Preparing the Sleep Lab for Sound Playback

All testing rooms in the Massachusetts General Hospital sleep laboratory have monitoring equipment to measure brain activity and heart reactions during sleep. For this noise protocol, one room was specially fitted with a surround sound speaker system and a computer programmed to present dynamic replay of the 14 hospital sounds in random order at the rising noise levels .

Selecting Sleepers http://sleep.med.harvard.edu/research/recruitment

Twelve adult subjects were selected from numerous responses to outreach advertisements for research participation (see website above). Each participant was carefully screened for normal hearing and for any health problems that could interfere with sleep. Hospital Institutional Review Board and informed consent procedures were followed as required.

Three Nights in the Sleep Lab

After one quiet adaptation night, sounds were introduced on the two following nights during bouts of stable sleep. Sounds were initiated at 40 dBA and presented every 30 seconds in rising 5 dB increments until an arousal was observed or 70 dBA was reached. On all three nights the subjects were physiologically monitored to indicate and record the depth and stages of their sleep. This monitoring allowed for identification of arousals (transition to a lighter sleep stage or full awakening) that followed immediately upon exposure to the sounds.

Probability of Arousal

Arousal information on all 12 subjects was combined for each of three sleep stages separately: light sleep (NREM2), deep sleep (NREM3), and dream sleep (REM). The percentage of subjects aroused by each of the 14 sounds at each decibel step level was then plotted to form sleep stage specific arousal probability threshold curves. These combined results confirmed that, at levels commonly experienced by patients, the selected hospital sounds significantly disrupted sleep.

Intravenous (IV) alarm and phone signals, designed to be alerting, showed the greatest impact on sleep with between 88% and 94% of subjects aroused at the lowest sound level of 40 dBA. Human voices, for which some awareness during sleep could be adaptive, were also very arousing; between 70% and 75% of subjects were aroused at 40 dBA. This finding is consistent with patient complaints of sleep disturbance by night staff conversations. No differences were seen between responses to conversations conveying good or bad prognoses. Like speech, other sounds which had shifting contours, (Snoring, Towel dispenser, Ice dispenser, Door closing and Toilet flush) produced higher probability of arousal, between 35% and 73% of subjects at 40dBA, than those which shifted less over the 10 second exposure periods. During dream sleep (REM), there was less differentiation of sounds from each other in terms of arousal impact than in light sleep (NREM2) or deep sleep (NREM3). Light sleep is the most relevant for evaluating noise levels because greater time is spent in that sleep stage for the adult population.

Implications:

For Hospital Design and Construction

This project contributes scientific evidence validating the provision of minimum acoustic standards recently established in the 2010 edition of the Guidelines for the Design and Construction of Health Care Facilities. The results also underscore the need for innovation in materials and equipment to improve acoustics in healthcare environments, positively impacting health outcomes through improved sleep, privacy, and communication.

For Acoustic Science

Self-report of satisfaction by the user is a common assessment tool for acoustic appraisal of occupied buildings. However, self-report is not an adequate technique for assessing sleep environments. During sleep, memory is not fully engaged, and subjects often do not accurately remember arousal experiences. Even individuals who are aware of having limited sleep underestimate the deficits in attention that result.

Our evidence-based arousal probability threshold curves offer a clearer picture of vulnerability to sleep disruption from noise. Accurate databases contributing to acoustic assessments of sleep environments are critical for protection of the health and functioning of noise-exposed residents.

For Sleep Medicine

This research provides a methodology for bringing real acoustic elements into the sleep laboratory, using fully monitored subjects to ascertain the probabilities for arousal.

While this project focused on healthcare facilities, disruption of sleep by noise occurs in many environments, such as crowded urban neighborhoods and beneath airplane flight-paths. More than just an annoyance factor, the health impacts over time of disrupted sleep can be serious, including changes in glucose metabolism, elevated blood pressure, increased inflammation, and higher risk of industrial and auto accidents. Clear recognition of the role of noise in disrupting sleep is necessary for policy decisions and code enactments to protect public health.

For Collaboration and Innovation

Reflecting broadly on this innovative sleep and acoustics project, it is clear that solving complex modern problems calls for greater permeability among disciplines. To engender successful cross- disciplinary collaborations, foundations and touch points in knowledge bases and standards of practice should be made more explicit. These encompass critical concepts and heuristics, vocabulary and acronyms, reporting mechanisms, and compensation systems. In addition greater transparency is needed regarding conflicts of interest and sponsorship disclosures, client confidentiality requirements and human subject research protections. There are also subtle, often unspoken, characteristics of professional cultures, including the degrees to which competitive individualism is endorsed, help seeking and error detection are valued, hierarchy can be leveled, excellence defined, and real consensus achieved.