Study finds staff and visitor voices are major source of excessive noise in hospital ICUs

The major contributor to excessive noise levels in hospital intensive care units (ICUs) is the voices of staff and visitors, a study from the University of Portsmouth has found.

A soundscape analysis conducted in three separate ICUs in hospitals in Iraq found that over 50 percent of the cumulative acoustic noise came from conversations. 

Lead author of the study, Awwab Althahab, from the University’s School of Electrical and Mechanical Engineering, said: “Noise levels in ICUs are much higher than what’s recommended by the World Health Organisation (WHO), which is a growing concern in hospitals worldwide.

“With the advancement of technology and the addition of more equipment in ICUs, the environments have become much louder. 

“These noise levels pose a major risk to human health as they can affect both patients and staff psychologically and physiologically. And previous research has shown that patients often recall noise as a prominent and sometimes distressing aspect of their ICU stay.”

Due to the significant consequences of acoustic noise, the WHO recommends that noise levels in ICUs should not exceed 35 decibels (dBA) during the day and 30 dBA at night, with no peaks exceeding 45 dBA. 

The study developed and deployed an intelligent and cost-effective recording system, which not only measured and monitored the noise levels and recorded sound events but also measured other environmental parameters, including temperature, humidity, and light intensity.

This system was successfully deployed across three ICUs for comprehensive data collection. It was mounted approximately one metre behind a patient bed in two respiratory ICUs and one neonatal ICU. It recorded acoustic data continuously for varying numbers of days, based on the consent provided by the ICU heads.

The authors accurately classified the collected noise events to identify the major noise sources, while the nursing staff helped annotate the classified audio events, such as ventilator alarms, humidifier alarms, patient monitor alarms and oxygen pumps.

Mr Althahab added: “By accurately classifying noise events, we've identified speech as a predominant contributor across all observed ICUs. At the same time, care was taken to intelligently remove speech from the collected data for confidentiality reasons.”

This research showed that noise levels exceeded the WHO standard limits for a significant amount of time. One ICU was considerably louder than the other two - with the minimum level reaching 41dBA and the maximum nearing 82 dBA. 

Mr Althahab explained that his study leverages deep neural networks to not only detect but also classify ICU noise events, offering a more sophisticated and cost-effective solution.

He said: “Traditionally, noise monitoring in ICUs has depended on expensive sound level metres, which fail to distinguish and identify the sources of noise. 

“The precise measurements and comprehensive intelligent analysis achieved by the developed measurement system and acoustic event classification system mark a significant leap forward in understanding and managing ICU noise, as well as serve as a starting point for further research in this area.

“Addressing this problem is crucial for improving patient outcomes, reducing healthcare costs and promoting a more productive environment for healthcare workers.”

Mr Althahab is keen to collaborate with UK hospitals, so he can keep working on interventions in this area. 

The paper is published in Acoustics Australia