endpoint. Spironolactone in this population rests on TOPCAT, which missed its primary endpoint overall.
Does the Medicare Part D out-of-pocket cap protect heart failure patients?
Not for the standard regimen. Four-pillar therapy generates roughly $605 to $1,506 in annual patient out-of-pocket cost, below the $2,100 catastrophic cap for 2026. The cap engages only when a high-cost agent such as finerenone enters, and finerenone alone consumes the full cap. For most heart failure patients the binding cost is the SGLT2 inhibitor copayment rather than the annual ceiling.
What is the PQA heart failure measure, and why does its specification matter?
PQA convened a technical expert panel in April 2026 to develop a health plan measure titled Evidence of Guideline-Directed Medical Therapy in Persons with Heart Failure, spanning ARNI, ACE inhibitors and ARBs, beta blockers, MRAs and SGLT2 inhibitors. PQA measures feed Part D quality programs, so the specification written now determines what plans are held to later. Two design choices matter most, namely how the measure identifies ejection fraction phenotype and whether it counts MRAs at the class level or the agent level.
Can claims data identify heart failure phenotype?
Not reliably. Ejection fraction lives in the echocardiogram report rather than in a claim. ICD-10 separates systolic (I50.2x), diastolic (I50.3x) and combined (I50.4x) heart failure, but the unspecified codes I50.9 and I50.1 remain billable and are heavily used, the systolic and diastolic descriptors are physiologic labels rather than ejection fraction thresholds, and mildly reduced ejection fraction has no distinct code at all.
Is there a measure design that avoids the phenotype problem?
One approach appears in the 2026 discharge cohort, which measured adherence and persistence against the regimen each patient was actually discharged on rather than against a fixed four-pillar standard. The prescribing decision encodes the phenotype, which keeps the measure claims-native and removes the need for ejection fraction. Its blind spot is undertreatment, so pairing it with a separate completeness measure that does require clinical data would cover both failure modes.
References
- Keykhaei M, Rashedi S, Greene SJ, et al. Cost Offset With Quadruple Therapy for Heart Failure. JAMA Cardiology. 2026. PubMed 42268625. doi:10.1001/jamacardio.2026.1552.
- Bessette LG, Magnani JW, Brooks MM, et al. Initiation, Adherence, and Persistence to Guideline-Directed Medical Therapy After Heart Failure Hospitalization. JAMA Internal Medicine. 2026. PubMed 42475075. doi:10.1001/jamainternmed.2026.2908.
- Mukhopadhyay A, Adhikari S, Li X, et al. Prior Authorization Requirements and Prescription Fill Patterns Among Patients With Heart Failure. JACC: Advances. 2026. PubMed 41581386. doi:10.1016/j.jacadv.2025.102583.
- Essien UR, Singh B, Swabe G, et al. Association of Prescription Co-payment With Adherence to GLP-1 Receptor Agonist and SGLT2 Inhibitor Therapies in Patients With Heart Failure and Diabetes. JAMA Network Open. 2023. PubMed 37261826. doi:10.1001/jamanetworkopen.2023.16290.
- Solomon SD, McMurray JJV, Vaduganathan M, et al. Finerenone in Heart Failure with Mildly Reduced or Preserved Ejection Fraction (FINEARTS-HF). New England Journal of Medicine. 2024. PubMed 39225278. doi:10.1056/NEJMoa2407107.
- Pfeffer MA, Claggett B, Assmann SF, et al. Regional Variation in Patients and Outcomes in the TOPCAT Trial. Circulation. 2015. PubMed 25406305. doi:10.1161/CIRCULATIONAHA.114.013255.
- Centers for Medicare & Medicaid Services. Medicare Drug Price Negotiation Program, Negotiated Prices for Initial Price Applicability Year 2026. CMS fact sheet.
- Centers for Medicare & Medicaid Services. Contract Year 2027 Medicare Advantage and Part D Final Rule. 2026. CMS fact sheet.
- Pharmacy Quality Alliance. PQA Is Launching a Heart Failure Technical Expert Panel. 2026. PQA announcement.
- US Food and Drug Administration approval of finerenone for heart failure with left ventricular ejection fraction of 40 percent or greater, July 2025.
- Payer and patient cost estimates derived from Pennsylvania State Maximum Allowable Cost schedules, March 2026, and the CMS Maximum Fair Price fact sheet. OneAnother Health analysis. Rounded market snapshot, variable by plan and member benefit.
All views, analyses, and frameworks presented here reflect independent professional judgment informed by more than two decades of experience across payer strategy, clinical transformation, and health system operations. They do not represent the views or positions of any current or former employer or affiliated organization.