If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address correspondence to: Joshua D. Brown, Department of Pharmaceutical Outcomes and Policy, University of Florida College of Pharmacy, 1225 Center Dr. HPNP #3320, Gainesville, FL 32606.
Institute for Pharmaceutical Outcomes and Policy, University of Kentucky College of Pharmacy, Lexington, KY, USADepartment of Pharmaceutical Outcomes and Policy, University of Florida College of Pharmacy, Gainesville, FL, USA
Considerable interest exists among health care payers and pharmaceutical manufacturers in designing outcomes-based agreements (OBAs) for medications for which evidence on real-world effectiveness is limited at product launch.
Objectives
To build hypothetical OBA models in which both payer and manufacturer can benefit.
Methods
Models were developed for a hypothetical hypercholesterolemia OBA, in which the OBA was assumed to increase market access for a newly marketed medication. Fixed inputs were drug and outcome event costs from the literature over a 1-year OBA period. Model estimates were developed using a range of inputs for medication effectiveness, medical cost offsets, and the treated population size. Positive or negative feedback to the manufacturer was incorporated on the basis of expectations of drug performance through changes in the reimbursement level. Model simulations demonstrated that parameters had the greatest impact on payer cost and manufacturer reimbursement.
Results
Models suggested that changes in the size of the population treated and drug effectiveness had the largest influence on reimbursement and costs. Despite sharing risk for potential product underperformance, manufacturer reimbursement increased relative to having no OBA, if the OBA improved market access for the new product. Although reduction in medical costs did not fully offset the cost of the medication, the payer could still save on net costs per patient relative to having no OBA by tying reimbursement to drug effectiveness.
Conclusions
Pharmaceutical manufacturers and health care payers have demonstrated interest in OBAs, and under a certain set of assumptions both may benefit.
Growing focus on value in health care has led to a number of initiatives that are designed to shift the reimbursement system to better align costs of services to value in real-world settings [
]. In recent years, there has been increased interest in manufacturer and payer agreements that tie reimbursement to product performance. These are generally referred to as performance-based risk-sharing arrangements or outcomes-based agreements (OBAs) [
]. In an OBA, performance in a defined patient population is tracked over a specified period of time in a defined population or at the individual patient level, and the amount or level of reimbursement is determined on the basis of the outcomes achieved [
Performance-based risk-sharing arrangements—good practices for design, implementation, and evaluation: report of the ISPOR Good Practices for Performance-Based Risk-Sharing Arrangements Task Force.
]. Both pharmaceutical manufacturers and payers have been motivated to develop such agreements to more closely align price and value. Recent headlines regarding higher cost new medicines have also likely been a catalyst for interest in OBAs [
]. In the United States, attempts to structure OBAs have been few and far between, because the contracting parties often struggle to align on and define the core metrics used to assess health outcomes under the contract [
]. In addition, specific details on existing deals are limited because of the proprietary nature of these agreements between individual payers and manufacturers [
Given the difficulty observed in developing successful OBAs, the primary objectives of this study were to build models to better understand key variables having the greatest impact on outcomes, costs, and cost sharing, and to propose design elements for the development of OBAs that may improve each party’s willingness to negotiate such agreements in the future.
Methods
Hypercholesterolemia was chosen as a case example for the model given recent innovations in treatment paradigms that are anticipated to have a large budget impact on payers [
]. Hypercholesterolemia is associated with multiple clinical outcomes that can be measured including surrogate end points (low-density lipoprotein cholesterol [LDL-C] measurements and goals) as well as “hard” clinical outcomes (i.e., acute myocardial infarction [MI] and stroke). Appropriate outcomes and target populations for implementing a hypercholesterolemia OBA were determined on the basis of review of end points from published literature and prescribing information for available products [
]. Excel-based models were developed to demonstrate the impact that various parameters could have on OBAs. An internal project advisory board provided input via firsthand accounts of successes and limitations of previous OBAs from both manufacturer and payer perspectives. This input was incorporated into the model development to address areas in which OBAs may be improved.
Data on hypercholesterolemia incidence were extracted from the Humana Research Database (Louisville, KY). The patient population included patients enrolled in Medicare or commercial plans at Humana Inc. with an index diagnosis of familial hypercholesterolemia or a history of atherosclerotic cardiovascular disease (ASCVD), defined as MI, stroke, angina, peripheral arterial disease, or revascularization procedures [
2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines.
Two model structures were developed. The first modeled patients reaching a goal LDL-C reduction on the basis of observed LDL-C reductions in pivotal clinical trials of proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitors, a novel class of medications indicated for hypercholesterolemia (model 1) [
]. As part of the hypothetical OBA structure, patients with LDL-C values at baseline would be followed up at 12 weeks after treatment initiation. Given that PCSK9 inhibitors have been shown to be highly effective at lowering the LDL-C, the percentage of patients meeting LDL-C goal reductions at 12 weeks (“first-line successes”) was set as 80% (ranging from 70% to 90%) [
]. The patients who did not meet LDL-C goal reductions were classified into one of three groups: 1) failed goal and terminated therapy (40% of first-line failures, i.e., those 20% not meeting initial LDL-C goals), 2) met goal after a dose change and an additional 12 weeks of therapy (40% of first-line failures), and 3) failed goal after an additional 12 weeks of therapy (20% of first-line failures). For this OBA structure, the manufacturer was to absorb financial responsibility for all PCSK9 inhibitor medication costs until the LDL-C goal was achieved, whereas the payer was responsible for medication costs afterward. Thus, the manufacturer was responsible for 12 weeks of therapy costs for 80% of the population and 24 weeks for the remaining 20% of the population (those requiring additional therapy before reaching goal or terminating therapy). Thereafter, the payer would reimburse the manufacturer at the negotiated price. Total cost of participation (i.e., responsibility for medication costs) was calculated for both manufacturers and payers at a 1-year time period.
The second model structure (model 2) included ASCVD-related outcomes occurring after treatment initiation and up to 1 year of follow-up. An expected rate reduction of these events was modeled for three different scenarios: 1) when the use of the medication was associated with a predefined outcome event rate goal, 2) when the outcome event rate reduction exceeded the predefined expectations (“overperformance”), and 3) when the outcome event rate failed to meet expectations (“underperformance”). The baseline event rate was determined on the basis of the manage care organization’s population. Expected reductions of 15% (range 10%–20%) in outcome events were estimated on the basis of recently published data for the Further Cardiovascular Outcomes Research with PCSK9 Inhibition in Subjects with Elevated Risk (FOURIER) trial [
]. For the payment structure in model 2, a tiered system was developed that linked reimbursement to the three outcome event rate scenarios described previously. Reimbursement for the product was reduced by 30% to 70% if the medication underperformed, and allowed for increased reimbursement of 105% to 125% if the medication overperformed. Reimbursement rates remained flat if the reduction in event rates met expectations.
Medication costs were assumed to be $1,167 per month, or $14,000 annually, on the basis of estimated prerebate costs for PCSK9 inhibitors currently on the market [
]. Costs were calculated with an assumed adherence of 100% to therapy. Although this is unlikely in real-world situations, we assumed the simplest case considering that no real-world adherence data are available for our case example of PCSK9 inhibitors. Moreover, lower adherence rates would result in proportionally lower cost estimates (e.g., 80% adherence would result in 80% of medication costs). Because of the lack of real-world evidence on the exact relationship between adherence levels and effectiveness, the effect of lower adherence on effectiveness could be assumed to be proportional as well. Many existing OBAs have reportedly used predefined cohort inclusion criteria specifying minimum adherence to therapy to assess the agreed upon outcomes of interest. Estimated medical offsets due to a reduced rate of ASCVD events were also included in the payer value calculations in model 2. These offsets were calculated on the basis of annual hospitalization rates for MI, stroke, and unstable angina, as well as coronary artery bypass graft and percutaneous coronary intervention procedures, taken from the Humana Research Database. The derived rate estimates were then multiplied by the incremental 1-year cost of care for each type of event, including follow-up outpatient care, on the basis of available literature (see Appendix Table 3 in Supplemental Materials found at doi:10.1016/j.jval.2017.07.009) [
]. For each model, payer costs were calculated as total costs per treated patient, and payments to manufacturers were defined as the total reimbursement paid to the manufacturer.
In each model, two simulations were included: “without OBA” and “with OBA.” We assumed that payers who implemented an OBA with a pharmaceutical manufacturer would deploy fewer barriers to access the medication because the uncertainty regarding the medication’s value was shared between both parties. Thus, a scenario in which no OBA was present, formulary restrictions, previous authorizations, higher patient co-payments, or other barriers would be in place, which meant that fewer patients would have access to the medication. These considerations incorporate either of the following two general scenarios: 1) a medication is first to market or is unique in its class or 2) a medication has one or more comparators, with the manufacturer competing for market share. In scenario 1, restrictions to market access for a given patient population may be due to the nonformulary status of the drug. For scenario 2, tiered formularies may influence market share, and a relative position versus competitors is also important to consider. These models do not incorporate scenarios in which market access restrictions limit certain patient populations from accessing the medication; rather, these models focus on the market influences for a general population for whom the medication is indicated.
In the “with OBA” scenario, a tiered reimbursement structure was deployed. Manufacturers would receive less reimbursement when patients failed to meet the OBA goal, contractual reimbursement when an outcome was met, and higher reimbursement via a bonus system when the medication performance exceeded expectations. Sensitivity analyses were conducted for key parameters including the size of the treated population through improved market access, the effectiveness of the medication in achieving the desired outcome, and the amount of the tiered negative or positive reimbursement payments using a range of values for each parameter discussed earlier. OBA contracts were developed under a 1-year time horizon, and implications beyond the 1-year time horizon are discussed. Model inputs, sensitivity ranges, and sources are presented in Table 1.
Table 1OBA model inputs, sensitivity ranges, and sources
Among the 20% who did not meet the LDL-C goal after 12 wk, it was assumed that some discontinued treatment (40%), some continued at a higher dose and met goal (40%), and some continued at a higher dose and did not meet goal (20%).
Negotiated medication cost between payer and manufacturer. Actual payment is determined by the proportion of patients meeting or failing goal per LDL-C outcome definition.
The total cost of care, taken by multiplying incidence by the annual cost of care and summing for all CV events, was $9,220,234. Costs of care were incremental from the Bonafede study and represented quarterly cost estimates for the first quarter after an event plus three additional quarters of costs to equal 12 mo of follow-up costs. Confidence intervals for a sensitivity range were not used because of the narrow range and negligible effect on OBA results (~$50 per-patient costs).
Assumed that the population size would be increased with an OBA in place because of improved market access.
† Among the 20% who did not meet the LDL-C goal after 12 wk, it was assumed that some discontinued treatment (40%), some continued at a higher dose and met goal (40%), and some continued at a higher dose and did not meet goal (20%).
‡ The total cost of care, taken by multiplying incidence by the annual cost of care and summing for all CV events, was $9,220,234. Costs of care were incremental from the Bonafede study and represented quarterly cost estimates for the first quarter after an event plus three additional quarters of costs to equal 12 mo of follow-up costs. Confidence intervals for a sensitivity range were not used because of the narrow range and negligible effect on OBA results (~$50 per-patient costs).
In the LDL-C goal model (model 1) without an OBA in place, the payer’s cost per treated patient was static and the amount paid to the manufacturer was the simple product of the drug cost and the number of patients treated (Table 2). For a treated population of 500 without an OBA, this translated to a total population cost of $5,040,000 (range $4,935,000–$5,145,000), or $10,080 per treated patient (range $9870–$10,290), being paid to the manufacturer (see Table 2; range values reflect different assumptions regarding population size and efficacy of treatment). With an OBA in place, some risk was taken on by the manufacturer, because the treatment cost was paid by the manufacturer until patients reached the goal. This reduced the payer’s costs to $7840 per treated patient (range $6510–$9170). After adjusting for those who failed treatment, the manufacturer’s total reimbursement was $7,840,000 (range $6,510,000–$9,170,000). The model indicated that this increase in overall manufacturer reimbursement persisted as long as the size of the treated patient population, that is, the increase in market access, was at least 28% (52%–11%), for example, from 780 (range 660–900) patients to 1000 patients in the base model, larger with the OBA versus without the OBA (Fig. 1).
Table 2Base-case and sensitivity results for LDL-C goal attainment outcome (model 1)
Fig. 1Change in patient population needed for manufacturer reimbursement breakeven point with an OBA versus without an OBA in model 1. Note. Dashed horizontal lines represent the manufacturer reimbursement with an OBA, including 1000 treated patients, calculated with base-case and sensitivity thresholds for medication effectiveness from model 1 (Table 2). Worst case and best case represent scenarios in which the manufacturer is responsible for treatment failures. Solid diagonal lines represent reimbursement by varying the number of patients treated without an OBA. The intersections of lines of the same color represent the breakeven points when an increase in the number of patients through an OBA results in the same reimbursement despite some of the medication costs paid by the manufacturer. For example, with the base case, the OBA would be beneficial to the manufacturer so long as the treated population increased from 780 to 1000 patients with an OBA, an increase of 28% (intersection of green lines). OBA, outcomes-based agreement.
The ASCVD outcome model (model 2) included a tiered reimbursement strategy for a medication underperforming or overperforming in the treated population for ASCVD-related clinical outcome event rates (Table 3). For the base case, in which 1000 patients were treated with a 15% reduction in ASCVD events (worst case <10%, best case >20%), the payer cost per patient was $14,000 ($7,000,000 for the total population). Annual health care costs for ASCVD events were estimated to be $9,220,234, with net costs after medical offsets of $7,837,199 if a 15% reduction in outcome event rates was achieved (Table 3). Taking into account the lower health care costs associated with fewer incurred ASCVD events because of treatment, this translated to a net cost per patient of $12,617 (range $13,170–$12,064).
Table 3Base-case and sensitivity results for model 2 with outcome event reduction tied to reimbursement
Simulation
Treated population
Rate reduction
Expected rate reduction of 15% for cardiovascular events including myocardial infarction, stroke, unstable angina, and percutaneous coronary intervention/coronary artery bypass graft procedures with a range of 10%–20%.
Based on the total (acute and follow-up) cost of care for prevented cardiovascular disease events and procedures [8] (see Appendix Table 3 in Supplemental Materials).
($)
Payer net per-patient cost ($)
Without OBA
500
15
100
7,000,000
14,000
691,518
12,617
1000
15
100
14,000,000
14,000
1,383,035
12,617
500
<10
100
7,000,000
14,000
414,911
13,170
500
>20
100
7,000,000
14,000
968,125
12,064
With OBA
1000
15
100
14,000,000
14,000
1,383,035
12,617
1000
<10
50
7,000,000
7000
829,821
6170
1000
>20
115
16,100,000
16,100
1,936,249
14,164
1000
<10
30
4,200,000
4200
829,821
3370
1000
<10
70
9,800,000
9800
829,821
8970
1000
>20
105
14,700,000
14,700
1,936,249
12,764
1000
>20
125
17,500,000
17,500
1,936,249
15,564
Note. Boldfaced figures represent the base case without OBA and a reduced market access population size.
OBA, outcomes-based agreement.
Expected rate reduction of 15% for cardiovascular events including myocardial infarction, stroke, unstable angina, and percutaneous coronary intervention/coronary artery bypass graft procedures with a range of 10%–20%.
† Percent of total drug payment by the payer to the manufacturer.
‡ Based on the total (acute and follow-up) cost of care for prevented cardiovascular disease events and procedures
With an OBA in place, the treated population was increased to 1000 patients through assumptions regarding improved market access as a driving motivation for a manufacturer to enter an OBA. If the medication failed to meet expectations, the manufacturer received 50% (worst case 30%, best case 70%) lower reimbursement for drug costs, resulting in the total manufacturer reimbursement of $7,000,000 (range $4,200,000–$9,800,000). Net cost per patient for the payer, taking into account the reduced drug cost and the reduced cost associated with fewer ASCVD events, was $6170 (range $3370–$8970). When the medication exceeded expectations, the manufacturer received increased reimbursement of 115% (105%–125%). This payment strategy raised the total manufacturer reimbursement to $16,100,000 (range $14,700,000–$17,500,000) with per-patient payer costs of $14,164 (range $12,764–$15,564).
In this example, under an agreement of 125% reimbursement paid to the manufacturer when the medication exceeds outcome expectations, the payer incurred higher net cost per patient compared with when an OBA was not in place even when considering medication offsets (Fig. 2). The manufacturer reimbursement remained higher than the “without OBA” reimbursement threshold of $7,000,000 so long as the maximum decrease in reimbursement remained higher than 50% in the case of a doubling of the treated population (Fig. 2). If there was a greater than 50% reduction in reimbursement, manufacturer reimbursement would be lower than under the scenario without an OBA in place with a smaller treated population but no decrement in reimbursement because of underperformance. Ultimately, the factors having the greatest impact on reimbursement and costs were the number of patients treated and drug effectiveness (Fig. 3).
Fig. 2Breakeven thresholds of positive and negative reimbursements for payers and manufacturers in OBAs. Note. Solid horizontal lines represent the manufacturer reimbursement (red) and payer cost per patient (blue) without an OBA, including 500 treated patients, calculated with base case (50% rate reduction) medication effectiveness in model 2. Dashed lines are the total reimbursement and costs at varying percentages of reimbursement on the basis of the medication performance in the treated population. The intersection of same colored lines represents the breakeven points (denoted by the vertical black lines) for an overperforming product (payer breakeven 104% bonus payment to manufacturer) and an underperforming product (manufacturer breakeven 50% reduced reimbursement penalty). OBA, outcomes-based agreement.
Fig. 3Sensitivity analysis of key parameters in OBAs and impact on manufacturer reimbursement and payer cost compared with base case. Note: Diagram shows change in reimbursement or costs from the base case for the best-case scenario (red) and the worst-case scenario (blue) for each of the parameters. OBA, outcomes-based agreement.
Recent interest in OBAs has emerged largely from payers’ desire to reduce the risks associated with high-cost pharmaceuticals with uncertain real-world outcomes and effectiveness [
]. By using OBAs, payers can shift some risk to manufacturers with the potential incentive of improving market access for the manufacturer’s product, thus improving patient access to novel therapies and potentially generating important real-world evidence. Nevertheless, the use of OBAs in the United States remains nascent in practice and there remains considerable interest among payers and manufacturers in understanding how to structure successful OBAs in the United States [
]. It will be of considerable interest to observe the actual financial impact of these OBAs and to track over time the value gained by both parties. Incorporating the use of clinical outcomes (in this case, ASCVD events) versus surrogate outcomes (LDL-C) as part of the OBA structure should also be evaluated for future innovation and overall performance for both parties. As manufacturers continue to publish data demonstrating the impact of PCSK9 inhibitors on actual ASCVD events, the effectiveness measures can continue to be updated, relying on ASCVD event reduction rather than on surrogate outcomes. How these new findings impact future OBA structure will be of continued interest.
In this study, models were developed to examine important considerations in structuring an OBA. Specifically, we sought to understand which variables had the largest impact on the dollar flow between the payer and the manufacturer on the basis of learnings from the project team and the advisory board. Two common themes emerged for OBAs including the motivation to share risk across both parties as well as inclusion of positive feedback to the manufacturer. This bidirectional reimbursement concept was subsequently modeled as a percent change in the payment for either underperforming or overperforming on the basis of the specified outcome of the OBA. This concept stemmed from discussions with the project’s advisory board in reference to experiences with OBAs that focused only on off-loading risk to manufacturers with no positive reinforcement for product successes. Such past practice may limit pharmaceutical manufacturers’ desire to explore OBAs because of the increased risk. Similarly, managed care organizations are unlikely to offer reimbursement bonuses because of prescription drug costs in the US health care system. From our models, we were able to observe breakeven thresholds for both parties when an OBA would not have detrimental budgetary impact compared with the base case in which an OBA does not exist. This is important to establish in negotiations of an OBA so that some perception of value is present to bring both parties to the table. In this model, the threshold for the payer allows a 4% higher reimbursement paid to the manufacturer for a high-performing product. This threshold could be higher, for more effective medications, on the basis of higher rate reductions in outcome events, which would generate more medical cost offsets and allow for a larger positive reimbursement policy to the manufacturer.
We further observed that changing assumptions around the number of treated patients had a significant impact on the economic impact of the OBA to the manufacturer. If the size of the treated population does not increase with the OBA, there is no direct motivation for a manufacturer to enter such an agreement in which the reimbursement would be subject to new risk, but access to the medication is not improved. Thus, OBAs make inherently more sense for medications that are subject to formulary management through adverse tier placement, previous authorizations, or step edits whether in absolute changes for the medication or in relative standing compared with competitor medications. From a payer perspective, focusing on medications with high drug costs and/or high prevalence will be of the most interest to reduce the overall budget impact [
]. Furthermore, an increase in the size of a treated population would also prevent more outcome events, thereby benefiting both the payer and the patient. Our models assumed a homogeneous patient population with or without the OBA, that is, similar risk of outcome events. Conceivably, access restrictions may limit medication use to those patient groups with the highest risk, and each additional increase in access may introduce patients with incrementally lower risk. Further consideration in OBA design is needed to incorporate how formulary management directs medication use to certain risk groups and whether those additional patients who are granted access through an OBA are representative or would have lower medical offsets.
Because our models encompassed only a 1-year time horizon, they did not account for the potential longer term ramifications of these agreements and the real-world evidence they could generate. If a product performed well in the population, it would likely have positive, long-term market access implications, potentially further increasing market share and revenues for the life cycle of the medication. Conversely, underperformance of a product in an OBA may jeopardize future market access through increased formulary management by the payer, maybe even beyond the previous level the OBA was intended to overcome. From a payer’s perspective, evidence generated through an OBA could drive further competition within the therapeutic class. For other high-cost medication classes, for example, specialty oncology products, this evidence can help inform formulary decision making, decreasing costs by allowing more effective medications to be placed in preferred tiers. These considerations have been modeled in depth in a previous study that considered formulary decision making by the payer in terms of net monetary benefit of a drug rebate as well as pricing motivation for the manufacturer [
]. Our study added consideration of medical offsets for the payer as a result of the medication’s effectiveness, and factors these into the value calculation of an OBA. From this analysis, payers are likely to always benefit from OBAs via reduced drug costs, especially in therapeutic classes with marginal differences in effectiveness between competitors. Nevertheless, incorporating a positive incentive to the manufacturer could increase payer costs dependent on the overall effectiveness and total medical cost offsets. In the example of PCSK9 inhibitors, this increase in cost to the payer rises to a threshold of 5% when the highest effectiveness levels are achieved. Manufacturers can also benefit and receive positive feedback through increased reimbursement for sharing the risk with the payer when the medication is ultimately demonstrated to be more effective than expected. More flexibility in this scheme is possible for medications with a combination of higher medical cost offsets and larger margins of effectiveness.
Another study surveyed key opinion leaders and acknowledged that OBAs are underutilized with limitations related to the costs and additional effort to implement the OBA [
]. Participants in that study noted other relevant considerations such as data quality and availability, measurement of meaningful and relevant real-world outcomes, and potential regulatory or legal implications. In general, a formal collaboration agreement between manufacturers and payers could help to reduce uncertainty in establishing OBAs. The cost to evaluate an OBA was considered to be equivalent to conducting an observational study including research personnel, data, and so forth, for outcomes that can be measured using claims data. Costs associated with patient- or clinician-reported outcomes would inherently be higher, but there were no real-world examples of OBAs that have used such an agreement. The cost to administer the agreement at the plan level is established given that the payer would be managing the therapeutic class and need only to make adjustments to the formulary.
This work is limited in considering only one type of innovative contracting agreement in OBAs. The work by Walker et al. [
Performance-based risk-sharing arrangements—good practices for design, implementation, and evaluation: report of the ISPOR Good Practices for Performance-Based Risk-Sharing Arrangements Task Force.
] established a taxonomy of coverage options including outcomes-based, nonoutcomes-based, and evidence-generating agreements. Our study considered only outcomes-based contracts given that, by name, these are the only forms of agreement that measure outcomes and tie those to reimbursement. Other forms of coverage options are negotiations around product price and not necessarily measurable outcomes, more typical of those used in settings outside the United States. Similarly, coverage with evidence development generally involves payers allowing market access at a reduced medication price so that evidence can be collected on a new medication for future decision making, but without payment being tied to the attainment of an outcome [
]. Although our models are specified for a single therapeutic area, the concepts of identifying the potential change in market access with an OBA, breakeven thresholds for both parties, and overall general structure can be extrapolated and applied to multiple therapeutic areas by altering the assumptions around incidence rates, treatment effectiveness (i.e., event rate reduction), and the associated pharmaceutical and medical costs of treatment.
Conclusions
Although a few OBAs have been established in the United States to date, more are likely to be initiated because of the desire to better align costs and performance associated with a medicine. By structuring agreements with shared risk between the payer and the pharmaceutical manufacturer, positive and negative incentives via reimbursement levels, and careful consideration of potential breakeven points, payers may incentivize manufacturers to enter OBAs, decrease risk in the short-term, and provide new evidence to inform long-term formulary management. Manufacturers entering OBAs can benefit from improved patient access while supporting evidence generation for the real-world effectiveness and safety of their product. By incorporating these concepts into an OBA, along with a formal collaboration agreement between the payer and the pharmaceutical manufacturer, more OBAs and future innovation may be possible. Further evaluations of obstacles and benefits of real-world OBAs are needed to establish the true short-term and long-term implications for both parties.
Source of financial support: This study was funded by Pfizer Inc. and conducted in collaboration with Comprehensive Health Insights, a wholly owned subsidiary of Humana Inc. J. D. Brown receives fellowship funding from both Humana Inc. and Pfizer Inc. There are no other financial disclosures.
Performance-based risk-sharing arrangements—good practices for design, implementation, and evaluation: report of the ISPOR Good Practices for Performance-Based Risk-Sharing Arrangements Task Force.
2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines.
☆Conflicts of interest: R. Sheer, L. Sudharshan, and M. Pasquale are all employees of Comprehensive Health Insights Inc. F. Brownfield is an employee of Humana Inc. K. Axelsen, P. Subedi, D. Wiederkehr, and S. Kamal-Bahl are employees and shareholders of Pfizer Inc. There are no other perceived or actual conflicts of interest.