Impact of rapid methicillin-resistant Staphylococcus aureus polymerase chain reaction testing on mortality and cost effectiveness in hospitalized patients with bacteraemia: a decision model

Abstract

Background: Patients hospitalized with Staphylococcus aureus bacteraemia have an unacceptably high mortality rate. Literature available to date has shown that timely selection of the most appropriate antibacterial may reduce mortality. One tool that may help with this selection is a polymerase chain reaction (PCR) assay that distinguishes methicillin (meticillin)-resistant S. aureus (MRSA) from methicillin-susceptible S. aureus (MSSA) in less than 1 hour. To date, no information is available evaluating the impact of this PCR technique on clinical or economic outcomes.

Objective: To evaluate the effect of a rapid PCR assay on mortality and economics compared with traditional empiric therapy, using a literature-derived model.

Methods: A literature search for peer-reviewed European (EU) and US publications regarding treatment regimens, outcomes and costs was conducted. Information detailing the rates of infection, as well as the specificity and sensitivity of a rapid PCR assay (Xpert MRSA/SA Blood Culture PCR) were obtained from the peer-reviewed literature. Sensitivity analysis varied the prevalence rate of MRSA from 5% to 80%, while threshold analysis was applied to the cost of the PCR test. Hospital and testing resource consumption were valued with direct medical costs, adjusted to year 2009 values. Adjusted life-years were determined using US and WHO life tables. The cost-effectiveness ratio was defined as the cost per life-year saved. Incremental cost-effectiveness ratios (ICERs) were calculated to determine the additional cost necessary to produce additional effectiveness. All analyses were performed using TreeAge Software (2008).

Results: The mean mortality rates were 23% for patients receiving empiric vancomycin subsequently switched to semi-synthetic penicillin (SSP) for MSSA, 36% for patients receiving empiric vancomycin treatment for MRSA, 59% for patients receiving empiric SSP subsequently switched to vancomycin for MRSA and 12% for patients receiving empiric SSP for MSSA. Furthermore, with an MRSA prevalence of 30%, the numbers of patients needed to test in order to save one life were 14 and 16 compared with empiric vancomycin and SSP, respectively. The absolute mortality difference for MRSA prevalence rates of 80% and 5% favoured the PCR testing group at 2% and 10%, respectively, compared with empiric vancomycin and 18% and 1%, respectively, compared with empiric SSP. In the EU, the cost-effectiveness ratios for empiric vancomycin- and SSP-treated patients were Euro 695 and Euro 687 per life-year saved, respectively, compared with Euro 636 per life-year saved for rapid PCR testing. In the US, the cost-effectiveness ratio was $US 898 per life-year saved for empiric vancomycin and $US 820 per life-year saved for rapid PCR testing. ICERs demonstrated dominance of the PCR test in all instances. Threshold analysis revealed that PCR testing would be less costly overall, even at greatly inflated assay prices.

Conclusions: Rapid PCR testing for MRSA appears to have the potential to reduce mortality rates while being less costly than empiric therapy in the EU and US, across a wide range of MRSA prevalence rates and PCR test costs.