Sleepless and Disconnected: How Poor Rest and Weak Social Ties Undermine Nurse Safety

Nurses stand at the front line of patient safety, and the smallest lapse in their attention or judgment can carry serious consequences for the people in their care. A new mixed-methods study published in BMC Nursing by Jie Peng of the First People’s Hospital of Yulin and colleagues offers one of the most detailed pictures yet of how two everyday forces—sleep quality and social connection—shape whether nurses follow safe practices at work. Drawing on survey data from 418 nurses and in-depth interviews with 23 more, the research maps the psychological pathways that link a restless night to a skipped protocol, and it suggests that a nurse’s social network may be one of the most powerful buffers available.

The study is grounded in the Job Demands–Resources model, a widely used framework in occupational health psychology that describes how workplace conditions split into two broad categories. Job demands are the physical, psychological, or organizational aspects of work that consume energy: heavy workloads, time pressure, emotional strain, and, in this formulation, poor sleep. Job resources are the elements that help workers cope and stay engaged, ranging from autonomy and feedback to social support. The model predicts that demands deplete performance while resources sustain it, and that the two interact. What makes the new study distinctive is that it treats a nurse’s personal social network—an ego network in the language of social network analysis—as a measurable job resource, and then traces how that resource interacts with sleep, itself framed as a job demand.

Between June and December 2024, the researchers recruited 418 nurses through convenience sampling and administered a face-to-face survey. Sleep quality was measured with the Pittsburgh Sleep Quality Index, a validated instrument that scores seven components of sleep, including duration, latency, disturbances, and daytime dysfunction, with higher scores indicating worse sleep. Social position was captured through ego-network degree centrality, a metric that counts how many direct connections a person maintains within their network; a nurse with high degree centrality is one who can call on many colleagues, friends, and contacts directly. Perceived social support was assessed with the Perceived Social Support Scale, which measures how supported a person feels, and safety behavior was recorded with the Nurses’ Safety Behavior Questionnaire.

The statistical core of the study was structural equation modeling, performed with AMOS 26.0, a technique that allows researchers to test whether an entire web of hypothesized relationships fits the observed data simultaneously. The results were striking. Ego-network degree centrality was positively associated with safety behavior, with a standardized coefficient of 0.317, meaning nurses embedded in larger, denser webs of direct contacts reported consistently safer practices. Sleep quality moved in the opposite direction: higher scores on the Pittsburgh index, indicating poorer sleep, were negatively associated with safety behavior, with a coefficient of −0.184. Both effects were statistically significant at p < 0.001, indicating they are very unlikely to be products of chance.

But the most interesting findings emerged when the researchers examined mediation—the question of how poor sleep translates into unsafe behavior. They found that poor sleep was associated with lower ego-network degree centrality, with a coefficient of −0.168, and with lower perceived social support, with a coefficient of −0.008. In other words, nurses who slept badly tended to occupy weaker positions in their social networks and to feel less supported. Each of these two variables, in turn, was linked to lower safety behavior, meaning that the network position and the feeling of support each carried part of the effect of sleep on safety. Together they formed a serial mediation pathway: poor sleep eroded network centrality, which eroded perceived support, which in turn reduced safety behavior. This chained pathway carried a coefficient of −0.018 and accounted for 8.11 percent of the total association between sleep quality and safety behavior in the unadjusted model.

Eight percent may sound modest, but in the arithmetic of occupational safety it is meaningful. The majority of the sleep–safety link remains direct, suggesting that exhausted nurses make more errors through fatigue itself—slowed reaction times, lapses in attention, degraded decision-making. Yet the mediated portion reveals a second, subtler route: sleep loss appears to quietly shrink a nurse’s social world. This is a pattern familiar from sleep science more broadly. Sleep-deprived people are known to be less socially engaged, less accurate at reading facial expressions, and more likely to withdraw from interaction. A nurse running on fragmented sleep may skip the break-room conversation, decline the shift handover chat, and gradually drift to the periphery of the ward’s informal network—losing precisely the collegial checking and information-sharing that keep small mistakes from becoming large ones.

To understand the mechanisms behind these numbers, the team turned to qualitative methods. Between January and February 2025, they purposively selected 23 nurses for semi-structured interviews and analyzed the transcripts with directed content analysis using NVivo 12.6, taking the Job Demands–Resources model as the predetermined framework. This approach starts with theory-driven categories and refines them against the data, allowing the researchers to identify which specific job demands and resources nurses themselves connected to safe practice. The interviews reinforced the quantitative picture: participants described demands—among them fatigue and poor sleep—that drained their capacity for careful work, and resources, particularly supportive relationships, that restored it. The qualitative strand also pointed to psychological responses accompanying poor sleep that the nurses associated with lower safety behavior, suggesting that the pathway from a bad night to a risky shift runs partly through mood and cognition.

The methodological design deserves attention. The study was reported according to GRAMMS guidelines, the standard for mixed-methods research, and it combined the breadth of a survey with the depth of interviews in a deliberate sequence: the quantitative model first established which pathways existed and how much variance they explained, and the qualitative phase then explored why. Ethics approval came from the First People’s Hospital of Yulin, and the work was conducted in line with the Declaration of Helsinki, with anonymous questionnaires and informed consent from all participants. The research was supported by the Humanistic Nursing Committee of the China Association for Life Care and the Yulin City Scientific Research and Technology Development Plan Project.

Like all studies, this one has limits worth keeping in view. The cross-sectional design captures a single moment in time, so it cannot prove that poor sleep causes weaker networks or reduced safety behavior—only that the associations are real and patterned in the way the model predicts. Convenience sampling means the 418 participants may not represent all nurses, and the setting in Guangxi, China, may shape both sleep patterns and workplace culture in ways that differ elsewhere. Self-reported sleep and safety behavior are also vulnerable to bias, since nurses may underreport lapses. The authors themselves frame their conclusions carefully: within the Job Demands–Resources framework, sleep quality may be considered a job demand related to safety behavior, while network centrality and perceived support act as resource factors that mediate the association.

Even with those caveats, the practical implications are hard to ignore. Hospitals invest heavily in safety protocols, checklists, and training, yet this study suggests that two of the strongest levers on nurse safety lie outside the protocol binder: how well nurses sleep and how connected they feel. Interventions that protect sleep—smarter shift scheduling, shorter consecutive night shifts, nap facilities, and education on sleep hygiene—could pay a double dividend, directly reducing fatigue-related errors and indirectly preserving the social ties that sustain safe behavior. Likewise, efforts to strengthen ward-level networks, from buddy systems to structured peer support, may function as genuine safety infrastructure rather than soft benefits. As health systems worldwide grapple with nursing shortages and burnout, the message of this research is clear: a well-rested, well-connected nurse is not just a happier employee, but a measurably safer one, and the two conditions feed each other in ways that hospital leaders can no longer afford to overlook.

Subject of Research: The pathways linking nurses' sleep quality, ego-network centrality, perceived social support, and safety behavior within the Job Demands–Resources model

Article Title: Social networks, sleep quality, and nurses’ safety behavior: a mixed-methods study

Article References: Peng, J., Zhu, X., Huang, H., Feng, X., Lai, L., Li, J., Liu, X., Li, X., Du, H., Li, D., Huang, K., & Huang, Y. (2026). Social networks, sleep quality, and nurses’ safety behavior: a mixed-methods study. BMC Nursing. https://doi.org/10.1186/s12912-026-05478-y

Image Credits: AI Generated

DOI: 10.1186/s12912-026-05478-y

Keywords: nursing, sleep quality, safety behavior, social networks, Job Demands–Resources model, perceived social support, structural equation modeling, mixed methods, occupational health, Pittsburgh Sleep Quality Index, ego-network centrality, patient safety