by Daniel Kantor, Michelle Brauer and Jason Shafrin
As drug prices continue to rise, many patients, payers and policymakers are asking whether the treatment benefits justify the cost. One method favored by health technology assessment (HTA) bodies for making this decision in a formal manner is to conduct a cost effectiveness analyses (CEA). However, there is no single method to conduct CEA that is universally agreed upon. For instance, what constitutes value varies depending on whether it is measured from the patient, payer or societal perspective. In some cases value is limited to measures of direct patient health outcomes—such as safety and efficacy—while others analyses consider other factors such as treatment effects on patient productivity, educational outcomes, and burden on caregivers. Consequently, conclusions about the cost-effectiveness of RMS treatments vary widely.
In September 2016, the Second Panel of Cost-Effectiveness in Health and Medicine (Second Panel) published guidelines that offer a potential common ground in how to conduct CEA. This begs the question: to what extent have previous cost effectiveness studies followed the Second Panel’s recommended best practices? In this article, we attempt to answer this question for the case of disease modifying treatments (DMTs) in multiple sclerosis.
A short history of recent innovations to treat multiple sclerosis
Over the past 20 years, a number of new treatments have been developed that improve the prognosis of patients with multiple sclerosis a debilitating neurological disease that causes mobility impairment and disability in patients. In the early 1990’s and 2000’s, injectable BRACE treatments (Betaseron, Rebif, Avonex, Copaxone, Extavia) were developed to treat relapsing forms of multiple sclerosis (RMS) and became the standard of care to reduce disability and flare ups in symptoms, known as relapses More recently, non-BRACE treatments (Aubagio, Gilenya, Lemtrada, Ocrevus, Tecfidera, Tysabri, and Zinbryta) were developed and offer further benefits to patients. Because of these innovative therapies, an uninsured patient diagnosed with RMS today can expect significantly better outcomes than those diagnosed 20 years ago. A network meta-analysis performed by CADTH, found that whereas non-BRACE therapies reduced relapse rates by 15%-30% compared to a placebo, non-BRACE therapies reduced relapses by 50%-80%.
While patient outcomes have improved dramatically, financial pressures have been building on both payers and patients. BRACE treatments originally costing $8,000 to $11,000, now cost approximately $60,000 per year. Newer treatments for RMS and newer RMS treatments entering the market cost 25-60% more than existing treatments.
Heterogeneity in assessments of value
To determine how well previous cost effectiveness studies have followed the Second Panel’s guidelines on measuring value from a societal perspective, we review five cost-effectiveness analyses for RMS treatments. These studies were selected from peer reviewed publications and English speaking HTA agencies Studies identified CEA performed in the US, UK, and Canada by ICER NICE, CADTH, and publications by Noyes et al 2011 and Lee et al 2012.
In these studies, direct clinical benefits to patients were always captured by measuring outcomes needed for pharmaceutical approval, however, other clinical outcomes important to patients and cognitive measures that assess disease progression were captured inconsistently. While all studies accounted for quality of life, there was no consensus on which aspects of quality of life should be considered. For example, all studies included depression, mobility, and pain, however only 3/5 included fatigue and 2/5 included bowel and bladder control. Similarly, measures of cognitive function (3/5) and patient preferences such as route (3/5) and frequency of administration (1/5) were captured inconsistently, and novel value components—such as the “peace of mind” value of innovative treatment to healthy individuals—were excluded.
Similarly, although direct medical costs were included in all five studies, there were inconsistencies in addressing non-medical costs and the impact on other stakeholders. All CEA captured costs of treatment, hospitalization and office visits to health professionals, yet non-medicals cost of equipment and home improvements often required by RMS patients and their families were only included in 3/5 studies. Because of the extreme impairment and cognitive declines associated with RMS, patients are frequently unable to work or must work part time. This lost productivity was also only captured in 3/5 assessments. Costs borne by non-patients, such as caregiver burden (3/5) and the cost of support services provided by non-medical organizations (2/5) were captured inconsistently and the impact on children of patients with MS was omitted.
Novel components of value are excluded from all analysis
Recent research indicates that standard CEA may be ignoring the full treatment benefit to society. For instance, the “peace of mind” value to healthy individuals is ignored in standard CEA. Although the value of new RMS treatments to non-patients is modest a per person basis, when multiplied over all non-patients, the impact is large. A substantial proportion of benefit is derived from “peace of mind” value – the value of treatment to risk-averse non-patients due to the positive probability that they could have multiple sclerosis and would value the treatment in the future. A recent study of RMS therapies suggests that incorporating insurance value almost doubles the treatment value.
Evidence of other novel value components is limited in RMS, however there is evidence of additional value from studies of other diseases. For example, if RMS therapies can slow disease progression, patients’ health benefits may be underestimated due to the treatment’s option value. In other words, patients may be able to delay disability progression sufficiently in order to have improved responses from any new treatments for RMS that soon come on line. Studies in other severe diseases show that this option value may provide an additional 25% of value. Many HTA also fail to incorporate the existence value, the altruistic value of having the treatment available for others who need it. By incorporating these novel value components, HTA can develop a more comprehensive assessment of treatment benefits and costs.
Additional methodological challenges
Although the analysis above describes the heterogeneity in the value components included in the cost effectiveness analysis, different stakeholders vary in how they aggregate these value components. For instance, employers may care more about productivity than the public payers, while patients may care more about caregiver burden than private health insurers. Even within a given stakeholder group, aggregating preferences is problematic. Arrow’s impossibility theorem demonstrates that aggregating preferences across individuals is problematic and has been shown to be unsolvable in political contexts. Thus, despite the advantages of aggregating preferences into a single measure of value, this measure will inevitably obscure heterogeneity both across and within different stakeholder groups.
Even if it was possible to agree upon a single measure of value for all stakeholders, other methodological constraints exist that prevent HTA from comprehensively evaluating treatment benefit. Clinical trials rarely report outcomes other than those used for regulatory approval. Even if other outcomes were available, experiences in clinical trials may not be representative of real world experiences. Furthermore, patient populations in clinical trials may not match populations in the real world. Despite the fact that effectiveness may vary dramatically for patients with multiple chronic conditions, few RCTs are powered to identify efficacy specifically for these patients. Additionally, clinical trials for newer therapies use more sensitive MRI technology that leads to earlier detection of lesions in healthier patients, whereas clinical trials for older therapies detect lesions later in patients with severe disease states. This phenomenon—known as the Will Rodgers phenomenon—has been well-documented in the case of multiple sclerosis diagnosis over time. If sicker patients are more likely to experience faster disability progression, older clinical trial may demonstrate a larger proportional decrease in disease progression simply due to the patient population included in the trial. This fact makes comparisons of trials that occur over a long time horizon more difficult. Lastly, some studies measuring option value, value of hope and existence value are not available among the RMS patient population.
A path forward
Technology assessment provides a useful way of synthesizing information, but patients, payers and policymakers should not fixate on a single number given the limitations of availability data and the various methodological approaches for measuring value. Despite these limitations, we recommend that any future technology assessments follow the recommendations from the Second Panel on Cost Effectiveness and take the following approach:
- Consider multiple perspectives on value including the patient, payer and societal perspective
- Describe the components of value explicitly
- Create interactive tools that allow the user—whether it be a patient, payer, or policymaker—to determine which components of value are most important to them
The goal of technology assessment in a U.S. market ultimately is not to impose a top-down, single value for all stakeholders. Rather, technology assessment should be used to inform stakeholders to facilitate pricing negotiations and prescribing decisions.
Disclaimer: The research on heterogeneity in assessments of value described in this article was funded by the Novartis Pharmaceutical Company.