Clinical Utility and Cost-Effectiveness of Molecular Diagnostic-Guided Antimicrobial Therapy in Sepsis: A Retrospective Cohort Study
Description
This study aims to determine the clinical utility and cost-effectiveness of molecularly-guided antimicrobial therapy in sepsis. Sepsis is a leading cause of morbidity and mortality worldwide, with a high burden of disease and death reported in various contexts, and is recognized by the WHO as a public health priority, highlighting the need to improve its timely diagnosis and management. According to the Sepsis-3 consensus, sepsis is defined as life-threatening organ dysfunction caused by a dysregulated host response to infection, and early identification of the etiological agent and antimicrobial resistance mechanisms is essential for optimizing treatment. In this context, rapid molecular diagnostic tests, such as multiplex PCR, allow for the early detection of pathogens and resistance genes, reducing the time to appropriate therapy. Several studies have demonstrated that their implementation is associated with improved clinical outcomes and optimized antimicrobial use, in addition to being a cost-effective strategy, particularly when integrated with antimicrobial stewardship programs. These tools also contribute to the advancement of precision medicine in infectious diseases. This was an observational and retrospective study conducted in a private level III-1 hospital in Lima, which included 200 patients who met the criteria for Sepsis-3, had been hospitalized for more than 48 hours, and presented positive results in blood cultures and molecular tests (2019-2023), excluding cases of COVID-19, bacteria of low clinical relevance, and contaminated samples. The key words used were: Sepsis, Length of Stay, Mortality, Cost-Effectiveness Analysis. This study was approved by the Ethics Committee of the University of Tacna in Perú (Resolution FACSA-CEI/004-02-2025). The study was conducted in accordance with the principles of the Declaration of Helsinki.
Files
Steps to reproduce
An ad hoc Excel spreadsheet was created with all the variables and possible scenarios. Values of 0 for "no" and 1 for "yes" were assigned to dichotomous nominal variables, and values of 0, 1, and 2 were used for polytomous nominal variables, as described in the database legend. Data were collected at a private healthcare facility with inpatient and intensive care units. Information was gathered at the end of hospitalization, and patients were required to meet the inclusion criteria: hospitalization for more than 48 hours and a positive culture result. The statistical analysis was performed using SPSS software version 27. From a population of 485, the estimated sample size was 215, determined with Epidat software version 4.2 with 80% power for cohort sample calculations. A descriptive analysis was carried out, presenting categorical variables through frequencies and percentages, and numerical variables with measures of central tendency and dispersion. The distribution of continuous variables was assessed using the Shapiro–Wilk and Kolmogorov–Smirnov tests. For outcome frequencies lower than 10%, the Odds Ratio (OR) was calculated; in general, a p value < 0.05 was considered significant, and a 95% confidence interval (CI) was used. In the inferential analysis, to compare variables between groups, we used the Chi-square test, Fisher’s exact test for categorical variables, and the Mann–Whitney U test for numerical variables, after confirming their distribution. As measures of association, the crude relative risk (RR) was calculated, and the adjusted risk was estimated using log-binomial regression. Confounding variables were identified through the Mantel–Haenszel method and log-binomial regression; multivariate analyses were also performed using negative binomial regression and generalized linear models. To control for disease severity, patients were stratified according to their severity (ICU admission, mechanical ventilation, guided treatment for MM, Urinary tract infection, E.coli infection, hospitalary days, among others) to assess whether the benefits of molecular tests varied according to disease severity. All these steps were used to arrive at the results that we will present in the study.