In recent years, policymakers and payers have been focused increasingly not only on how new health technologies improve overall health, but also whether they can reduce health disparities. Oftentimes, however, discussions around health disparities are conducted on a qualitative basis with litlted quantitative analysis of tradeoffs between equity and effectiveness.
Distributional cost effectiveness analysis (DCEA) is one approach for measuring not only how new health technologies improve health for the average person, but also whether the new technology is likely to exacerbate or attenuate any existing health disparities. While DCEA is theoretically attractive to academics, there are a number of challenges of implementation in practice. This is the topic of an interesting commentary by Meunier et al. (2022).
One of the key questions is over which groups should health disparities be measured. Race? Income? Education? What if a new technology reduces health disparities across race but increases then across different income groups? One solution is to use various indices of health disparities. For instance, the United Kingdom often uses the Index of Multiple Deprivation (IMD) to measure health disparities. This approach incorporates 7 dimensions of deprivation including: income, employment, education, health, crime, housing, and living environment. In the US, the CDC uses the Social Vulnerability Index (SVI) to measure health disparities. While these indices are empirically attractive, for payers and policymakers they are a bit confusing. Reducing health disparities across race, income or education groups is a laudable goal that can be shared with constituents; reducing health disparities across various disparities indices (e.g., IMD or SVI) is harder to understand and may be less politically attractive option even it is more scientifically robust.
Once groups are defined, Meunier et al. (2022) note a number of other limitations related to data. Information on treatment effectiveness across groups is not always reported and even it is many clinical trials are not powered to measure effectiveness by socioeconomic status. Even if these data were collected in clinical trials, different groups may be more or less disadvantaged depending on the country and likely the clinical trial will not have sufficient sample size to examine efficacy or safety by subgroup by country. Information on access to treatment or medication adherence by group also may not be available, particularly at drug launch.
Additionally, additional value may occur when treatments are developed for diseases that disproportionately impact disadvantaged groups. However, disease prevalence by group are likely only available for the most common diseases; this is particularly when new treatments are indicated for specific disease subtypes. Epidemiological data–where it exists–may need to be linked across data sets which may be time and resource intensive.
In part because of these limitations, the authors wisely identify a number of areas of future research:
- Identify top-priority areas of health disparities. Do payers and policymakers care most about reduction in health disparities across racial groups? Education? Income? Other? Which areas of disparity are most important: access? outcomes? adherence? Clearly identifying these priorities can help with future data development and DCEA implementation.
- Improve data collection. Collecting more data will help to implement DCEA in a more robust manner. For instance, few electronic health records collect race, income or education data consistently, although CMS and other quality measurement groups are pushing to improve the collection of equity-relevant characteristics
- Inclusiveness of clinical trials. Clinical trials should strive to be as inclusive as possible in clinical trials. The authors note that ” White patients represented 76% of participants in clinical trials that supported the US Food and Drug Administration approval of new drugs between 2015 and 2019 based on a 2020 analysis by FDA, despite just 62% of the US population being White. These numbers may not be unreasonable if treatments often target patients who are older and older individuals–in the US at least–are more likely to be White. However, over time, the older population will increasingly look more diverse and drug manufacturers may want to oversample from more diverse group so their evidence base is future-proof as the US becomes more racially diverse.
DCEA is a useful tool for estimating tradeoffs between equity and effectiveness, but to implement DCEA effectively in practice, more work is needed.