Development and validation of a psychometric tool for assessing the readiness and feasibility of CRE screening in LMICs: A cross-sectional instrument validation study

Abstract

Background & Objectives

Carbapenem-resistant Enterobacterales (CRE) pose a critical threat in Intensive Care Units (ICUs) within low- and middle-income countries (LMICs), with prevalence exceeding 20% and mortality reaching 50%. Despite high clinical demand, routine screening remains unstandardized in Vietnam due to resource constraints and a lack of validated tools. This study aimed to develop and validate a psychometric instrument to evaluate healthcare workers’ (HCWs) perspectives across three domains: barriers, feasibility and readiness.

Methods

A cross-sectional validation study was conducted using total population census sampling (n=97, total front-line staff) across three hospitals in Vietnam (Viet Tiep, Kien Thuy, and Thuy Nguyen). Following expert panel review (S-CVI/Avg = 0.89) and a 10.3% pilot test (n=10), construct validity and reliability were evaluated via Exploratory and Confirmatory Factor Analysis (EFA/CFA). Group differences were analyzed using Fisher’s Exact and Chi-square tests.

Results

The barriers subscale was excluded due to low internal consistency (alpha < 0.60), yielding a refined 16-item, two-factor instrument (6.1:1 observation-to-item ratio). The final tool demonstrated excellent reliability (alpha: 0.84–0.89). CFA confirmed a robust model fit for the Feasibility and Readiness structure [GFI/CFI = 0.95; RMSEA = 0.062]. HCWs reported high Readiness but moderate Feasibility. Significant occupational divergences emerged: 88.9% of nurses viewed sampling as simple (Item R4) compared to only 45.8% of laboratory technicians (p = 0.003). Furthermore, an experience-perception paradox was identified (p < 0.05); mid-career staff (5–10 years) reported the lowest feasibility scores (50.0% high score) compared to junior and senior cohorts.

Conclusion

The psychometric validation resulted in a methodologically sound 16-item tool tailored for LMICs. The divergence between clinical readiness and laboratory feasibility highlights critical implementation bottlenecks. Sustainable transition to routine CRE surveillance requires resource-optimized targeted workflows to reduce frontline workloads, alongside peer-led governance that leverages senior expertise to mitigate mid-career burnout.

1. Introduction

The global emergence of Carbapenem-resistant Enterobacterales (CRE) represents one of the most significant challenges to modern medicine, often described as a “triple threat” due to its multi-drug resistance, high transmissibility, and association with severe clinical outcomes. ¹ In Intensive Care Units (ICUs), CRE infections are particularly devastating, with reported mortality rates ranging from 40% to 50%. ²

To mitigate the silent spread of these pathogens, the WHO and the CDC emphasize Active Surveillance Culture (ASC). Operationally, ASC targets high-risk populations, including patients admitted directly to ICUs, those transferred from long-term care facilities, or individuals epidemiologically linked to known CRE-positive index cases. The protocol dictates that screening occurs immediately upon admission, followed by periodic (e.g, weekly) intervals if the initial screening is negative. The standard specimen types utilize rectal or perirectal swabs, as stool carriage provides the highest diagnostic yield. Upon sample collection, implementation procedures require preemptive contact precautions or patient isolation until laboratory confirmation via selective chromogenic media or molecular assays is finalized. It aims to identify asymptomatic carriers and prevent cross-transmission before outbreaks occur.^3,4

Beyond universal CDC-recommended culture surveillance, alternative screening strategies exist along a spectrum of cost and diagnostic speed. These include targeted clinical risk-score screening—which tests only patients with pre-specified risks like prior antibiotic exposure or recent hospitalization—and universal molecular screening using rapid Polymerase Chain Reaction (PCR) assays. ⁵ While molecular testing offers a rapid turnaround time of under 2 hours compared to traditional 24-to-48-hour cultures, its implementation costs are often prohibitive for resource-limited facilities. ⁶ Conversely, risk-score screening reduces laboratory overhead but routinely misses up to a third of asymptomatic carriers, highlighting the need for a localized evaluation of strategy feasibility before deployment. ⁵

Although implementing ASC requires upfront institutional investment, its clinical utility in preventing healthcare-associated infections (HAIs) and improving patient outcomes is well documented. Robust clinical trials have shown that active surveillance, paired with targeted isolation, can reduce the relative risk of hospital-acquired CRE acquisition by over 40% to 50%. ⁷ By uncovering the hidden reservoir of asymptomatic colonized carriers, ASC allows clinical teams to implement contact precautions early. This directly cuts down on cross-transmission events and subsequent lethal bloodstream infections, ultimately lowering the crude mortality rate associated with secondary multi-drug resistant sepsis in critical care environments. ⁸

Consequently, the routine deployment of ASC reflects a stark global divide. In many HICs, active surveillance is systematically integrated into national guidelines or mandated during regional outbreaks, such as Israel’s successful country-wide containment mandate. ⁸ Conversely, in LMICs, routine facility-wide ASC is exceptionally rare—frequently restricted to temporary, localized research protocols or completely absent due to the prohibitive recurrent costs of laboratory reagents, infrastructure limitations, and specialized personnel shortages.^4,9

Literature linking routine ASC to a reduced aggregate ICU length of stay operates through an indirect, herd-protection mechanism rather than a direct therapeutic effect on the colonized individual. Because a positive screening result often reflects asymptomatic gastrointestinal colonization rather than active infection, the screening itself does not alter the index patient’s immediate clinical trajectory. Instead, the immediate isolation of that carrier prevents the silent cross-transmission of CRE to vulnerable, uninfected patients nearby. ¹⁰ In critical care settings, preventing secondary healthcare-associated CRE infections—which independently prolong hospital stays by an average of 6 to 15 days per patient—dramatically minimizes the aggregate bed-day burden and lowers the average length of stay across the entire unit. ¹¹

While the baseline prevalence of CRE colonization or infection in ICUs within high-income countries (HICs) remains highly localized and generally suppressed between 1% to 5% (Tacconelli E et al., 2014)., settings In low- and middle-income countries (LMICs), including Vietnam, frequently experience hyper-endemic rates exceeding 20%. ³ This massive epidemiological disparity is driven by a combination of systemic vulnerabilities. In LMICs, severe resource constraints lead to suboptimal nurse-to-patient ratios, structural deficiencies in hand hygiene infrastructure, and limited isolation capabilities. This environment is further exacerbated by unstandardized or absent antimicrobial stewardship programs, which leads to widespread empirical overuse of broad-spectrum carbapenems, putting extreme selective pressure on hospital flora. ¹² Furthermore, the success of screening protocols relies heavily on the adherence and perceptions of front-line HCWs, who navigate the tension between clinical necessity and operational barriers daily. ¹³

While many studies have focused on the clinical and molecular epidemiology of CRE, there is a significant lack of standardized, validated instruments to assess the “readiness” and “feasibility” of screening from an organizational and behavioral perspective in LMICs. Previous international studies have identified several critical determinants influencing the success of CRE screening programs. Research on Readiness has shown that healthcare workers’ clinical knowledge and their perception of CRE as a high-priority threat are fundamental to policy adherence. ¹⁴ Regarding Barriers, existing literature frequently highlights high operational costs, laboratory turnaround times, and the psychological burden on patients under contact precautions as primary obstacles in resource-limited settings. ¹⁵ Conversely, the Benefits of routine screening—such as a significant reduction in HAI rates and shorter durations of intensive care—have been well-documented in high-income countries. ⁸ However, most of these studies utilized qualitative interviews or non-validated surveys, which fail to provide a quantifiable and reproducible measure of how these factors interact within the complex hierarchy of LMIC hospitals. There remains a distinct lack of an integrated tool that can simultaneously assess these dimensions through the lens of both professional readiness and practical feasibility.

While various institutional surveys attempt to capture healthcare workers’ perceptions of infection control, many lack rigorous psychometric evaluation. Psychometric validation is not merely a statistical exercise; it is a fundamental prerequisite for establishing construct validity and ensuring generalizability. ¹⁶ Without evaluating structural invariance and item-to-factor stability, a survey instrument risks measuring confounding psychological noise rather than the actual target constructs, such as readiness or feasibility. By purging unstable items and proving internal consistency, psychometric purification ensures that the instrument behaves predictably across different clinical cohorts.^16,17 This structural integrity allows administrators to safely generalize findings beyond the initial sample and design evidence-based interventions tailored to verified behavioral and logistical bottlenecks.

In summary, while HICs possess the financial cushion and laboratory infrastructure to sustain highly sensitive active surveillance networks, LMICs face a unique convergence of high CRE endemicity and severe operational restrictions. ³ Existing implementation frameworks designed in HICs cannot be seamlessly transposed onto LMIC systems without risking operational failure or severe waste of clinical resources. ¹⁸ A significant gap remains in our understanding of how front-line healthcare workers in low-resource settings perceive the structural feasibility and baseline readiness of these aggressive protocols. Developing and psychometrically validating a localized, robust assessment tool is a vital first step to bridge this gap, offering a data-driven path to design sustainable infection control interventions where they are needed most.

Therefore, to address this gap, the primary aim of this study is to develop and psychometrically validate a localized assessment tool to evaluate the HCW’s readiness and feasibility of CRE screening. By employing Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA), we aim to provide a statistically robust tool for measuring barriers, feasibility, and readiness. Furthermore, this research seeks to identify key professional factors—such as years of experience—that influence clinical adoption, thereby providing a strategic roadmap for hospital administrators to implement effective and sustainable CRE active surveillance in Vietnam and similar LMIC contexts.

2. Methods

2.1. Study design and setting

A cross-sectional, multi-center instrument validation study was conducted between 15^th October, 2025 and 15^th March 2026. This cross-sectional validation study was conducted across three distinct public hospitals in Haiphong, Vietnam, representing different levels of the regional healthcare hierarchy: Viet Tiep Hospital, Kien Thuy District Hospital, and Thuy Nguyen District Hospital. Viet Tiep Hospital is a tertiary-level, 1800-bed regional referral center with a dedicated 153-bed ICU, serving an approximate annual volume of 124 526 inpatient admissions. In contrast, Kien Thuy and Thuy Nguyen hospitals are secondary-level district facilities. Kien Thuy Hospital operates at a 164 bed capacity with a 18 bed general critical care unit (annual admission volume of 6750), while Thuy Nguyen Hospital maintains 640 beds with a 23-bed critical care unit (annual admission volume of 23277). All targeted departments frequently manage patients at high risk for multi-drug resistant infections but lack structural budgets for routine molecular surveillance.

Ethical approval was obtained from the IRB of Viet Tiep Hospital (Ref No: 09/GCN-HĐĐĐ) on 07th October 2025. Following this official ethical clearance, written informed consent was acquired from all participants, and data recruitment commenced subsequently on 15th October, 2025.

2.2. Participants and sampling

The study population comprised front-line healthcare workers (HCWs)—specifically doctors and nurses—working in Intensive Care Units (ICUs) and infectious disease departments. We determined the minimum required sample size based on guidelines for multivariate research and EFA. According to Hair, a minimum observation-to-variable ratio of 5:1 is required, with a preferred ratio of 10:1. ¹⁹ Given that our survey instrument consists of 21 items, a sample size of 105 participants was the targeted minimum to ensure sufficient statistical power and the stability of the factor structure. However, as the total population of eligible front-line HCWs across the three study hospitals was 97, we employed a total population sampling (census) method, recruiting all eligible staff to maximize the sample size and statistical reliability.

The study population comprised front-line HCWs—specifically physicians, nurses, and laboratory technicians—working in ICUs and infectious disease departments across the three participating medical centers. To ensure sample homogeneity, strict inclusion criteria dictated that eligible participants must: (1) be full-time registered clinical or laboratory staff directly assigned to critical care or infectious disease divisions, and (2) possess a minimum of one year of active clinical practice within the facility. Exclusion criteria applied to: (1) temporary personnel, rotating interns, or student trainees, and (2) staff on extended medical or administrative leave during the five-month data collection window.

The initial 21-item questionnaire pool was systematically developed by a core multidisciplinary research team consisting of three infectious disease physicians, two infection control nurse specialists, and an epidemiologist proficient in healthcare-associated infection (HAI) behavioral metrics. The item generation process began with an exhaustive review of global literature regarding carbapenem-resistant organism containment, paired with the structural domains of the Consolidated Framework for Implementation Research (CFIR). Items were specifically mapped onto three intended conceptual domains: Barriers (5 items), Feasibility (9 items), and Readiness (7 items), phrased specifically to fit the operational realities of low-resource clinical environments.

To ensure a standardized frame of reference during data collection, all participants were explicitly instructed via a standardized introductory script to evaluate the readiness and feasibility parameters under a specific, proposed Active Surveillance Culture (ASC) protocol. This assumed screening framework was modeled directly on CDC guidelines: (1) Target Population: mandatory universal screening for all patients admitted directly to the ICU; (2) Timing: an initial sample collected immediately within 12 hours of admission, followed by periodic weekly intervals for patients whose baseline samples were negative; and (3) Workflow: preemptive contact precautions maintained until laboratory clearance.

2.3. Instrument development

The development of the CRE Screening - Barriers Feasibility, and Readiness Assessment Tool followed a three-phase process:

• 1. Item generation: An initial pool of items was generated based on a literature review. This included three theoretical domains: Barriers (B), Feasibility (F), and Readiness (R). (See Appendix 1 for detail contents).

• The development of the CRE Screening – Barriers, Feasibility, and Readiness Assessment Tool followed a three-phase methodology:Item Generation: An initial pool of 21 items was generated following an extensive review of contemporary literature and mapped onto the structural domains of the Consolidated Framework for Implementation Research (CFIR), spanning three intended theoretical domains: Barriers (5 items), Feasibility (9 items), and Readiness (7 items). Content Validity: To eliminate design bias and establish rigorous content validity, an independent panel of five external experts (completely unaffiliated with the core questionnaire design group) evaluated the initial item pool. The panel comprised two external senior professors of clinical microbiology, a ministry-level public health policy advisor, and two independent hospital epidemiology directors from distinct provinces. Using a 4-point ordinal relevance scale, Item-Level Content Validity Index (I-CVI) and Scale-Level Content Validity Index (S-CVI/Avg) were calculated. Items with I-CVI ≥0.78 were retained. The full 21-item matrix demonstrated an S-CVI/Avg of 0.89 (Table 1), indicating strong initial content validity prior to regional rollout. Pilot Testing & Instrument Validation: Because this psychometric instrument was newly developed to address the structural absence of validated surveillance tools in LMICs, field-level pilot testing was conducted. To eliminate item bias, content validity was formally evaluated by an independent panel of five external experts completely unaffiliated with the core questionnaire design group. The panel evaluated the items using a 4-point ordinal relevance scale, establishing an acceptable Scale-level Content Validity Index (S-CVI/Avg) of 0.89, confirming strong expert validation prior to field deployment. Following panel approval, a field-level pilot test was conducted within a sub-cohort comprising 10.3% of the total target population size (n=10 HCWs). This pilot phase focused exclusively on assessing linguistic clarity, readability, and item comprehension. Minor syntactic modifications were integrated based on participant feedback to ensure unambiguous field interpretation before initiating the final cross-sectional validation rollout (n=97).

• 2. Content Validity: To eliminate design bias and establish rigorous content validity, an independent panel of five external experts was convened to calculate Item-Level and Scale-Level Content Validity Indices (I-CVI and S-CVI). Crucially, these five evaluators were entirely independent of, and unaffiliated with, the core questionnaire development team. The panel comprised two external senior professors of clinical microbiology, a ministry-level public health policy advisor, and two independent hospital epidemiology directors from distinct provinces. Using a 4-point ordinal relevance scale, items yielding an I-CVI ≥ 0.80 were retained, ensuring that the full 21-item matrix possessed strong initial panel validity before regional rollout. The cut-off of I-CVI and Index (S-CVI/total items) according to Polit & Beck as following ²⁰ :

  • • I-CVI ≥ 0.78: Good content validity
  • • I-CVI < 0.78: Necessary to consider revising or removing the item
  • • S-CVI/total items ≥ 0.80: Instrument with acceptable content validity

Table 1.

Item and scale-level content validity index.

Factors 21	Item	Number of experts scoring ≥ 3	Total experts	I-CVI	S-CVI/Avg 5
Barriers	B1	4	5	0.80	0.89
	B2	4		0.80
	B3	4		0.80
	B4	4		0.80
	B5	4		0.80
Feasibility	F1	5		1.00
	F2	5		1.00
	F3	5		1.00
	F4	5		1.00
	F5	4		0.80
	F6	4		0.80
	F7	5		1.00
	F8	4		0.80
	F9	5		1.00
Readiness	R1	4		0.80
	R2	5		1.00
	R3	4		0.80
	R4	5		1.00
	R5	4		0.80
	R6	4		0.80
	R7	5		1.00

²⁰World Health Organization. Guidelines for the prevention and control of carbapenem-resistant Enterobacteriaceae, Acinetobacter baumannii and Pseudomonas aeruginosa in health care facilities. Geneva: World Health Organization; 2017.

⁵Polit DF, Beck CT. The content validity index: are you sure you know what’s being reported? Critique and recommendations. Res Nurs Health. 2006;29(5):489–497. doi: 10.1002/nur.20148.

S-CVI/total items ≥ 0.90: Instrument with excellent content validity3. Instrument validation: The structural design and validation of the CRE Screening Assessment Tool progressed through a systematic, multi-phase methodology. Because this specific psychometric instrument was newly developed by our research team to address the structural absence of validated surveillance tools in LMICs, rigorous validation steps were mandatory. In the first phase, a pool of 21 items was generated following an extensive review of contemporary literature and mapped onto the foundational constructs of the CFIR, spanning three intended theoretical domains: Barriers (5 items), Feasibility (9 items), and Readiness (7 items).

2.4. Data collection

Data were collected using a structured self-administered questionnaire divided into two sections:

• • Demographics: Clinical role, and years of professional experience (<5, 5–10, >10 years).

• • Psychometric Scales: Items measuring, Factor B, F and R, scored on a 5-point Likert scale ranging from 1 (Strongly Disagree) to 5 (Strongly Agree).

2.5. Statistical analysis

Data were analyzed using SPSS 26.0 and AMOS 24.0.

The validation process involved:

• • Reliability Analysis: Internal consistency was evaluated using Cronbach’s alpha. Items or factors with Cronbach’s alpha < 0.60 were considered for removal to ensure the instrument’s robustness.

• • Construct Validity: EFA was used to identify the underlying factor structure.

• • CFA was used to verify the final multi-factor model. Model fit indices included: chi-square/df < 3, GFI > 0.9, CFI > 0.9, and RMSEA < 0.08.

For determining potential associated factors with chi-square, the mean scores of the responses were interpreted using the neutral midpoint of 3.0 as the cut-off point, where scores above 3.0 represent agreement and scores below 3.0 represent disagreement. This interpretation is consistent with established measurement practices in Likert-scale analysis (Allen & Seaman, 2007; Cohen et al., 2018). ²²

3. Results

3.1. Demographic characteristics of study cohort

A total of 97 healthcare workers completed the survey. The majority of participants were nurses (45, 46.4%), followed by doctors (28, 28.9%), and technicians (24, 24.7%). In terms of professional experience, (12, 12.4%) had less than 5 years, (38, 39.2%) had 5–10 years, and (47, 48.4%) had over 10 years of working.

3.2. Content validity of instrument items

The content validity of the 21-item instrument was evaluated by five experts. The Item-level Content Validity Index (I-CVI) for all items ranged from 0.8 to 1.0. Items with an I-CVI score ≥ 0.78 were considered acceptable. All 21 items met the inclusion criteria, resulting in a Scale-level Content Validity Index (S-CVI/total items) of 0.89, indicating acceptable content validity of the survey instrument.

3.3. Reliability and factor elimination

The initial reliability analysis evaluated the internal consistency of three theoretical constructs: Factor B, F and R. Results indicated that Factors F and R exhibited strong internal consistency, with Cronbach’s alpha coefficients of 0.92 and 0.85, respectively. In contrast, Factor B demonstrated suboptimal psychometric properties, characterized by a Cronbach’s alpha below 0.60 and inadequate item-total correlations. Consequently, Factor B was excluded from subsequent analyses to maintain the overall psychometric integrity of the instrument. And the second reliability analysis was performed on two factors with Cronbach’s alpha coefficients of 0.97 – High reliability (See Table 2).

Table 2.

Cronbach’s Alpha of Questionnaire by 2 factors.

Code	Items of questionnaire	Cronbach’ Alpha of each factor	Cronbach’ Alpha of questionnaire
F1	Bedside CRE screening can be routinely implemented in the ICU	0.93	0.97
F2	The benefits of CRE screening outweigh the costs and effort involved
F3	CRE screening should be integrated into the ICU admission protocol
F4	With clear CRE screening procedures and training, the department can implement it effectively
F5	CRE screening should be maintained long-term rather than just implemented for research purposes
R1	Bedside CRE screening allows for early detection of patients carrying CRE	0.95
R2	CRE Screening results facilitate early and timely isolation decisions
R3	CRE screening helps reduce the risk of CRE cross-contamination in the ICU
R4	The bedside sampling procedure is relatively simple and easy to perform
R5	CRE screening does not significantly disrupt care and treatment activities
R6	CRE screening helps guide physicians toward more rational antibiotic use
R7	CRE screening contributes to improving the effectiveness of infection control in the department

3.4. Construct validity: Confirmatory factor analysis (CFA)

Consequently, during further Structural Equation Modeling (SME), factor B was excluded from the final model due to low reliability and poor reliability. Testing of the remaining CRE screening feasibility and readiness constructs showed the chi-square/df = 1.76, an RMSEA value of 0.062 and CFI value of 0.958. While these indices mathematically align with standard acceptable fit thresholds, they must be interpreted with strict caution as they represent an initial model fit derived from the same cohort used for exploratory analysis, which inherently optimizes fit characteristics. (Figure 1).

Figure 1.

Measurement model for the dimension of CRE screening implementation.

3.5. Factors influencing perceptions

The CRE implementation assessment demonstrated a clear disparity between professional readiness and feasibility. While HCWs exhibited high levels of readiness for CRE screening (Factor R; Mean: 3.5–3.7), indicating strong alignment with clinical standards, the perceived feasibility of the intervention remained significantly lower (Factor F; Mean: 3.0–3.2).

The cut-off 3.0 is used for Chi-square tests for 2 independent variables (jobs and years of working) to readiness and feasibility. The result only revealed that years of professional experience significantly influenced attitudes toward CRE screening readiness (p=0.02, calculated via Fisher’s Exact Test to adjust for the small n=12 junior subgroup) (See Table 3).

Table 3.

Associated factors to HCW’s perception on CRE implementation (n=97).

Characteristic factors		Low score		High score		p-value
		n	%	n	%
Feasibility
Job role	Doctor	11	39.3	17	60.7	0.50
	Nurse	16	35.6	29	64.4
	Technician	12	50.0	12	50.0
Years of working	< 5	3	25.0	9	75.0	0.22 a
	5 - 10	19	50.0	19	50.0
	> 10	17	35.2	30	63.8
Readiness
Job role	Doctor	8	28.6	20	71.4	0.10
	Nurse	14	31.1	31	68.9
	Technician	13	54.2	11	45.8
Years of working	< 5	3	25.0	9	75.0	0.02* a
	5 - 10	20	52.6	18	47.4
	>10	12	25.5	35	74.5

(*) p<0.05 in bold for statistical significant;

^aNote. Fisher’s Exact Test was applied to the “Years of working” analyses because the small sample size of the junior subgroup (n=12) resulted in expected cell frequencies below 5, violating standard Pearson Chi-Square test assumptions.

To better understand implementation bottlenecks, a granular item-level comparison was performed across professional roles (Doctors, Nurses, and Technicians) for key operational indicators. Statistically significant differences emerged regarding the perceived simplicity of execution. For item R4 (‘The bedside sampling procedure is relatively simple and easy to perform’), agreement was highest among nursing staff (88.9%) and physicians (71.4%), who directly manage or execute clinical swabbing at the bedside. Conversely, laboratory technicians demonstrated a significantly lower agreement rate of only 45.8% (p = 0.003, Fisher’s Exact Test). This indicates that while clinical staff view sample acquisition as non-disruptive, laboratory personnel perceive the expansion of universal surveillance as a substantial operational burden on laboratory processing capacity. Similarly, for item F1 (‘Routine implementation is viable’), technicians expressed the lowest overall feasibility scores (50.0%) compared to nurses (64.4%) and doctors (60.7%) (See Table 4).

Table 4.

Occupational comparison of healthcare workers’ perceptions on key operational indicators (n=97).

Professional category	Item R4: Low/Neutral agreement (≤3.0)	Item R4: High agreement (>3.0)	p-value
Physicians (n=28)	8 (28.6%)	20 (71.4%)	0.003*
Nurses (n=45)	5 (11.1%)	40 (88.9%)
Technicians (n=24)	13 (54.2%)	11 (45.8%)

(*) p<0.05

4. Discussion

4.1. Partial validation of the CRE implementation instrument

The initial objective of this study was to develop a comprehensive three-factor psychometric tool. However, this aim was only partially realized due to the necessary statistical exclusion of Factor B (Barriers) during psychometric cleansing. The necessity to psychometrically exclude the ‘Barriers’ subscale during the exploratory and confirmatory factor analysis is an instructive finding with distinct structural implications. Statistically, the low internal consistency (alpha < 0.60) signifies that the items designed to capture clinical barriers did not form a distinct or stable psychological construct in our sampled cohort. In an accessible sense, this implies that front-line healthcare workers do not conceptualize operational bottlenecks (such as workload or lab turnaround times) as a standalone theoretical domain; instead, they naturally blend these obstacles into their overall assessment of a protocol’s practical Feasibility. This aligns with the psychometric principles suggested by Boateng, where items with low reliability often reflect constructs that are culturally or contextually irrelevant. ¹⁶ In the Vietnamese clinical context, “barriers” may not be perceived as a standalone psychological construct, practical obstacles may have been statistically conflated with or captured within the “Feasibility” items by the respondents. This suggests that HCWs evaluate obstacles based on their perceived self-efficacy to overcome them, a phenomenon observed in other health behavior studies in Asia. ²³ There are clear advantages to this exclusion: it structurally simplifies the instrument from a complex three-factor model to a highly focused, rapid-to-administer two-factor tool (Readiness and Feasibility). This leaner 16-item architecture significantly reduces survey fatigue among busy ICU clinicians, increasing long-term compliance when tracking hospital sentiment. However, a notable limitation is the loss of granular diagnostic power. By absorbing specific obstacles into the broader feasibility domain, the refined tool cannot isolate individual structural barriers (such as specific patient psychological resistance or material supply chain failures). It acts as a reliable macro-level gauge of implementation viability rather than a fine-grained troubleshooter of specific workflow hurdles. This study did not collect concurrent qualitative data to explore the participants’ cognitive processing of these items, the exact behavioral mechanism behind this low reliability remains unverified. Future revisions of the tool must employ mixed-methods designs, incorporating cognitive interviews, to successfully reconstruct and isolate the operational barrier construct in LMIC settings.

4.2. The readiness-feasibility gap

We observed a marked discrepancy between high Readiness (3.5–3.7) and moderate Feasibility (3.0–3.2). This “implementation gap” is a common theme in infection control literature. Our findings echo Lin. Y, where HCWs demonstrated strong theoretical knowledge but faced “paralysis” during execution due to systemic constraints. ¹⁴ Furthermore, the moderate feasibility scores reflect the concerns raised by Dickson regarding the hidden costs of screening, such as increased workload and the need for specialized isolation rooms, which are often at a premium in overcrowded LMIC ICUs.^15,24

This disparity suggests that despite a clear professional mandate for implementation, significant concerns regarding resource allocation and workload burden act as critical barriers to the successful adoption of these screening protocols.

Our findings reveal a distinct implementation mismatch: while HCWs exhibit high professional Readiness and ideological willingness to adopt CRE screening, they are constrained by low practical Feasibility. To bridge this gap within resource-constrained LMIC frameworks, healthcare administrators must transition away from high-overhead universal screening models toward resource-optimized, targeted surveillance workflows. First, institutions should implement automated, electronic risk-score algorithms built directly into the admission triage sheets. By screening only high-risk subsets—such as patients transferred from long-term facilities, those with exposure to carbapenems within the past 90 days, or individuals requiring mechanical ventilation—hospitals can capture up to 70–80% of carriers while reducing total laboratory reagent volume and associated costs by more than half. Second, to alleviate the workflow disruption reported by clinical staff, hospitals should adopt batched testing schedules or train a dedicated, mobile infection-control phlebotomy/swabbing cadre. This protects front-line ICU nurses from compounding task burdens, transforming a chaotic bedside disruption into a predictable, standardized daily protocol.

4.3. The experience-perception paradox (The 5–10 year of working group)

The most striking finding was the “U-shaped” trend in perceptions, with mid-career staff (5–10 years) reporting the lowest feasibility scores with the potential reasons as: (1) Workload Sensitivity: Research by Zimmerman suggests that mid-career nurses often bear the heaviest clinical load. ¹² Unlike junior staff who are highly supervised or senior staff in management, this group is responsible for the manual execution of screening (swabbing, labeling, transporting). Their skepticism likely stems from “policy fatigue,” where new protocols are viewed as additional burdens rather than clinical improvements ²⁵ ; (2). Expertise vs. Cynicism: While Schwaber emphasized the benefits of national mandates, our data suggests that mandates alone may not suffice for mid-career HCWs. ⁸ Studies in organizational psychology indicate that employees with 5–10 years of experience often transition from “idealistic beginners” to “realistic practitioners,” making them more critical of top-down policies that lack clear resource allocation ²⁶ ; (3) Seniority and Big-Picture Awareness: The high support from staff with >10 years of experience aligns with findings that senior HCWs possess higher “situational awareness” and are more likely to prioritize long-term hospital safety over short-term workload increases. ²⁷

The ‘experience–perception paradox’ identified in Table 3 reveals that mid-career staff (5–10 years of experience) express the lowest overall perception of screening feasibility (50.0%) compared to their junior and senior peers. This group represents the operational core of the ICU—they take on the heaviest shift loads, manage the most critically ill patients, and directly absorb the logistical frictions of new mandates. Consequently, their lower feasibility scores reflect acute operational fatigue rather than a lack of knowledge.

To resolve this behavioral bottleneck, interventions must move away from generic didactic lectures—which senior staff already understand and junior staff passively accept—and focus on peer-led governance and workflow co-design. Administrators should appoint influential mid-career clinicians as ‘Infection Control Champions,’ actively compensating or reducing their clinical shift load in exchange for leading the screening rollout. Involving this critical cohort in the co-creation of the daily swabbing schedule ensures that procedures do not conflict with heavy morning medication rounds, turning an imposed top-down policy into a bottom-up, sustainable clinical routine.

4.4. Limitations of the study

Despite the study’s robust psychometric validation, several limitations should be noted. First, a notable constraint of this study is the absolute sample size (n=97). While we used a total population census method across three target hospitals to capture all available front-line staff, this sample size initially fell below the preferred 5:1 ratio (n=105) required for a 21-item instrument. In psychometric testing, inadequate sample sizes can lead to unstable factor loadings or artificial cross-loadings. However, during the validation process, the instrument was optimized by dropping the 4 items of the “Barriers” factor due to suboptimal reliability. This resulted in a final, refined questionnaire containing 16 items. For these 16 items, our final sample provides an observation-to-variable ratio of 6.1:1, which fulfills the strict minimum methodological requirement (≥ 5:1) established by Hair. ¹⁹ This ratio, paired with superior empirical model fit indices (CFI = 0.958, RMSEA = 0.062), strongly supports the structural stability of the tool. Nevertheless, the small absolute sample size remains a major limitation, and care should be taken when generalizing these structural dynamics to larger or highly distinct LMIC healthcare systems. Second, a major methodological limitation is that both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) were performed sequentially on the same sample of 97 participants. Due to the census design, splitting the sample was not structurally viable. This approach likely introduces capitalization on chance and artificially inflates the model fit indices (CFI = 0.958, RMSEA = 0.062), meaning these parameters cannot be viewed as an independent verification of the scale’s stability. Consequently, the psychometric properties reported here must be treated as a preliminary framework, and future external validation using a distinct, larger sample is mandatory to verify true structural invariance. Third, the exclusion of the “Barriers” factor due to low internal consistency indicates that the current items may not fully capture operational obstacles in the Vietnamese context, warranting future qualitative refinement. Additionally, our reliance on self-reported data introduces potential social desirability bias, as findings represent subjective perceptions rather than objective clinical realities. Finally, the cross-sectional design provides only a single-point snapshot; therefore, future longitudinal research is essential to evaluate how these perceptions evolve following the implementation of a screening program.

5. Conclusion

This study partially achieved its objective by developing and validating a reliable 16-item with two-factor psychometric tool (comprising readiness and feasibility) rather than the intended three-factor framework. Our findings reveal a significant gap between the high professional readiness of healthcare workers and the moderate practical feasibility of screening protocols. Notably, the “mid-career” group (5–10 years of experience) emerged as the most skeptical demographic, highlighting a critical target for hospital administrators.

Recommendations: CRE containment requires transitioning from generalized guidelines to localized, highly specific operational strategies. First, to mitigate the acute workload burden on front-line healthcare workers, institutions must abandon un-targeted universal screening in favor of electronic, automated risk-score algorithms integrated directly into admission triage sheets. By restricting mandatory swabbing to a high-risk subset (e.g., patients with recent broad-spectrum antibiotic exposure or transfers from long-term care), hospitals can dramatically compress total laboratory processing volume while scheduling bedside collection during batched, non-urgent care windows to prevent nursing workflow fragmentation. Second, the extensive clinical expertise of senior healthcare professionals must be systematically leveraged by appointing them as institutional ‘Infection Control Champions’ within formal hospital steering committees. Rather than burdening them with repetitive bedside swabbing, senior staff should lead peer-mentoring networks and clinical case reviews, translating their high baseline readiness into active, bottom-up clinical governance that guides junior staff. Finally, fostering a supportive organizational culture requires concrete administrative structural changes. Hospital management must institute a transparent, non-punitive ‘no-blame’ compliance auditing system for isolation errors, shifting the workplace culture from fear of administrative penalties to collaborative problem-solving. This should be paired with mandatory monthly, inter-departmental safety huddles between clinical ICU wards and microbiology laboratories, ensuring that resource constraints are openly evaluated and that executive leadership visibly backs the material procurement needs of front-line staff.

Appendix. Barriers, Feasibility and Readiness Questionnaire

Item	Content	Linkert’s scale
		1 (less) → 5 (more)
		1	2	3	4	5
B1	CRE screening Increases workload for healthcare staff	​	​	​	​	​
B2	Shortage of personnel to perform routine CRE screening	​	​	​	​	​
B3	Time constraints during the care of critically ill patients	​	​	​	​	​
B4	Difficulties in coordination between clinical and microbiology departments for CRE test	​	​	​	​	​
B5	CRE screening increases testing costs	​	​	​	​	​
B6	There is a notable absence of standardized and unified operational guidelines for screening.	​	​	​	​	​
B7	Healthcare workers have received inadequate training regarding CRE screening procedures.	​	​	​	​	​
B8	Clinical teams face suboptimal cooperation from patients or family members during specimen collection.	​	​	​	​	​
B9	Turnaround times are delayed, meaning results are not returned as rapidly as expected.	​	​	​	​	​
F1	Bedside CRE screening can be routinely implemented in the ICU	​	​	​	​	​
F2	The benefits of CRE screening outweigh the costs and effort involved	​	​	​	​	​
F3	CRE screening should be integrated into the ICU admission protocol	​	​	​	​	​
F4	With clear CRE screening procedures and training, the department can implement it effectively	​	​	​	​	​
F5	CRE screening should be maintained long-term rather than just implemented for research purposes	​	​	​	​	​
R1	Bedside CRE screening allows for early detection of patients carrying CRE	​	​	​	​	​
R2	CRE Screening results facilitate early and timely isolation decisions	​	​	​	​	​
R3	CRE screening helps reduce the risk of CRE cross-contamination in the ICU	​	​	​	​	​
R4	The bedside sampling procedure is relatively simple and easy to perform	​	​	​	​	​
R5	CRE screening does not significantly disrupt care and treatment activities	​	​	​	​	​
R6	CRE screening helps guide physicians toward more rational antibiotic use	​	​	​	​	​
R7	CRE screening contributes to improving the effectiveness of infection control in the department	​	​	​	​	​

Note. Item R1 specifically evaluates the early detection of asymptomatic CRE ‘carriers’ rather than active clinical infections. In critical care environments, patients with active CRE infections are routinely identified through standard clinical diagnostic cultures triggered by systemic symptoms (such as fever or leukocytosis). In contrast, active surveillance screening is uniquely designed to reveal the otherwise invisible reservoir of asymptomatic gastrointestinal colonization (carriage). Distinguishing asymptomatic carriers is the core mechanism of epidemiological containment, allowing staff to implement preemptive isolation before horizontal transmission occurs.

Item R4 (‘The bedside sampling procedure is relatively simple and easy to perform’), participants were explicitly instructed to base their perceptions on the routine collection of rectal or perirectal swabs. This specimen type was selected as the operational baseline because rectal swabbing is the universally accepted standard for detecting carbapenemase-producing Enterobacterales gastrointestinal carriage, requiring minimal technical training compared to invasive clinical sampling but requiring specific logistical coordination at the ICU bedside.

ORCID iDs

Ngo Anh The https://orcid.org/0000-0002-8010-4031

Nguyen Thi Thu Huyen https://orcid.org/0009-0001-3726-6772

Nguyen Duc Thanh https://orcid.org/0000-0002-9872-9611

Tran Chi Thanh https://orcid.org/0000-0001-6582-4471

Tran Thi Lien https://orcid.org/0009-0003-3152-2848

Ethical considerations

Ethical approval was obtained from IRB of Viet Tiep Hospital (Ref No: 09/GCN-HĐĐĐ) on 07th October 2025 and written informed consent was acquired from all participants.