Presentation
Short-term Environmental Changes Impair Cognitive Performance of ICU Nursing Staff
SessionPoster Session 2
DescriptionBackground: It has been well documented that occupational demands within a hospital environment, specifically the intensive care unit (ICU), have led to burnout and fatigue. Additionally, much of the literature has shown that inadequate conditions within particular healthcare settings have a ripple effect that impacts the staff and eventually amplifies the risk of compromising patient care quality. Suboptimal physical conditions, particularly related to a lack of control over thermal and lighting conditions, have been shown to heighten stress and fatigue within clinical settings. Within the hospital setting, several studies have examined the effects of indoor environmental quality on patients. Much less work, however, has focused on the impact of environmental quality on staff. While it is recognized that high-paced environments impact both physical and psychological wellbeing, there is still a lack of understanding of the additional impact on cognitive performance in ICU staff due to indoor environmental quality. This study aimed to explore the effects of indoor lighting and thermal conditions in ICUs on mood, alertness, and cognitive performance in nursing staff. This exploratory study aimed to provide a quantitative link between specific environmental parameters and measurable changes in human performance.
Methods: For this pilot study, nurses (n=13) were recruited from one ICU unit and participated for three consecutive, same-shift (day vs night) days. During participation, participants were also labeled as “no float” to ensure they stayed in the ICU during the duration of data collection.
Throughout the study, indoor environmental quality (air temperature, relative humidity, and dewpoint) was continually measured using a commercial environment monitoring device placed at the main nurse’s station in the unit with data sampling occurring every minute. Lighting, cognitive, and wellbeing data was collected during each three-day participation period. Environmental light exposure (light intensity, color temperature, and circadian impact) was captured using a wearable light tracking sensor that sampled every 15 seconds. At approximately hour five and hour 10 of their 12-hour shifts, participants were asked to complete a three-task cognitive battery (abstract matching, digital symbol substitution, and NBACK), a subjective mood questionnaire (Positive and Negative Affect Schedule), and a one-question test of subjective alertness (Stanford Sleepiness Scale) all via iPad. The purpose of this approach, combining objective environmental monitoring with subjective and objective cognitive assessments, enabled a comprehensive analysis of the study variables.
Because of the differences in sample rates among the variables, all data was aggregated to mean and slope within different timeframes (15 minutes, 1 hour, ~5 hours) before the start of each round of the cognitive battery. Within each timeframe, the environmental and lighting variables were standardized and grouped with theoretically meaningful composites representing the environmental and lighting domains. For each domain, we calculated both average exposure levels (mean) and temporal trends (slope) within each time frame. A cognitive performance composite score was also created by standardizing and combining individual test scores for each test period.
Linear mixed effects modeling was employed to analyze the relationship between environmental, lighting, and cognitive variables while accounting for repeated measures within individual participants. Models were fitted using composite predictors, with participant as a random intercept and shift type (day or night), study day (1, 2, or 3), and test window (first test of shift or second test of shift) as fixed-effect controls. The optimal timeframe was identified by examining which timeframe yielded significant composite predictors. Significant composite predictors were then systematically decomposed. This hierarchical approach balanced exploratory flexibility while controlling Type 1 error inflation.
Results: We found that immediate environmental changes, particularly rising air temperature in the 15 minutes before testing, significantly impaired nurses’ cognitive performance. Among the three aggregation timeframes tested, only the 15-minute model yielded a significant composite score predictor, the environmental slope composite score (β = -.30, p < .01). This suggests that immediate pre-test environmental conditions are more strongly associated with cognitive performance than longer-term exposure patterns. Specifically, the significance of the environmental slope composite score suggests that as air temperature, humidity, and dewpoint trend upwards, cognitive performance decreases.
Decomposition analysis of the environmental slope composite revealed air temperature slope (β = -.30, p < .01) and dewpoint slope (β = -.20, p = .03) as primary drivers, while relative humidity slope (β = -.09, p = .36) showed no significant association with the cognitive composite score. The final parsimonious model including only air temperature slope along with the control variables and random effects achieved predictive performance equivalent to the full composite model (ΔAIC = 6.9) with reduced complexity.
The cognitive composite was also decomposed to assess whether lighting and environmental factors impact each cognitive task differently. Digital Symbol Substitution showed the highest environmental sensitivity, with optimal prediction requiring all three environmental slope variables plus the additional composite predictors. The model also showed significantly better predictive ability than the composite-only model (ΔAIC = 6.9, p < .01). These results suggest that processing speed may be the most sensitive to fluctuations in air temperature, relative humidity, and dewpoint. NBACK exhibited a dual sensitivity pattern, with performance being impacted by both lighting trends (light intensity slope: β = -137.68, p < .01) and environmental levels (air temperature mean: β = 182.64, p = .01). This suggests that working memory is sensitive to changes in lighting as well as sustained air temperature. Abstract matching showed no significant associations with lighting or environmental variables across all models.
Conclusions: Despite our small sample size, the findings of this study provide strong evidence that short-term dynamic changes in the ICU environment, especially air temperature, can affect the cognitive performance of healthcare staff. As such, strategies focused on creating stable and controlled environmental conditions, rather than depending on individual resilience, could be a key intervention to support the well-being of nursing staff. Next steps for this project involve using a similar hierarchical modeling approach to understand the impact of environmental and lighting factors on subjective well-being and alertness. Future work should focus on larger sample sizes and different types of medical staff and clinical environments to develop a well-rounded understanding of the impact the work environment has on employees. This can be translated into evidence-based design recommendations for new or renovated hospitals. Overall, this work establishes a foundation for an environmentally focused approach to improving healthcare delivery through evidence-based design recommendations.
Methods: For this pilot study, nurses (n=13) were recruited from one ICU unit and participated for three consecutive, same-shift (day vs night) days. During participation, participants were also labeled as “no float” to ensure they stayed in the ICU during the duration of data collection.
Throughout the study, indoor environmental quality (air temperature, relative humidity, and dewpoint) was continually measured using a commercial environment monitoring device placed at the main nurse’s station in the unit with data sampling occurring every minute. Lighting, cognitive, and wellbeing data was collected during each three-day participation period. Environmental light exposure (light intensity, color temperature, and circadian impact) was captured using a wearable light tracking sensor that sampled every 15 seconds. At approximately hour five and hour 10 of their 12-hour shifts, participants were asked to complete a three-task cognitive battery (abstract matching, digital symbol substitution, and NBACK), a subjective mood questionnaire (Positive and Negative Affect Schedule), and a one-question test of subjective alertness (Stanford Sleepiness Scale) all via iPad. The purpose of this approach, combining objective environmental monitoring with subjective and objective cognitive assessments, enabled a comprehensive analysis of the study variables.
Because of the differences in sample rates among the variables, all data was aggregated to mean and slope within different timeframes (15 minutes, 1 hour, ~5 hours) before the start of each round of the cognitive battery. Within each timeframe, the environmental and lighting variables were standardized and grouped with theoretically meaningful composites representing the environmental and lighting domains. For each domain, we calculated both average exposure levels (mean) and temporal trends (slope) within each time frame. A cognitive performance composite score was also created by standardizing and combining individual test scores for each test period.
Linear mixed effects modeling was employed to analyze the relationship between environmental, lighting, and cognitive variables while accounting for repeated measures within individual participants. Models were fitted using composite predictors, with participant as a random intercept and shift type (day or night), study day (1, 2, or 3), and test window (first test of shift or second test of shift) as fixed-effect controls. The optimal timeframe was identified by examining which timeframe yielded significant composite predictors. Significant composite predictors were then systematically decomposed. This hierarchical approach balanced exploratory flexibility while controlling Type 1 error inflation.
Results: We found that immediate environmental changes, particularly rising air temperature in the 15 minutes before testing, significantly impaired nurses’ cognitive performance. Among the three aggregation timeframes tested, only the 15-minute model yielded a significant composite score predictor, the environmental slope composite score (β = -.30, p < .01). This suggests that immediate pre-test environmental conditions are more strongly associated with cognitive performance than longer-term exposure patterns. Specifically, the significance of the environmental slope composite score suggests that as air temperature, humidity, and dewpoint trend upwards, cognitive performance decreases.
Decomposition analysis of the environmental slope composite revealed air temperature slope (β = -.30, p < .01) and dewpoint slope (β = -.20, p = .03) as primary drivers, while relative humidity slope (β = -.09, p = .36) showed no significant association with the cognitive composite score. The final parsimonious model including only air temperature slope along with the control variables and random effects achieved predictive performance equivalent to the full composite model (ΔAIC = 6.9) with reduced complexity.
The cognitive composite was also decomposed to assess whether lighting and environmental factors impact each cognitive task differently. Digital Symbol Substitution showed the highest environmental sensitivity, with optimal prediction requiring all three environmental slope variables plus the additional composite predictors. The model also showed significantly better predictive ability than the composite-only model (ΔAIC = 6.9, p < .01). These results suggest that processing speed may be the most sensitive to fluctuations in air temperature, relative humidity, and dewpoint. NBACK exhibited a dual sensitivity pattern, with performance being impacted by both lighting trends (light intensity slope: β = -137.68, p < .01) and environmental levels (air temperature mean: β = 182.64, p = .01). This suggests that working memory is sensitive to changes in lighting as well as sustained air temperature. Abstract matching showed no significant associations with lighting or environmental variables across all models.
Conclusions: Despite our small sample size, the findings of this study provide strong evidence that short-term dynamic changes in the ICU environment, especially air temperature, can affect the cognitive performance of healthcare staff. As such, strategies focused on creating stable and controlled environmental conditions, rather than depending on individual resilience, could be a key intervention to support the well-being of nursing staff. Next steps for this project involve using a similar hierarchical modeling approach to understand the impact of environmental and lighting factors on subjective well-being and alertness. Future work should focus on larger sample sizes and different types of medical staff and clinical environments to develop a well-rounded understanding of the impact the work environment has on employees. This can be translated into evidence-based design recommendations for new or renovated hospitals. Overall, this work establishes a foundation for an environmentally focused approach to improving healthcare delivery through evidence-based design recommendations.
Event Type
Poster Presentation
TimeTuesday, March 244:45pm - 6:15pm EDT
LocationRhinelander Gallery
Hospital Environments