Making Sense of ED Mortality Statistics
Making Sense of ED Mortality Statistics
By Kathy Chen, MA
The Chicago School of Professional Psychology
The topic of eating disorders (EDs) and mortality can be particularly sensitive because of the controversial findings in the research and perhaps their implications. The common belief appears to be that EDs, especially anorexia nervosa (AN), are associated with a low rate of survival. Many factors influence the results of research regarding EDs and mortality; yet, the more confusing point seems to be the ways in which these results are recorded. Therefore, reviewing the research and examining the methods that lead to a study’s results could provide mental health professionals as well as friends and families with a more in-depth understanding of the relationship between mortality and EDs.
The first step to understanding the relationship between mortality and EDs is to study the specific parts of information presented in the research. First, there are multiple definitions used to determine the rate of death in EDs (Neumärker, 2000). For instance, the term “mortality rate” is different from the term, “standardized mortality ratio.” The mortality rate is usually expressed as the number of deaths per 100,000 of the population, whereas the standardized mortality ratio (SMR) is the number of observed deaths divided by the number of expected deaths in a specific population. In addition, when authors write of mortality rate, they usually refer to the crude mortality rate, which includes the number of deaths out of the total number of people studied during a specific amount of time. Let me give you a simple example. Let’s imagine we are researchers who have 100 people in our study. If we check back with these people in 10 years and find that 5 of them have passed away, then the crude mortality rate would be 5 percent. To calculate the SMR, let’s again think of a simple example. We are researchers who have 100 people in our study. This time these 100 people are adolescent females with AN. We observe that 50 of them have died within a certain amount of time. However, let’s say that the expected number of deaths for adolescent females without an ED is 10. Therefore, the SMR in this hypothetical example would be 5. This result means that the sample of adolescent females with AN has a level of mortality that is 5 times greater than the average adolescent female population without an ED. Though these terms may seem similar, they can nonetheless impact the ways in which readers interpret these results. Thus, when one reads research findings, one would likely want to be aware of the different ways that mortality is measured for the purpose of having a more comprehensive understanding of the research findings.
Second, mortality typically varies depending on the specific ED, such as AN, bulimia nervosa (BN), or eating disorder not otherwise specified (EDNOS), which is affected by the type and edition of the particular diagnostic manual. Specifically, some studies have been conducted via the Diagnostic and Statistical Manual of Mental Disorders (DSM) while others via the International Classification of Diseases (ICD). The research results can even vary depending on the edition of the classification system, as seen with the inclusion of EDNOS in the DSM-IV, which was previously absent in the DSM-III-R. For example, Korndörfer and colleagues (2003) used the DSM-III-R criteria for AN when conducting their study, which also seemed to include criteria associated with EDNOS because EDNOS was grouped with AN in the DSM-III-R. These authors’ results indicated no significant differences in long-term survival between those afflicted with AN and the general population in Rochester, Minnesota. However, the inclusion of characteristics of EDNOS likely lowered the mortality rate for AN in this study. Specifically, the EDNOS diagnosis typically involves disordered patterns of eating and means that the individual does not meet criteria for a specific ED, such as AN or BN. Therefore, the results from this study may be interpreted as AN having a similar mortality rate as the general population; however, a critical view of the research methods may suggest otherwise.
Finally, the populations that are sampled can significantly impact research findings. In particular, results will vary depending on whether or not the participants are from treatment facilities focused on ED treatment versus a general treatment setting. For example, Birmingham and colleagues (2005) found an SMR of 10.5 for the sample of AN individuals in a specialized ED program. In contrast, Korndörfer and colleagues (2003) found an SMR of 0.71 in AN patients who were not referred to specialized care. Though there were other factors influencing this significant discrepancy in SMR value, the population that was sampled could play an important role in the findings of a study. Therefore, many factors can influence the results of research surrounding mortality and EDs, which seem to contribute to greater difficulty understanding this topic.
As with all studies, one must keep a critical view of these research findings to gain a deep and comprehensive understanding of EDs and mortality. Many studies employ different methods to study mortality and EDs. Thus, producing generalized statements about this topic may actually be misleading. EDs are a serious physical and mental health concern. We need to give even greater attention to the relationship between eating disorders and mortality.
Birmingham, C., Su, J., Hlynsky, J., Goldner, E., & Gao, M. (2005). The mortality rate from anorexia nervosa. International Journal of Eating Disorders, 38(2), 143-146.
Korndörfer, S., Lucas, A., Suman, V., Crowson, C., Krahn, L., & Melton, J. (2003). Long-term survival of patients with anorexia nervosa: A population-based study in Rochester, Minn. Mayo Clinic Proceedings, 78, 278-284.
Neumärker, K. (2000). Mortality rates and causes of death. European Eating Disorders Review, 8, 181-187.