How Valid is a Case Fatality Rate (CFR) When Monitoring Cholera Care?
This post originally appeared on the CORE Group M&E Working Group list serv.
Many serious diseases report a case fatality rate (CFR) to illustrate the severity of the infection or to monitor the effectiveness of treatment. However, when examining a CFR, one may need to dig a bit deeper to understand how the rate is determined. As with any rate, the CFR is calculated from a numerator and a denominator. Generally, the denominator is the number of people with the disease who are diagnosed as having the disease, often counted at the hospital or clinic. The numerator is the number of people, among those who are diagnosed, who die:
If left untreated, severe cholera has a CFR of about 50%. However, if cholera patients receive prompt and high quality treatment, all cholera patients should survive. Thus, for cholera, the CFR is used as a measure of the quality of care since timely and appropriate treatment should prevent death. Therefore, a generally accepted bench mark for quality of care is a CFR of <1%.
Calculating the CFR for cholera
When cholera outbreaks occur, countries are supposed to report the number of cases and the CFR to the World Health Organization (WHO). These numbers are then published in the Weekly Epidemiological Record (1). It seems logical that the CFR would be a useful metric, however, it is not clear if the CFR is calculated in the same way across countries, or that the numerator and the denominator are even from the same population which is required when calculating a valid rate. In fact, as illustrated below, in most cases, neither the numerator nor the denominator are known with accuracy.
Understanding the cholera CFR denominator
When calculating the CFR for cholera, the denominator is the number of patients who are diagnosed as having cholera. This diagnosis however, is based on a clinical definition of cholera. According to the WHO, “In an area where there is a cholera epidemic, a patient aged 5 years or more who develops acute watery diarrhea is considered a cholera case” (2). However, the clinical diagnosis lacks specificity since cholera cases are rarely confirmed using a reliable laboratory test, and in fact, many of patients who present with cholera-like symptoms do not have cholera confirmed by laboratory tests. Rather, their illness is due to other causes. For example, in Bangladesh only about 30% of suspected cholera cases are confirmed to be cholera. However, in many African outbreaks, the proportion of such patients may be up to 80%. It is possible to correct for these uncertainties by testing a random sample of patients to make an adjustment for the over-counting and the under-counting of cases, but this is rarely done. Regardless, the variability of the true numbers for the denominator is quite broad. Another problem relates to the exclusion of younger children in the WHO definition. In fact these young children young may have cholera, but if these children with true cholera are not counted, this reduces the number of the denominator. Thus, the over-counting of cases among older patients, and the undercounting of cases among young children creates considerable uncertainty when calculating the CFR.
Understanding the cholera CFR numerator
The numerator, or the number of deaths from cholera, would seem to be an objective number that could be verified; however, even here there is uncertainty. Cholera patients who die may have died after reaching the treatment center, or alternatively, they might have died at home, in transit, or at a different health facility. For some data sets, only the deaths occurring in specific hospitals are counted, but in others, an attempt is made to include all cholera deaths, even those who do not reach the hospital. Unfortunately, the number of patients who die outside of the hospital is often not known, especially among populations living in remote and underserved areas. In fact, these community deaths are only detected through surveys after an outbreak and could exceed the numbers of those who die in the hospital. If community deaths are included in the numerator and the denominator only includes those clinically diagnosed with cholera, the CFR is not truly a rate since the numerator and denominator are not from the same population. Unfortunately, reports to WHO generally do not indicate the methods used to count the number of deaths.
Why is understanding the uncertainties of CFR important?
First, if you see a CFR <1%, the data used for the numerator and denominator need to be clear in order to understand if the rate is accurate or reflects an overestimate of cholera cases thereby falsely decreasing the rate and underestimating the true CFR. Second, if the CFR is <1% among hospital cases, there could be a false assumption that the situation is being managed well. However, this assumption could only be verified once it’s known how many deaths are being accounted for in the community. Finally, teasing out the numbers to understand where deaths are occurring is critical to reducing cholera deaths. A high CFR does not necessarily indicate what intervention will be most effective. For example, if the majority of deaths are occurring in a hospital, clearly improvements are needed to improve clinical care. Whereas, if deaths are occurring largely outside the hospital, the intervention may need to focus on issues regarding access to transportation to a clinic or increasing awareness among community members and health workers of the importance of prompt care seeking. Therefore, I recommend caution when reviewing reports of CFR, especially when comparing these figures between different countries and different settings. Understanding the validity of the numbers used to calculate a CFR and the way in which they are calculated is essential when deciding how the information should be used to inform public health programs. An accurate CFR is very useful for individual treatment centers when the actual number of cases and deaths are known with certainty. However, when evaluating the disease burden of cholera in a country, CFR may be a misleading number when neither the numerator nor the denominator is well defined. Additional efforts are needed to identify methods for estimating overall mortality, to identify the circumstances where cholera patients are dying and to develop strategies to prevent these deaths. When counting these deaths, countries should list, not only the death as being due to cholera, but also define where the death occurred and the circumstances which prevented adequate care.
- Cholera, 2015. Wkly Epidemiol Rec 2016; 91(38):433-440.
- Glossary of Terms for Cholera and Cholera Vaccine Programs: https://www.stopcholera.org/sites/cholera/files/glossary_of_terms_for_cholera_and_cholera_vaccine_programs.pdf.