Recommendations for Methicillin-Resistant Staphylococcus aureus Prevention in Adult ICUs: A Cost-Effectiveness Analysis*

Objective: Patients in the ICU are at the greatest risk of contracting healthcare-associated infections like methicillin-resistant Staphylococcus aureus. This study calculates the cost-effectiveness of methicillin-resistant S aureus prevention strategies and recommends specific strategies based on screening test implementation. Design: A cost-effectiveness analysis using a Markov model from the hospital perspective was conducted to determine if the implementation costs of methicillin-resistant S aureus prevention strategies are justified by associated reductions in methicillin-resistant S aureus infections and improvements in quality-adjusted life years. Univariate and probabilistic sensitivity analyses determined the influence of input variation on the cost-effectiveness. Setting: ICU. Patients: Hypothetical cohort of adults admitted to the ICU. Interventions: Three prevention strategies were evaluated, including universal decolonization, targeted decolonization, and screening and isolation. Because prevention strategies have a screening component, the screening test in the model was varied to reflect commonly used screening test categories, including conventional culture, chromogenic agar, and polymerase chain reaction. Measurements and Main Results: Universal and targeted decolonization are less costly and more effective than screening and isolation. This is consistent for all screening tests. When compared with targeted decolonization, universal decolonization is cost-saving to cost-effective, with maximum cost savings occurring when a hospital uses more expensive screening tests like polymerase chain reaction. Results were robust to sensitivity analyses. Conclusions: As compared with screening and isolation, the current standard practice in ICUs, targeted decolonization, and universal decolonization are less costly and more effective. This supports updating the standard practice to a decolonization approach.

http://ift.tt/2t0y3LK