ASA Lay Language Papers
161st Acoustical Society of America Meeting


Measuring How Acoustical Environments Affect Staff in Health-Care Facilities?

Hind Sbihi - hind.sbihi@ubc.ca
Murray Hodgson - murray.hodgson@ubc.ca
George Astrakianakis - george.astrakianakis@ubc.ca
School of Environmental Health
University of British Columbia
Vancouver, BC, V6T1Z3 Canada

Pamela Ratner - pam.ratner@ubc.ca
School of Nursing
University of British Columbia
Vancouver, BC, V6T1Z3 Canada

Popular version of paper 5aAA6
Presented Friday morning, May 27, 2011
161st Meeting on the Acoustical Society of America, Seattle, WA

Acoustical environments in healthcare facilities pose serious concerns, as they are becoming increasingly poor. Measured noise levels have increased from 57 dB(A) in 1960 to 72 dB(A) in 2004 for hospital day-time exposure, and 42 dB(A) in 1960 to 60 dB(A) in 2004 for night-time exposure (Busch-Vishniac et al. 2005). Noise is potentially hazardous not only to hearing, but also to normal physiological and psychological functioning (Babisch 1998; Westman and Walters 1981). Moreover, a poor acoustical environment can interfere with concentration and clear communication between caregivers, resulting in poor speech privacy and impeding task performance, increasing risk of patient-care errors.
With the increasing magnitude of the exposure, and the demonstrated effects of noise exposure, noise pollution in hospitals is attracting scrutiny. The bulk of the research has studied effects on patients; more investigation on the staff is warranted in light of preliminary findings that showed that the complex hospital soundscape contributes to stress and burnout in staff, a known risk factor for job dissatisfaction and absenteeism (Topf and Dillon 1988). Such outcomes may significantly degrade the work environment for a profession that already has staff-retention problems. For those who stay, long-term exposure to noise has been linked to chronic disease through repeated dysregulation of the normal stress response.

Given the aging demographic trend, which puts more demand on a profession already suffering from shortages, it is of the utmost importance to better understand the impact of environmental stressors such as noise on staff.

Studies characterizing healthcare-setting noise have focused primarily on hospitals. However, much of that research has been devoted to intensive-care units (neonatal, hematological, neurological, etc.), leaving a significant knowledge gap for long-term care units (LTC).

In a local Richmond (BC) LTC, following complaints made by staff, the background noise was measured on all floors; it ranged from 53 to 83 dB(A), failing the requirement of 35 dB(A) suggested by the WHO, and the less stringent 55 dB(A) recommended by the US-EPA.

The objective of our research is to investigate the association between the acoustical characteristics of the working environment and increased stress levels in staff, and to understand which set of acoustical requirements are likely to make the workplace a healthy environment, while ensuring better health-service delivery. The deleterious contribution of noise to stress can be thought of as the cumulative effect of a direct pathway;the noise-induced stress;and an indirect pathway, where stress is mediated by patient physical and/or verbal aggression. Using a multi-perspective approach that combines qualitative and quantitative methods, the research agenda has three components:
(1) to determine what type of healthcare facilities within the local health authority are potentially most in need of investigation;
(2) to consolidate the tools necessary to perform exposure assessments and investigate relevant study outcomes (verbal communication/privacy, stress, violence); and,
(3) to select a study population and perform an ecological cross-sectional study of the acoustical environments and suspected noise-related adverse effects.
The first two components have been completed in two pilot studies, referred to as Phase I. This phase encompassed the acoustical evaluation of, and perceived noise assessment in, four sites from three types of healthcare-delivery settings;acute (AC), community (CHC) and long-term care (LTC);as well as the development of novel acoustical parameters and physiological-stress measures adapted to healthcare (i.e. not intrusive on daily work activities).

Pilot Study #1

The first pilot study in Phase I compared three sites (AC, CHC and LTC) in Vancouver (BC) with 25 staff that participated in our assessment. This evaluation consisted of objective measurements in the form of noise dosimetry, noise monitoring, and of verbal-communication quality, as well as subjective assessment based on a noise-perception scale that we developed. Figure 1 shows the perceived noise effects reported by staff in the different types of healthcare facilities; it indicates that LTC consistently ranked highest for perceived negative effects which ranged from difficulty talking/hearing to distraction and annoyance. Conversely, LTC staff reported that their acoustical environments did not promote positive effects such as productivity and privacy, as these ranked consistently lower than in AC and CHC.


Noise dosimetry on 19 nursing staff revealed a statistically significant higher exposure in LTC nurses (Lex= 75 dB(A)); compared to staff in other types of facility (68.7 and 70 dB(A) in CHC and AC, respectively); traditional measures of noise (Leq using different weighting scales) did not show significant differences. However, Table 1 gives an interesting insight into a novel noise metric describing the acoustical qualities of the working environment, namely Peakiness;the difference between peak and average sound levels;which describes the intermittent nature of the acoustical environment;Peakiness discriminates between the three types of facilities, while the ;traditional; acoustical measures do not.
Table 1 - Objective acoustical evaluation in Pilot Study #1.


Unit

Duration

Leq

Leq, A

Leq,C

Peakiness

Acute care

14 hours

59.9

55.6

62.2

1.7

Long-term care

day

61.4

56.9

63.2

3.5

Long-term care

evening

60.3

55.6

62.2

4.2

Community healthcare (site1)

8 hours

51.4

52.2

63.0

3.9

Community healthcare (site2)

8 hours

52.6

50.5

59.7

2.4

A major finding of Pilot Study #1 was the identification of long-term care (LTC) as the work environment with the poorest acoustical conditions compared to those encountered in acute and community care, based on the selected sample population.

Pilot Study #2

14 nursing staff working either morning, evening or night shift in the long-term-care facility where the initial complaints originated from, were enrolled in the second pilot study, each for a three-day period including one day off work. With the feasibility of the acoustical measurements tested and the noise-perception scale developed, the second pilot study focused on testing physiological measurements of stress and refining the noise-perception scale.

Stress is a complex interaction stemming from physical and psychological responses to external triggers such as noise. The autonomic nervous system provides the body's first protective response to noise stimuli, by immediately releasing epinephrine and norepinephrine, to target the heart and muscles in preparation for eminent danger. An endocrine response may follow by releasing the hormone cortisol as a prolonged reaction to prepare the body for further threats. This inflammatory response, if frequent or prolonged, has been shown to be an increased risk factor for atherosclerosis.

We examined non-invasive methods to sample for noise-induced stress biomarkers: salivary cortisol and heart-rate variability. Study participants were asked to take salivary samples on four pre-determined occasions and to wear wireless Polar; Heart Rate monitors (Polar RS800, Polar Canada), set to datalog heart rate at 5-second intervals for three days, concurrently with salivary sampling. Logged heart-rate data were extracted using Polar; Precision Performance Software and further processed to extract heart-rate variability indices using Kubios HRV software.

Over the duration of the study, different psychological-stress measurement scales were administered. These scales included: (1) a study questionnaire with our Noise Perception Scale, the Maslasch Burnout Inventory scale, and a validated scale for hospital noise;the Disturbance Due to Hospital Noise Scale (DDHNS; Topf et al. 1998); and (2) a daily diary that was filled in every day by workers, to capture more transient stress and affect/mood that could modify the noise-stress relationship as well as aggressive events and behaviours. Work-related stress has been extensively studied in healthcare. In order to disentangle stress due to noise from work stress, we included a validated scale relevant to our study population (Sundin et al. 2007), to adapt it to the job specificities of nurses and nursing assistants. To further control for other stress confounders, the study questionnaire also encompassed a noise-sensitivity scale (WNS-6B; Kishikawa H et al., 2006) and a depression scale (CES-D).

The noise-monitoring results in Figure 2 show that levels in common areas (nursing stations, main lounge) of the long-term-care facility were as high as those on one of the busiest main roads of Vancouver (BC).


We examined the associations between the perceived health effects of the workplace acoustical environments as measured by the Noise Perceived Scale and different acoustical metrics, both traditional and novel (e.g. Occurrence Rate, defined as the proportion of time for which noise is above a pre-established threshold). Table 2 shows only significant associations, and demonstrates that peak measurements and Occurrence Rates above a given noise threshold were the acoustical measures that correlated best with negative perception of noise-related health effects.

Table 2; Significant Spearman rank p-values for acoustical parameters and perceived negative effects of noise.

Annoyance

Distraction

Stress

Fatigue

Tension/ Headache

Hearing

Talking

Leq

0.26

0.40

Leq,A

0.32

Lpeak

0.26

0.27

0.37

0.37

0.34

Occurrence Rate, Lpeak>80 dB

 

0.21

 

 

 

 

 

Occurrence Rate, Lpeak>90 dB

 

 

 

0.37

 

 

 

Occurrence Rate, Lpeak>100 dB

 

0.78

0.40

0.46

0.51

 

 

In assessing the feasibility of the sampling, we concluded that biological markers (salivary cortisol and heart-rate variability) can be collected during an employee's scheduled working hours with minimal disruption to both staff activity and resident routines. Furthermore, these markers are sensitive enough to detect differences between work and non-work days (see Table 3).

The cortisol samples were used to compute the total diurnal output by integrating the area under the curve generated using all 4 samples, this is AUCg. The second cortisol metric is the area under the curve with respect to the slope (AUCi) and indicates the reactivity of the endocrine system.

Heart rate variability can be computed either on the time domain (simple descriptive statistics, e.g. SDANN) or the frequency domain (using Fourier transform of the series of recorded heart beat intervals, e.g. VLF: Very Low Frequency). Frequency analysis of the HRV is more suited to physiological interpretation as the low-frequency power of HRV reflects both sympathetic and parasympathetic activity, and the normalized LF/HF ratio, is considered to mirror the sympathovagal balance (i.e. ability to control stress reactions)

Table 3 – Differences in biomarker of stress levels during work and non-work days among 14 participants.


Biological marker

Work Day
Mean (SD)

Day Off
Mean  (SD)

p-value

AUCg (nmol/L)

4.8 (3.7)

3.4 (2.2)

.007

AUCi (nmol/L)

3.2 (2.4)

2.2 (1.6)

.002

SDANN (ms)

58.7

78.0

.02

VLF (%)

23.2

26.9

.04

LF (%)

65.5

61.7

.04

LF/HF

8.01 (0.9)

6.99 (0.9)

.20

Table 4 - Regression coefficients between acoustical descriptors and cortisol indices.


Acoustical descriptor

Cortisol parameter

Coefficient

p-value

Peakiness

AUCg

-0.41

.06

Occurrence Rate, Leq,C>80 dB

AUCg

-0.13

.05

As seen in Table 4, the endocrine activity is significantly associated with acoustical indices describing the intermittent nature of the sound environment. On the other hand, for the autonomous nervous system monitored using HRV (see Table 5), a smaller SDANN indicates a decrease in parasympathetic axis, which has been shown to relate to fatigue, and a higher ratio LF/HF indicates an increase in stress and tension. These parameters were also significantly associated with an intermittent quality of the facility’s soundscapes.
Table 5 – Regression coefficients between acoustical descriptors and HRV indices.


Acoustical descriptor

HRV index

coefficients

p-value

Leq

SDANN

-1.80

.02

Peakiness

-4.10

.05

Occurrence Rate, Lpeak>80 dB

-1.10

.03

Occurrence Rate, Leq,C>60 dB

-0.46

.02

L10

-1.73

.04

Lpeak

LF/HF

0.92

<.01

Peakiness

1.41

<.01

Occurrence Rate, Lpeak>100 dB

0.6

.01

Of great interest is the fact that the same association between HRV indices that indicate stress, tension and fatigue and acoustical descriptors of intermittence is also found between these acoustical descriptors and stress, tension and fatigue reported in the Noise Perception Scale (Table 2).
This concurrence provides a form of validation for the scale we developed. 
Further validation of the scale was sought by performing an analysis to examine how our scale correlated with the burnout scale (MBI). Associations were found with annoyance, fatigue, difficulty talking and stress.

The last step was to examine the external validity of the burnout measured among the 14 participants by correlating the DDHNS with MBI. Only the Personal Accomplishment subscale of the MBI showed significant association. Given the small sample size of this pilot study, we can conclude that a certain level of external validity (the ability to draw inferences to a larger population) of the scales was obtained.

Conclusion

Decreased productivity in the workplace has an estimated cost of $11 billion, largely attributed to fatigue, loss of concentration and impaired decision making. Research has linked environmental stressors—in particular chronic noise exposure—to these adverse outcomes. However, formal assessment of this causality is needed to guide effective control measures. The planned second phase of the project—an ecological study—will provide such an assessment, by examining the impact of acoustics in 15 LTC sites in the lower mainland of British Columbia. Our Phase 1 pilot studies have allowed the development of tools for measuring both exposures and outcomes of interest in the full study.  Furthermore, the Phase 1 pilot studies have laid the foundation for procedures and approaches for obtaining participation and compliance—in Phase 1, we had 100% participation with no loss to follow up, and only 15% non-adherence to the biomarker sampling protocol.

Ultimately, this study will support the contention that targeted acoustical renovations are necessary to improve healthcare facilities, and create safer and healthier environments for staff, patients and visitors. The sustainability of our health-care system involves the retention of an experienced and healthy workforce, available to provide care.  Unfortunately, health-care workers exhibit some of the highest rates of illness and absenteeism. The results of this study will lead to recommendations for appropriate interventions to improve mental-health outcomes, with the ultimate goal of retaining healthy and productive human resources for a sustainable health-care system.

Acknowledgments

- Staff at UBC Hospital Urgent Care and Purdy Pavilion
- Staff at North Shore Community Center (special thanks to Jeannie Law)
- Staff at Minoru Residence (special thanks to Kathy Wong)
- All nurses who participated in the study
- Maureen Haddock, Vancouver Coastal Health