From April 2010 to May 2011, of the 145 (55%) eligible patients screened, 106 (73%) consented to the sleep study. Patients in the study were predominantly African American (63%) and female (51%), with a mean (SD) age of 66 (12) years and a median (interquartile range) length of stay of 4 (3–5) days. Subjects had relatively few comorbidities: 20% had chronic obstructive pulmonary disease; 16%, congestive heart failure; 23%, diabetes; and 36%, hypertension. One-third of subjects reported excessive daytime sleepiness on the Epworth Sleepiness Scale. Forty-eight percent of subjects were classified as “good sleepers” with a global PSQI score of 5 or lower.
For 92 patients with sound data, patient room noise levels over 155 hospital days universally exceeded the WHO recommendations for average noise level (Leq, 48.0 dB [95% CI, 47.2–49.5 dB] vs WHO, 30 dB; P < .001) and peak noise level approached that of a chain saw (Lmax, 80.3 dB [95% CI, 78.5–82.2] vs WHO, 40 dB; P < .001]). While all noise measures were lower between 11 PM to 7 AM, all nights exceeded recommendations for maximum noise level (mean Lmax, 69.7 dB [95% CI, 68.1–71.3 dB]) and 94% of the nights exceeded recommendations for average noise level (mean Leq, 38.2dB [95% CI, 36.9–39.4 dB]). More than 42% of patients (n=39 [42.5%]) reported noise disruptions of sleep, which was also associated with higher maximum noise levels (Lmax, 82.6 vs 78.6 dB; P =.04).
For 85 patients with sleep data, actigraphy data demonstrated that patients slept significantly less in the hospital than their self-reported baseline sleep (314 minutes vs 382 minutes; P=.002). Mean sleep efficiency when hospitalized was also low (73.3% [95% CI, 70.3%–76.4%]), with 52% of actigraphically recorded nights below the normal lower boundary of 80% efficiency for adults.
In multivariate analyses, patients exposed to the loudest tertile of average nighttime noise levels (mean Leq, 50 dB [95% CI, 47.8–52.0 dB]) slept significantly less (−76 minutes [95% CI, −134 to −18 minutes]; P=.01) than patients exposed to the quietest tertile of average nighttime noise (mean Leq, 43.3 dB [95% CI, 41.1–44.8 dB]) (). The most common sources of noise disruption reported by patients were staff conversation (65%), roommates (54%), alarms (42%), intercoms (39%), and pagers (38%).
Figure Predicted values of sleep duration across tertiles of average nighttime noise level from a multivariate linear regression model (85 patients; 117 nights) shows a difference of 76 minutes in sleep duration between quietest and loudest tertile (P=.01). (more ...)