Assessment of medical expenditures for sepsis:differentiating between cases with and without ruled-out diagnoses

Setting public health priorities requires precise estimation of the burden of disease, including disease-specific medical expenditure. Information on multiple and ruled-out diagnoses on health insurance claims (HICs) has been ignored in traditional analyses of disease-specific medical expenditures in Japan. This study reviewed 448 inpatients with at least one diagnosis of sepsis on their HICs, who were insured by corporate health insurance organizations making claims on services provided from April 2006 to March 2007 in Japan. Subjects in whom sepsis-related diagnoses were specified as "ruled-out" were compared with subjects in whom sepsis-related diagnoses were classified as "not-ruled-out" (i.e., subjects in whom sepsis was considered possibly or likely present). Direct medical expenditure, length of stay (LOS), cost per day, cost of antibiotics, and proportion of administered cephalosporin and carbapenems were significantly higher in subjects classified as not-rule-out. When using health insurance claims in Japan, the statistics of medical expenditures and LOS are influenced by procedures performed to rule out a diagnosis, as well as those performed to treat a confirmed diagnosis of sepsis.