FROM MMWR
Despite challenges in estimating sepsis-related mortality in the U.S. because of its complex clinical nature and variety of underlying causes, estimates based on administrative claims data may be more accurate than those derived from death certificates, according to a report published April 8 in the Morbidity and Mortality Weekly Report.
Dr. Lauren Epstein of the division of healthcare quality promotion at the National Center for Emerging and Zoonotic Infectious Diseases, and her colleagues, compared U.S. sepsis-related mortality estimates from different sources. Deaths attributable to diagnoses corresponding to ICD-10 diagnosis codes A40 (streptococcal septicemia) and A41 (other septicemia) from 1999 to 2014 were extracted from the CDC WONDER (Wide-ranging Online Data for Epidemiologic Research) database. Administrative claims data using various combinations of the ICD-9-CM administrative codes for primary or secondary infection and organ dysfunction to identify severe sepsis from 2004 to 2009 were extracted from the largest all-payer, publicly available inpatient database in the United States, the Nationwide Inpatient Sample. ( MMWR. 2016 Apr 8;65[13]:342-5 ).
Of the roughly 2.5 million death certificates listing sepsis as a cause of death, 22% identified sepsis as the underlying cause of death during the time period assessed. The results of the comparison demonstrated that the estimated range of sepsis-related mortality based on death certificate data was lower than that obtained using administrative claims data (ranges, 146,000-159,000 and 168,000-381,000, respectively). These results indicate that the annual estimate based on administrative claims data during the time period assessed was 15%-140% higher than the estimate based on death certificate data.
To explain the difference between the estimated sepsis-related mortality ranges, the authors said that while both death certificate and administrative claims data are important sources of public health information, they are each associated with limitations that can affect such estimates. For example, the authors said that death certificate certifiers may be more prone to record sepsis as an immediate cause of death, resulting in lower estimates of sepsis-related mortality based on underlying causes of death. Regarding administrative claims data, the authors said that such data cannot be all inclusive, as only those sepsis-related deaths occurring in health care facilities are captured.
The authors said that their results highlight the need for a better defined and more reliable sepsis surveillance system that should be based on objective clinical data. This approach would allow for increased accuracy in the tracking of sepsis trends in the United States, as well as an improved system for gauging the impact of sepsis awareness and prevention efforts.
No funding sources or conflicts of interest were reported.
