1. Background
Healthcare-associated infections (HAIs), historically termed nosocomial infections, are defined by the World Health Organization (WHO) as infections acquired during hospital care that were neither present nor incubating at admission. They typically manifest 48 - 72 hours after admission or shortly after discharge (1-3). Globally, HAIs pose a serious threat to patient safety and are a major indicator of healthcare quality (4). At any given time, more than 1.4 million hospitalized individuals worldwide are affected by these infections, which contribute to substantial morbidity, prolonged hospital stays, antimicrobial overuse, and increased mortality (5-10). Although HAIs have substantial economic consequences in high-income countries (11-13), developing health systems bear an estimated 25% higher burden because of infrastructure and resource constraints (14).
Up to one-third of HAIs are preventable through consistent adherence to standard infection control protocols and strict hand hygiene (15). Transmission frequently occurs via the contaminated hands of healthcare personnel or through improper handling of clinical equipment and medical waste (1); therefore, healthcare workers serve as the first line of defense (16). Nurses constitute the largest group of clinical professionals and have the most continuous, direct bedside contact with patients (17-18). Their adherence to infection control measures is therefore a key determinant of whether pathogen transmission is interrupted or cross-contamination occurs in clinical settings (18).
Effective infection prevention depends on the interplay of knowledge, attitude, and practice (KAP) (19). Within behavioral frameworks such as Kelman's theoretical model, adequate knowledge provides a conceptual foundation for adherence to procedures, whereas a favorable professional attitude helps translate awareness into sustained clinical behavior (5, 20-22). Nevertheless, studies in diverse international settings have identified substantial differences in nurses' KAP profiles (21, 23-26). For example, an Iranian survey by Soleimani et al. found predominantly average levels of knowledge (79.8%) and practice (62.8%), along with notable gaps in adherence (27). Studies in other regional settings have also reported discrepancies between theoretical knowledge and self-reported adherence (26).
These differing findings indicate that infection control behaviors are strongly influenced by institutional infrastructure, organizational safety culture, and local clinical training programs (16, 28). Evolving clinical challenges and the need for updated baseline data from Iranian tertiary care centers underscore the importance of assessing current KAP levels to design targeted, evidence-based in-service training programs.
2. Objectives
This study aimed to assess knowledge, attitudes, and self-reported practices regarding infection control and to identify associated sociodemographic predictors among clinical nurses at teaching hospitals in Iran.
3. Methods
3.1. Study Design and Setting
This descriptive cross-sectional study was conducted in 2024 among clinical nurses employed at three major teaching hospitals affiliated with Qazvin University of Medical Sciences: Velayat, Bu-Ali, and Shahid Rajaei. The sample size was calculated using Cochran's single-proportion formula:
In this formula, Z1-α/2 = 1.96 corresponds to a 95% confidence level, p = 0.37 represents the expected proportion with satisfactory infection control practice based on previous evidence, 1-p = 0.63, and d = 0.05 is the margin of precision. The calculation yielded a baseline sample size of 358 nurses. To account for potential clustering across hospital departments and shifts (design effect = 1.2) and anticipated incomplete questionnaires, the final target sample size was set at 431 nurses.
A proportional stratified systematic random sampling strategy was used. Complete staff registries were obtained from the nursing management offices of the three hospitals. The sample was allocated proportionally across the centers: 162 nurses from Velayat Hospital, 142 from Bu-Ali Hospital, and 127 from Shahid Rajaei Hospital. The sampling interval (k = N/n) was calculated for each institution. The first participant was selected using a random integer between 1 and k, and every kth nurse thereafter was approached systematically across departments and shifts. The inclusion criteria were at least three months of clinical nursing experience, at least a bachelor's degree in nursing, and the provision of written informed consent. Nurses who left more than 20% of questionnaire items unanswered or withdrew from participation were excluded.
3.2. Data Collection Instruments
3.2.1. Demographic Checklist
Demographic information was collected using a researcher-developed checklist that covered age, sex, marital status, employment status, work experience, hospital department, and shift type.
3.2.2. Knowledge, Attitude and Practice of Clinical Staff About Nosocomial Infections Instrument
We used the validated Persian instrument developed by Soleimani et al. (27), which comprises 45 items across three 15-item domains: knowledge (dichotomous responses; score range 0 - 15), attitude (3-point Likert scale; score range 0 - 30), and self-reported practice (5-point Likert scale; score range 0 - 60). The practice domain assesses self-reported behavior rather than directly observed behavior. Following the original validation, scores were classified into three mutually exclusive tiers based on the percentage of the total score: ≤ 33.3%, 33.4% - 66.7%, and > 66.7%. The corresponding categories were low, moderate, and high knowledge, with scores of 0 - 5, 6 - 10, and 11 - 15; negative, neutral, and positive attitudes, with scores of 0 - 10, 11 - 20, and 21 - 30; and poor, moderate, and good practice, with scores of 0 - 20, 21 - 40, and 41 - 60, respectively. Construct validity was originally confirmed by confirmatory factor analysis (root mean square error of approximation [RMSEA] = 0.075; comparative fit index [CFI] = 0.89; Content Validity Index [CVI]/content validity ratio [CVR]: 0.80 - 1.00). The original reliability coefficients were 0.75 for knowledge (Kuder-Richardson formula 20 [KR-20]), 0.71 for attitude, and 0.82 for practice; the corresponding test-retest coefficients were 0.81, 0.84, and 0.92 (27). In the present study, a 10-member expert panel reconfirmed qualitative content validity. In our sample (n = 430), internal consistency was 0.82 for knowledge (KR-20), 0.79 for attitude, and 0.86 for practice. Test-retest reliability over 14 days (n = 30) yielded correlation coefficients of 0.81, 0.78, and 0.88, respectively.
3.3. Statistical Analysis
Data were analyzed using SPSS version 25.0 (IBM Corp., Armonk, NY, USA). Categorical data were expressed as frequencies and percentages, and continuous scores as the mean ± standard deviation (SD). Normality was confirmed using the Kolmogorov-Smirnov and Shapiro-Wilk tests (P > 0.05), with skewness and kurtosis within ± 1.0. Univariate and backward stepwise multivariable linear regression analyses were used to identify predictors of KAP; variables with univariate P < 0.20 were entered as candidates. Model assumptions were assessed using probability-probability plots for residual normality and scatterplots for homoscedasticity. The absence of multicollinearity was assessed using a variance inflation factor (VIF) < 2.0 and tolerance > 0.50. Analyses included 430 complete records, and statistical significance was set at a two-tailed P < 0.05.
4. Results
Across the three teaching hospitals, 455 eligible nurses were approached, of whom 24 declined because of clinical workload constraints. Of the 431 returned questionnaires, one was excluded during data auditing because of substantial missing data (> 20%), leaving a final analyzed sample of 430 nurses (response rate = 94.5%). Among the participants, 303 (70.5%) were female and 127 (29.5%) were male; 169 (39.3%) were single and 261 (60.7%) were married. Table 1 summarizes the participants’ demographic and occupational characteristics.
Table 1.
Demographic and Occupational Characteristics of the Participating Nurses (N = 430)
| Variables | No. (%) |
|---|---|
| Age (y) | |
| 20 - 24 | 38 (8.8) |
| 25 - 29 | 147 (34.2) |
| 30 - 34 | 86 (20.0) |
| 35 - 39 | 79 (18.4) |
| 40 - 44 | 66 (15.3) |
| ≥ 45 | 14 (3.2) |
| Sex | |
| Female | 303 (70.5) |
| Male | 127 (29.5) |
| Marital status | |
| Single | 169 (39.3) |
| Married | 261 (60.7) |
| Work experience (y) | |
| < 1 | 61 (14.2) |
| 1 - 4 | 111 (25.8) |
| 5 - 10 | 120 (27.9) |
| 11 - 15 | 49 (11.4) |
| 16 - 20 | 56 (13.0) |
| > 20 | 33 (7.7) |
| Hospital unit | |
| Operating room | 37 (8.6) |
| Dialysis | 35 (8.1) |
| Emergency | 58 (13.5) |
| CCU/ICU | 105 (24.4) |
| Internal | 110 (25.6) |
| Surgery | 85 (19.8) |
| Shift status | |
| Fixed | 60 (14.0) |
| Rotating | 370 (86) |
| Employment status | |
| Permanent | 223 (51.9) |
| Contractual | 207 (48.1) |
Most nurses had moderate knowledge (67.4%). Nearly half reported positive attitudes (49.5%), whereas 48.4% reported neutral attitudes and 2.1% reported negative attitudes. Most reported good practice (86.3%), whereas 13.3% reported moderate practice and 0.5% reported poor practice (Table 2). At the item level, the highest proportion of correct knowledge responses concerned Ambu-bag disinfection (K11: 79.5%), whereas indications for negative-pressure rooms had the highest error rate (K13: 78.4%) (Table 3). For attitude items, agreement was highest for immediate reporting of sharps injuries (N6: 90.0%), whereas disagreement was highest with the statement that hand hygiene training is unnecessary for all personnel (N11: 78.1%) (Table 4). Table 5 presents item-level distributions of self-reported practice across clinical procedures.
Table 2.
Overall Levels of Knowledge, Attitude, and Self-reported Practice Regarding Nosocomial Infection Control Among Nurses (N = 430)
| Variables | No. (%) |
|---|---|
| Knowledge | |
| Low | 37 (8.6) |
| Moderate | 290 (67.4) |
| High | 103 (24.0) |
| Attitude | |
| Negative | 9 (2.1) |
| Neutral | 208 (48.4) |
| Positive | 213 (49.5) |
| Practice | |
| Poor | 2 (0.5) |
| Moderate | 57 (13.3) |
| Good | 371 (86.3) |
Table 3.
Distribution of Responses to Individual Knowledge Items Regarding Nosocomial Infection Control (N = 430) a
| Question | False Answer | True Answer |
|---|---|---|
| K1 | 108 (25.1) | 322 (74.9) |
| K2 | 146 (34.0) | 284 (66.0) |
| K3 | 246 (57.2) | 184 (42.8) |
| K4 | 161 (37.4) | 269 (62.6) |
| K5 | 162 (37.7) | 268 (62.3) |
| K6 | 318 (74.0) | 112 (26.0) |
| K7 | 142 (33.0) | 288 (67.0) |
| K8 | 105 (24.4) | 325 (75.6) |
| K9 | 219 (50.9) | 211 (49.1) |
| K10 | 259 (60.2) | 171 (39.8) |
| K11 | 88 (20.5) | 342 (79.5) |
| K12 | 132 (30.7) | 298 (69.3) |
| K13 | 337 (78.4) | 93 (21.6) |
| K14 | 144 (33.5) | 286 (66.5) |
| K15 | 143 (33.3) | 287 (66.7) |
a Values are expressed as No. (%).
Table 4.
Distribution of Responses to Individual Attitude Items Regarding Nosocomial Infection Control (N = 430) a
| Question | Disagree | Neutral | Agree |
|---|---|---|---|
| N1 | 37 (8.6) | 34 (7.9) | 359 (83.5) |
| N2 | 41 (9.5) | 36 (8.4) | 353 (82.1) |
| N3 | 61 (14.2) | 41 (9.5) | 328 (76.3) |
| N4 | 32 (7.4) | 26 (6.0) | 372 (86.5) |
| N5 | 157 (36.5) | 47 (10.9) | 226 (52.6) |
| N6 | 16 (3.7) | 27 (6.3) | 387 (90.0) |
| N7 | 47 (10.9) | 16 (3.7) | 367 (85.3) |
| N8 | 50 (11.6) | 26 (6.0) | 354 (82.3) |
| N9 | 120 (27.9) | 41 (9.5) | 269 (62.6) |
| N10 | 280 (65.1) | 33 (7.7) | 117 (27.2) |
| N11 | 336 (78.1) | 24 (5.6) | 70 (16.3) |
| N12 | 307 (71.4) | 28 (6.5) | 95 (22.1) |
| N13 | 190 (44.2) | 34 (7.9) | 206 (47.9) |
| N14 | 78 (18.1) | 22 (5.1) | 330 (76.7) |
| N15 | 39 (9.1) | 45 (10.5) | 346 (80.5) |
a Values are expressed as No. (%).
Table 5.
Distribution of Responses to Individual Self-Reported Practice Items Regarding Nosocomial Infection Control (N = 430) a
| Question | Never | Rarely | Sometimes | Often | Always |
|---|---|---|---|---|---|
| P1 | 1 (0.2) | 10 (2.3) | 10 (2.3) | 36 (8.4) | 373 (86.7) |
| P2 | 1 (0.2) | 18 (4.2) | 36 (8.4) | 131 (30.5) | 244 (56.7) |
| P3 | 23 (5.3) | 64 (14.9) | 34 (7.9) | 133 (30.9) | 176 (40.9) |
| P4 | 9 (2.1) | 43 (10.0) | 45 (10.5) | 146 (34.0) | 187 (43.5) |
| P5 | 1 (0.2) | 16 (3.7) | 44 (10.2) | 137 (31.9) | 232 (54.0) |
| P6 | 13 (3.0) | 24 (5.6) | 55 (12.8) | 129 (30.0) | 209 (48.6) |
| P7 | 84 (19.5) | 28 (6.5) | 38 (8.8) | 81 (18.8) | 199 (46.3) |
| P8 | 8 (1.9) | 11 (2.6) | 30 (7.0) | 111 (25.8) | 270 (62.8) |
| P9 | 14 (3.3) | 12 (2.8) | 74 (17.2) | 124 (28.8) | 206 (47.9) |
| P10 | 7 (1.6) | 10 (2.3) | 38 (8.8) | 100 (23.3) | 275 (64.0) |
| P11 | 7 (1.6) | 20 (4.7) | 27 (6.3) | 72 (16.7) | 304 (70.7) |
| P12 | 8 (1.9) | 34 (7.9) | 68 (15.8) | 84 (19.5) | 236 (54.9) |
| P13 | 34 (7.9) | 51 (11.9) | 79 (18.4) | 99 (23.0) | 167 (38.8) |
| P14 | 5 (1.2) | 19 (4.4) | 30 (7.0) | 81 (18.8) | 295 (68.6) |
| P15 | 28 (6.5) | 27 (6.3) | 51 (11.9) | 101 (23.5) | 223 (51.9) |
a Values are expressed as No. (%).
In univariate linear regression, male nurses had significantly lower mean scores than female nurses across all three domains: knowledge (B = -0.27, P = 0.011), attitude (B = -0.25, P = 0.017), and self-reported practice (B = -0.41, P < 0.001). Married nurses had significantly higher knowledge scores than single nurses (B = 0.27, P = 0.005), and nurses working rotating shifts had higher knowledge scores than those working fixed shifts (B = 0.39, P = 0.005). Work experience of 11 - 15 years (B = 0.57, P = 0.003) and 16 - 20 years (B = 0.45, P = 0.013) were positively associated with knowledge compared with < 1 year of experience. Nurses working in surgery units had higher practice scores than those working in the operating room (B = 0.39, P = 0.040) (Table 6).
Table 6.
Univariate Linear Regression Analysis of Factors Associated with Nurses' Knowledge, Attitude, and Self-Reported Practice Scores (N = 430) a
| Variables | Crude β coefficient | ||
|---|---|---|---|
| Knowledge | Attitude | Practice | |
| Age | 0.06 (-0.004, 0.13); P = 0.066 | 0.03 (-0.03, 0.10); P = 0.323 | 0.07 (0.00, 0.14); P = 0.047 |
| Sex | |||
| Female | 1 | ||
| Male | -0.27 (-0.47, -0.06); P = 0.011 | -0.25 (-0.45, -0.05); P = 0.017 | -0.41 (-0.61, -0.21); P < 0.001 |
| Marriage status | |||
| Single | 1 | ||
| Married | 0.27 (0.08, 0.47); P = 0.005 | -0.08 (-0.27, 0.11); P = 0.402 | 0.12 (-0.07, 0.31); P = 0.219 |
| Work experience (y) | |||
| < 1 | 1 | ||
| 1 - 4 | -0.06 (-0.36, 0.25); P = 0.706 | -0.18 (-0.49, 0.14); P = 0.267 | -0.26 (-0.57, 0.05); P = 0.106 |
| 5 - 10 | 0.20 (-0.09, 0.51); P = 0.187 | -0.11 (-0.42, 0.19); P = 0.447 | 0.01 (-0.29, 0.32); P = 0.947 |
| 11 - 15 | 0.57 (0.20, 0.94); P = 0.003 | 0.07 (-0.30, 0.45); P = 0.701 | 0.21 (-0.16, 0.58); P = 0.227 |
| 16 - 20 | 0.45 (0.09, 0.81); P = 0.013 | 0.08 (-0.27, 0.45); P = 0.641 | 0.22 (-0.14, 0.58); P = 0.227 |
| > 20 | -0.17 (-0.58, 0.25); P = 0.427 | -0.23 (-0.65, 0.19); P = 0.294 | 0.11 (-0.31, 0.53); P = 0.559 |
| Unit of the hospital | |||
| Operating room | 1 | 1 | |
| Dialysis | -0.39 (-0.84, 0.07); P = 0.097 | -0.17 (-0.64, 0.28); P = 0.448 | -0.15 (-0.61, 0.29); P = 0.495 |
| Emergency | 0.18 (-0.23, 0.59); P = 0.390 | -0.05 (-0.46, 0.36); P = 0.787 | -0.14 (-0.55, 0.26); P = 0.483 |
| CCU/ICU | 0.15 (-0.21, 0.53); P = 0.404 | -0.17 (0.54, 0.20); P = 0.371 | 0.20 (-0.16, 0.57); P = 0.274 |
| Internal | 0.16 (-0.20, 0.54); P = 0.368 | -0.18 (0.55, 0.19); P = 0.333 | -0.09 (0.46, 0.27); P = 0.617 |
| Surgery | -0.11 (-0.50, 0.27); P = 0.554 | 0.13 (-0.25, 0.52); P = 0.498 | 0.39 (0.02, 0.78); P = 0.040 |
| Shift status | |||
| Fixed | 1 | 1 | 1 |
| Rotating | 0.39 (0.12, 0.66); P = 0.005 | 0.18 (-0.09, 0.45); P = 0.193 | -0.18 (-0.46, 0.08); P = 0.175 |
a Values are expressed as (95% CI). Reference categories: Female (sex), Single (marital status), Permanent (employment status), < 1 year (work experience), Operating room (unit of the hospital), Fixed (shift status). Abbreviations: B, unstandardized regression coefficient; CI, confidence interval.
In multivariable linear regression adjusting for confounders, male sex remained an independent negative predictor across all domains: knowledge (B = -0.25, P = 0.015), attitude (B = -0.28, P = 0.008), and practice (B = -0.44, P < 0.001). In the knowledge model, rotating shifts (B = 0.45, P = 0.002) and clinical experience of 11 - 15 years (B = 0.53, P = 0.009) and 16 - 20 years (B = 0.43, P = 0.032) retained independent positive associations. Attitude was positively predicted by age (B = 0.14, P = 0.002) and rotating shifts (B = 0.29, P = 0.040). Assignment to surgery units independently predicted higher practice scores (B = 0.39, P = 0.038). Full parameter estimates, 95% confidence intervals, and model summaries are presented in Table 7.
Table 7.
Multivariable Linear Regression Models Identifying Independent Predictors of Nurses' Knowledge, Attitude, and Self-reported Practice Regarding Infection Control (N = 430) a
| Independent Variables | Adjusted β | 95% Confidence Interval | P-Value |
|---|---|---|---|
| Model 1: Predictors of Knowledge | |||
| Sex | |||
| Female (Reference) | 1.00 | — | — |
| Male | -0.25 | -0.45 to -0.05 | 0.015 |
| Shift status | |||
| Fixed (Reference) | 1.00 | — | — |
| Rotating | 0.45 | 0.16 to 0.73 | 0.002 |
| Work experience (y) | |||
| < 1 (Reference) | 1.00 | — | — |
| 1 - 4 | -0.06 | -0.35 to 0.25 | 0.716 |
| 5 - 10 | 0.17 | -0.14 to 0.49 | 0.279 |
| 11 - 15 | 0.53 | 0.13 to 0.93 | 0.009 |
| 16 - 20 | 0.43 | 0.04 to 0.82 | 0.032 |
| > 20 | -0.03 | -0.47 to 0.42 | 0.906 |
| Model 2: Predictors of Attitude | |||
| Age (y) | |||
| continuous | 0.14 | 0.05 to 0.22 | 0.002 |
| Sex | |||
| Female (Reference) | 1.00 | — | — |
| Male | -0.28 | -0.48 to -0.07 | 0.008 |
| Employment type | |||
| Contractual (Reference) | 1.00 | — | — |
| Permanent | 0.39 | 0.16 to 0.62 | 0.001 |
| Shift status | |||
| Fixed (Reference) | 1.00 | — | — |
| Rotating | 0.29 | 0.01 to 0.58 | 0.040 |
| Model 3: Predictors of Practice | |||
| Sex | |||
| Female (Reference) | 1.00 | — | — |
| Male | -0.44 | -0.65 to -0.23 | <0.001 |
| Hospital unit | |||
| Operating room (Reference) | 1.00 | — | — |
| Dialysis | -0.22 | -0.67 to 0.22 | 0.326 |
| Emergency | -0.11 | -0.51 to 0.29 | 0.586 |
| CCU/ICU | 0.10 | -0.26 to 0.47 | 0.584 |
| Internal medicine | -0.18 | -0.54 to 0.18 | 0.328 |
| Surgery | 0.39 | 0.02 to 0.77 | 0.038 |
a A dash (—) in reference rows denotes the reference category. Abbreviations: β, standardized regression coefficient; CI, confidence interval.
4.1. Variable Selection and Modeling
Candidate variables with P < 0.20 in univariate screening (Table 6) were entered into outcome-specific multivariable linear regression models using backward elimination to obtain the final parsimonious models. In Model 1 (knowledge), marital status was initially entered but was excluded from the final model because its association was attenuated and non-significant (P > 0.05) after adjustment for work experience and shift type. In Model 3 (practice), age was adjusted for but was no longer significant in the presence of unit type and sex.
4.2. Collinearity Diagnostics
Multicollinearity was evaluated in all models. Variance inflation factors ranged from 1.08 to 1.84 across all included predictors, below the conservative threshold of 2.5, and tolerance values were > 0.54. These findings indicated no multicollinearity.
5. Discussion
This study assessed nurses' KAP regarding nosocomial infection control and found that most nurses had moderate knowledge. This finding is consistent with studies from Palestine, China, and Saudi Arabia (29-31), suggesting that knowledge gaps occur across settings. Knowledge of infection control may be influenced by access to educational resources, educational structures, staffing shortages, and heavy workloads. The studies cited above and the present study may therefore reflect similar challenges. However, our finding differs from studies reporting good knowledge of nosocomial infections among most healthcare workers (1, 21, 23). These differences may reflect their broader study populations, which included physicians, dentists, nurses, pharmacologists, radiologists, and optometrists. Nurses have extensive patient contact and provide much of the direct care; therefore, their knowledge levels may differ from those of professionals with less patient contact. A Nigerian study reported that more than half of healthcare workers lacked adequate knowledge of nosocomial infection prevention (32). That study included physicians and nurses from internal medicine, obstetrics, and pediatrics units, whereas the present study included nurses from internal medicine, surgery, dialysis, emergency, ICU, and CCU units. Nurses working in intensive care units usually have stronger scientific and practical skills than other nurses (33), which may help explain the differences between the studies. Nursing administrators and managers should identify knowledge gaps and promptly organize in-service training to address them.
Nearly half of the nurses (49.5%) had a positive attitude toward nosocomial infection control, consistent with findings from Africa (23, 34) and China (30). Abalkhail et al. and Gezie et al. also reported positive staff attitudes (21, 35), although the proportions were higher than those in the present study. Differences in professional settings, managerial support for infection control programs, and healthcare workers' participation in infection control decisions may explain this variation. In Sarani et al.'s study of Iranian nurses, only 33% had a good attitude toward infection control (36), a lower proportion than that in the present study. Although both studies were conducted in Iran, our substantially larger sample provided greater statistical power [430 vs. 145 nurses in Sarani et al.'s study (36)]. Differences in assessment instruments and regional access to education in Zabol and Qazvin may also explain the findings. The proportion with positive attitudes (49.5%) should be interpreted alongside the 48.4% with neutral attitudes, which warrants attention from managers and policymakers. From a clinical perspective, neutral attitudes may pose a risk to patient safety. These nurses may not fully understand the importance of infection control, or existing processes and policies may not sufficiently support their commitment. Managers and policymakers should strengthen the patient safety culture and use qualitative studies to explore the reasons underlying neutral attitudes.
Most nurses reported good practice in controlling nosocomial infections, consistent with findings from China and Saudi Arabia (5, 21, 30). Although those studies included diverse healthcare professionals, institutional infection surveillance systems and structured protocols may contribute to these similar findings. In contrast, healthcare workers in Ethiopia reported poorer practices (1, 23), possibly reflecting differences in institutional resources, staff-to-patient ratios, and access to infection control supplies. Because our assessment used a self-administered questionnaire rather than direct observation or objective infection surveillance, favorable ratings reflect perceived adherence and may be influenced by social desirability bias rather than verified bedside compliance.
Economic conditions and patient volume may also affect staff practice and help explain differences between studies. Staff training may be another contributing factor. The two Ethiopian studies also included participants with diplomas, whereas the present study included only nurses with at least a bachelor's degree. According to Kelman's theory, knowledge, attitude, and practice influence one another (20). Das et al. also reported that repeated training significantly improved infection control practices among healthcare personnel in India (37). Educational level is therefore one factor affecting practice. Periodic evaluation of nurses' practice may help prevent deterioration and identify areas requiring improvement. Addressing these areas may strengthen the patient safety culture and reduce the adverse effects of nosocomial infections.
Female nurses scored significantly higher than male nurses in knowledge, attitude, and practice, consistent with previous studies (38-39). Conversely, studies in Saudi Arabia (31) and Poland (40) found no significant differences or reported contrary patterns. These discrepancies may reflect differences in nursing workforce composition, sampling frameworks, or ward assignments. Although female nurses may demonstrate greater adherence to some hygiene protocols in certain clinical settings (39), explanations for sociodemographic or behavioral differences remain hypothetical because our cross-sectional study did not assess personal motivation, baseline clinical training exposure, or daily task allocation. Our multivariable findings that permanent employment, longer clinical experience, and rotating shifts positively predicted KAP scores are also consistent with recent Iranian evidence from Shajari Sales et al. In their study of midwifery personnel, 73.5% were classified as having good practice and 56.8% had moderate knowledge. Employment status and specialized training were significantly associated with knowledge, and non-permanent staff were more vulnerable (41). Healthcare administrators should conduct formal in-service training-needs assessments, particularly for contractual and less-experienced personnel, to assess gaps in adherence objectively and ensure that infection control education reaches all nursing staff.
5.1. Study Limitations
The study was conducted in three tertiary teaching hospitals in a single city, Qazvin, Iran, using a cross-sectional design. Caution is therefore needed when generalizing the findings to non-academic or private hospitals or to nationwide healthcare settings; multicenter and longitudinal studies are recommended. The self-report format was another limitation because perceived practice may not reflect observed bedside adherence or infection outcomes and may be affected by social desirability bias. The researchers sought to address this limitation by ensuring anonymity and explaining the importance of accurate responses. Future studies should supplement questionnaires with direct observational audits. Nurses' mental and physical condition and workload may also affect the results; the researchers therefore attempted to administer the questionnaires when participants were in a relatively stable condition. Corrections for multiple testing were not applied to the exploratory univariate analyses, which may increase the risk of type I error; these findings consequently warrant cautious interpretation.
5.2. Conclusions
Reducing the adverse effects of nosocomial infections requires adequate knowledge, favorable attitudes, and appropriate infection control practices among nurses. Most nurses in this study had moderate knowledge, nearly half had positive attitudes, and most reported good practice. Managers and educational policymakers should address the moderate knowledge levels through in-service training. The attitude findings suggest that existing educational and safety policies are not sufficiently effective and that motivational interventions are needed. Improving knowledge and promoting positive attitudes among nurses with neutral attitudes may support more effective infection control.
Remaining weaknesses in clinical practice should be identified and addressed through intervention studies and direct observational audits. Qualitative studies are recommended to identify barriers to positive attitudes. Longitudinal studies should assess the effects of management policies, interventions, and training programs.
