The application of the mortality‐to‐incidence ratio for the evaluation of cancer care disparities globally

We read with interest the letter by Cordero-Morales et al. Our purpose in writing our original article1 was to draw attention to disparities in cancer screening and treatment. This was done in relation to the mortality-to-incidence ratio (MIR) in Organization for Economic Co-operation and Development (OECD) countries characterized by high-quality health-related data. Cordero-Morales et al raised methodologic and conceptual issues that we address herein. First, the authors contend that the MIR is limited to future predictive capacity based on its correlation with survival (ie, as 1-MIR).2 However, the objective of that article by Asadzadeh et al was to assess the usefulness of the MIR as a proxy for long-term survival across multiple cancer types. Cordero-Morales et al claim a conceptual linkage between 5-year survival2 with disparities in care and future predictive capacity, for which they provide neither empirical evidence nor relevant citations. Indeed, there is a separate literature regarding the application of the MIR in the setting of disparities. This has been done across multiple cancer sites in South Carolina3 and Georgia,4 in Australia,5 in federally qualified health centers in the United States,6 and in examining the Human Development Index in gastrointestinal cancers.7 Second, Cordero-Morales et al contend that our demonstration of a qualitative linkage between divergent MIR values and the presence or absence of national colorectal screening measures is insufficient. They cite the work of Batouli et al,8 which shows an association between a higher concentration of external beam radiotherapy devices and lower MIRs, as a more appropriate substantiation of the association between MIR and cancer care. However, the presence and implementation of colorectal screening guidelines is a broader indicator of cancer-associated health infrastructure than more advanced treatment modalities such as devices that are generally used for treating patients with later-stage cancers. Third, Cordero-Morales et al take issue with using the World Health Organization health rankings as a means of evaluating a cancer-related outcome. As a logical extension of the previous point, we note that cancer care is not limited to cancer specialty resources alone. Competent cancer screening depends on a nonspecialist primary care infrastructure as a critical bridge to the appropriate identification of at-risk patient populations. The findings of our study1 corroborate the principle that cancer care lies on a continuum that does not begin or end with cancer specialists or advanced treatment modalities. Thus, the World Health Organization health rankings, with their emphasis on health care delivery and equity, provide a clear and durable conceptual link to cancer care. Fourth, a concern was raised regarding the temporal relationship (ie, ordering) between the incidence and mortality data. For cancers with highly variable survival times (including colorectal cancer), the choice of the “denominator period” is not straightforward. To examine the effect of altering the denominator period to vary alignment with the average incidence-to-mortality time interval, we conducted sensitivity analyses using data from the South Carolina Central Cancer Registry, whose collection of incidence data began in 1996.3, 9 MIRs were parallel across 8 different 5-year periods beginning in 1996 across all combinations of sex and race for all anatomic sites, including colorectal cancer. Therefore, although it is true that the choice of the denominator period matters less for very virulent cancers from which most people die quickly, it also appears that it is true for less deadly cancers, largely because rates are relatively stable over even moderate periods of time. High-quality cancer registries in areas in which health care delivery is reasonably good, as in OECD countries, produce rates that are consistent over time. Fifth, alternatives to the MIR require special permissions and other logistical and legal complexities needed for data linkage. For example, data regarding cause and date of death are not usually released directly to researchers and permissions may be very difficult to obtain. This, along with the movement of individuals in mobile societies, greatly complicates the ability to define survival time and may even lead to “immortal patients” (ie, patients with cancer who are most likely deceased, but for whom a death record could not be found). Use of the MIR obviates this problem. We agree that “understanding the actual mediating mechanisms” should be an important goal of work focused on use of the MIR. However, it also is important to recognize the past literature on the MIR and cancer care disparities, as well as the significant role of broader health infrastructure in supporting the appropriate identification and treatment of at-risk patients with cancer. In doing so, the MIR can be even more productively applied as a means to help reduce and eventually eliminate these substantial global disparities in care. Vasu Sunkara, MD, PhD Saint Vincent Hospital Erie, Pennsylvania James R. Hébert, ScD Cancer Prevention and Control Program Department of Epidemiology and Biostatistics Arnold School of Public Health University of South Carolina; Department of Family and Preventive Medicine University of South Carolina School of Medicine Columbia, South Carolina