Sock it Toumi

Since taking office in January, US President Barack Obama has taken a number of major decisions to control healthcare expenditure. Two of these should provide potentially significant savings with little investment:

The opening of the US market to follow-on biologics, or biosimilars, with estimated savings ranging from $12 billion over 10 years to $378 billion over the next two decades;^1-2

The lifting of the ban in the 2003 Medicare prescription drug act that prevents negotiation of the prices of medicines purchased for Medicare, and which incurs estimated extra costs as high as $30 billion annually.³

The efficiency of Mr Obama's third decision, to establish the Federal Co-ordinating Council for Comparative Effectiveness Research (FCCCER), is more doubtful. Mr Obama has dedicated $1.1 billion in the economic stimulus package for federal agencies to develop information on the comparative effectiveness of medicines. This decision aims to provide evidence to support value-driven decision-making. Such research is thought to be a potential source of cost-savings, permitting better care decisions to be made and an improved allocation of resources by Medicare, health insurers and even private individuals. The Congressional Budget Office has estimated an impact on healthcare spending of up to five times the total investment, not all of which would be visible as cost savings.

Two main arguments support Mr Obama's decision to create the FCCCER:

It is difficult to establish comparisons between products that have not been compared in head-to-head clinical trials. Such studies are rare because they are not required by the US FDA for products to be granted marketing authorisations. Furthermore, they present substantial risks to the drug manufacturers who have to pay for them. Bristol-Myers Squibb's PROVE-IT trial comes to mind.

Clinical trial data provide evidence in selected populations under experimental conditions ("internal validity") that might not reflect the true benefit of products under usual conditions of clinical practice ("external validity"). Pragmatic studies are expected to include a larger patient population to allow a combination of interventions and to reflect usual practice. They are expected to help patients and prescribers to select the most appropriate drug for individual cases.

Four research areas may generate data to allow comparative effectiveness studies: large-scale pragmatic trials, the use of databases, new quantitative methodologies and the design of complex quantitative models. No single area is favoured, although it is likely that large-scale pragmatic trials will capture most of the budget.

Yet many large clinical trials have failed to significantly affect medical practice. We can cite the following government-sponsored trials as examples:

CATIE, a National Institute of Mental Health (NIMH) trial in more than 1,400 patients comparing newer atypical antipsychotics with each other and with a conventional antipsychotic for the treatment of schizophrenia;

ALLHAT, the largest-ever antihypertensive trial and the second-largest lipid-lowering trial, with more than 42,000 patients enrolled; and

STAR-D, a seven-year NIMH trial aiming to define strategies to address treatment-resistant depression after failure to respond to a standard trial of treatment with an antidepressant medication.

The impact of these studies on actual clinical practice is debatable. One year after the CATIE results were published, 82% of clinicians treating schizophrenia indicated that they were unfamiliar with the results or that the results had not led them to change their clinical practice.⁴

In the two years following the publication of ALLHAT, the use of thiazide-type diuretics rose from 30.6% of the eligible population (uncomplicated hypertensive patients) to 39.4%, before falling back to 36.5% of patients.⁵ While the improvement was statistically significant, it remains far below the estimated total population that could benefit from such treatment.

STAR-D, meanwhile, did not provide conclusive results to alter medical practice.⁶ Thus large-scale trials are controversial, with debatable results and sometimes marginal impacts.

Following lobbying from the drug and medical devices industries, the Obama administration has clarified the potential use of comparative effectiveness research on reimbursement. The official conference report related to the bill specifies that lawmakers do not intend research to be used to mandate coverage and reimbursement policies for private or public payers.⁷ This makes real impact of comparative effectiveness research on pricing and reimbursement even less likely.
It would have been more effective to use taxes to provide an incentive for the development of Bayesian mixed-treatment comparisons and "complex systems" theory, applied to the field of public health sciences.

Bayesian comparisons can effectively address the lack of head-to-head comparisons between drugs. This well validated and recognised methodology remains underexploited in healthcare decision-making. Bayesian analyses can be used to make rapid and inexpensive comparisons that rank the effectiveness of alternative treatment options for a given indication. Furthermore, use of the technique will encourage the pharmaceutical industry to generate its own direct head-to-head evidence early in the drug development process.

Complex system analysis is already widely used in the physical sciences, meteorology, behavioural sciences and economics. Health outcomes and healthcare definitely obey the definition of complex systems, a new field of science that studies how the different parts of a system give rise to its collective behaviours, and how the system interacts with its environment. It is not possible to capture the actual outcome of a health intervention without integrating all the variables that contribute to it, such as healthcare policies, patient behaviour, the services offered, the societal attitude toward the disease and the insurance coverage. Such variables are currently treated as confounding factors whereas they are in fact an integrated part of the observed outcome.

New mathematical methods to support decision-making for complex systems have been developed, but they remain rarely used in public health. None of the most frequently used statistical methodologies, including propensity scoring (adjusting for differences in population sampling or confounding variables), addresses the complex system field in healthcare.

Clustering the assessment of the effectiveness of drugs assumes that the various health outcome components of an intervention can be analysed in isolation and that they are simply cumulative. Simultaneous analysis of all the factors that drive drug-related outcomes requires the use of complex systems sciences. This is the only way to support proper decision-making.

The cost of an indirect comparison performed by a consulting company ranges from $200,000 to $300,000. Such indirect comparisons are already used by European health technology assessment agencies, such as NICE in the UK.⁸ Investing $500,000 in complex system analysis would allow significant progress in analysing through different methodologies the results of studies already available. A lot of information that has already been generated and stored in data warehouses could be subject to this new analysis and would certainly yield new insights and deliver new benefits.

Without channelling effort into promoting these new methodologies it is likely that tax spending will continue to feed large and expensive pragmatic studies that bring little valuable information to support clear decision-making. Instead, alternative cost-effective methodologies should be encouraged through incentives for the development of comparative effectiveness. This would require a change of paradigm in drug effectiveness assessment. Yet this might well be the true challenge for the FCCCER if it wants to make best use of both the millions allocated to it and the historic opportunity to place comparative effectiveness research on the agenda of the life sciences sector.

References

6. S Hatcher. 'The STAR*D trial: the 300 lb gorilla is in the room, but does it block all the light?' Evidence-Based Mental Health, 11 (4), 97-99 DOI: 10.1136/ebmh.11.4.97, 2008

Professor Mondher Toumi is the chair of market access at the University of Lyon, France, and founder and president of Creativ-Ceutical, a strategic pricing and market access consultancy. Email: mondher.toumi@univ-lyon1.fr.