Point: The Power of Data to (Mis)Guide Policy to Improve Cancer Health Equity

Health expenditures have been an ongoing topic of debate in many health policy spheres. This issue of JNCCN includes an article, “Wealth, Health Expenditure, and Cancer: A National Perspective” (see page 972), that aims to provide a national perspective on the complicated interplay of wealth, health expenditures, and their relationship to cancer screening. In this ecological study, the investigators calculated Spearman's rank correlations to determine the association between financial indicators (gross domestic product [GDP] and health expenditure per capita) and cancer outcomes at the state level. The study found that only breast cancer mortality/incidence ratios were significantly correlated with health expenditure; all-cancer and colorectal cancer (CRC) outcomes were not. The investigators suggest that “findings represent evidence of regional disparities in terms of financial and cancer outcomes, indicating a potential inefficient allocation of resources in the efforts against cancer.”

We, as scientists, have the opportunity to move the needle on cancer health equity. Some may argue that within this opportunity lies an underlying social responsibility to consider how science affects public conversations and policy decisions that shape our present and future society. The messages we send to consumers in our publications are scrutinized not only by the readers of scientific journals but also by the wider media. Media then frame the messages and subsequently distill them to policymakers, other leaders, and, most importantly, communities, including those who are differentially impacted by these messages.

Our concern with this article is that in the absence of additional angles and levels of analyses, these ecological data provide a limited viewpoint that does little to advance the groundwork necessary for future policy development and policy implementation. We caution the readers in interpreting these data out of context. This article presents data that predate the Affordable Care Act (ACA) implementation and the financial downturn experienced in the United States in 2008 and 2009. Studies have yet to fully illuminate the impact of the downturn and recovery from it and ACA implementation. The state-level data included in the study are also focused on broad health expenditures and not those specific to cancer. Variation within a state, especially among rural, suburban, and urban contexts and populations, are absent. Although conducting analyses at a more micro-level, such as county or city, would generate valuable findings, we understand this was outside the scope of this article's intent. Therefore, as a whole, these data may not be indicative of the “here and now,” and we caution readers eager to use this study to extrapolate to the local level or to cancer-related health expenditures.

Concrete illustrations of how this study's findings may fall short of depicting the interplay of wealth and expenditures as it impacts cancer outcomes can be found in cities where local data contrast with state trends. For example, Chicago has a significant black:white mortality disparity in breast cancer. Where one lives does impact health, and in Chicago's 77 community areas, there are major wealth and health expenditure differences in resource allocation at a community area level that can drive poor cancer outcomes. This picture might not be clear from reading the article.

Every time we design, conduct, and publish a study, we need to be thoughtful about what the end goal message is and how that message can be construed in the larger policy arena. The unintended potential consequence of an article with a “national perspective” orientation is to fuel the mindset, and in a sense give permission, for states that have high health expenditures to cap or decrease spending for certain health issues. Increased spending does not necessarily improve quality of care, but capping or cutting spending on health care does not necessarily solve problems either. Furthermore, many on-the-ground programs such as patient navigator/community health worker programs, advocacy groups, and community organizations impact cancer outcomes, yet are not reflected in the data presented. Many such programs and groups must find ways to support their efforts outside of state and federal funds. Policymakers can benefit from nuanced studies that examine what works best in different contexts to tailor health expenditures in a more precise manner.

With respect to CRC, we agree with the authors that, “it is important for states and federal agencies to conduct system-wide efforts against barriers to CRC screening programs by limiting disparities, increasing access, and improving compliance.” Cancer inequities such as those in CRC screening may certainly stem from several factors, as a large body of research highlights. However, a bundled view that high expenditures are not correlated with better CRC outcomes does not necessarily guide policymakers on which allocations are effective or which populations are impacted and to what extent. The authors state that, “It is important to match observations to the level of decision-making. As such, assessing these associations at the higher system level of the allocative model is warranted to guide health policies but could not justify individual-level inferences.” However, health policies created from the broad generalities framed in this article have the potential to misguide readers if presented without additional perspectives or discussion.

Alongside a discussion of wealth and health expenditures, we should have dialogue on related topics of resource allocation, use, and the importance of “on the ground” efforts. The literature on community health workers, patient navigators, and other community-engaged or -partnered interventions, for example, consistently calls for connectivity between local community members, local wrap-around or enabling services, and the local health care teams, including community health centers and hospitals. These types of partnerships in conjunction with local health departments, state and federal policymakers, and other supporting agencies are critical to optimizing resources and improving cancer outcomes such as CRC and breast cancer screening. Indeed, the saying that it takes a village may apply to improving cancer inequities. We believe that examining state-level data about wealth and health expenditures at one time point falls short of contributing to current trends and addressing the policy issues at stake.

We are all in this together. We must work together to pose research questions, design studies, conduct analyses, and then present research that will contribute to advancing the state of the state in cancer health inequities, so that stakeholders as diverse as policymakers, scientists, payers, clinicians, and communities—especially those most differentially impacted by health inequities—can benefit from the research. This is not to say that studies should not be presented if they have no direct benefit to stakeholders. Rather, this article is a good example of how we as a scientific community can be mindful of end points and potential policy implications, especially when framing data within a national perspective. Otherwise we risk creating or exacerbating messaging that guides policies that further leave behind and obfuscate the experiences of disenfranchised communities and populations.