Patients and staff in hospitals, including intensive care units (ICUs) can be exposed to high levels of acoustic noise. In many cases those levels can be significantly above the levels recommended by the World Health Organisation (WHO), affecting both patients and staff working on those units. A first step towards reducing the ICU noise is to monitor and analyse noise levels in the ICU. In most studies performed to date, the analysis of noise and noise levels has been done manually, via human
interpretation of raw recorded results. This paper uses singular spectrum analysis (SSA) to process and analyse sound pressure levels (SPLs) recorded in ICUs in hospitals in Spain using a dedicated data logging system. SSA algorithm decomposes time-series, in this case SPL time-series measured over a period of time, into a number of components. Those components can then be analysed and interpreted individually or merged with some other components with similar characteristics to be analysed as a group. This approach reveals some interesting characteristics of the SPL time-series and could potentially help in predicting as well as preventing the high SPLs in the coming periods.
