Setting up an outcomes guarantee for pharmaceuticals: new approach to risk sharing in primary care

Ensuring maximum benefit from expenditure on drugs is one way that health authorities could improve their use of NHS resources. Chapman and colleagues describe a project which should help achieve this

To obtain maximum health benefits from a drug it is necessary to follow the pharmaceutical company's recommendations for use. However, variations in clinical prescribing practice and patient compliance mean that drugs are not always used under optimum conditions. One way of achieving maximum benefit could be to set up an “outcomes guarantee,” in which a pharmaceutical company and prescribing stakeholders (such as health authorities and primary care or hospital trusts) agree on the outcomes that they would expect from a drug in a given indication. If the drug fails to fulfil expectations, the pharmaceutical company refunds the health service for the cost of the drug. This encourages the pharmaceutical company to promote responsible prescribing and ensures that healthcare resources are not wasted on ineffective treatments. The National Institute for Clinical Excellence's shared risk policy for interferon beta in multiple sclerosis, and the subsequent commentary, has raised awareness of the potential for such schemes.^1–5

In 1999, North Staffordshire Health Authority, Parke-Davis (now Pfizer), and Keele University entered a pilot collaboration to provide an outcomes guarantee for statins in lowering blood cholesterol concentrations. The health authority had identified cardiovascular disease as a local priority⁶ and was looking for ways to disseminate best practice without making excessive demands on itself or its general practitioners. The project enabled the health authority to work openly with a pharmaceutical company to achieve common goals and to share the risks of disseminating good clinical practice. In this article, we describe how the project, which was launched in 2000, was set up.

Summary points

• Under an outcomes guarantee, drug manufacturers agree to refund the health service if a drug fails to meet agreed performance targets when used under appropriate conditions

• Such guarantees encourage pharmaceutical companies to promote responsible prescribing for their drugs and ensure that healthcare resources are not wasted on ineffective treatments

• In this pilot project, the principles of an outcomes guarantee were readily accepted by the pharmaceutical company, health authority, and, importantly, general practitioners

• Wider use of outcomes guarantees could improve clinical behaviour to the benefit of all concerned

Identifying a suitable agent

Treatment with statins was chosen for the project because

• Cholesterol concentration was a high priority for treatment

• Prescribing of statins was increasing rapidly and was a burden on healthcare resources

• Consensus on recommended practice and clear guidelines were available^7–10

• There was an objective and measurable outcome of treatment

• An interested pharmaceutical partner was available.

Treatment of high cholesterol concentrations was particularly relevant in North Staffordshire because its death rates from coronary heart disease were significantly higher than those elsewhere in the West Midlands and across the whole of England and Wales.¹¹

Identifying key stakeholders and developing the pilot

As a first step, Keele University defined the concept of an outcomes guarantee for a pharmaceutical product and submitted a proposal for testing its practicability to the health authority. The Parke-Davis Foundation agreed to fund the project through an educational grant; Pfizer has subsequently assumed this sponsorship. To bring to light any potential conflicts of interest, all parties were asked to identify their key motives for participating in the project (table 1).

Table 1.

 Stakeholder analysis from a project steering meeting in June 1999

Objective	Drug company	Health authority	Primary care group	Practices	Keele
Develop a model for outcomes guarantee:
Competitive advantage	√√	•	•	•	•
Reduce costs of health care	•	√√	√√	√√	•
Academic kudos	•	•	•	•	√√
Obtain comparative data on outcomes for different statins	√√	√	√	√•	√
Develop ways of joint working	√√	√√	√√	√	√√
Increase health of local population	√•	√√	√√	√√	√•
Contain cost and manage market growth	•	√√	√√	√	•
Prioritise statins towards patients with highest need	•	√√	√√	√	•
Achieve consistent prescribing practice in line with HA policy	•	√√	√ (√√)	√	•
Increase research funding	•	•	•	•	√√
Become centre of excellence in outcomes guarantee	√√	•	√	•	√√
Link in outcomes guarantee work with NICE	√√	•	√•		√√
Publications	√√	√	√	√	√√

√√ Primary objective for participating, √ secondary objective (beneficial spin-off), • neutral to organisation's objectives. 

HA=health authority, NICE=National Institute for Clinical Excellence. 

We then established a core project steering group, which negotiated various project guarantees and undertakings (box B1). The steering group also agreed on inclusion criteria and an intervention programme and developed a project outline for wider review and discussion at a consultation meeting. This meeting was attended by interested colleagues from primary care groups in North Staffordshire, the health authority, the local medical audit advisory group, and the hospital trust.

Box 1.

Summary of project undertakings

Keele University

• To ensure academic rigour
• To balance the interests of Parke-Davis/Pfizer and the health authority

Health authority

• To encourage prescribing of statins in line with national guidelines
• To partner Parke-Davis/Pfizer to promote use of statins to agreed cholesterol target levels within an agreed target population

Parke-Davis/Pfizer

• To adhere to health authority guidelines
• Not to use the outcomes guarantee project as a platform to promote atorvastatin in preference to other statins
• To align with other initiatives on coronary heart disease in North Staffordshire in so far as is practical

Negotiating terms and conditions of supply

The outcomes guarantee was aimed at protecting the health service from paying for a drug if it did not work—for example, because it was inappropriately prescribed or was not as effective as claimed. We therefore needed to agree on criteria to measure the outcome. The claimed performance for atorvastatin was defined in terms of a matrix derived by Parke-Davis/Pfizer from the results of clinical trials of their drug and was scrutinised by the steering group.

The matrix indicates the percentage of participants expected to reach target low density lipoprotein concentrations at each dose, depending on their baseline concentrations (table 2). The outcomes guarantee is a measure of the claimed effectiveness of the drug against agreed performance targets. However, patients who do not comply with treatment will not reach the target. We agreed to assume a high level of compliance, and the matrix allowed for non-compliance by 20% of patients before reimbursement payments were triggered. The terms of the agreed guarantee are given in box B2.

Table 2.

 Cumulative percentages of patients expected to reach low density lipoprotein target of <3.0 mmol/l according to baseline cholesterol concentration and dose of statin. Numbers in parentheses are the cumulative percentages at which reimbursement is triggered^*

Atorvastatin (mg)	Baseline low density lipoprotein cholesterol (mmol/l)
	Mild (>3.0-4.8)	Moderate (>4.8-5.6)	Severe (>5.6-6.0)	Complex (>6.0-6.5)
10	89 (71)	36 (29)	14 (11)	8 (7)
20	98 (78)	69 (55)	37 (30)	23 (18)
40	99 (79)	81 (65)	60 (48)	45 (37)
80	100 (80)	90 (72)	73 (58)	60 (48)

^*Expected values were adjusted by 20% to reflect non-compliance. 

Box 2.

Terms of outcome guarantee

1. The outcome relates to the achievement of a low density lipoprotein cholesterol concentration target of <3 mmol/l.

2. All patients must be treated to target or maximum dose before being entered into the study population.

3. The guarantee relates to the total study population reaching the agreed target and not to individual patients.

4. Any financial rebate will be determined at the end of the study and will be based on the cost differential between different doses of atorvastatin shown in table 2. Other related costs (including clinic times or costs of alternative or adjuvant therapy) are not included.

5. The guarantee does not depend on the level of atorvastatin prescription during the study. There will be no other compensation linked to changes in prescribing habits of atorvastatin.

6. The guarantee relates only to the appropriate licensed use of atorvastatin, as stated in the summary of product characteristics.

7. The guarantee is applicable only to first line use of atorvastatin; patients whose treatment is switched to atorvastatin are not eligible unless their cholesterol concentration before treatment with the first statin has been clearly documented to be at the level required by the project steering group.

8. These terms and conditions are limited to the recruitment and fieldwork stage of study.

9. The 20% adjustment for non-compliance may be revised by the project steering group at the review stage.

Audit and intervention programme

Keele University devised an audit and intervention programme based on the joint British recommendations on prevention of coronary heart disease in clinical practice.¹⁰ This was approved by the steering group. The aim was to encourage participating practices to carry out the audit and intervention using practice staff, with the help of appropriate support material. Any overtime costs incurred by staff at participating practices were reimbursed through Innovex (funded by Parke-Davis/Pfizer).

Scientific merit and ethics committee approval

The project was designed in three phases, which were approved by local scientific merit and ethics committees (ethical approval was sought for phases 2 and 3 only). The key issue was to ensure that the project did not direct or provide an incentive to the general practitioner to prescribe a particular statin. Phase 1 was an audit of practice registers to identify patients meeting national guidelines for treatment with a statin. Phase 2 involved treatment with a statin and subsequent analysis of the drugs that patients had been prescribed. Phase 3 involved calculation of the effectiveness of one drug, atorvastatin, against its claimed benefit, with a refund to the health authority of any wasted resources under the terms of the outcomes guarantee.

Recruitment of practices

We identified primary care practices through informal discussion with all primary care groups in North Staffordshire Health Authority. We formally invited each practice within interested primary care groups (on behalf of the primary care groups, the health authority, and Keele University) and held introductory meetings with each practice that expressed an interest.

Next steps

The audit and intervention stages of the project have been completed, and an independent qualitative evaluation of the acceptability of the model to the key stakeholders has been carried out. Keele University managed the data collection and has conducted quantitative and health economic analyses. Results of these evaluations and analyses will be published later.

Discussion

The principles of establishing an outcomes guarantee were readily accepted by the stakeholders, and the approach was approved by the local scientific merit and ethics committees. Importantly, general practitioners quickly perceived the benefit, for themselves and the local population, of promoting appropriate prescribing of drugs to patients who need them most. Recruitment was therefore swift and comprehensive (for example, over 90% within three months in central Stoke). In the wider context, all stakeholders share the aim of improving the management of coronary heart disease within the framework of national guidelines. We believe that the ultimate goal—refinement of clinical behaviour—is achievable through the outcomes guarantee project.

An outcomes guarantee has the potential to ensure predictable health gains for a given drug expenditure. This may be particularly relevant in the future as the high costs of new medicines and concern about their inappropriate use could result in resistance to uptake. Disseminating new drugs under the terms of such a guarantee provides reassurance to both parties; the company is more likely to get its drug to those who need it most, and the NHS has a reassurance of return on investment. The evaluation of this pilot project will determine the overall acceptability of the model to stakeholders and whether it can be reproduced or transferred within the United Kingdom or internationally.