Introduction
Ventilator-capable skilled nursing facilities (vSNFs) have been increasingly associated with high rates of endemic and emerging multidrug-resistant organisms (MDROs), particularly Candidozyma auris. Reference McKinnell, Singh and Miller1,Reference Rossow, Ostrowsky and Adams2 Understanding whether MDRO prevalence differs among residents in ventilator-designated beds (vBeds) versus non-ventilator-designated beds (non-vBeds) within vSNFs may inform infection prevention strategies specifically for these units.
Methods
We conducted MDRO prevalence surveys in seven vSNF nursing homes with dedicated ventilator units in Orange County, California, from June 2021 to September 2024. At each assessment, 50 residents were randomly selected, and any refusal were replaced to maintain 50 sampled residents per facility survey. Facility nurses obtained swabs from the bilateral nares, skin (bilateral axilla/groin), and perirectal area using standardized protocols.
Skin and perirectal samples were cultured for Staphylococcus aureus (both methicillin-resistant [MRSA] and methicillin-susceptible [MSSA]), vancomycin-resistant Enterococci (VRE), extended-spectrum β-lactamase producers (ESBLs), carbapenem-resistant Acinetobacter baumannii (CRAB), carbapenem-resistant Enterobacterales (CRE), and C. auris. Nares samples were cultured for S. aureus and C. auris. Samples were processed within six hours of collection using previously described standard microbiologic methods. Reference McKinnell, Singh and Miller1 MDRO positivity was defined as growth of MRSA, VRE, ESBLs, CRE, CRAB, or C. auris from any body sample.
Demographic and clinical characteristics, including care type (postacute vs long-stay), lucidity (yes/no; defined as being alert and oriented to person, place and time), bed-bound status (yes/no), wounds, any medical device, MDRO history on admission (available retrospectively for all but C. auris), and current antibiotic use were collected from nursing home records and where applicable, confirmed visually at the time of sampling. Standard infection prevention activities, including Enhanced Barrier Precaution (EBP) for residents with wounds, invasive devices, or MDRO colonization, was applied to both vBed and non-vBed locations. This study and related data collection and resident sampling were approved as a Quality Assurance Performance Improvement activity at each vSNF, and data analysis was approved by the Institutional Review Board at the University of California, Irvine. Specimen collection was conducted as an operational project without written informed consent, with allowances for resident refusal.
We compared vBed vs non-vBed characteristics using Pearson’s χ2 tests or Fisher’s exact tests (when expected counts <5) and Mann-Whitney U tests. Multivariable generalized linear mixed-effects models with random intercepts for facility and resident were used to estimate adjusted odds ratios (aORs) and 95% CIs. Variables with collinearity were excluded from the multivariable models. Among clinically related variables, those reflecting overlapping aspects of resident-level dependency and acuity were also excluded to avoid redundant adjustment within the multivariable models.
To enhance interpretation of modeled odds ratios, predicted probabilities were also calculated at fixed age and sex (details in Supplementary Table 2). All statistical analyses were conducted using SAS/STAT 15.3 (SAS, Cary, NC).
Results
Across 13 prevalence surveys at seven vSNFs, we obtained 650 sampling episodes from 590 unique residents (some residents were sampled on multiple surveys). Of residents selected for sampling, 5.8% refused swabbing and were replaced. vBed residents were younger, less likely to be lucid/alert, and more likely to have medical devices and be incontinent of stool. A greater proportion of vBed residents had a history of each MDRO (Table 1).
Demographic and clinical characteristics of vSNF residents by bed type
Table 1. Long description
The table presents a comparison of demographic and clinical characteristics of vSNF residents by bed type, with data from 590 unique residents. It includes columns for total residents, vBeds residents, and Non-vBeds residents, along with P values indicating statistical significance. The table has 28 rows and 5 columns, with headers for age, gender, race, nursing home days of stay at swabbing, care type, lucid/alert status, bed bound status, current wound, devices, incontinence, body mass index, diabetes medication, current antibiotic use, contact precautions, and history of MDRO on admission. Notable trends include younger age, less likelihood of being lucid/alert, and higher prevalence of medical devices and incontinence among vBed residents. Additionally, a greater proportion of vBed residents had a history of each MDRO.
IQR, interquartile range; MDRO, multidrug-resistant organism; MRSA, methicillin-resistant Staphylococcus aureus; VRE, vancomycin-resistant Enterococci; ESBL, extended-spectrum β-lactamase-producing organisms; CRE, carbapenem-resistant Enterobacterales; CRAB, carbapenem-resistant Acinetobacter baumannii; C. auris, Candidozyma auris (formerly Candida auris); vBed, ventilator-designated bed; non-vBed, non-ventilator-designated bed. Data represent clinical and demographic characteristics at the first swab culture of each resident (N = 590 residents). Continuous variables (age, body mass index, length of stay) were compared using the Mann–Whitney U test. Categorical variables were compared using Pearson’s χ2 tests; when any expected cell count was <5, Fisher’s exact tests were used. All tests were two-sided.
a Non-missing counts were 589 for age, 381 for race, 589 for care type, 588 for bed bound, 486 for body mass index, and 453 for CRAB history. Percentages and P values are calculated using non-missing denominators. All other variables were based on 590 residents.
b Other race includes American Indian or Alaska Native, Native Hawaiian or Other Pacific Islander and Multiracial. Missing race is not shown as a separate category.
c History of C. auris on admission could not be reliably abstracted.
Percent MDRO colonization was significantly higher among vBed vs non-vBed residents (79.0% vs 68.2%; P = .004). ESBL (48.1% vs 28.2%, P < .001), CRE (9.5% vs 2.3%, P < .001), CRAB (13.3% vs 3.6%, P < .001), and C. auris (38.6% vs 15.2%, P < .001) colonization was more common in vBed residents, whereas MRSA (35.2% vs 45.5%, P = .01) and MSSA (6.2% vs 20.7%, P < .001) was more common in non-vBed residents (Supplementary Table 1).
Multivariable model results are shown in Figure 1 and Supplementary Table 3. vBed status was associated with higher odds of ESBL and C. auris colonization, but lower odds of MRSA and MSSA. Wounds were consistently associated with increased MDRO colonization, whereas being lucid/alert was protective for some organisms. Tracheostomy/ventilator use and gastrointestinal device use were excluded from the multivariable models because of collinearity with vBed status. Other clinically related status, including bed-bound status, urinary catheter use, and incontinence status, were not retained because they reflected overlapping aspects of resident dependency and acuity already represented by retained covariates.
Frequencies of organism-specific MDRO positivity and adjusted odds ratios for factors associated with MDRO colonization among vSNF nursing home residents. Unit of analysis is each resident point prevalence sampling episode (n = 650); an episode is organism-positive if any site is cultured positive. Adjusted odds ratios (aOR) are derived from multivariable generalized linear mixed models clustering for residents and facilities. Final models included age (per decade), sex, wounds, antibiotic use at the time of swabbing, alertness level (lucid/alert vs not), care type (postacute vs long-stay), ventilator designation (vBed vs non-vBed), and prior MDRO history for MRSA, VRE, and ESBL models. Note that history of C. auris on admission was not able to be reliably abstracted. Tracheostomy/ventilator use and gastrointestinal device use were excluded due to collinearity with vBed status. Forest plots are displayed on a base-10 logarithmic scale; the dotted vertical line indicates aOR = 1. Confidence intervals extending beyond the plotted range are truncated with an arrow. Bold values indicate statistically significant associations (p < 0.05). Boxes are used to emphasize the impact of vBed status, the primary exposure of interest. aOR, adjusted odds ratio; MRSA, methicillin-resistant Staphylococcus aureus; VRE, vancomycin-resistant Enterococci; ESBL, extended-spectrum β-lactamase-producing organisms; C. auris, Candidozyma auris.
Figure 1. Long description
The table presents data on the frequencies of organism-specific multidrug-resistant organism (MDRO) positivity and adjusted odds ratios (aOR) for factors associated with MDRO colonization among ventilator-capable skilled nursing facility (vSNF) residents. The unit of analysis is each resident point prevalence sampling case (n=650). A case is organism-positive if any site is cultured positive. Adjusted odds ratios are derived from multivariable generalized linear mixed models clustering for residents and facilities. The table includes data for MRSA, VRE, ESBL, and C. auris positivity. Key factors considered are age, sex, wounds, antibiotic use, alertness level, care type, ventilator designation, and prior MDRO history. Forest plots are displayed on a base-10 logarithmic scale, with the dotted vertical line indicating an aOR of 1. Confidence intervals extending beyond the plotted range are truncated with an arrow. Statistically significant associations (p>0.05) are highlighted in bold. Boxes emphasize the impact of ventilator-designated bed (vBed) status, the primary exposure of interest.
Predicted probabilities demonstrate substantial additive risks by bed type and resident characteristics (Supplementary Table 2). For example, for vBed residents who had wounds, were receiving antibiotics, were not lucid, and were not receiving postacute care, the predicted probability of any MDRO colonization was 93.5% and of C. auris colonization was 41.9%. In contrast, for residents in non-vBeds who had no wound, were not receiving antibiotics, were lucid, and were receiving postacute care, the predicted probability of any MDRO colonization was 59.8% and that of C. auris colonization was 9.2%.
Discussion
This study identified distinct MDRO colonization patterns between vBed and non-vBed residents within vSNFs nursing homes. vBed residents were significantly more likely to harbor Gram-negative MDROs (ESBL, CRE, CRAB) and C. auris, whereas non-vBed residents were more likely to harbor MRSA. Reference Feldman, Kassel and Cantrell3–Reference Donlan5 The greater predominance of Gram-negative MDRO may be explained by the shift of oropharyngeal flora from Gram-positive to Gram-negative organisms in mechanically-ventilated patients. Reference Feldman, Kassel and Cantrell3 The greater predominance of C. auris may be related to endotracheal tube biofilm formation. Reference Sherry, Ramage and Kean4,Reference Donlan5
This extensive MDRO carriage in vSNFs, and particularly vBed residents, occurred despite the stated use of EBP for residents with wounds, indwelling medical devices, or known MDRO colonization or infection. 6 Due to ventilation devices, all vBed residents would have required EBP. Thus, EBP and usual hand hygiene and environmental cleaning appear insufficient for preventing MDRO spread in vSNFs, particularly in vBed areas. Our data suggest that additional approaches are needed, which could include rigorous campaigns to enhance adherence with routine hand hygiene, environmental cleaning, and EBP. However, given the extensive MDRO burden, it is likely that expanded approaches are necessary for MDRO containment, such as decolonization Reference Miller, McKinnell and Singh7 and intensive antibiotic stewardship. 8
This study has several limitations. First, we did not measure adherence to hand hygiene, cleaning, or EBP. Second, because admission sampling did not occur, the admission history of an MDRO may not be a complete representation of preexisting MDRO status. Third, C. auris history at admission could not be reliably extracted and was not included in adjusted models. Fourth, small numbers of CRE and CRAB prevented adjusted analyses.
In conclusion, MDRO colonization burden was remarkably high in both vBed and non-vBed locations of vSNF nursing homes, with particularly high prevalence of Gram-negative MDROs and C. auris in vBeds. These findings support the need for additional prevention approaches beyond current EBP to reduce MDRO burden in this high-acuity nursing home population.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/ice.2026.10483.
Acknowledgements
Conflict of interest and financial disclosures: All authors completed the ICMJE disclosure form. Huang, Singh, Kleinman, Gussin, and Bittencourt report institutional research support from federal agencies. Huang and Singh report honoraria and travel support from the Kuwait Ministry of Health. Huang, Singh, and Kleinman report that antiseptic products were provided by Xttrium Laboratories for federally funded studies conducted at their institutions. Kim reports no conflicts of interest.
We thank Gabriel Gadia, BS, Justina Bui, BS, Jahan M. Hosseinian, BS, and Raheeb Saavedra, AS, for their assistance with data collection and study coordination.
Open access
