What if health care economics became more personalized?
Twenty-three years ago, Herceptin® (trastuzumab) became the first FDA-approved therapy to treat HER2-overexpressing breast cancer. It will go down in history as the first medicine approved with the use of a companion diagnostic—the first “personalized medicine.” Herceptin was the kick-starter for designing the right drug for the right patient at the right time. Flash forward to 2021—recent estimates indicate that more than 230 biomarkers are being used with approved targeted medicines across many different disease areas. In general, medicines provide a higher value within their target population than in the broader patient population. With a growing cache of real-world data, can we apply this high-value approach to all therapies regardless of their status as personalized medicines?
Within a health care system, estimating the relative value of therapies is crucial when determining the allocation of the limited available resources. Traditionally, relative value is assessed through cost-effectiveness analysis, which compares the incremental costs and clinical benefit of a new therapy to the standard of care. The system then determines if it is willing to pay for the incremental clinical benefit of the new therapy compared to the standard of care.
Cost-effectiveness analyses are primarily based on an average patient. Clinical and economic inputs in the models are often the mean or median outcomes captured through pivotal clinical trials—intended to represent a typical patient. However, a clinical trial patient, by design, may not represent a true “typical” patient that physicians encounter in the everyday, real world of clinical practice. They are certainly not representative of you or me. In today’s environment where everything in our lives is trending toward personalization, including medicine, shouldn’t market access and reimbursement decisions be personalized as well?
Patient clusters with different levels of response to therapy may be categorized by similar demographic or clinical characteristics. The typical patient for the different patient clusters would yield different relative value outcomes in cost-effectiveness analysis. In precision medicine, the ability to identify patient clusters or individuals with a higher likelihood and tighter range of response to a therapy allows for an increase in the clinical benefit compared to the broader population—ultimately increasing the relative value of the therapy. Identification of patient clusters gets us closer to individualized market access and reimbursement decisions, which would bring much-needed efficiencies to the health care market. Can we go deeper than the subgroups already identified in trial populations to identify more specific patient clusters?
Using large real-world EMR and claims datasets we can. These datasets allow researchers to dig deeper than possible in clinical trials due to the number of patients and long-term follow-up available. They also provide stakeholders with a more accurate representation of patient characteristics and outcomes outside of the clinical trial setting. Patient similarity analysis identifies clusters of patients with similar characteristics and outcomes. These clusters allow for more accurate health economic modeling as the typical patient becomes more representative of all individuals in the cluster. As the data evolves, cost-effectiveness models can shift from one-off exercises to real-time dynamic models that are continuously updating based on new data. The risk of statistical uncertainty is also reduced over time as the dataset expands.
At Real Chemistry, we believe real-world data and analytics can make the health care system more efficient and ensure innovative medicines get to the patients who can most benefit from them. Personalized health economics allows stakeholders to confidently support real-time individualized reimbursement and become more efficient resource allocators.
