Noise within hospital wards: a survey method

Noise pollution is one of the most disturbing environmental factors. It directly affects people's lives, especially vulnerable people such as children, old-age, and hospital patients. Hospitals are complex buildings with manifold spaces and facilities where different users live with different sensitivity to noise. Several studies highlight the discomfort that patients, staff, and visitors experience in hospitals due to inadequate noise levels. All this leads to stress, sleep disorders, interference in communication, aggression, and annoyance. Noise exposure in hospitals is due to several sound sources. Each varies throughout the day, and their detection could be challenging. The present study proposes a clustering technique to analyze long-term sound level meter monitoring in an Italian hospital (Pisa, Tuscany). Equivalent short-term levels obtained with an interval time of 0.1 s are used as the database to exploit the occurrences curves of different chunks of the monitoring. Then, through a validated method, the spectra of different kinds of sound sources are reconstructed via the Gaussian Mixture Model. Results are compared with near-field measurements of single sources.