There are many within the industry who believe that conducting an economic evaluation (cost-effectiveness or cost-utility analyses) of a new therapy alongside a pivotal registration trial is the way to go. With the increasing importance of health economics data needed to expedite reimbursement and gain broad access to patients, many brands can’t afford to be slapped with a ‘limited use’ label or wait years to generate data.
So it is imperative that organisations start to think about the need for this data well in advance of the launch of a new product. And I mean well in advance. Some organisations have already recognised the need for a change in their way of thinking.
The advantages are clear. We have the pragmatic and financial ease of collecting economic data without having to fund an economic evaluation as a stand-alone trial and the personnel and processes for data collection are already in place. The nature of the trial design can also be a benefit (not always) for an economic evaluation in that the concepts of ‘blinding’ and ‘randomisation’ are highly sought after and desirable for an economic evaluation.
However, there are pitfalls to watch out for in this approach that bear mentioning. Firstly, can we really make an economic decision based on the patient population in one clinical trial? This is, perhaps, the single greatest argument against using trial data for an economic evaluation. The patients in these trials are ‘pristine’. The rigorous inclusion and exclusion criteria do not mimic the types of patients that we see in waiting rooms. Is it, then, fair to make funding and reimbursement decisions on this segment of patients? Another pitfall is the fact that both clinicians and patients in clinical trials are highly motivated and committed. Clinicians receive the latest drugs and devices with which to monitor, diagnose and treat trial patients and the patients themselves are, on average, more committed to following protocols than the average patient.
Additionally, we must be careful about protocol-driven costs that would not normally occur in the everyday treatment of a disease. For example, a trial protocol stipulates that patients on an investigational therapy require monitoring on day one, day 14 and day 28. These ‘extra’ monitoring and testing costs may not be seen in the real world. Highly committed patients and investigators may stick to the protocol but outside a trial there is great variability in monitoring and testing. What about adherence and compliance to therapy? In a trial setting, strict compliance and adherence to therapy is enforced and routinely is higher than what one would see outside a trial setting. This poses two problems. Because compliance is higher in a trial setting, are we including drug costs that are higher than expected or observed in the real world? Secondly, because compliance is lower in the real world (as compared to a trial population) and lack of compliance leads to complications in disease management, are we missing the cost of managing complications due to poor compliance by using clinical trial outcomes upon which to build our economic evaluation?
Cost and resource use is another element of the trial setting that needs to be examined. Typically, a teaching hospital will have different costs and practice patterns associated with a particular disease management approach than you might find in a community setting.
Inadequate length of follow-up and inappropriate clinical alternatives for comparison are two additional elements that can render economic data alongside a clinical trial challenging to accept. Ideally, economic evaluations require as long a follow-up period and look at end-points like ‘survival’. Clinical trials, of course, cannot run in perpetuity and use surrogate markers as end-points. For example, ‘cholesterol’ might be a surrogate marker for ‘death from heart disease’. As a result, clinical trials only run long enough to show a meaningful difference between two (or more) therapies. However, disease progresses but data collection has stopped. And sometimes we compare a new therapy to placebo or other ‘standard of care’ which itself has not been economically evaluated.
Can we really make an economic decision based on the patient population in just one clinical trial?
What about switching therapy and the actual out-of-pocket cost of drug therapy? To truly mimic what happens in the real world, it would be ideal to have patients pay for their therapy and give them the freedom to switch therapy due to adverse events or lack of efficacy. In a clinical trial, the switching of therapy is not allowed and trial drug is provided free. What impact do these elements play in the overall economics?
There is no perfect system – that much is clear. And each jurisdiction evaluates economic data differently, so the challenges of economic evaluations alongside clinical trials are not going away soon. Working with payers and regulators upfront, however, to understand what success ‘looks like’ for an economic evaluation is a critical element in this puzzle. Otherwise, we are simply rolling the dice and hoping for the best.
