The study, published in BMC Nursing by Elizabeth A. M. Michels of the Else Kröner Fresenius Center for Digital Health at TUD Dresden University of Technology and colleagues, took an observational, multimethod approach to a problem that has long frustrated occupational health researchers. Surveys about job demands and resources can tell us how nurses feel about their work in general, but they blur the very moments that accumulate into exhaustion, sickness absence, and ultimately the workforce shortages that hospitals around the world are struggling to contain. What the Dresden team wanted to know was simpler and sharper: which specific events during a normal shift actually correspond to an acute stress response in the body?

To answer that question, the researchers equipped up to 85 nurses with wearable electrocardiography devices that recorded heart rate variability continuously across an eight-hour shift. Heart rate variability, and in particular a measure called the root mean square of successive differences between heartbeats, or RMSSD, is a well-established window into the autonomic nervous system. When the sympathetic branch of that system ramps up in response to a stressor, the natural beat-to-beat variation in heart rhythm tends to decline. The researchers used the natural logarithm of this measure, lnRMSSD, as their physiological outcome, allowing them to detect moment-to-moment shifts in autonomic state as the shift unfolded.

Physiology alone, however, cannot say what caused a dip in heart rate variability. That is where the second layer of the design came in. Trained observers shadowed the nurses throughout their shifts, continuously logging the work events that occurred and rating the nurses’ momentary stress on a visual analogue scale. This ecological momentary assessment strategy, pairing event-level observation with ambulatory physiological monitoring, meant that every logged event could be linked to the cardiac state recorded at the same time. The researchers also controlled for physical activity, measured in metabolic equivalents of task, which is a crucial statistical adjustment in a job that involves a great deal of walking, lifting, and rushing; without it, a racing heart from climbing stairs could easily be mistaken for a racing heart from stress.

The events were sorted into three broad categories: clinical care burden, covering direct patient care demands; organizational burden, covering administrative and systemic frictions; and positive events, the moments of satisfaction, connection, and success that punctuate clinical work. In total, the analysis drew on 1,617 event-level observations, a dataset rich enough to support multilevel models that properly account for the fact that many events nest within each individual nurse. The findings were strikingly consistent for the observer-rated stress measure. Clinical care burden was associated with higher momentary stress than the other event categories, while positive events were associated with the lowest stress of all.

The physiological picture added an important nuance. Clinical care burden was linked to lower lnRMSSD than organizational burden, indicating a stronger autonomic stress signature when nurses were absorbed in demanding patient care than when they were wrestling with paperwork and system obstacles. Interestingly, no significant differences emerged between positive events and the other categories on the physiological side, suggesting that the recovery signal of a good moment may not register as clearly in cardiac data as the strain of a difficult one registers in an observer’s judgment. Stress, in other words, is not a single thing: what an observer sees and what the heart does can diverge in informative ways.

The study also probed whether individual characteristics shaped these patterns. Chronic stress, measured before the shifts, was associated with higher observer-rated stress during work, consistent with the idea that baseline strain colors how demanding each moment feels. An exploratory association appeared between sex and heart rate variability, with male nurses showing higher lnRMSSD values on average, a finding the authors flag as exploratory and one that echoes known sex differences in cardiac autonomic control. Age, by contrast, did not emerge as a significant factor in these analyses. These person-level results remind us that event effects unfold against the backdrop of who the nurse is and what they carry into the ward.

One of the most practically interesting parts of the study concerned the patient call system, the ubiquitous network of buttons and buzzers through which hospital patients summon help. The researchers examined whether call events differed from non-call work events, and found no significant differences in either observer-rated stress or heart rate variability. But when they dug into the reasons behind the calls, a pattern appeared. Calls related to comfort and service issues were rated as less stressful than calls involving acute medical needs or calls requiring care and mobilization support. The alarm itself, it seems, is not the stressor; what lies behind the alarm is what matters. No significant HRV associations were found for call reasons, keeping this a story about perceived demands rather than cardiac responses.

Methodologically, the study is a proof of concept as much as a set of findings. The authors report that combining event-based assessment with ambulatory physiological monitoring is feasible in routine clinical practice, which is no small claim in an environment where every extra device and every extra observer competes with patient care for attention and time. The multilevel modeling framework, using restricted maximum likelihood estimation and centering within clusters to separate within-person event effects from between-person differences, represents the current standard for this kind of intensive longitudinal data. The result is a method with high ecological validity: it captures stress where and when it actually happens, rather than asking nurses to reconstruct it retrospectively at the end of a draining shift.

Why does this matter beyond the wards of one Dresden hospital? Because targeted interventions need targets. If clinical care burden, rather than organizational friction, is the most consistently stress-associated category of everyday work, then efforts to protect nurses’ health, whether through staffing ratios, task redesign, teamwork support, or recovery breaks timed to the rhythm of patient care, can be aimed at the moments that physiology confirms are the hardest. With stressful working conditions contributing substantially to nurses’ sickness absence and to the global nursing shortage, precision about which events drive acute stress is not an academic luxury. It is the difference between broad, expensive wellness programs and focused changes to the actual texture of the working day. The Dresden team’s work, supported by the German Federal Ministry of Research, Technology and Space as part of the SEMECO project, suggests that the technology to see that texture clearly, a small wearable and a careful observer, is already within reach of routine practice.

Subject of Research: Psychophysiological measurement of acute stress responses to everyday work events in hospital nursing

Article Title: Identifying stress-relevant events in everyday nursing practice using psychophysiological outcomes – an observational multimethod study

Article References: Michels, E. A. M., Rominger, C., Schwerdtfeger, A. R., Kurze, G. F., Klein, J., Steinhäuser-Meerz, J. L., Tolle, H., & Wekenborg, M. K. (2026). Identifying stress-relevant events in everyday nursing practice using psychophysiological outcomes – an observational multimethod study. BMC Nursing, 25(1), Article 899. https://doi.org/10.1186/s12912-026-05434-w

Image Credits: AI Generated

DOI: 10.1186/s12912-026-05434-w

Keywords: nursing, occupational stress, heart rate variability, ecological momentary assessment, wearable sensors, multilevel modeling, clinical work events, hospital workforce, autonomic nervous system, job demands-resources, Identifying, stress-relevant