Estimating the Cost of Nosocomial Candidemia in the United States

Sir—Candida species are the fourth most common cause of bloodstream infections in the United States [1] and are associated with an attributable mortality of up to 38% [2]. Recent estimates published in this journal suggest that each case of nosocomial candidemia results in excess costs of $34,123 per Medicare patient and $44,536 per private insurance patient, with national costs of nosocomial candidemia exceeding $200 million per year in the United States [3]. However, since that report was generated, our understanding of the epidemiology of nosocomial candidemia has significantly improved. In the recent report estimating national costs of nosocomial candidemia, an incident case of candidemia was identified through International Classification of Diseases, 9th revision (ICD-9), discharge diagnoses listed in the Nationwide Inpatient Sample (NIS) database [3]. Until that time, this method was the best mechanism for estimation of incidence, since population-based estimates of candidemia were lacking. One limitation of ICD-9 codes is that they have been developed for billing purposes, and billing personnel typically use no more than 15 ICD-9 codes for a given hospitalization. Other investigators have found that identifying incident cases of infection by this method can result in underestimation [4]. Potential for underestimation may be especially pertinent when patients generate numerous diagnoses during a complicated hospitalization, as is the case with many patients with nosocomially acquired candidemia. A recently published population-based survey demonstrated that the incidence of nosocomial candidemia in the United States is 8/100,000 population [5], nearly 4 times the rate estimated in the recent cost analysis [3]. The accuracy of this estimate is supported by results from another large population-based study from Connecticut and the Greater Baltimore area that found a nearly identical rate of candidemia (8.1/100,000) [6]. Given that the incidence of candidemia is much higher than previously estimated, excess medical costs incurred as a result of this infection are ∼4 times previous estimates and would exceed $800 million per year in the United States. When one considers lipid formulations of amphotericin B, which were not widely prescribed during the time of the initial cost estimate, a growing population of immunocompromised patients at risk for Candida bloodstream infections, and costs of empiric therapy directed toward patients at high risk for candidiasis without proven invasive infection, costs of candidemia may be still higher. In summary, the excess cost in the United States attributable to candidemia is significantly greater than previously held, approaching $1 billion per year.