ICU Structure and Care Processes Explain Much of VAP and CLABSI Variation in Brazilian ICUs: Findings from a 50‑Unit IMPACTO‑MR Nested Cohort

Highlight

– A multicenter nested cohort within the IMPACTO‑MR platform quantified VAP and CLABSI rates across 50 Brazilian adult ICUs and identified ICU‑level modifiable factors associated with lower infection burdens.
– Hospital‑level factors and processes of care accounted for roughly 40–45% of between‑hospital variability in VAP and CLABSI rates.
– Specific practices linked to reduced infections included prevention protocols, hand hygiene training, flexible family visitation, improved nurse staffing, use of single‑use gowns and alcohol, nurse‑led sedation titration, respiratory‑therapist‑led weaning, and dentist availability.

Background

Healthcare‑associated infections (HAIs) such as ventilator‑associated pneumonia (VAP) and central line‑associated bloodstream infection (CLABSI) remain major causes of morbidity, mortality, and cost in intensive care units (ICUs). Their incidence is typically higher in low‑ and middle‑income countries (LMICs) than in high‑income settings because of differences in staffing, infrastructure, surveillance, and adherence to prevention practices. Understanding which institution‑level attributes drive higher or lower HAI rates is essential to prioritize interventions that are feasible and scalable in resource‑constrained settings.

Study design

This investigation was a nested multicenter cohort study within the IMPACTO‑MR platform conducted in 50 Brazilian adult ICUs. The study prospectively collected individual patient data from September 2019 through December 2021 and paired these with cross‑sectional hospital‑level data collected at baseline. All patients with at least 48 hours of exposure to invasive mechanical ventilation (MV) were considered at risk for VAP; those with at least 48 hours of central venous catheter (CVC) exposure were considered at risk for CLABSI. Surveillance and case definitions followed guidance from the Brazilian regulatory agency. The analysis used negative binomial or Poisson multiple regression to model rates, adjusting for patient‑level covariates and including hospital random effects to quantify between‑hospital variability. Hospital‑level fixed effects were used to quantify the proportion of variability explained by measured institutional factors.

Key findings

Population and basic incidence:

Across 75,164 ICU admissions, 19,108 patients were at‑risk for VAP (≥48 hours MV), contributing 244,059 MV‑days. The VAP incidence rate was 6.03 per 1,000 MV‑days (95% CI, 5.73–6.35).
There were 26,560 patients at‑risk for CLABSI with 375,078 CVC‑days. The CLABSI incidence rate was 1.63 per 1,000 CVC‑days (95% CI, 1.51–1.77).

Between‑hospital heterogeneity:

The median rate ratio (MRR) for hospital random effects was 4.39 (95% CI, 2.72–6.06) for VAP and 3.53 (95% CI, 2.30–4.76) for CLABSI. These MRRs indicate large inter‑hospital differences in infection risk after adjustment for measured patient factors.
Measured hospital‑level fixed effects explained 39.9% (95% CI, 33.6–46.1%) of between‑hospital variability for VAP and 44.7% (95% CI, 35.0–54.5%) for CLABSI—substantial proportions attributable to institutional factors captured in the dataset.

Institutional factors associated with lower infection rates (adjusted analyses):

Broad prevention protocols (likely bundles and written policies) were associated with reduced rates of both VAP and CLABSI.
Hand hygiene training was associated with reductions in both VAP and CLABSI.
Flexible family visitation policies were associated with lower rates of both infections—an intriguing finding that may reflect multiple indirect mechanisms (staffing ratios, family engagement in care, patient mobilization, or cultural/contextual factors) and warrants further study.
Specific to CLABSI: better nursing staffing ratios, availability of single‑use gowns, and wall‑mounted or point‑of‑care alcohol (hand rub) were associated with lower CLABSI rates.
Specific to VAP: nurse‑led sedation titration, respiratory‑therapist‑led weaning protocols, and availability of an on‑site dentist were associated with lower VAP rates—each aligning with the biological plausibility of reducing ventilator exposure and oral colonization/inoculation.

Clinical and policy relevance:

The absolute rates observed—6.03 VAPs/1,000 MV‑days and 1.63 CLABSIs/1,000 CVC‑days—provide contemporary benchmarks for Brazilian adult ICUs operating within this multicenter platform. The large MRRs highlight that where a patient receives care substantially influences their HAI risk, beyond individual clinical severity. Importantly, nearly half of the between‑hospital variability was explained by measured, and therefore potentially modifiable, institutional characteristics and processes.

Expert commentary and interpretation

These findings provide pragmatic, actionable intelligence for ICU leaders and health system policymakers in LMICs. Several observations deserve emphasis:

Processes of care matter: The association of nurse‑led sedation protocols and respiratory therapist involvement with lower VAP supports the benefit of empowering bedside providers to act on weaning and sedation decisions—measures that reduce ventilator exposure time and aspiration risk.
Basic infection prevention fundamentals retain high value: Hand hygiene training and availability of alcohol‑based hand rubs, plus single‑use gowns, were tied to lower CLABSI rates. These are relatively low‑technology investments with high expected returns when widely implemented and sustained.
Staffing remains central: Nurse staffing ratios independently associated with CLABSI reduction underscores longstanding evidence that adequate nurse staffing improves surveillance, line care, and timely removal of devices.
Unexpected but plausible associations: Flexible family visitation correlated with lower infection rates—this may be a marker of a more patient‑centered, well‑resourced culture or reflect indirect benefits of family engagement. The availability of on‑site dentistry is biologically plausible for VAP reduction by controlling oral pathogens and maintaining oral hygiene in mechanically ventilated patients.

Limitations to keep in mind:

Observational design cannot establish causality. Residual confounding and unmeasured institutional characteristics (e.g., local microbiology, antimicrobial stewardship intensity, or real‑time adherence to bundles) may influence associations.
Surveillance definitions followed Brazilian regulatory guidance, which may differ from other surveillance systems; therefore, direct rate comparisons to NHSN or other registries should be cautious.
The study period (2019–2021) overlaps the COVID‑19 pandemic, a time of marked strain on ICUs, changing case‑mix, and altered infection prevention priorities; although adjusted models included patient‑level factors, pandemic effects may distort baseline rates and practices.
Cross‑sectional capture of hospital‑level variables at baseline cannot account for temporal changes in structure or process during the study period.

Implications for practice and policy

For ICU directors and health system leaders in LMICs, the actionable priorities that emerge are:

Invest in and standardize prevention protocols (device insertion and maintenance bundles) and ensure implementation fidelity through training and audit.
Prioritize hand hygiene infrastructure and regular staff training; ensure ready access to alcohol hand rub at the point of care.
Evaluate and improve nurse staffing where feasible—evidence here and elsewhere supports meaningful patient‑level benefits.
Empower bedside clinicians with nurse‑driven sedation protocols and protocols allocating respiratory therapist responsibility for weaning to shorten ventilator time safely.
Consider incorporating dental/oral care programs for ventilated patients as part of VAP prevention strategies.

Conclusion

The IMPACTO‑MR nested cohort across 50 Brazilian ICUs demonstrates that a large proportion of inter‑hospital variability in VAP and CLABSI rates is explained by measurable institutional characteristics and processes of care. Low‑technology, high‑impact measures—prevention protocols, hand hygiene training and supplies, improved nursing ratios, and role‑specific care protocols—are associated with reduced HAI burden and should be prioritized in LMIC ICU quality improvement agendas. Future research should prospectively evaluate implementation strategies, measure fidelity, and link process changes to patient‑centered outcomes in cluster randomized or stepped‑wedge designs to strengthen causal inference.

Funding and clinicaltrials.gov

For details on funding sources, sponsoring institutions, and any trial registration, refer to the original publication: Besen BAMP et al., Crit Care Med. 2025 Dec 1;53(12):e2540‑e2551 (doi: 10.1097/CCM.0000000000006881).

Reference

Besen BAMP, Dietrich C, Pinheiro CCG, Silva DP, de Mattos RR, Spadoni CCDS, Reis LFL, Nunes Neto PA, Paciência LEM, Caser EB, Ferreira Fernandes CC, Fernandes VF, Ferronato BR, Urbano HCA, Grion CMC, Medeiros EML, Silva ACD, Golin NA, Lima VP, Boschi E, Machado AS, Foernges RB, de Oliveira Junior LC, Silva EM, Pereira FC, Lisboa TC, Nassar AP Jr, Pereira AJ, Veiga VC, Arns B, Marsola G, Machado FR, Cavalcanti AB, Azevedo LCP, Tomazini BM; Impact of Infections by Antimicrobial-Resistant Microorganisms in Patients Admitted to Adult ICUs in Brazil: Platform of Projects to Support the National Action Plan for the Prevention and Control of Antimicrobial Resistance (IMPACTO-MR) investigators and Brazilian Research in Intensive Care Network (BRICNet). Institutional Risk Factors Associated With Healthcare-Associated Infections in Brazilian ICUs: A Nested Cohort Within the IMPACTO-MR Platform. Crit Care Med. 2025 Dec 1;53(12):e2540-e2551. doi: 10.1097/CCM.0000000000006881. Epub 2025 Sep 24. PMID: 40990605.