Relating clustered noise data to hospital patient satisfaction

Area of Science:

• Healthcare acoustics
• Machine learning in healthcare

Background:

• Hospital noise is a persistent issue affecting patient experience and satisfaction.
• Conventional noise metrics may not fully capture the complexity of hospital acoustic environments.
• Understanding noise patterns is crucial for improving patient well-being.

Purpose of the Study:

• To apply unsupervised machine learning clustering to analyze hospital acoustic environments.
• To investigate the relationship between noise patterns and patient perception.
• To identify novel acoustic metrics relevant to patient satisfaction.

Main Methods:

• Acoustic measurements were taken in nine patient rooms across three hospital units over 24 hours.
• Unsupervised machine learning (k-means clustering) was used to categorize noise into 'active' and 'non-active' states.
• Patient satisfaction data was correlated with novel acoustic metrics derived from clustering.

Main Results:

• Two distinct noise clusters, 'active' and 'non-active', were identified using k-means clustering.
• Traditional metrics like equivalent sound pressure level (LAeq) did not significantly correlate with patient perception.
• Lower patient satisfaction was significantly associated with higher 'Active Sound Levels', higher 'Total Percent Active', and lower 'Percent Quiet at Night'.

Conclusions:

• Statistical clustering provides deeper insights into hospital acoustic environments than conventional metrics.
• Specific noise patterns, such as increased active noise and reduced nighttime quiet, negatively impact patient satisfaction.
• Novel metrics derived from clustered noise data offer a more accurate assessment of the acoustic impact on patient experience.