Prevalence of bacterial contamination and operating thresholds (ROC-Youden) for cell phones and keyboards in Univalle Hospitals, Cochabamba, Bolivia

Abstract

Background: The introduction of technological devices such as cell phones (CP) and computer keyboards (KB) has streamlined clinical work; at the same time, however, it has introduced surfaces that facilitate the transmission of healthcare-associated infections caused by pathogens with a high potential for antimicrobial resistance. Aim: To quantify the prevalence of microbiological positivity in hospital CP and KB to establish decision thresholds and assess geographical heterogeneity (south vs. north). Materials and Methods: A cross-sectional observational study (N=29) using cultures in blood agar and EMB agar. The prevalence was assessed, and survey predictors (frequency, timing, cleanliness) were used. ROC curves were estimated to determine the optimal cutoff point (Youden’s index), and penalized multivariate logistic regression was applied to estimate adjusted odds ratios (OR). Results: Overall growth in blood agar was extremely high (CP: 97%; KB: 90%). The main isolates were coagulase-negative Staphylococcus (CP: 62%; KB: 52%) and Gram-positive bacilli (CP 52%; KB 55%). ROC analyses revealed a clear dose-response pattern and clinically interpretable thresholds. In the multivariate model, greater frequency of use was associated with a significant increase in risk (OR > 1), while the use of isopropyl alcohol had a significant protective effect (OR < 1). Discussion: CP and KB serve as potential vectors. The integration of ROC-Youden thresholds and multivariate models facilitates the conversion of evidence into concrete actions. The primary operational recommendations are to standardize disinfection using isopropyl alcohol and to integrate Youden‑derived thresholds (e.g., time‑ or use‑based limits) as triggers for mandatory cleaning with routine compliance audits. The moderate sample size and the lack of species classification limit specific clinical inferences. It is advisable to conduct multicenter, longitudinal studies that include typing and the evaluation of resistance profiles.