Skip to main content

Universal vs Targeted Chlorhexidine Bathing and Nasal Decolonization in Hospitalized Patients.

JAMA network open2025-03-10PubMed
Total: 76.0Innovation: 7Impact: 9Rigor: 7Citation: 9

Summary

Using a decision-analytic model informed by the cluster-randomized ABATE trial, targeted chlorhexidine bathing and nasal decolonization for patients with medical devices was cost-effective from both payer and hospital perspectives across broad scenarios. Universal decolonization minimized HOB events but required higher incremental costs per event averted and may be preferable only in high-device or low-adherence settings.

Key Findings

  • In the base case, standard of care was least effective and most costly; targeted decolonization was least costly.
  • Universal decolonization achieved the fewest HOB events but had ICERs of $119,700 (payer) and $126,600 (hospital) per HOB averted versus targeted.
  • Targeted decolonization was cost-effective across broad scenarios; universal may be preferred in high device-prevalence units or with low targeted adherence.

Clinical Implications

Adopt targeted decolonization for general medical/surgical units as the default, reserving universal decolonization for high-device units or when targeted adherence is poor; align infection prevention with payer and hospital willingness-to-pay thresholds.

Why It Matters

This analysis directly informs hospital policy by quantifying cost-effectiveness trade-offs for decolonization strategies aimed at reducing hospital-onset bloodstream and fungal infections, a key sepsis prevention lever.

Limitations

  • Model outcomes depend on assumptions (adherence, costs, effect sizes) and willingness-to-pay thresholds
  • Generalizability may vary by unit case mix and implementation fidelity

Future Directions

Prospective implementation studies comparing targeted vs universal strategies by unit type, with real-world adherence, resistance ecology, and equity outcomes.

Study Information

Study Type
Cohort
Research Domain
Prevention
Evidence Level
III - Decision-analytic model informed by a large cluster-randomized trial; no new randomized allocation
Study Design
OTHER