Contribution of alarm noise to average sound pressure levels in the ICU

Objectives: To explore sound levels, alarm frequencies and the association between alarms and sound levels. Design: A single center observational cross-sectional study. Setting: Four intensive care units. Main outcome measures: Contribution of alarms: red (life threatening), yellow (indicate excess of limits) and blue (technical) to sound pressure levels dB(A) at nursing stations. Results: Mean sound pressure levels differed significantly between day (56.1 ± 5.5), evening (55.1 ± 5.7) and night periods 53.6 ± 5.6; p < 0.01. 175,996 alarms were recorded of which 149,764 (85%) were yellow, 18,080 (10%) were red and 8,152 (5%) were blue. The mean sound levels without alarms (background) is 56.8 dB(A), with only red: 56.0 dB(A), only yellow: 55.6 dB(A), only blue: 56.0 dB(A) and mixed alarms: 56.3 dB(A). Yellow alarms (b = −0.93; 95% CI: −1.26 to −0.6; p < 0.001) were weakly but significantly associated with mean sound levels and lead to a slight decrease in noise level (1 dB), Red alarms (b = −0.3; 95% CI: −1.237 to 0.63; p = 0.52). The R Square of the model with all alarms was 0.01 (standard error of estimate, 6.9; p < 0.001). Conclusions: Sound levels were high during all day-periods. Alarms exceeding limits occurred most frequently. However, the contribution of alarms to sound levels measured at the nursing station is clinically limited.