The COVID-19 pandemic significantly impacted healthcare professionals, causing increased workload, staff shortages, and stressful working conditions. Nurses faced heightened anxiety, fatigue, and fear due to exposure risks and insufficient resources. Prioritizing patient safety and reducing adverse events is crucial, as unsafe healthcare is a leading cause of death and disability globally. Patient Safety Culture (PSC), encompassing values, attitudes, and behaviors related to safety, is a key indicator of healthcare safety. While pre-pandemic studies showed high PSC levels in some countries, the pandemic negatively impacted PSC, particularly in units with high patient loads. The nursing practice environment (NPE), encompassing organizational characteristics affecting nursing practice, is also critical, impacting nurse satisfaction, job performance, and patient care quality. Studies have shown NPE's predictive role in PSC, although the predictive value of NPE subscales remains inconsistent. This study aimed to determine the association between NPE, safety perception, and PSC during the COVID-19 pandemic in Peru.

Existing literature highlights the significant impact of the COVID-19 pandemic on healthcare workers' well-being and working conditions. Studies showed increased stress, anxiety, and fatigue among nurses due to increased workload, staff shortages, and concerns about infection. The importance of patient safety and the use of PSC as a measure of safety culture in healthcare settings are well-established. Pre-pandemic research indicated varying levels of PSC across different countries and healthcare settings. However, studies conducted during the pandemic reveal a decline in PSC in many areas, indicating the negative impact of the pandemic on safety culture. The influence of nurse staffing levels, working hours, and workload on patient safety has also been studied. Furthermore, the literature acknowledges the significant role of the nursing practice environment (NPE) in promoting positive patient safety outcomes and nurse well-being. Prior research suggests a strong correlation between NPE and PSC, although the predictive value of specific NPE subscales has yielded mixed results. This research gap prompted the current study.

This quantitative, non-experimental, correlational, and cross-sectional study was conducted during the second wave of the COVID-19 pandemic in Peru. The study population consisted of 211 nurses working in various healthcare settings (primary healthcare facilities, public and private hospitals, and military hospitals). Non-probability snowball sampling was used to recruit participants. Data were collected using an online self-administered questionnaire in Google Forms. The Practice Environment Scale of the Nursing Work Index (PES-NWI) was used to assess the NPE, comprising five subscales: nurse participation in hospital affairs, nursing foundations for quality of care, nurse manager ability, leadership, and support of nurses, staffing and resource adequacy, and collegial nurse-physician relations. The Hospital Survey on Patient Safety Culture (HSOPSC) version 2.0 measured PSC, encompassing ten dimensions of patient safety culture. Safety perception was assessed using a single question about the rating of patient safety in the work unit. Descriptive statistics, Shapiro-Wilk test for normality, Spearman's correlation coefficient for correlations between variables, and two stepwise regression models (one with NPE subscales and sociodemographic data, the other with the total NPE scale) to predict PSC were used. Data analysis was performed using R software. Ethical approval was obtained from the Institutional Research Ethics Committee of the Universidad Norbert Wiener.

The study included 211 nurses, 91.5% of whom were female. The majority of participants (79.6%) were aged between 25 and 44. 45.5% reported a favorable NPE and 61.1% reported a neutral PSC. Spearman's correlation showed statistically significant positive correlations between all NPE subscales and PSC (p < 0.001). Regression analysis revealed that safety perception and three NPE subscales significantly predicted PSC (adjusted R-squared=0.50). Specifically, nurses perceiving good, very good, and excellent patient safety reported significantly higher PSC scores. The nurse manager's ability, leadership, and support of nurses subscale was a significant predictor of PSC. In another regression model (adjusted R-squared=0.51), the total NPE scale, safety perception, and workplace (military hospital vs. other settings) were significant predictors. Nurses in military hospitals reported higher PSC than those in other settings. The overall PSC level was reported as neutral by most nurses.

The findings support the significant role of NPE and safety perception in shaping PSC. The predictive role of NPE aligns with studies in various contexts. The importance of nurse manager support and leadership in fostering PSC is consistent with the leader-member exchange theory. The discrepancy with studies showing other NPE subscales as more significant predictors might be attributed to the strong control effect of safety perception in this study. The higher PSC among nurses in military hospitals might relate to organizational structure and resource allocation, but further research is needed. The neutral PSC levels reported might reflect the challenges faced by nurses during the pandemic, contrasting pre-pandemic findings in Peru. The favorable NPE perception might indicate that despite the pandemic, many hospitals in Peru maintained reasonable working conditions for nurses, though this warrants further investigation.

This study confirms that safety perception, workplace, and NPE predict PSC, explaining 51% of the variance. While the relationship between specific NPE subscales and PSC is complex, the findings highlight the importance of strong leadership focused on safety, enhanced management skills, interprofessional collaboration, and incorporating nurse feedback for continuous improvement in creating a positive safety culture, especially in resource-constrained settings.

The cross-sectional design limits causal inferences. Self-reported data may be susceptible to bias. The sample, while larger than the calculated minimum, might not fully represent all Peruvian nurses. Future longitudinal studies with larger, more diverse samples are recommended to confirm these findings and investigate causal relationships more thoroughly.