A cross-sectional study on knowledge, attitudes, and practices of occupational protection among central sterile supply department nurses: current status, associated factors, and an exploratory path analysis

Abstract

Background

This study aimed to investigate knowledge, attitudes, and practices (KAP) regarding occupational protection among Central Sterile Supply Department nurses.

Methods

A cross-sectional online survey was conducted among a convenience sample of Central Sterile Supply Department nurses recruited through a province-level CSSD professional network in Zhejiang Province, China, between 4 January and 1 March 2026. A self-designed questionnaire with acceptable reliability (Cronbach’s α = 0.725) was used to assess demographic characteristics and the three KAP dimensions.

Results

A total of 306 participants were included; most were female (95.8%), aged > 45 years (41.5%), and had a bachelor’s degree or above (85.3%). Mean knowledge, attitude, and practice scores were 14.77 ± 1.73 (possible range: 0–18), 37.06 ± 2.87 (possible range: 8–40), and 45.90 ± 3.52 (possible range: 10–50), respectively. Path analysis revealed that attitudes showed the strongest direct association with practices (β = 0.542, P = 0.014). In multivariable analysis, higher attitude scores (β = 0.667, 95% CI: 0.549–0.784, P < 0.001) and ≥ 3 training sessions/year (β = 0.930, 95% CI: 0.062–1.798, P = 0.036) were associated with better practices, whereas 11–20 years (β=−1.249, P = 0.049) and > 30 years (β=−1.773, P = 0.009) of nursing experience were associated with lower practice scores, suggesting experience-related variations rather than a direct negative impact. Attitudes correlated strongly with practices (r = 0.584, P < 0.001), whereas knowledge showed no significant correlation with attitudes (r = 0.088, P = 0.125).

Conclusions

In this convenience sample, CSSD nurses in Zhejiang Province showed favourable self-reported occupational protection attitudes and practices, but exhibited substantial and critical item-level deficits in knowledge regarding bloodborne exposure management and post-exposure prophylaxis timing. Implementing a mandatory, high-frequency (≥ 3 sessions per year), scenario-based occupational protection training programme linked to annual performance evaluation, alongside targeted modules for senior nurses, is recommended to improve protective practices.

Clinical trial number

Not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12912-026-05006-y.

Background

Healthcare-associated infections (HAIs) and occupational exposures among healthcare workers (HCWs) remain significant global health challenges. It is estimated that approximately 7% of hospitalized patients in high-income countries and up to 15% in low- and middle-income countries acquire HAIs, contributing substantially to morbidity, mortality, and healthcare costs [1]. Meanwhile, healthcare workers are frequently exposed to occupational hazards, with over 2 million exposures occurring annually worldwide, leading to the transmission of bloodborne pathogens such as hepatitis B virus, hepatitis C virus, and human immunodeficiency virus [2]. In the context of infection prevention and occupational safety, the Central Sterile Supply Department (CSSD) plays a critical role in hospital infection control, as it is responsible for the cleaning, disinfection, sterilization, and distribution of medical instruments across clinical departments [3]. Unlike the practice in many western healthcare systems where these tasks are delegated to specialized technicians, in China, registered nurses who have completed standardized undergraduate or junior college nursing curricula are routinely stationed in the CSSD. As an integral part of their professional nursing role, these individuals directly perform and oversee the entire processing lifecycle, including the manual and automated cleaning, decontamination, packaging, high-temperature or chemical sterilization, and sterile distribution of reusable medical devices. Due to the nature of their work, CSSD nurses are routinely exposed to multiple occupational risks, including sharps injuries, chemical exposure from disinfectants, and various physical hazards associated with high-temperature equipment and complex working environments [3, 4]. These cumulative hazards not only threaten the health and safety of CSSD personnel but may also adversely affect infection prevention and control outcomes, as occupational exposures and adherence to preventive practices are closely associated with healthcare-associated infections [5].

The Knowledge, Attitudes, and Practices (KAP) framework is widely used to assess and explain health-related behaviors [6]. It suggests that knowledge forms the basis of behavior change, while attitudes and beliefs act as key factors influencing practice, and knowledge alone does not necessarily lead to behavioral change [7]. In infection control and occupational safety, KAP-based studies have been applied to assess healthcare workers’ awareness, perceptions, and adherence to protective measures [8, 9]. Previous studies have demonstrated that higher levels of knowledge and more positive attitudes are associated with better compliance with infection prevention practices [10]. However, existing research has largely focused on general infection control behaviors, hand hygiene, or specific technical procedures rather than occupational protection as a comprehensive system [11, 12]. Studies specifically targeting CSSD nurses remain limited, particularly in China, despite their exposure to complex and multifactorial occupational hazards. Furthermore, most studies tend to examine knowledge, attitudes, and practice independently, with limited exploration of their interrelationships and underlying mechanisms. As a result, the pathways linking knowledge, attitudes, and practice in occupational protection among CSSD nurses, as well as key influencing factors, remain insufficiently characterized.

To accurately evaluate these professionals, it is necessary to understand the standardized professional tier system within Chinese nursing practice. Registered nurses in China are categorized by educational qualifications—principally three-year junior college diplomas or four-year bachelor’s degrees—and progress through a regulated clinical hierarchy of professional titles: from entry-level Nurse/Junior Nurse, through the mid-level Nurse-in-Charge, to the advanced ranks of Associate Chief Nurse and Chief Nurse. Therefore, this study aims to assess the knowledge, attitudes, and practices of occupational protection among a convenience sample of CSSD nurses in Zhejiang Province, China, by examining both the internal relationships among these three KAP dimensions and the external, modifiable predictors of protective practice. A structured questionnaire was used to evaluate key domains based on established infection control guidelines. To address the ceiling effects that limit variability in this sample, the analysis of both internal KAP pathways and external organizational or experiential factors was framed as an exploratory investigation. This dual approach allows us to report the internal KAP pathway as a finding in its own right, while simultaneously generating robust hypotheses regarding external factors that require validation in independent samples with wider score distributions. The findings may provide evidence to inform targeted training strategies and improve occupational safety and infection control in healthcare settings.

Methods

Study design and participants

A cross-sectional study was carried out among Central Sterile Supply Department nurses across multiple hospitals in Zhejiang Province, China, between January 4, and March 1, 2026. The study was conducted in accordance with the Declaration of Helsinki and was reviewed and approved by the Ethics Committee of Zhejiang Provincial People’s Hospital, the lead institution, under the rapid (expedited) review procedure (Approval No. ZRY-IRB-2025-Research-489). All participants provided electronic informed consent on the initial page of the digital questionnaire before proceeding to any survey items.

The inclusion criteria were as follows: registered nurses currently working in the CSSD, at least 1 year of CSSD work experience, and voluntary participation in the study. The exclusion criteria and sequential screening counts were as follows: refusal to provide informed consent (n = 1), an incorrect response to the trap question (n = 3), logical inconsistencies in the questionnaire responses, such as reporting 0 years of CSSD experience, 0 years of total nursing experience, a starting work age younger than 18 years, or conflicting assignments across the department’s physical layouts (n = 25), and test or otherwise invalid questionnaires (n = 1). In Chinese hospital standards, the CSSD is strictly divided into three functional zones based on infection control barriers: the decontamination area (dirty zone), the inspection-packaging-sterilization area (clean zone), and the sterile storage area (sterile zone).

Questionnaire and quality control

The questionnaire (Appendix) was developed with reference to the World Health Organization’s global report on infection prevention and control [13] and published evidence on occupational exposure among nurses in the Central Sterile Supply Department [14]. After the initial draft was completed, three senior experts in infection prevention and nursing management reviewed the instrument, and revisions were made based on their recommendations. A pilot test with 65 participants then yielded a Cronbach’s α of 0.754, indicating acceptable internal consistency prior to formal administration.

The final instrument consisted of four sections. Section I collected demographic and occupational information with 11 items, including specific questions regarding the participants’ regular rotation or permanent placement within the aforementioned three functional zones (decontamination, inspection-packaging-sterilization, and sterile storage areas). Section II assessed knowledge of occupational protection using 17 items, one of which served as an attention-check item (trap question) used solely to screen out inattentive respondents and was excluded from scoring. The remaining 16 items were scored to yield a total knowledge score ranging from 0 to 18: 14 single-answer items (true/false or single-choice) were scored 1 point each for a correct answer, and 2 multiple-choice items were scored up to 2 points each (2 points if all correct options were selected, 1 point for incomplete selection, and 0 points if any incorrect option was selected). Consequently, although the attention-check item was removed and only 16 substantive items remained, the maximum knowledge score is 18, because the two multiple-choice items each contribute up to 2 points (14 single-answer items × 1 point + 2 multiple-choice items × 2 points = 18). Section III evaluated attitudes toward occupational protection using 8 items scored on a 5-point Likert scale (ranging from 1 for strongly disagree to 5 for strongly agree, with reverse-scored items recoded accordingly), with a possible total score of 8 to 40. Section IV measured self-reported practices through 10 items on a 5-point Likert scale (ranging from 1 for never to 5 for always), of which three were reverse-scored, giving a total practice score range of 10 to 50. In this study, achieving a score greater than 80% of the maximum possible value in each domain was operationally defined as demonstrating adequate knowledge, a favourable attitude, and proactive self-reported practice [8, 10]. In the formal survey, the full instrument retained acceptable internal consistency, with a Cronbach’s α of 0.725. Internal consistency was additionally examined for each dimension according to its measurement scale. Because the 14 single-answer knowledge items are dichotomously scored, the Kuder–Richardson 20 (KR-20) coefficient was used for the knowledge dimension and was 0.27, whereas Cronbach’s α was 0.66 for the attitude dimension (8 Likert items) and 0.57 for the practice dimension (10 Likert items). The two partial-credit multiple-choice knowledge items were not included in the KR-20, which by definition applies only to dichotomous items. The low KR-20 for the knowledge dimension reflects its structure rather than measurement error: most knowledge items had very high correct-response rates (several exceeding 97%) and therefore minimal item variance, and the items assess content-heterogeneous, discrete facts rather than a single unidimensional construct. This composite is accordingly best interpreted as a knowledge checklist, and the corresponding measurement caveat is noted in the Limitations.

Prior to data collection, all participants were informed of the study’s purpose, risks, and confidentiality protections. Accessing the survey automatically directed respondents to an introductory landing page containing a formal consent statement; participants were required to provide electronic informed consent by checking a mandatory confirmation box before the system would permit them to proceed to the subsequent survey items. Participants were enrolled through convenience sampling. The questionnaire was delivered electronically via the Wenjuanxing platform, which generated a dedicated QR code disseminated through a province-level professional WeChat group for head nurses of Central Sterile Supply Departments in Zhejiang Province, which consists of approximately 400 members working across multiple hospitals. To protect data integrity and reduce duplicate entries, each IP address was permitted only one submission, and responses to all items were required before the questionnaire could be submitted. Returned questionnaires were then audited by the research team, who reviewed each record for completeness, internal coherence, and logical consistency; entries failing these checks were excluded from the final analysis.

Statistical analysis

Data analysis was conducted using SPSS version 27.0 (IBM Corp., Armonk, NY, USA) and AMOS version 26.0. To clarify how demographic characteristics were treated in the exploratory path model, all multi-categorical background variables were handled as ordinal (ranked) variables with ordered integer coding (1, 2, 3, 4…) reflecting ascending rank. Specifically, these demographic and occupational variables were coded as follows: Education (0 = Junior college or below, 1 = Bachelor’s or above); Professional title (coded as an ordered rank variable: 1 = Nurse / Junior nurse, 2 = Nurse-in-Charge, 3 = Associate chief nurse, 4 = Chief nurse/Other, reflecting an ascending clinical hierarchy); Total nursing experience (1 = ≤ 10 years, 2 = 11–20 years, 3 = 21–30 years, 4 = > 30 years); Training frequency (0 = < 3 sessions, 1 = ≥ 3 sessions); Training duration (0 = < 3 h, 1 = ≥ 3 h); Occupational exposure (0 = No, 1 = Yes). Continuous variables are presented as means and standard deviations (SD), whereas categorical variables are expressed as n (%). Differences in knowledge, attitude, and practice scores between two groups were compared using the Mann–Whitney U test, while comparisons among three or more groups were performed using the Kruskal–Wallis H test. Correlations among knowledge, attitude, and practice were examined using Spearman’s rank correlation analysis. In addition, an exploratory path analysis was used to explore the associations among knowledge, attitude, and practice. Covariates that showed statistically significant differences  (P< 0.05) in the univariate analyses (including Professional title, Total nursing experience, Training frequency, Training duration, and Occupational exposure) were incorporated into the exploratory path model as control variables. Model fit was evaluated using the chi-square/degrees of freedom ratio, root mean square error of approximation, incremental fit index, Tucker–Lewis index, comparative fit index, and the model degrees of freedom. The standardised root mean square residual (SRMR) is not generated by AMOS as part of its standard fit output and is therefore not reported; the absolute and incremental indices listed above were used instead. An initial full model including all candidate paths was first estimated; non-significant covariate paths  (P> 0.05) were then trimmed to obtain the final model, and the full initial model is reported in Supplementary Table S1. The stability of the path estimates and the significance of the direct and indirect effects were assessed using bias-corrected bootstrapping with 5,000 resamples, which does not rely on the assumption of multivariate normality. Modification indices were inspected to identify any potential model misspecification; they did not indicate any theoretically justifiable additional paths, and consequently no paths were added or otherwise modified on the basis of modification indices. Because the final model was obtained by data-informed trimming of non-significant paths, it should be regarded as exploratory and potentially data-driven, and the trimmed structure requires confirmation in independent samples. Furthermore, with the total practice score as the continuous dependent variable, all baseline characteristics and knowledge and attitude scores were first analyzed using univariate linear regression. Variables with P < 0.05 in univariate analysis were selected as independent variables to construct the final multivariable linear regression model using the enter method. Variance inflation factor (VIF) was used to assess multicollinearity among the independent variables, with VIF < 5 indicating no significant multicollinearity. A two-sided P value of less than 0.05 was considered statistically significant. Notably, the path coefficients are reported as standardised values to compare the relative strength of associations within the exploratory model, whereas the regression coefficients are reported as unstandardised values to quantify the expected absolute change in the practice score per unit change of each predictor. The sample size was adequate for the planned analyses: with 306 participants and 10 predictor terms entered into the final multivariable regression model, the data provided approximately 31 participants per predictor, well above the commonly recommended minimum of 10–15 subjects per variable; for the exploratory path model, which included 14 estimated parameters, this corresponded to approximately 22 cases per estimated parameter and satisfied conventional adequacy rules for path analysis (at least 10–20 cases per estimated parameter and a minimum total sample of 200).

Results

Sample characteristics and baseline features

Initially, this study collected a total of 336 questionnaires. Based on the exclusion criteria, 30 were excluded (1 declined participation, 3 failed the trap question, 25 had abnormal working years, and 1 was a test questionnaire). Ultimately, 306 valid questionnaires remained, with an effective rate of 91.07%. The majority were female (95.8%), aged > 45 years (41.5%), and held a bachelor’s degree or above (85.3%). Professional titles were distributed as nurse-in-charge (48.0%) and associate chief nurse (24.5%). Most nurses worked across all three functional zones (82.7%) and had at least five years of CSSD experience (66.0%). Their mean ± SD knowledge, attitude, and practice scores were 14.77 ± 1.73, 37.06 ± 2.87, and 45.90 ± 3.52, respectively. Knowledge scores differed significantly by work area only (P = 0.004). Attitude scores varied significantly with professional title (P = 0.002) and occupational exposure history in the past year (P = 0.026). Practice scores showed significant associations with professional title (P = 0.003), total nursing experience (P = 0.024), training frequency (P < 0.001), training duration (P < 0.001), and occupational exposure history (P = 0.014) (Table 1).

Table 1.

Baseline characteristics of participants and comparison of KAP scores across groups (N = 306)

	N (%)	Knowledge	P	Attitude	P	Practice	P
N=306		14.77 ± 1.73		37.06 ± 2.87		45.90 ± 3.52
Age (years)			0.787		0.138		0.494
≤ 45	179 (58.5)	14.79 ± 1.76		37.25 ± 2.85		45.92 ± 3.70
> 45	127 (41.5)	14.74 ± 1.70		36.80 ± 2.89		45.87 ± 3.26
Sex			0.615		0.124		0.275
Male	13 (4.3)	14.46 ± 1.98		38.00 ± 2.94		47.00 ± 2.71
Female	293 (95.8)	14.78 ± 1.72		37.02 ± 2.87		45.85 ± 3.54
Education			0.061		0.984		0.407
Junior college or below	45 (14.7)	14.36 ± 1.80		37.02 ± 3.01		46.33 ± 3.25
Bachelor’s or above	261 (85.3)	14.84 ± 1.71		37.07 ± 2.85		45.82 ± 3.56
Professional title			0.537		0.002		0.003
Nurse / Junior nurse	59 (19.3)	14.81 ± 1.85		37.95 ± 2.69		46.95 ± 3.36
Nurse-in-Charge	147 (48.0)	14.83 ± 1.74		37.03 ± 2.81		45.79 ± 3.49
Associate chief nurse	75 (24.5)	14.77 ± 1.52		36.29 ± 2.96		44.96 ± 3.68
Other	25 (8.2)	14.32 ± 1.97		37.44 ± 2.87		46.84 ± 2.72
Work area			0.004		0.964		0.622
Single zone	53 (17.3)	14.25 ± 1.67		36.96 ± 3.04		45.96 ± 3.82
All three zones	253 (82.7)	14.88 ± 1.72		37.08 ± 2.84		45.88 ± 3.46
Years in CSSD			0.142		0.720		0.611
< 5 years	109 (35.6)	15.02 ± 1.55		36.92 ± 3.04		46.06 ± 3.46
≥ 5 years	197 (64.4)	14.63 ± 1.81		37.14 ± 2.78		45.81 ± 3.55
Total nursing experience			0.329		0.291		0.024
≤ 10 years	44 (14.4)	14.36 ± 1.98		37.45 ± 2.89		47.30 ± 2.95
11–20 years	94 (30.7)	14.94 ± 1.76		37.14 ± 2.95		45.61 ± 3.63
21–30 years	112 (36.6)	14.85 ± 1.47		36.72 ± 2.90		45.75 ± 3.56
> 30 years	56 (18.3)	14.66 ± 1.93		37.30 ± 2.66		45.57 ± 3.48
Training frequency (past year)			0.431		0.088		< 0.001
< 3 sessions	167 (54.6)	14.69 ± 1.72		36.80 ± 2.98		45.17 ± 3.55
≥ 3 sessions	139 (45.4)	14.86 ± 1.75		37.38 ± 2.71		46.76 ± 3.28
Training duration (past year)			0.567		0.113		< 0.001
< 3 h	158 (51.6)	14.73 ± 1.68		36.87 ± 2.84		45.24 ± 3.67
≥ 3 h	148 (48.4)	14.82 ± 1.78		37.27 ± 2.90		46.59 ± 3.21
Occupational exposure (past year)			0.821		0.026		0.014
No	231 (75.5)	14.77 ± 1.73		37.29 ± 2.76		46.15 ± 3.49
Yes	75 (24.5)	14.76 ± 1.73		36.36 ± 3.09		45.11 ± 3.49

Note: K = Knowledge (maximum score: 18), A = Attitude (maximum score: 40), P = Practice (maximum score: 50); intergroup comparisons were performed using the Mann-Whitney U test or the Kruskal-Wallis H test; P < 0.05 indicates a statistically significant difference. “Occupational exposure (past year)” specifically refers to a self-reported sharps/needlestick injury or blood/body-fluid exposure occurring in the preceding 12 months

Distribution of KAP Scores

The distribution of scores across the three KAP dimensions is presented in Fig. 1. The score rates were 82.1% (14.77/18) for knowledge, 92.7% (37.06/40) for attitudes, and 91.8% (45.90/50) for practices. Notably, the knowledge dimension exhibited substantial individual variability and heterogeneity rather than a uniform high level (Fig. 1).

Fig. 1.

Distribution of KAP scores. Note: The violin plot illustrates the density distribution, while the boxplot shows the interquartile range. The blue dashed line indicates the mean value. Percentages in parentheses represent the score rate (mean/full score) for each dimension

Responses to knowledge items

The analysis of individual knowledge items revealed critical and clinically meaningful deficits in specific areas. The most pronounced deficit concerned post-exposure prophylaxis for HIV: only 13.7% of participants correctly recognized that prophylaxis should be initiated within one hour after exposure to blood or body fluids (Q12). Additionally, knowledge regarding the spectrum of bloodborne pathogens encountered when handling contaminated instruments was insufficient, with a correct response rate of 55.2% (Q15). A comparable deficit was observed for the selection of appropriate protective equipment in the decontamination area, where only 61.8% of participants identified the complete recommended PPE combination (Q16). Another notable gap was observed for hydrogen peroxide plasma sterilization safety, where 76.5% correctly understood the need for protective goggles and ventilation (Q5) (Table 2; Fig. 2).

Table 2.

Responses to knowledge items (N = 306)

Q	Question	Correctness N (%)
1	Are the requirements for personal protective equipment (PPE) the same in the decontamination area, inspection/packaging/sterilization area, and sterile storage area in the CSSD? (Correct answer: No)	89.5%
2	During the use of PPE, do face shields/goggles need to be frequently replaced, cleaned, or disinfected? (Correct answer: No)	77.8%
3	When using chlorine-containing disinfectants, waterproof isolation gowns, face shields, protective masks, and protective gloves should be worn.	95.8%
4	During ethylene oxide sterilization, professional protective masks should be worn, and adequate indoor ventilation should be ensured.	89.2%
5	After hydrogen peroxide plasma sterilization, when unloading items from the sterilizer, protective goggles should be worn and adequate indoor ventilation should be ensured.	76.5%
6	When operating high-temperature sterilization equipment (e.g., steam sterilizers/autoclaves), it is safe not to wear heat-resistant gloves. (reverse item; Correct answer: No).	95.1%
7	CSSD staff working in high-noise areas do not need to wear hearing protection. (reverse item; Correct answer: No)	85.0%
8	Dust or multi-enzyme detergents may cause respiratory irritation; therefore, masks or face shields should be worn.	99.0%
9	When cleaning instruments, if gloves are damaged, work can continue. (reverse item; Correct answer: No)	98.0%
10	After high-temperature disinfection, instruments may be handled directly with bare hands. (reverse item; Correct answer: No)	97.4%
11	If chemical liquid(s) splash(es) into the eyes, an eyewash station should be used immediately to flush the eyes for at least 15 min.	97.7%
12	After exposure to blood or body fluids, within what time frame should HIV post-exposure prophylaxis (PEP) be initiated? (Within 1 h) ★	13.7%
13	The first action that should be taken after a sharps injury: Immediately express blood from the wound and rinse with running water	99.7%
14	Cleaning stage is most prone to cross-contamination	95.8%
15	When handling blood-contaminated instruments, which of the following pathogens may be encountered? (Multiple choices allowed, HBV/HCV, HIV, Treponema pallidum, etc.★	55.2%
16	Which of the following protective equipment is suitable for the decontamination area? (Multiple choices; correct answer: waterproof isolation gown, medical protective mask, face shield, disposable gloves) ★	61.8%

★ indicates items with relatively low correct response rates (< 80%), representing key knowledge gaps (Q2: 77.8%, Q5: 76.5%, Q12: 13.7%, Q15: 55.2%, Q16: 61.8%). The total knowledge score ranges from 0 to 18: the 14 single-answer items were scored 1 point each and the 2 multiple-choice items (Q15 and Q16) up to 2 points each (14 × 1 + 2 × 2 = 18). The attention-check (trap) item was used only to screen out inattentive respondents and was excluded from scoring; the maximum is therefore 18

Fig. 2.

Ranked correct response rates of knowledge items (N = 306). Note: Red column indicates items with relatively low correct response rates (< 80%), representing key knowledge gaps (Q2: 77.8%, Q5: 76.5%, Q12: 13.7%, Q15: 55.2%, Q16: 61.8%)

Responses to attitude and practice items

Overall, participants demonstrated favorable attitudes toward occupational protection. However, a minority of nurses exhibited suboptimal attitudes in several key areas. Although the reverse-coded item “It is acceptable to simplify protective procedures when work is busy” (A5) was appropriately rejected by 91.8% (49.4% disagree, 42.5% strongly disagree), 7.5% (4.6% strongly agree, 2.9% agree) endorsed such unsafe shortcuts, and an additional 0.65% remained neutral. Similarly, for another reverse-coded item regarding PPE hindering work efficiency (A7), 12.1% (7.2% strongly agree, 4.9% agree) concurred that protective measures negatively affect efficiency, while 4.9% were uncertain. Regarding confidence in performing all occupational protection procedures (A2), 1.6% were neutral, indicating a lack of full assurance. Furthermore, a small proportion expressed doubts about the adequacy and quality of provided PPE (A3: 0.3% disagreed, 2.9% neutral) and about the effectiveness of training and assessments (A8: 1.0% neutral, 0.3% strongly disagreed) (Table 3).

Table 3.

Distribution of responses to attitude and practice items (N = 306)

Item	Question	Always/Strongly agree	Often/Agree	Sometimes/Neutral	Rarely/Disagree	Never/Strongly disagree
Part III — Attitude (Attitude Dimension, N  = 306, maximum score 40, mean 37.06 ± 2.87)
A1	I believe that strictly implementing occupational protection measures can effectively protect myself and patients from infection.	288 (94.1)	17 (5.6)	—	—	1 (0.3)
A2	I am confident in my ability to competently and correctly perform all occupational protection procedures.	256 (83.7)	44 (14.4)	5 (1.6)	—	1 (0.3)
A3	I believe that the personal protective equipment (PPE) provided by my department is sufficient in quantity and meets quality standards.	249 (81.4)	49 (16.0)	7 (2.3)	1 (0.3)	—
A4	I support and comply with the occupational protection regulations and assessment mechanisms in the CSSD.	260 (85.0)	43 (14.1)	3 (1.0)	—	—
A5▼	When work is busy, it is acceptable to simplify or omit some less important protective procedures. (reverse item)	14 (4.6)	9 (2.9)	2 (0.7)	151 (49.3)	130 (42.5)
A6	I am willing to proactively correct improper protective behaviors of myself or my colleagues.	233 (76.1)	68 (22.2)	1 (0.3)	—	4 (1.3)
A7▼	I believe that wearing PPE and adhering to occupational protection measures negatively affect my work efficiency. (reverse item)	22 (7.2)	15 (4.9)	15 (4.9)	139 (45.4)	115 (37.6)
A8	I believe that training and assessments can effectively improve the level of occupational protection.	232 (75.8)	71 (23.2)	2 (0.7)	—	1 (0.3)
Part IV — Practice (Practical Dimension, N  = 306, maximum score 50, mean 45.90 ± 3.52)
P1	When working in the decontamination area, I always wear complete PPE (e.g., waterproof isolation gown, protective goggles/face shield, double gloves).	262 (85.6)	39 (12.8)	5 (1.6)	—	—
P2	I perform proper hand hygiene after each contact with contaminated items or after removing PPE.	285 (93.1)	19 (6.2)	2 (0.7)	—	—
P3	I always dispose of used PPE and medical waste into designated containers according to proper classification.	299 (97.7)	7 (2.3)	—	—	—
P4	When handling sharps during cleaning, I operate cautiously and follow the principle of “no recapping and no bending.”	256 (83.7)	26 (8.5)	1 (0.3)	4 (1.3)	19 (6.2)
P5▼	I continue working even when ventilation equipment is malfunctioning. (reverse item)	15 (4.9)	12 (3.9)	47 (15.4)	99 (32.4)	133 (43.5)
P6	In the past six months, I have actively or passively participated in training or assessments related to occupational protection.	176 (57.5)	72 (23.5)	44 (14.4)	12 (3.9)	2 (0.7)
P7	I proactively monitor for skin injuries or physical discomfort and report them in a timely manner.	213 (69.6)	62 (20.3)	21 (6.9)	9 (2.9)	1 (0.3)
P8▼	When handling high-temperature equipment or hazardous chemicals, I sometimes neglect certain safety procedures. (reverse item)	12 (3.9)	8 (2.6)	22 (7.2)	98 (32.0)	166 (54.2)
P9	After experiencing a sharps injury or chemical exposure, I always take appropriate measures promptly.	277 (90.5)	22 (7.2)	3 (1.0)	2 (0.7)	2 (0.7)
P10▼	I sometimes use PPE without checking for damage or contamination. (reverse item)	12 (3.9)	3 (1.0)	16 (5.2)	88 (28.8)	187 (61.1)

Note ▼ indicates reverse-scored items; scores were calculated after reverse coding. Percentages are based on valid responses; “—” indicates zero responses

Nurses demonstrated high adherence to fundamental occupational protection practices, including proper hand hygiene (P2: 93.1% “Always”), correct waste classification (P3: 97.7% “Always”), and prompt post‑exposure measures (P9: 90.5% “Always”). The least positive practices were observed in active engagement with training and in consistent sharps handling. P6 (active or passive participation in occupational protection training or assessments in the past six months) had the lowest “Always” rate (57.5%), with 14.4% responding “Sometimes” and 4.6% “Rarely” or “Never”. P7 (proactive monitoring of skin injuries or physical discomfort and timely reporting) showed a moderate “Always” rate of 69.6%, leaving room for improvement. Additionally, P4 (cautious sharps handling following “no recapping and no bending”) had an “Always” rate of 83.7%, but concerningly, 6.2% reported “Never” following this principle, and another 1.3% “Rarely” did so, indicating a subgroup with persistently unsafe practices (Table 3).

Correlation and exploratory path analysis of KAP

Spearman correlation analysis showed that attitudes were strongly positively correlated with practices (r = 0.584, P < 0.001). Knowledge showed no significant correlation with attitudes (r = 0.088, P = 0.125), no significant correlation was found between knowledge and practices (r = 0.052, P = 0.364) (Fig. 3).

Fig. 3.

Spearman correlation matrix of KAP dimensions and scatter plot of attitudes vs. practices Note: The left panel shows a heatmap of correlation coefficients (green indicates positive correlation; red indicates negative correlation). The right panel presents a scatter plot of attitudes versus practices, with the red line representing the fitted regression line (r = 0.584, P < 0.001)

The exploratory path analysis model demonstrated a good fit: CMIN/DF = 1.850 (df = 14), RMSEA = 0.054, IFI = 0.955, TLI = 0.917, CFI = 0.953. These structural findings indicated that the data only partially supported the classical Knowledge→Attitude→Practice sequence. The model results indicated that educational level was positively associated with knowledge (β = 0.180, P = 0.008). Knowledge was positively associated with attitudes (β = 0.169, P = 0.010); this small significant effect emerged only after simultaneously adjusting for education, professional title, and training frequency, indicating a suppression effect when these covariates were controlled. Attitudes showed the strongest direct association with practices (β = 0.542, P = 0.014). In addition, while the total effect of knowledge on practice was non-significant (β = 0.106, P = 0.123), its indirect effect through attitude was significant (β = 0.092, P = 0.010), suggesting that these path effects are adjustment-dependent and should be interpreted with caution. The frequency of training in the past year showed significant positive links with both attitudes (β = 0.139, P = 0.007) and practices (direct association: β = 0.172, P = 0.012). Professional title was negatively associated with attitudes (β = −0.217) and practices (indirect association: β = −0.118). Years of nursing experience showed a direct negative association with practices (β = −0.135, P = 0.005) (Fig. 4). The significance of the direct and indirect effects was evaluated using bias-corrected bootstrapping (5,000 resamples); the standardised total, direct, and indirect effects with their 95% bootstrap confidence intervals are summarised in Table 4.

Fig. 4.

Exploratory path analysis model of occupational protection KAP among CSSD nurses Note: Values represent standardized path coefficients (β); **P < 0.01, *P < 0.05. Dashed lines indicate indirect pathways. Non-significant paths were removed from the final model

Table 4.

Standardized total, direct, and indirect effects from the exploratory path model, with bias-corrected bootstrap 95% confidence intervals

Model paths	Standardized Total effects		Standardized direct effects		Standardized indirect effects
	β (95%CI)	P	β (95%CI)	P	β (95%CI)	P
Education→Knowledge	0.180 (0.067,0.302)	0.008	0.180 (0.067,0.302)	0.008
Education→Attitude	0.030 (0.007,0.089)	0.003			0.030 (0.007,0.089)	0.003
Professional title→Attitude	-0.217 (-0.304,-0.103)	0.016	-0.217 (-0.304,-0.103)	0.016
Training frequency→Attitude	0.139 (0.043,0.251)	0.007	0.139 (0.043,0.251)	0.007
Knowledge→Attitude	0.169 (0.046,0.294)	0.010	0.169 (0.046,0.294)	0.010
Education→Practice	0.019 (0.001,0.059)	0.055			0.019 (0.001,0.059)	0.055
Professional title→Practice	-0.118 (-0.179,-0.064)	0.009			-0.118 (-0.179,-0.064)	0.009
Training frequency→Practice	0.248 (0.136,0.337)	0.009	0.172 (0.076,0.253)	0.012	0.075 (0.022,0.140)	0.009
Knowledge→Practice	0.106 (-0.023,0.232)	0.123	0.014 (-0.078,0.102)	0.784	0.092 (0.023,0.170)	0.010
Attitude→Practice	0.542 (0.438,0.629)	0.014	0.542 (0.438,0.629)	0.014
Total nursing experience→Practice	-0.135 (-0.241,-0.053)	0.005	-0.135 (-0.241,-0.053)	0.005

Note: Estimated by maximum likelihood with bias-corrected bootstrapping (5,000 resamples). β = standardised coefficient. The total, direct, and indirect “effects” reported here denote the standard statistical decomposition of associations estimated under a cross-sectional design and do not imply causal relationships among the variables

Multivariate linear regression analysis of factors influencing practice

The multivariable model included the variables that were significant in univariate analysis (P < 0.05); all variance inflation factors were < 5, indicating no significant multicollinearity. In the multivariate linear regression analysis, a higher attitude score (β = 0.667, 95% CI: [0.549–0.784], P < 0.001) and attending three or more training sessions per year (β = 0.930, 95% CI: [0.062–1.798], P = 0.036) were independently associated with better practice scores. Conversely, total nursing experience of 11–20 years (β = − 1.249, 95% CI: [–2.491 to − 0.007], P = 0.049) and more than 30 years (β = − 1.773, 95% CI: [–3.095 to − 0.450], P = 0.009) were independently associated with lower practice scores. Knowledge score, professional titles, and training duration were not significant in the final model (all P > 0.05) (Table 5).

Table 5.

Univariate and multivariate linear regression analysis of factors associated with practice scores (N = 306)

Variable	Univariate linear regression		Multivariate linear regression
	β (95% CI)	P	β (95% CI)	P
Knowledge score	0.241 (0.008, 0.474)	0.043	0.049 (− 0.144, 0.241)	0.619
Attitude score	0.699 (0.584, 0.813)	< 0.001	0.667 (0.549, 0.784)	< 0.001
Age (years)
≤ 45	Reference
> 45	−0.291 (− 1.108, 0.526)	0.484
Sex
Male	Reference
Female	−0.472 (− 2.337, 1.393)	0.619
Education
Junior college or below	Reference
Bachelor’s or above	−0.134 (− 1.295, 1.027)	0.821
Professional title
Nurse / Junior nurse	Reference		Reference
Nurse-in-Charge	−1.454 (− 2.580, − 0.327)	0.012	−0.066 (− 1.135, 1.002)	0.903
Associate chief nurse	−2.440 (− 3.691, − 1.189)	< 0.001	−0.719 (− 1.997, 0.560)	0.269
Other	−1.400 (− 2.986, 0.186)	0.083	0.026 (− 1.309, 1.360)	0.970
Work area
Single zone	Reference
All three zones	−0.020 (− 1.103, 1.064)	0.971
Years in CSSD
< 5 years	Reference
≥ 5 years	−0.038 (− 0.901, 0.824)	0.930
Total nursing experience
≤ 10 years	Reference		Reference
11–20 years	−1.195 (− 2.507, 0.118)	0.074	−1.249 (− 2.491, − 0.007)	0.049
21–30 years	−0.995 (− 2.270, 0.280)	0.126	−0.619 (− 1.932, 0.694)	0.354
> 30 years	−1.833 (− 3.225, − 0.442)	0.010	−1.773 (− 3.095, − 0.450)	0.009
Training frequency (past year)
< 3 sessions	Reference		Reference
≥ 3 sessions	1.838 (1.046, 2.630)	< 0.001	0.930 (0.062, 1.798)	0.036
Training duration (past year)
< 3 h	Reference		Reference
≥ 3 h	1.260 (0.456, 2.065)	0.002	0.217 (− 0.659, 1.093)	0.626
Occupational exposure (past year)
No	Reference
Yes	−0.757 (− 1.711, 0.198)	0.120

Note: Variables with P < 0.05 in univariate analysis were entered into the multivariable model; the variance inflation factor was used only as a collinearity diagnostic (all VIF < 5). Bold values indicate significance in the multivariable analysis (P < 0.05). β = unstandardised coefficient; 95% CI = 95% confidence interval

Discussion

Occupational protection among this convenience sample of CSSD nurses in Zhejiang Province was characterized by overall favourable self-reported attitudes and practices but insufficient knowledge in critical exposure-response content, with attitudes showing the strongest direct association with protective practice. Future interventions must establish a mandatory, high-frequency (≥ 3 sessions per year), scenario-based occupational protection training programme linked to annual performance evaluation. This framework should integrate specific modules on bloodborne exposure response and differentiated refresher strategies for senior nurses to systematically sustain protective behaviours.

A closer examination of the knowledge dimension shows that deficits were concentrated in areas requiring rapid decision-making, particularly in relation to bloodborne exposure and sterilization safety. Evidence from comparable healthcare settings indicates that workers commonly demonstrate adequate general knowledge while showing specific deficiencies in post-exposure procedures, particularly those involving strict timing or protocol-dependent actions [1]. Guidelines for occupational exposure management emphasize that timely and appropriate responses, specifically initiating HIV PEP within the first golden hour, are essential for reducing transmission risk [2]. However, this striking deficit (13.7% correctness) likely stems from the infrequent real-world application of HIV PEP protocols within the CSSD context compared to front-line clinical wards, leading to cognitive rustiness. Furthermore, limited local dissemination of updated occupational exposure guidelines and a lack of specific, time-critical assessment modules contribute to this gap [15]. This pattern aligns with literature showing that general theoretical knowledge does not automatically translate into rapid, protocol-dependent decision-making under time pressure, indicating that knowledge retention alone is insufficient unless reinforced by continuous, scenario-based simulation.

The relationships among knowledge, attitudes, and practices further clarify this issue. In the present study, attitudes showed a stronger association with practices than knowledge did, and knowledge had only a limited relationship with practice outcomes. This structural layout directly demonstrates that the classical KAP sequence has limited explanatory power in this specific setting, given that knowledge was not associated with attitude or practice at the zero-order level [11]. Furthermore, the substantial ceiling effects observed in the attitude (92.7%) and practice (91.8%) dimensions restricted score variability, which may have artificially attenuated these inter-dimensional associations. As an alternative interpretation supported by the data, attitude functions as the immediate proximal correlate of practice, while organisational factors such as training frequency and experiential factors such as nursing experience serve as the principal correlates of protective behaviours [11]. The exploratory pathway structure identified in this study only partially supports the classical Knowledge→Attitude→Practice sequence. While zero-order Spearman correlations showed no significant link between knowledge and either attitude or practice, the exploratory path analysis revealed a small significant association between knowledge and attitude (β = 0.169, P = 0.010) and an indirect association between knowledge and practice via attitude (β = 0.092, P = 0.010). This statistical discrepancy suggests a suppression effect that emerges only after adjustment for education, professional title, and training frequency, meaning these effects are adjustment-dependent and require cautious interpretation. Ultimately, the findings highlight that organisational and experiential factors, rather than knowledge alone, were the principal correlates of protective practice in this setting.

The distribution of attitude items indicates that most nurses recognized the importance of occupational protection, but a small proportion still expressed reservations about procedural requirements, particularly under conditions of high workload. Such views are well documented in clinical environments where efficiency pressures frequently create tension with safety protocols [16]. Qualitative studies have shown that healthcare workers sometimes perceive protective measures as time-consuming or disruptive to workflow, which can lead to selective adherence [17]. In this study, the presence of even a small subgroup holding such views is relevant, as attitudes were identified as a key determinant of practice. This implies that attitudinal variation, even if limited in scope, may contribute to observable differences in behaviour.

Practice-related findings also revealed uneven patterns. While adherence to routine procedures such as hand hygiene and waste classification was high, weaker performance was observed in areas involving proactive behaviour, such as participation in training, monitoring of physical discomfort, and consistent sharps handling. Healthcare workers are known to comply more consistently with standardized, observable procedures than with behaviours requiring self-initiation or sustained vigilance [18]. Studies on occupational safety have also reported that underreporting of minor injuries or symptoms is common, particularly in settings where such events are perceived as part of routine work [19]. These findings indicate that improving occupational protection requires attention not only to procedural compliance but also to behavioural engagement and reporting practices.

The analysis of influencing factors provides further insight. Educational level was associated with knowledge but did not directly predict practice outcomes. Formal education is recognized to enhance theoretical understanding without reliably translating into improved procedural adherence in clinical settings [11]. In structured healthcare environments, behaviour is often shaped more by organizational norms and supervision than by individual educational background. This may explain why education contributed indirectly through knowledge and attitudes rather than exerting a direct effect on practice.

Work area differences were also observed in relation to knowledge. Nurses working across multiple functional zones demonstrated higher knowledge levels, which may reflect broader exposure to different procedures and risks. Increased task diversity has been associated with greater familiarity with a broader range of occupational hazards [20]. However, this did not translate into differences in attitudes or practices, indicating that exposure alone is insufficient to influence behaviour without corresponding reinforcement.

When interpreting these factors, it is critical to balance the standardised path analysis findings with the unstandardised regression outcomes to avoid selectively emphasising any single variable. While attitude showed the strongest relative strength within the path model, training frequency demonstrated a highly potent and independent absolute contribution in the regression analysis, reinforcing the concept that structured institutional interventions are just as critical as individual psychological readiness. Professional title showed associations with attitudes and practices in baseline comparisons, and pathway analysis suggested a negative relationship with these outcomes. In associational terms, greater professional seniority was related to lower attitude and practice scores; rather than implying that experience causally erodes protective behaviour, this pattern may reflect alternative explanations, including selection bias, changes in job responsibilities that accompany promotion, and cohort effects [21]. Studies on occupational exposure have also reported that experienced healthcare workers may be less likely to report incidents or follow updated protocols consistently [19]. In the present study, these effects were not retained in the final regression model, suggesting that they may operate indirectly through other variables, particularly attitudes.

Years of nursing experience showed a consistent negative association with practice. Long-term occupational experience has been linked to normalization of risk and progressively reduced adherence to safety protocols [19]. This pattern has been observed in studies of needlestick injuries and occupational exposure, where more experienced workers sometimes demonstrate lower compliance with reporting and preventive measures [19]. The persistence of this association after adjustment indicates that it is unlikely to be explained solely by confounding factors.

Training frequency emerged as one of the most important factors associated with both attitudes and practices. Nurses who participated in training more frequently showed better performance, and this effect remained significant in the multivariable analysis. Repeated, structured training has consistently demonstrated greater effectiveness than sporadic instruction in sustaining infection prevention behaviours [11]. Studies have also shown that scenario-based and simulation-based approaches can enhance practical skills and decision-making [22]. In the context of CSSD work, where tasks are repetitive but risks are variable, regular reinforcement may help maintain attention to safety procedures.

In contrast, training duration did not show an independent association with practice after adjustment. Training effectiveness has been shown to depend more on frequency and content relevance than on session length alone [22]. Short, repeated sessions that focus on practical scenarios may be more effective than longer, less frequent sessions. This finding supports the need for training models that prioritize continuity and applicability.

The results also have implications for organizational practice. Weaknesses in training participation, symptom monitoring, and sharps handling suggest that current systems may not fully support consistent engagement with occupational protection. Organizational factors — including workload, staffing levels, and resource availability — are well-established determinants of adherence to safety practices [16]. Studies of hospital management have emphasized the role of leadership and institutional support in promoting infection prevention behaviour [23]. Based on these findings, healthcare administrators should implement a mandatory, high-frequency (≥ 3 sessions per year) training framework that transitions from traditional lectures to scenario-based simulations linked directly to annual performance evaluations. Furthermore, institutional programs must incorporate targeted modules on bloodborne exposure management and implement differentiated refresher training paths tailored specifically to senior nurses to address these experience-associated variations.

This study has several limitations. First, the cross-sectional design precludes causal inference and cannot exclude reverse causality among knowledge, attitudes, and practices. Second, convenience sampling through a province-level head-nurse WeChat group may have introduced selection bias: respondents were reached via head nurses participating in this professional network and may not be representative of all Central Sterile Supply Department nurses. Third, the affiliated hospital and hospital tier were not collected at the individual level, so institution-level representativeness and between-hospital comparisons could not be assessed. These factors limit representativeness, indicating that the findings require validation in larger, multi-province, multi-centre studies. Fourth, data were collected using a self-developed instrument; although it was reviewed by experts and pilot-tested, its psychometric properties were not comprehensively evaluated using state-of-the-art methods, such as confirmatory factor analysis of the full instrument. Fifth, the reliance on self-reporting may introduce recall bias and social desirability bias, as reflected in the ceiling effects where attitude and practice score rates both exceeded 90%, thereby limiting score variability. Sixth, treating ordered categorical demographic variables, such as the four levels of professional titles, directly as continuous ranked integers within the path analysis represents a methodological simplification that may not perfectly capture non-linear category effects. Finally, several unmeasured organisational confounders, such as ward management style, institutional safety culture, staffing levels, and daily workload, were not evaluated and could influence protective practices. Two further measurement and modelling caveats warrant mention. The internal consistency of the knowledge dimension was low (KR-20 = 0.27); this reflects the heterogeneous, content-based nature of the knowledge items and their limited variance rather than measurement error, but it means the knowledge composite is best interpreted as a checklist score rather than a unidimensional scale. In addition, because non-significant paths were trimmed from the path model in a partly data-driven manner, the final model carries a risk of overfitting, and the trimmed structure should be confirmed in independent samples.

In summary, occupational protection among CSSD nurses was characterized by acceptable attitudes and practices, but marked by significant heterogeneity and critical deficiencies in item-level knowledge, particularly concerning bloodborne exposure response and decontamination-area protective requirements. Future occupational protection programs should prioritize high-frequency, scenario-based training on bloodborne exposure management and establish targeted refresher pathways for senior nurses, so that knowledge gaps can be addressed in a timely manner and sustained improvements in real-world protective behaviours can be achieved.

Conclusion

In this provincial convenience sample, CSSD nurses in Zhejiang Province demonstrated generally good levels of occupational protection KAP; however, substantial gaps remain in knowledge related to emergency management of bloodborne exposure, particularly regarding the timing of HIV PEP initiation. The findings partially support the assumed KAP sequence, demonstrating that attitudes, alongside organisational training and experience-based factors, serve as the primary correlates of protective practice behaviors rather than knowledge alone. To bridge these gaps, healthcare administrators should implement a mandatory, high-frequency (≥ 3 sessions per year), scenario-based occupational-protection training programme, explicitly linked to annual performance evaluations, with targeted modules on bloodborne-exposure response and differentiated refresher training for senior nurses.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

HAIs

Healthcare-associated infections

CSSD

Central Sterile Supply Department

KAP

Knowledge, Attitudes, and Practices

SD

Standard deviations

Author contributions

Feifei Jing and Shuyan Li carried out the studies, participated in collecting data, and drafted the manuscript. Feifei Jing and Shuyan Li performed the statistical analysis and participated in its design. Feifei Jing and Shuyan Li participated in acquisition, analysis, or interpretation of data and draft the manuscript. All authors read and approved the final manuscript.

Funding

General Scientific Research Project of Zhejiang Provincial Department of Education, Project Number Y202559968.

Data availability

All data generated or analysed during this study are included in this published article.

Declarations

Ethics approval and consent to participate

All procedures were performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. The study was reviewed and approved by the Ethics Committee of Zhejiang Provincial People’s Hospital, the lead institution, under the rapid (expedited) review procedure (Approval No. ZRY-IRB-2025-Research-489). The approval covered the study protocol and the informed-consent documents for an online, minimal-risk, anonymous questionnaire survey of Central Sterile Supply Department nurses in Zhejiang Province. The Committee’s procedures follow the Measures for Ethical Review of Life-Science and Medical Research Involving Humans, the international ethical guidelines for health-related research, and GCP/ICH-GCP. All participants provided electronic informed consent before proceeding to the survey items.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.