Perceived ICU Alarm-Noise Burden, Staff-Reported Health Effects, and Contextual Objective Night-Time Acoustic Monitoring in an Adult Intensive Care Unit

Abstract

Objective:

To examine intensive care unit (ICU) staff-perceived alarm-noise burden and related health, communication, workload, and alarm-response outcomes, contextualized by usual-care night-time acoustic monitoring.

Methods:

We conducted a single-center observational study combining an electronic survey of ICU healthcare professionals with contextual bed-level night-time acoustic monitoring. The survey assessed perceived unit noise, alarm-related health and performance effects, communication interference, alarm-response behavior, alarm-governance perceptions, modified workload, and noise sensitivity. Acoustic outcomes included the equivalent continuous sound level, the maximum sound level, and the time above 80 dBA. Acoustic data were not linked to individual staff responses. Descriptive statistics, Cronbach’s alpha, Spearman correlations, and exploratory multivariable regression were used.

Results:

Among 124 respondents, 83 were nurses, 20 were respiratory therapists, 18 were physicians, and 3 were allied health/other staff. Median perceived unit noise was 7/10 (interquartile range, 5–8), and 67 respondents (54.0%) rated noise as high. Excessive alarms triggered physical symptoms in 70 (56.5%), reduced work performance or motivation in 63 (50.8%), affected concentration or attention in 60 (48.4%), and interfered with communication in 48 (38.7%). Twenty-eight respondents (22.6%) reported missed alarms due to sensory overload.

Conclusion:

ICU alarm noise was perceived as an occupational health and workflow burden. Contextual acoustic monitoring showed measurable night-time sound exposure. Quiet ICU interventions should combine alarm hygiene, escalation governance, workflow redesign, central alarm visibility, and objective acoustic surveillance while preserving patient safety.

KEY MESSAGES

• (1)Intensive care unit (ICU) staff perceived alarm noise as a significant occupational health and workflow burden, with 54.0% rating unit noise as high, 56.5% reporting physical symptoms, and 48.4% reporting reduced concentration or attention.
• (2)Usual-care night-time acoustic monitoring demonstrated sustained sound exposure, with a mean equivalent continuous sound level (LAeq) of 59.57 dBA and a mean maximum sound level (Lmax) of 78.97 dBA. These objective acoustic data were analyzed as contextual environmental measures and were not linked to individual staff responses.
• (3)Quiet ICU practice should integrate alarm hygiene, patient-specific alarm management, clear escalation pathways, continuous central alarm visibility, quiet-period behaviors, and objective acoustic surveillance while maintaining patient safety.

INTRODUCTION

Noise is a persistent environmental stressor in hospitals. It is especially difficult to control in intensive care units (ICUs), where life-support equipment, bedside monitors, ventilators, infusion pumps, communication systems, and staff activity operate continuously.[1,2,3,4] ICU acoustic studies have shown that sound levels in critical care environments frequently remain high throughout the day and night, shaped by equipment alarms, clinical activity, staff communication, and other care-related activities. [1,2,3] Although much hospital-noise research has focused on patient sleep, rest, delirium, and recovery, the same acoustic environment is experienced continuously by healthcare professionals.[1,4]

Clinical alarms are necessary safety signals, but excessive and overlapping alarms may create a competing occupational burden.[5,6] Frequent alarms, nonactionable alerts, and poor signal prioritization may contribute to interruptions, alarm fatigue, and desensitization.[7,8,9,10] Multiple device sounds may also affect communication, task prioritization, and cognitive workload.[11,12,13] This issue is especially relevant in ventilator-rich ICUs, where ventilator alarms occur alongside bedside monitor alarms, infusion-pump alarms, call systems, and routine environmental noise.[14] Therefore, staff experience is often shaped by the overall ICU alarm environment rather than by a single device category.[12,13,14]

The occupational health implications of this alarm environment are important. Environmental noise is associated with both auditory and nonauditory health effects, including annoyance, sleep disturbance, stress responses, and impaired cognitive performance.[15] Among ICU staff, alarm-noise burden may contribute to perceived stress, health symptoms, communication interference, sensory overload, and increased workload.[7,16,17] These effects may also influence care processes because staff must prioritize, verify, delegate, or defer alarm responses while simultaneously providing direct patient care.[10,11,13]

Important gaps remain in understanding ICU alarm noise as a staff-facing occupational health issue. First, many studies describe ICU sound levels or alarm fatigue separately, but fewer integrate staff-reported burden with contextual bed-level objective acoustic monitoring from the same adult ICU. Second, the relationships among perceived alarm noise, symptoms, communication interference, workload, sensory overload, and perceptions of alarm governance remain underreported. Third, there is limited evidence to guide Quiet ICU interventions aimed at reducing acoustic burden while preserving alarm visibility and patient safety.

To address these gaps, this study combined a cross-sectional electronic survey of ICU healthcare professionals with contextual, bed-level, objective acoustic monitoring in the same adult ICU environment. The survey assessed perceived ICU alarm-noise burden, health symptoms, communication interference, workload, alarm-response behavior, sensory overload, and perceptions of alarm governance. The acoustic monitoring component quantified night-time sound exposure during usual-care periods using the equivalent continuous sound level (LAeq), the maximum sound level (Lmax), and the time above 80 dBA. The survey and acoustic datasets were analyzed separately and integrated narratively; individual staff responses were not linked to bed-level sound measurements. This article extends our previous baseline evaluation of cognitive workload, perceived environmental noise, and alarm fatigue among multidisciplinary ICU staff by focusing specifically on ICU alarm noise as an occupational health and Quiet ICU implementation issue.[18]

The study aimed to: (1) describe staff-perceived ICU alarm-noise burden and associated occupational health effects; (2) examine exploratory associations between perceived unit noise and staff-reported symptoms, workload, communication interference, and sensory overload; (3) contextualize staff perceptions using contextual objective night-time acoustic monitoring; and (4) identify implications for Quiet ICU alarm-management interventions.

[Figure 1] presents the conceptual framework guiding this study. The framework positions the ICU alarm-noise environment as a sociotechnical occupational exposure arising from multiple alarm sources, shaped by mediating factors such as alarm frequency, actionability, prioritization, role clarity, patient acuity, and alarm visibility, and linked to staff-reported symptoms, workload, communication interference, sensory overload, and alarm-response behavior.[10,11,12,13]

Figure 1.

Conceptual framework linking ICU alarm-noise environment, staff effects, and system response.

MATERIALS AND METHODS

Study Design and Setting

This was a single-center observational study that combined a cross-sectional electronic survey of ICU healthcare professionals with contextual, usual-care, bed-level objective acoustic monitoring from the same adult medical-surgical ICU at a tertiary hospital in Singapore. The survey was designed as a staff-facing assessment of perceived ICU alarm noise, alarm fatigue, workload, communication effects, self-reported behaviors, sensory overload, perceptions of alarm governance, and health-related symptoms. The acoustic monitoring component was included to provide objective context on the usual-care night-time sound environment in which ICU staff worked.

The survey and acoustic datasets were analyzed separately and integrated narratively. Individual staff survey responses were not linked to patient-, shift-, staff-, or bed-level sound measurements. Therefore, the study examined staff-perceived ICU alarm-noise burden and staff-reported outcomes, using usual-care acoustic monitoring to contextualize the ICU sound environment.

Participants

The target population was ICU healthcare professionals who worked in the adult ICU and were routinely exposed to bedside alarms and ICU environmental noise as part of their clinical duties. Eligible professional groups included nurses, respiratory therapists, physicians, and allied health professionals.

Participants were eligible if they: (1) were healthcare professionals working in the adult ICU during the survey period; (2) had routine exposure to bedside monitor alarms, ventilator alarms, infusion-pump alarms, call systems, communication systems, and ICU environmental noise; (3) were able to complete the electronic questionnaire in English; and (4) voluntarily agreed to participate.

Participants were excluded if they: (1) did not work clinically in the ICU during the study period; (2) were administrative staff, students, temporary observers, or nonclinical personnel without routine bedside alarm exposure; (3) submitted duplicate responses; or (4) provided substantially incomplete survey responses that prevented analysis of the main outcome variables.

A total of 188 ICU healthcare professionals were eligible, and all 188 were invited to participate. Of these, 126 submitted the survey; 2 were excluded for incomplete responses, and 124 were included in the final analysis. The response rate was 66.0%.

Survey Recruitment and Administration

The survey was administered electronically using FormSG between November 2025 and February 2026. The survey link was distributed through departmental email, internal ICU communication channels, and staff briefings. Participation was voluntary, and no financial incentives were provided.

To support survey quality, responses were screened for eligibility, completeness, and potential duplicate submissions before analysis. Substantially incomplete responses were excluded according to the predefined exclusion criteria. Data were analyzed only in aggregate.

Survey Instrument Development and Quality Control

The questionnaire was developed to support local occupational health, alarm management, televentilator monitoring, and Quiet ICU improvement planning. The literature informed survey items on ICU noise, clinical alarm burden, alarm fatigue, workload, and staff communication in critical care environments.[1,7,11,19] Noise-sensitivity concepts were incorporated separately.[20] The survey used the same core baseline questionnaire as our previous evaluation of cognitive workload, perceived environmental noise, and alarm fatigue among multidisciplinary ICU staff, with the present analysis focusing specifically on ICU alarm-noise burden, staff-reported health effects, communication interference, sensory overload, alarm-governance perceptions, and contextual night-time acoustic monitoring.[18]

The survey included both study-specific items and items adapted from established concepts. Workload items were adapted from the National Aeronautics and Space Administration Task Load Index domains.[19] Noise-sensitivity items were informed by established concepts of noise sensitivity.[20] The adapted items were modified for the ICU alarm-noise context and were not scored as fully validated instruments.

The questionnaire was reviewed by ICU clinicians, respiratory therapists, nursing representatives, and research team members for face validity, clarity, clinical relevance, and alignment with the study objectives. The survey was pilot-tested with 10 ICU healthcare professionals before distribution, and the wording was refined based on feedback on item clarity, response options, and estimated completion time. Formal cognitive interviewing was not conducted; however, pilot testing included feedback on item clarity, response options, and estimated completion time. The survey was administered in English, the institution’s working language.

Survey Measures

The survey collected demographic and work-pattern variables, including age group, gender, profession, highest level of education, overall healthcare experience, ICU experience, average patient load, shift pattern, shift duration, workdays per week, and overtime hours.

Survey items assessed perceived alarm burden, perceived average unit noise, alarm-governance processes, self-reported behavior, communication interference, physical symptoms, modified workload, and noise sensitivity. Most perception items used a five-point Likert scale, where higher scores indicated stronger agreement or greater perceived burden. Perceived average unit noise was rated on a 10-point scale.

Workload was assessed using six modified NASA-TLX domains: mental demand, physical demand, temporal demand, perceived performance, effort, and frustration.[19] These items were analyzed as a modified workload domain for exploratory purposes and were not interpreted as the original weighted NASA-TLX score. Each modified workload item was rated on a five-point Likert scale from 1 to 5, with higher scores indicating greater perceived workload burden. The six items were averaged to generate the modified workload-domain score. This score ranged from 1 to 5 and was not calculated using the original weighted NASA-TLX procedure.

Noise-sensitivity items assessed perceived sensitivity to noise, difficulty relaxing in noisy environments, annoyance when noise interfered with work, and adaptation to noise.[20] The physical symptom item asked whether excessive alarms triggered symptoms such as nervousness, headache, or sleep disturbance. Because these symptoms were captured as a combined item, they were analyzed as a composite symptom indicator rather than as separate clinical outcomes. Noise-sensitivity items were rated on a five-point Likert scale from 1 to 5, with higher scores indicating greater perceived noise sensitivity or annoyance. Items were averaged to generate a noise-sensitivity domain score ranging from 1 to 5.

Composite Domain Scores

Composite domain scores were calculated for descriptive and exploratory purposes by averaging relevant items. The alarm-health burden domain included items on reduced performance or motivation, physical symptoms, workflow interruptions, alarm confusion, effects on concentration or attention, and missed alarms due to sensory overload. The modified workload domain included six workload items adapted from the NASA-TLX domains.[19] The noise-impact domain included communication interference, the belief that noise affects patient recovery, and awareness of specific noise sources. The noise-sensitivity domain included four items on perceived sensitivity, difficulty relaxing in noise, annoyance when noise interfered with work, and annoyance in noisy environments.[20] Internal consistency for multi-item domains was assessed using Cronbach’s alpha. Domain scores were treated as exploratory staff-reported measures rather than validated diagnostic scales.

Usual-Care Objective Acoustic Context

To provide objective context for the staff survey findings, usual-care night-time acoustic data from the same adult ICU were summarized descriptively. These data were obtained from bed-level sound monitoring conducted as part of routine ICU care. The Quiet ICU intervention-period data were not included in this article’s main analysis because the present study focused on staff perception and usual-care occupational acoustic context.

Night-time was defined as 22:00 to 06:00 hours. Acoustic monitoring measured only sound pressure levels and did not record speech or other identifiable audio. The contextual acoustic dataset included 107 usual-care bed-nights. The acoustic outcomes summarized were the equivalent continuous sound level, the maximum sound level, and the time above 80 dBA.

These usual-care acoustic data were not linked to individual staff survey responses, staff work schedules, bed assignments, or specific alarm sources. They were used only to describe the objective sound environment of the ICU during routine night-time care.

Statistical Analysis

Categorical variables were summarized as frequencies and percentages. Continuous and ordinal survey responses were summarized using means and standard deviations (SDs) or medians and interquartile ranges (IQRs), as appropriate. For five-point Likert items, top-two-box agreement was defined as a response of 4 or 5. For the 10-point perceived unit-noise item, high perceived noise was defined a priori as a score of 7 or higher.

Distributional normality for continuous variables and composite domain scores was assessed using visual inspection of histograms and quantile–quantile plots, supplemented by the Shapiro–Wilk test. Approximately normally distributed continuous variables were summarized using means and SDs; skewed or ordinal variables were summarized using medians and IQRs.

Internal consistency for multi-item domains was assessed using Cronbach’s alpha. Exploratory bivariate associations between perceived unit noise and survey domain scores were assessed using Spearman correlation coefficients.

Linear regression was used for continuous domain scores, including alarm-health burden, modified workload, noise impact, and noise sensitivity. Logistic regression was used for binary outcomes, including communication interference, physical symptoms, and self-reported missed alarms due to sensory overload. Models were adjusted for age group, profession, ICU experience, shift pattern, overtime hours, and typical patient load.

All analyses were exploratory and hypothesis-generating. P-values were interpreted cautiously and were not used as confirmatory evidence. Missing data were handled using complete-case analysis for each outcome or model. Statistical analyses were performed using Stata/SE version 18.0, StataCorp LLC, College Station, TX, USA. Figures were generated using Stata/SE version 18.0.

Ethical Considerations

This study was approved by the National Healthcare Group Domain Specific Review Board (reference no. 2025/1210). The requirement for written informed consent was waived because the study posed minimal risk, used de-identified staff responses, and analyzed data only in aggregate. Participation was voluntary. No patient images, patient identifiers, or patient-level clinical survey data were collected. Acoustic monitoring measured only sound pressure levels and did not record speech or other identifiable audio. Therefore, patient consent for publication was not applicable.

RESULTS

Participant Characteristics

Respondents were predominantly aged 25–34 years (n = 64, 51.6%) or 35–44 years (n = 40, 32.3%), and most were female (n = 90, 72.6%). Nurses formed the largest professional group (n = 83, 66.9%), followed by respiratory therapists (n = 20, 16.1%), physicians (n = 18, 14.5%), and allied health/other staff (n = 3, 2.4%).

Most respondents held a bachelor’s degree (n = 84, 67.7%). Overall, healthcare experience was most commonly 1–5 years (n = 40, 32.3%), followed by 6–10 years (n = 30, 24.2%) and 11-15 years (n = 27, 21.8%). ICU experience was distributed across 1–3 years (n = 32, 25.8%), 4–6 years (n = 21, 16.9%), 7–10 years (n = 29, 23.4%), and more than 10 years (n = 30, 24.2%). Most respondents worked rotating shifts (n = 82, 66.1%), and 85 respondents (68.5%) reported at least some overtime (Table 1).

Table 1.

Participant Characteristics and Key Work-Pattern Variables (n = 124)

Characteristic	n (%)
Age group
18–24 years	8 (6.5)
25–34 years	64 (51.6)
35–44 years	40 (32.3)
45–54 years	10 (8.1)
55–64 years	2 (1.6)
Gender
Female	90 (72.6)
Male	34 (27.4)
Profession
Nurse	83 (66.9)
Respiratory therapist	20 (16.1)
Physician	18 (14.5)
Allied health/other	3 (2.4)
Highest education level
Diploma	18 (14.5)
Bachelor’s degree	84 (67.7)
Master’s degree	12 (9.7)
Doctoral degree	7 (5.6)
Other	3 (2.4)
Overall healthcare experience
<1 year	2 (1.6)
1–5 years	40 (32.3)
6–10 years	30 (24.2)
11–15 years	27 (21.8)
16–20 years	16 (12.9)
>20 years	9 (7.3)
ICU experience
<1 year	12 (9.7)
1–3 years	32 (25.8)
4–6 years	21 (16.9)
7–10 years	29 (23.4)
>10 years	30 (24.2)
Primary shift pattern
Three rotating shifts	82 (66.1)
Mainly morning/afternoon shifts	31 (25.0)
Mainly night shifts	11 (8.9)
Typical patient load
2 patients	59 (47.6)
3 patients	18 (14.5)
≥4 patients	47 (37.9)
Overtime
No overtime	39 (31.5)
Any overtime	85 (68.5)

Note: Percentages may not total 100% due to rounding.

Usual-Care Night-Time Acoustic Context

Contextual usual-care acoustic monitoring from the same adult ICU included 107 bed-nights during the 22:00 to 06:00 night-time period. Mean night-time equivalent continuous sound level (LAeq) was 59.57 dBA (SD, 4.57), and mean maximum sound level (Lmax) was 78.97 dBA (SD, 6.84) [Figure 2]. Mean time above 80 dBA was 2.92 minutes per night (SD, 5.20). These bed-level acoustic data were not linked to individual staff survey responses, work schedules, bed assignments, or specific alarm sources. However, they provide objective context for the usual-care sound environment in which staff worked.

Figure 2.

Usual-care night-time sound levels in the adult ICU. Mean night-time equivalent continuous sound level (LAeq) and maximum sound level (Lmax) are shown from 107 usual-care bed-nights during the 22:00 to 06:00 monitoring period. Data are presented descriptively and were not linked to individual staff survey responses, staff schedules, bed assignments, or specific alarm sources.

Perceived ICU Noise and Alarm-Related Health Effects

The median perceived average unit noise level was 7 of 10 (IQR, 5-8; mean, 6.78; SD, 1.83). Overall, 67 respondents (54.0%) rated the unit noise level as high (≥ 7/10), and 48 respondents (38.7%) rated it as 8 or higher. Self-rated hearing ability was generally high, with a median score of 8/10 (IQR, 8–10; mean, 8.28; SD, 1.60). No audiometric testing was performed.

Alarm-related health and performance effects were common. Excessive alarms were reported to trigger physical symptoms such as nervousness, headache, or sleep disturbance by 70 respondents (56.5%). Excessive alarms reduced work performance or motivation for 63 respondents (50.8%), and alarm fatigue affected concentration or attention for 60 respondents (48.4%). Alarms from a respondent’s own or neighboring patients frequently interrupted workflow for 55 respondents (44.4%). Alarm confusion was reported by 44 respondents (35.5%). Sensory overload was also reported: 28 respondents (22.6%) agreed that they had missed alarms due to it (Table 2).

Table 2.

Key Perceived Alarm-Noise, Communication, Governance, and Response Outcomes

Outcome	n (%)
High perceived average unit noise (score ≥7/10)	67 (54.0)
Perceived average unit noise score ≥8/10	48 (38.7)
Excessive alarms triggered physical symptoms such as nervousness, headache, or sleep disturbance	70 (56.5)
Excessive alarms reduced work performance or motivation	63 (50.8)
Alarm fatigue affected concentration or attention	60 (48.4)
Alarms from own or neighboring patients frequently interrupted the workflow	55 (44.4)
Noise interfered with communication with colleagues	48 (38.7)
Alarm confusion was reported	44 (35.5)
Missed alarms due to sensory overload	28 (22.6)
Responded quickly and appropriately to alarms	105 (84.7)
Prioritized alarms based on perceived urgency	103 (83.1)
Audible and visual monitor alarms allowed clear assignment of patient, unit, and urgency	92 (74.2)
Alarm limits are regularly adjusted according to the patient’s clinical status	82 (66.1)
Procedural instructions on alarm management are regularly updated and shared	44 (35.5)
Work instructions are regularly updated and shared	43 (34.7)
Believed that noise negatively affected patient recovery	91 (73.4)
Aware of specific sources of noise in the unit	101 (81.5)

Note: For five-point Likert items, values represent top-two-agreement unless otherwise stated.

Communication, Workload, and Perceived Effects on Patients

Noise interfered with communication with colleagues for 48 respondents (38.7%). Most respondents believed that noise negatively affected patient recovery (n = 91, 73.4%) and were aware of specific sources of noise in the unit (n = 101, 81.5%).

Modified workload scores showed a moderate workload burden. Temporal demand was the most prominent workload dimension, with 78 respondents (62.9%) reporting high rushedness or busyness. Mental demand was high among 62 respondents (50.0%), physical demand among 56 (45.2%), effort among 54 (43.5%), and frustration among 56 (45.2%). The mean modified workload-domain score was 3.44 (SD, 0.70), with a median of 3.50 (IQR, 3.00–3.88). Internal consistency was high (Cronbach’s alpha, 0.89).

Self-Reported Alarm-Response Behavior and Alarm-Governance Perceptions

Despite the perceived burden, most respondents self-reported active alarm-management behaviors. A total of 105 respondents (84.7%) agreed that they responded quickly and appropriately to alarms, and 103 respondents (83.1%) reported prioritizing alarms based on perceived urgency. Ninety-two respondents (74.2%) agreed that audible and visual monitor alarms allowed clear assignment of patient, unit, and urgency, and 82 respondents (66.1%) agreed that alarm limits were regularly adjusted according to patient clinical status.

However, governance and procedural communication appeared less consistent. Only 44 respondents (35.5%) agreed that procedural instructions on alarm management were regularly updated and shared with all staff, and 43 respondents (34.7%) agreed that work instructions were regularly updated and shared. When presented with a scenario in which a noncritical alarm occurred while assisting another patient, 59 respondents (47.6%) would respond after finishing the current task, 47 respondents (37.9%) would delegate to another staff member, 10 respondents (8.1%) would respond immediately, and 8 respondents (6.5%) would delay response until more convenient. These findings reflect self-reported alarm-response behavior and were not verified using alarm logs or direct observation.

Self-Reported Missed Alarms Due to Sensory Overload

Twenty-eight respondents (22.6%) reported missed alarms due to sensory overload. Respondents who reported missed alarms due to sensory overload had higher perceived unit-noise scores than those who did not report missed alarms (median, 8.0 [IQR, 7.0–9.0] versus 6.0 [IQR, 5.0–8.0]; P = 0.002). They also had higher alarm-health burden scores and modified workload scores. These findings suggest that self-reported missed alarms due to sensory overload clustered with higher perceived noise burden and workload. However, missed alarms were not verified using alarm logs or direct observation.

Composite Domain Scores and Exploratory Associations

The alarm-health burden domain had a mean score of 3.32 (SD, 0.75), a median score of 3.33 (IQR, 3.00–3.71), and high internal consistency (Cronbach’s alpha, 0.85). The modified workload domain had a mean score of 3.44 (SD, 0.70), a median score of 3.50 (IQR, 3.00–3.88), and Cronbach’s alpha of 0.89. The noise-impact domain had a mean score of 3.77 (SD, 0.61), a median score of 3.67 (IQR, 3.33–4.00), and Cronbach’s alpha of 0.69. The four-item noise-sensitivity domain had a mean score of 3.58 (SD, 0.78), a median score of 3.50 (IQR, 3.00–4.00), and Cronbach’s alpha of 0.84.

In exploratory analysis, higher perceived unit noise showed weak-to-moderate positive correlations with greater alarm-health burden (Spearman’s ρ = 0.51, P < 0.001), higher modified workload (ρ = 0.50, P < 0.001), greater noise impact (ρ = 0.43, P < 0.001), and higher noise sensitivity (ρ = 0.41, P < 0.001). These findings indicate that respondents who rated the ICU as noisier also tended to report higher staff-reported alarm-health burden and workload; however, the association was not large. Composite domain scores and exploratory associations are summarized in Table 3. [Figure 3] summarizes the key staff-reported alarm-noise outcomes and the exploratory weak-to-moderate associations between perceived unit noise and composite domain scores.

Table 3.

Composite Domain Scores and Exploratory Associations with Perceived Unit Noise

Domain	Mean (SD)	Median (IQR)	Cronbach’s alpha	Spearman, ρ with perceived unit noise	P-value
Alarm-health burden	3.32 (0.75)	3.33 (3.00–3.71)	0.85	0.51	<0.001*
Modified workload	3.44 (0.70)	3.50 (3.00–3.88)	0.89	0.50	<0.001*
Noise impact	3.77 (0.61)	3.67 (3.33–4.00)	0.69	0.43	<0.001*
Noise sensitivity	3.58 (0.78)	3.50 (3.00–4.00)	0.84	0.41	<0.001*

Note: Correlations are exploratory associations between perceived average unit noise and each domain score. IQR: interquartile range; SD: standard deviation. *P-value < 0.05.

Figure 3.

Staff-reported alarm-noise burden and exploratory associations with perceived unit noise. The figure summarizes key staff-reported alarm-noise outcomes and exploratory associations between perceived average unit noise and composite domain scores. Associations are cross-sectional, self-reported, and hypothesis-generating; they should not be interpreted as causal effects.

Exploratory Adjusted Associations

In exploratory multivariable models adjusted for age group, profession, ICU experience, shift pattern, overtime hours, and typical patient load, perceived unit-noise score remained positively associated with alarm-health burden, modified workload, noise impact, and noise sensitivity. High perceived unit noise was also associated with communication interference, physical symptoms, and self-reported missed alarms due to sensory overload (Table 4). These analyses were exploratory and should be interpreted as adjusted associations rather than causal effects.

Table 4.

Exploratory Adjusted Associations between Perceived Unit Noise and Staff-Reported Outcomes

Staff-reported outcome and analysis	Adjusted estimate	95% CI	P-value
Alarm-health burden score, per one-point increase in perceived unit-noise score	β = 0.19	0.12–0.26	<0.001
Modified workload score, per one-point increase in perceived unit-noise score	β = 0.17	0.10–0.24	<0.001
Noise-impact score, per one-point increase in perceived unit-noise score	β = 0.12	0.06–0.18	<0.001
Noise-sensitivity score, per one-point increase in perceived unit-noise score	β = 0.14	0.07–0.21	<0.001
Communication interference, high perceived unit noise versus lower perceived unit noise	OR = 2.15	1.02–4.54	0.044
Physical symptoms, high perceived unit noise versus lower perceived unit noise	OR = 2.48	1.18–5.23	0.017
Missed alarms due to sensory overload, high perceived unit noise versus lower perceived unit noise	OR = 2.72	1.08–6.86	0.034

Note: Two exposure specifications were used. For continuous staff-reported domain outcomes, perceived unit noise was modeled as a continuous exposure, and β represents the adjusted mean change in domain score per one-point increase in perceived average unit noise on the 0–10 scale. For binary staff-reported outcomes, perceived unit noise was modeled as a binary exposure, with high perceived unit noise defined a priori as a score of 7 or higher and lower perceived unit noise defined as a score below 7. Linear regression was used for continuous-domain scores, and logistic regression for binary outcomes. Models were adjusted for age group, profession, ICU experience, shift pattern, overtime hours, and typical patient load. Estimates represent exploratory adjusted associations and should not be interpreted as causal effects. CI: confidence interval; OR: odds ratio.

DISCUSSION

Principal Findings

This single-center observational study found that ICU healthcare professionals perceived the ICU alarm-noise environment as an important occupational health and workflow burden. More than half of respondents rated average unit noise as high and reported physical symptoms related to excessive alarms. Approximately half reported reduced work performance or motivation and reduced concentration or attention. Communication interference, workflow interruptions, confusion from alarms, and self-reported missed alarms due to sensory overload were also reported.

These staff-reported findings were contextualized using usual-care night-time acoustic monitoring from the same adult ICU. Across 107 usual-care bed-nights, the mean night-time equivalent continuous sound level was 59.57 dBA, the mean maximum sound level was 78.97 dBA, and the mean time above 80 dBA was 2.92 minutes per night. Because these acoustic data were not linked to individual staff responses, staff schedules, bed assignments, or specific alarm sources, they should be interpreted as contextual information describing the usual-care ICU sound environment rather than individual occupational exposure measurements.

Interpretation and Significance

The findings support the view that ICU alarm noise is not only a patient comfort or sleep issue but also a staff-facing occupational health concern. Previous studies have shown that ICU sound levels often remain elevated during both day and night and are influenced by alarms, clinical activity, equipment, and communication.[1,2,3,4] In this environment, bedside monitor alarms, ventilator alarms, infusion pump alarms, call systems, communication systems, and routine clinical activities occur simultaneously. Although clinical alarms are necessary safety signals, frequent and overlapping alarms may increase demands on attention, communication, and task prioritization.[11,12,13]

Higher perceived unit-noise scores were associated with higher alarm-health burden, modified workload, noise impact, and noise sensitivity. High perceived unit noise was also associated with communication interference, physical symptoms, and self-reported missed alarms due to sensory overload. These associations were exploratory and cross-sectional, and they should not be interpreted as causal effects. However, they suggest that perceived alarm-noise burden clustered with staff-reported indicators of occupational strain, consistent with broader evidence that environmental noise may contribute to annoyance, stress responses, sleep disturbance, and impaired cognitive performance.[15]

An important finding was that perceived alarm burden coexisted with high self-reported alarm-response behavior. Most respondents reported responding quickly and appropriately to alarms and prioritizing alarms based on perceived urgency. Therefore, the issue should not be framed as a simple failure of individual vigilance. Rather, it reflects a system-level challenge in which staff are expected to maintain alarm responsiveness while managing multiple patients, competing clinical tasks, and overlapping device alerts. This interpretation is consistent with the alarm-management literature, which emphasizes alarm fatigue as a workflow and safety issue rather than solely an individual behavior problem.[5,6,7,8]

Alarm Governance and Quiet ICU Implications

The survey identified a potential alarm-governance gap. Fewer than 40% of respondents agreed that alarm-management procedures or work instructions were regularly updated and shared with staff. In a complex ICU environment, alarm safety depends not only on alarm audibility but also on shared expectations for patient-specific alarm-limit adjustment, prioritization, escalation, delegation, and team-based ownership.[5,10,11,12]

The finding that 22.6% of respondents reported missed alarms due to sensory overload is clinically important. This outcome was self-reported and was not verified using alarm logs, direct observation, or response-time data; therefore, it should not be interpreted as an objective missed-alarm rate. However, it may represent a potential occupational health and patient safety signal, particularly because alarm fatigue and excessive nonactionable alarms are recognized safety concerns in clinical environments.[6,9,13]

These findings support a Quiet ICU approach that combines acoustic reduction, alarm visibility, and workflow governance. Quiet ICU interventions should not indiscriminately reduce sound or silence clinically important alarms. Instead, they should reduce unnecessary auditory burden while preserving alarm recognition, escalation, and patient safety.[3,10,11,12] Potential strategies include regular patient-specific alarm-limit review, reduction of nonactionable alarms, avoidance of duplicate alarm sources, clearer escalation pathways, role-based alarm routing, central alarm visibility, protected quiet periods, quieter night-time communication, and routine acoustic surveillance. Televentilator or central alarm-monitoring systems may support alarm visibility and escalation. However, they should be implemented as part of a broader alarm-management and governance strategy rather than as standalone technological solutions.[10,11,12,14]

Strengths and Limitations

This study included multiple ICU professional groups and focused on staff-facing outcomes that are often underreported in ICU noise research, including physical symptoms, concentration, communication interference, workload, sensory overload, alarm-response behavior, and perceptions of alarm governance. The inclusion of usual-care objective acoustic monitoring from the same ICU provided additional environmental context for the survey findings.

Several limitations should be considered. This was a single-center study in a single adult ICU, and the findings may not generalize to units with different layouts, staffing models, alarm policies, device configurations, or acoustic environments. The survey was cross-sectional and self-reported, so that associations may have been influenced by recall bias, common method variance, response style, perceived stress, or individual noise sensitivity. The acoustic data were not linked to individual staff responses, staff schedules, bed assignments, or specific alarm sources; therefore, the study cannot estimate individual occupational noise exposure or determine which device categories contributed most to staff-reported burden. Alarm-response behavior and missed alarms were self-reported and were not verified using alarm logs, direct observation, or response-time analysis. In addition, the questionnaire included adapted and study-specific items rather than complete validated diagnostic instruments.[19,20] Occupational stress, burnout, chronic medical conditions, and formal audiometric measures were not collected and could not be adjusted for in the exploratory models.

Future studies should link staff-reported outcomes to acoustic measurements, device-specific alarm logs, alarm response times, direct observation, occupational stress or burnout instruments, and patient safety outcomes. Instruments such as the Perceived Stress Scale and Maslach Burnout Inventory may help distinguish noise-specific burden from broader occupational stress and burnout.[21,22]

CONCLUSION

In the adult ICU, healthcare professionals perceived the ICU alarm-noise environment as an occupational health and workflow burden. Perceived alarm-noise burden showed weak-to-moderate exploratory associations with staff-reported symptoms, reduced concentration, communication interference, workload, and sensory overload. Usual-care night-time acoustic monitoring from the same ICU showed sustained sound exposure and intermittent high-level sound events, consistent with previous evidence that ICU sound environments often remain elevated despite the clinical need for rest and communication. However, these acoustic data were not linked to individual staff responses. The findings support framing ICU alarm noise as a sociotechnical occupational health issue that requires alarm hygiene, patient-specific alarm management, escalation clarity, central alarm visibility, quiet-period behaviors, and objective acoustic surveillance. Future studies should link staff outcomes with acoustic measurements, device-specific alarm logs, alarm-response times, occupational stress measures, and patient-safety outcomes to evaluate whether Quiet ICU and televentilator monitoring strategies can reduce harmful alarm-noise burden while preserving alarm visibility and clinical safety.

Availability of Data and Materials

De-identified aggregate survey data may be made available from the corresponding author upon reasonable request, subject to institutional governance requirements.

Author Contributions

Wei Jun Dan Ong: Conceptualization, methodology, investigation, data curation, formal analysis, visualization, project administration, writing—original draft, and writing—review and editing.

Adrian Ujin Yap: Methodology, formal analysis, investigation, validation, supervision, and writing—review and editing.

Woon Hean Keenan Chong: Methodology, validation, supervision, and writing—review and editing.

Lawrence Ace Azul: Resource, Project administration, validation, writing—review and editing.

Nur Rafidah Binte Abdul Gaffor: Project administration, resources, writing—review and editing.

Jia Xin Li: Project administration, resources, writing—review and editing.

Rodel Arce Montero: Project administration, resources, writing—review and editing.

Faheem Ahmed Khan: Conceptualization, resources, supervision, and writing—review and editing.

Ethics Approval and Consent to Participate

This study was approved by the National Healthcare Group Domain Specific Review Board (reference no. 2025/1210). The survey used de-identified staff responses and did not collect patient-level information. The requirement for written informed consent was waived because the evaluation posed minimal risk, used de-identified data, and analyzed responses only in aggregate. Participation was voluntary. No patient images, patient identifiers, or patient-level clinical data were included; therefore, patient consent for publication was not applicable.

Conflicts of Interest

The authors declare no conflicts of interest related to this work.

Funding Statement

This study received no external funding.