Noise Health Home 

[Download PDF]
Year : 2012  |  Volume : 14  |  Issue : 60  |  Page : 230--236

The acoustic environment of intensive care wards based on long period nocturnal measurements

Hui Xie1, Jian Kang2,  
1 Faculty of Architecture and Urban Planning, Chongqing University; Key Laboratory of New Technology for Construction of Cities in Mountain Area, Ministry of Education, Chongqing University, Chongqing, China; School of Architecture, University of Sheffield, Western Bank, Sheffield, S10 2TN, UK
2 Key Laboratory of New Technology for Construction of Cities in Mountain Area, Ministry of Education, Chongqing University, Chongqing, China; School of Architecture, University of Sheffield, Western Bank, Sheffield, S10 2TN, UK

Correspondence Address:
Jian Kang
School of Architecture, University of Sheffield, Western Bank, Sheffield S10 2TN, UK


The patients in the Intensive Care Units are often exposed to excessive levels of noise and activities. They can suffer from sleep disturbance, especially at night, but they are often too ill to cope with the poor environment. This article investigates the acoustic environment of typical intensive care wards in the UK, based on long period nocturnal measurements, and examines the differences between singlebed and multibed wards, using statistical analysis. It has been shown that the acoustic environment differs significantly every night. There are also significant differences between the noise levels in the singlebed and multibed wards, where acoustic ceilings are present. Despite the similar background noises in both ward types, more intrusive noises tend to originate from the multibed wards, while more extreme sounds are likely to occur in the single wards. The sound levels in the measured wards for each night are in excess of the World Health Organization's (WHO) guide levels by at least 20 dBA, dominantly at the middle frequencies. Although the sound level at night varies less than that in the daytime, the nocturnal acoustic environment is not dependant on any specific time, thus neither the noisiest nor quietest period can be determined. It is expected that the statistical analysis of the collected data will provide essential information for the development of relevant guidelines and noise reduction strategies.

How to cite this article:
Xie H, Kang J. The acoustic environment of intensive care wards based on long period nocturnal measurements.Noise Health 2012;14:230-236

How to cite this URL:
Xie H, Kang J. The acoustic environment of intensive care wards based on long period nocturnal measurements. Noise Health [serial online] 2012 [cited 2020 Jul 9 ];14:230-236
Available from:

Full Text


The Intensive Care Unit (ICU) accommodates patients who are relatively confined to their environment, so that unlike ordinary ward patients, they have little respite from the ambient sounds. The level of noise and activity is also high for these patients, as major healthcare activities are common at all times of day and night. The significant problem of excessive noise in critical care environments has not been adequately addressed, but several international standards have been published to promote excellence in the care of critically ill patients. Those documents have been drafted by various international professional societies, such as, the European Society of Intensive Care Medicine, the American Society of Critical Care Medicine, and the World Federation of Societies of Intensive and Critical Care Medicine.

In the UK, the existing standards mainly relate to buildings, services, equipment, and nursing deployment. [1],[2] As the UK's main planning document, the Health Building Note (HBN) 27 provides guidance on the planning and design of the ICU, in order to facilitate good management and achieve costeffective running costs. [3] Although acoustically absorbent ceilings without microbiological hazards are recommended to reduce the noise in the ICU, there is still a lack of concern for the detailed consideration of acoustic characteristics. The latest Health Technical Memorandum 08-01 [4] does not have detailed acoustic design targets, due to insufficient field acoustic measurements. More importantly, this document is not specifically written for such care environments, hence, the unique features of critical care are not taken into account.

Sleep disturbance at night, is a common problem for intensive care patients. [5] As they tend to be too ill to cope with stress, the World Health Organization (WHO) has advised that the noise level should not exceed 35 dB L Aeq inside most rooms, where patients are treated or observed. The guideline values during the night are only 30 dB L Aeq in terms of sleep disturbance. [6] Unfortunately, many previous studies have shown that the average noise levels, up to 24 or 48 hours, within ICUs, are commonly significantly above the guideline values, ranging from 56 to 80 dBA. [7],[8],[9],[10],[11] Few researchers, however, have measured the acoustic environment in critical care for long periods, such as, over several days, and even fewer studies have identified whether the single-bed wards and multibed wards are statistically different with respect to noise levels. In particular, to establish appropriate standards and guidelines for intensive care environments, there is a recognized need for more data from on-site measurements.

The aim of this article, therefore, is to investigate the acoustic environment of typical intensive care wards based on long period measurements at night; and to examine the differences between singlebed and multibed wards, using statistical analysis, as well as the time variability of the acoustic environment.


Case study site

A typical ICU at the Critical Care Department of the Northern General Hospital in Sheffield was selected as the case study site. Previous studies have been conducted in this unit to systematically investigate the sound field of such a typical critical care environment. [5,12-14] There are two types of wards based on the number of beds, namely singlebed wards and multi-bed wards. Despite the popularity of singlebed wards in the recent hospital design, [15] a large number of multi-bed wards still exist in UK hospitals, especially those constructed in the past two decades. The sound fields of both single-bed and multibed wards are thus worthy of examination and analysis.

As a result of different levels of infection control, there were two types of ceilings in the studied ICU, which were acoustically absorptive and reflective. As measurements were strictly prohibited in the infection control wards without any suspended acoustic ceilings, after the patients settled, all the measurements were taken in the wards with acoustic ceilings. During the experiment period, a total of four single- bed wards and two four-bed wards were funded in the ICU by the local National Health Service (NHS) Foundation Trust. The dimensions of its single-bed and multi-bed wards were approximately 5.5 m × 4.8 m × 3 m and 13 m × 9.5 m × 3 m, as illustrated in [Figure 1]. It is noted that a multi-bed ward was over four times larger than a single-bed ward.{Figure 1}

Measurement procedure

The long period acoustic measurement focused on the actual operating environment of the Critical Care Department, and ISO 18233:2006 [16] was followed in the measurement. The temperature and relative humidity were 25 - 28°C and 50 - 55% respectively.

The measurement included one nine-hour pilot measurement in a single ward extending from approximately 6 p.m. to 3 a.m. of the following morning, and six nonconsecutive nocturnal measurements, for both singlebed and multibed wards, from 11:30 p.m. to 7 a.m. of the following morning. In other words, seven independent measurements were carried out. In order to be representative of the actual healthcare environment, suggestions were taken from the medical team regarding the selection of specific wards and time periods for the measurements, taking into account the varied patient conditions and the usage of medical equipment. Every ward occupied by patients could be selected from the ICU, whenever appropriate.

A precision Type 1 sound level meter 01 dB SOLO was used to record the Aweighted sound pressure level (SPL) for every second, L Aeq-1sec , during the measurement period. For every measurement in the single wards, the microphone was positioned at a height of 1.2 m, between the wall and bed, but much closer to patient's head and well away from the reflecting surfaces. Hence, the measured SPL was a good indication of the patient's noise exposure. The microphone position in the multiple wards, however, varied nightly due to their larger space, but the height was fixed at 1.2 m as in single wards, as displayed in [Figure 2]. For instance, Receiver O1 indicated the microphone location on the first measurement night of multi-bed wards. In particular, Receiver O2 on the second night of multibed ward measurement was initially set at the middle of the ward, but there was a high possibility of blocking the staff movement along the central aisle. O2 was therefore moved 1 m toward the right, away from the middle point.{Figure 2}

Data analysis

A Matlab programme was developed to analyze the measured data using a number of acoustic parameters, for instance, logarithmic average and standard deviation of the noise level over the entire measurement period L Aeq (dBA), maximum noise level L max (dBA), intrusive noise level L 5 (dBA) and L 10 (dBA), median noise level L 50 (dBA), background noise levels L 90 (dBA) and L 95 (dBA), logarithmic average noise level for each 10 minute period L Aeq-10mins (dBA), and the logarithmic average noise level of each 30 minute period L Aeq-30mins (dBA). It is noted that L 5 , L 10 , L 50 , L 90 , and L 95 are all statistical sound level indicators. Ln is the level of noise exceeded for n per cent of the specified measurement period. In other words, if N noise levels are obtained in a time period T with a given time interval and they are sorted in descending order, then L n is the (n*N /100)th noise level in the order. [17]

The SPSS statistical software was adopted to determine if the measured noise levels of single-bed wards differed significantly from the levels of multi-bed wards in terms of L Aeq-10mins , and if there were significant differences among the long period sound levels measured over the six nights. In order to determine whether parametric tests or non- parametric methods should be used, tests of normality for all the collected data were first performed.


Nine-hour pilot measurement in the single ward

The pilot study was originally intended to be a 24-hour measurement in a single ward. Unfortunately, due to the unexpected sound meter system crash, only the data for a nine hour measurement was successfully recovered.

For the continuous nine-hour period, a total of 32,374 values were recorded for L Aeq-1sec . For the sake of clarity of presentation, L Aeq-1sec was converted to L Aeq-10min and L Aeq-30mins' as presented in [Figure 3]. It can be seen that the noise levels were in the range of 45 to 70 dBA, and the daytime sound levels were normally higher than the night levels, as expected.{Figure 3}

The nosiest period over the nine-hour period was found to be between 10 p.m. and 11 p.m. The handover for the night shift was likely to be the main reason, as the number of staff doubled in number during the handover, and more noises, such as talking and moving might have been generated on account of this. The most peaceful environment during the pilot measurement period was after 0:30 a.m.

In order to determine the dominant frequencies in the critical care environment, [Figure 4] shows the unweighted sound levels measured at the 1 / 1 octave band ranging from 125 Hz to 4kHz for each 10 minutes. The sound levels at the selected frequencies varied approximately between 30 and 70dB over the nine-hour period, with the sound level of 1 kHz close to the peak value and 4 kHz down the bottom. It was also suggested from the spectrum analysis that middle frequencies, say 500 Hz and 1 kHz, dominated the noise environment, whereas, the influence of high frequencies was comparatively less.{Figure 4}

Based on the calculations using the Matlab programme and the data of one second interval, logarithmic averages, the corresponding standard deviations of L Aeq , and the unweighted sound pressure level (SPL) at a range of frequencies were obtained, as shown in [Table 1], where the differences between day and night time are also given. All the values are presented in three time periods, for comparison purposes, including the daytime (from 18:00 to 23:00 hours), nighttime (from 23:00 to 03:00 hours), and the whole nine-hour period.

In agreement with several recent case studies, [7],[10],[11],[18],[19],[20],[21] the sound level results obtained in the pilot measurement were all in excess of the WHO guideline levels, by at least 23 dBA. Not surprisingly, the daytime noise levels were over 5 dBA higher than those of the night-time. Probably due to the decreased human activities at night, less SPL variation over time was revealed, as reflected by its smaller standard deviation. With regard to the noise levels at different frequencies, it is interesting to note that both the highest sound levels and the most significant time variations appeared at the middle frequencies (500 Hz and 1 kHz). Moreover, the largest sound level differences between day time and night time were found at 1 kHz, followed by 2 kHz.{Table 1}

Nocturnal measurement in the single-bed and multi- bed wards

Since the pilot measurement confirmed that there was less fluctuation and complexity of the ward acoustic environment at night, the formal long period SPL study focused on the nocturnal measurement for not only single-bed wards, but also for multibed wards. The equivalent sound levels measured for each 10 minute period L Aeq-10mins , over six nights, in the two types of wards are illustrated in [Figure 5]. Although all the measurements commenced at similar nocturnal periods, the temporal patterns of the sound levels differed nightly. Taking the peak noise level as an example, the noisiest time period during Night 2 was around 4:30 a.m., but in contrast 4:30 a.m. was the quietest time on Night 3. This suggests that it is difficult to determine either a common noisy time period or relatively quiet time period in the ICU, even for the same type of ward.{Figure 5}

A number of acoustic indicators, as mentioned above, over six nights, are given in [Table 2] based on onesecond interval data, together with the averaged differences between single-bed and multi-bed wards. Corresponding to the results in the pilot measurement, every night the noise level exceeded the WHO recommended level by a large extent, [6] by more than 20 dBA. Night 1 measured in a single ward was the noisiest night, with a L Aeq of 54.64 dBA, followed by Night 6 in a multiple ward. The maximum sound level L max on each night was at least 75 dBA, intrusive enough to disrupt the patients' sleep. The highest maximum sound level L max was obtained on Night 4 of a multibed ward, although its L Aeq was actually the lowest of the six nights, indicating the uncertain relationship between the equivalent noise level and maximum noise level. With regard to the five statistical sound indicators (L 5 , L 10 , L 50 , L 90 , and L 95 ), the highest values were all found on Night 6, but the lowest values appeared on different night periods.{Table 2}

Comparing the indicator values of single-bed wards with those of multi-bed wards, it is interesting to note that the averaged L Aeq and L max of the three nights measured in the single wards were higher than those of the three nights in the multiple wards, and they were also less varied in terms of the standard deviation. In contrast, the averaged L 5 , L 10 , L 50 , L 90 , and L 95 of single-bed wards were lower than the statistical indicators of multi-bed wards, and more importantly, the differences between the single and multiple wards were gradually decreased according to the order of the above five indicators. These detailed comparisons suggested that the background nocturnal noise level of singlebed wards might be generally similar to that of multibed wards, but more intrusive noises were likely to be generated in the multi-bed wards at night. To explain the higher L Aeq of single-bed wards, a possible reason might be the greater occurrences of some extreme noises inside the singlebed wards, such as, annoying equipment alarms.

Although the differences between single-bed and multi-bed wards are clearly shown in [Table 2], it is important to ensure whether the differences are statistically significant or not, as discussed in the next section.

Statistical analysis for the noise level differences in the single-bed and multi-bed wards

The test of normality on the measured noise levels is important in terms of choosing the correct statistical methods. Both numerical and graphical approaches of assessing normality showed that the long-term sound levels L Aeq-10mins , over six nights, in the single and multiple wards, were not normally distributed. It is therefore evident that nonparametric statistical methods are more appropriate for this type of data.

After the normality tests, first a MannWhitney Test was applied, to analyze the differences between the sound levels measured in the singlebed wards and the multibed wards, in terms of LAeq-10mins . The grouping variables for this specific statistical test were the types of ward, namely, singlebed wards and multibed wards. [Table 3] presents the output of the MannWhitney Test for the sound level differences in the wards. It indicates that the measured long period noise levels in the multibed wards were higher than in the singlebed wards, in terms of rank, as a higher mean rank was found in multibed wards. Not to be confused with the results in [Table 2], it should be noted that, unlike the definition of LAeq , nonparametric methods are based on ranks, and they generally compare medians rather than means, thus, to a certain extent, the mean rank result is similar to L50 in [Table 2]. The results in [Table 3] also suggest that there are statistically significant differences between the noise levels in the single wards and the multiple wards (twotailed significance < 0.01). In other words, the acoustic environment in the ICU is significantly related to the types of ward.{Table 3}

The statistical significance of differences among the sound levels measured over six nights has been assessed using the KruskalWallis Test method, again in terms of L Aeq-10mins . The KruskalWallis Test is assumed to be a logical extension of the MannWhitney Test, as it permits the comparison of three or more independent groups. In this case, the measured nights have been defined as the grouping variables, and thus, six groups have been tested accordingly. As [Table 4] demonstrates, Night 6 and Night 1 have been regarded as the two noisiest nights of the overall six nights, due to their higher mean ranks of measured noise levels in L Aeq-10mins , in accordance with the L eq results shown in [Table 2]. It is further concluded that the sound levels in Critical Care differed significantly every night, regardless of the ward types, as the corresponding significance was obtained at a level lower than 0.01. On the other hand, this also verified the representativeness of the selected six nights and the most likely typical noise data collected for the long period measurements.{Table 4}


As demonstrated in the measurement results, the sound levels at night are not simply dependent on the time, but more relevant to each patient's specific condition, and the irregular and unpredictable activities in the ward, such as, patient admission, transfer, and discharge. Compared with the wards in the daytime, the human activities and environmental interventions at nocturnal wards are reduced to a lower degree. For instance, there are no handovers and no visitors after 11 p.m., and consequently the acoustic environment at night tends to be steadier and less fluctuating, as supported by the pilot measurement. Despite the fact that the remaining activities may exert more notable influence on the nocturnal environment, in terms of triggering the noises concurrently, they can immediately lead to a considerable increase of noise level in a short time.

A good example is the manual turning of patients in the ICU. The nursing staff need to turn the ventilated patients every few hours to lessen the risk of inducing ventilator-associated pneumonia and other related problems. [22] Usually, at least two nurses must carry out the turning of one patient. It is unrealistic to request the nurses to maintain total silence through the process. Apart from staff talking and objects moving, quite often the arterial lines inserted in the patient's body might become loose, which sets off loud warning sounds from the equipment, and thus, the noise level would sharply rise up during patient turning. [12] On the other hand, the frequency and time of patient turning are normally decided by the charge nurses on duty, and thus, the effects of the turning of patients on noise levels might differ on a daily basis, even for the same ward. This once again contributes to the low correlation of the nocturnal acoustic environment with the specific time, in ICU wards.


By utilizing the long period acoustic measurement, it was found that the hospital acoustic environment differed significantly every night, meanwhile, there were significant differences between the noise levels in the single-bed and multi-bed wards, where acoustic ceilings were present. Despite the similar background noises in both types of wards, more intrusive noises tended to originate from the multi-bed wards, while more extreme sounds were likely to occur in the single wards. The sound levels measured in the wards for each night were in excess of the WHO guideline levels by at least 20 dBA, dominantly at the middle frequencies. Although the sound level at night varied less than that in the daytime, the nocturnal acoustic environment was not dependant on any specific time. Thus, neither the noisiest nor quietest period could be determined in the wards. It is expected that the statistical analysis of the collected data will provide essential information for the development of relevant guidelines and noise reduction strategies.


The authors are indebted to the nurses and doctors in the Critical Care Department, Northern General Hospital for their help during the measurements. The financial supports of NIHR NHS Physical Environment Research Programme (B(09)09) and the Visiting Scholar Foundation of Key Laboratory of New Technology for Construction of Cities in Mountain Area In Chongqing University are gratefully acknowledged.


1Department of Health. Guidelines on admission to and discharge from intensive care and high dependency units. London; 1996.
2Intensive Care Society. Standards for Intensive Care Units. London; 1997.
3Department of Health. Health Building Note 27: Intensive Care Unit. London: HMSO; 1993.
4Department of Health Estates and Facilities Division. Specialist services: Health technical memorandum 08-01: Acoustics. London: Stationery Office Books; 2008.
5Xie H, Kang J, Mills GH. Clinical review: The impact of noise on patients' sleep and the effectiveness of noise reduction strategies in intensive care units. Crit Care 2009;13:208.
6Berglund B, Lindvall T, Schwela DH. Guidelines for community noise. Geneva: World Health Organization; 1999.
7Elliott RM, McKinley SM, Eager D. A pilot study of sound levels in an Australian adult general intensive care unit. Noise Health 2010;12:26-36.
8Tsiou C, Eftymiatos D, Theodossopoulou E, Notis P, Kiriakou K. Noise sources and levels in the Evgenidion Hospital intensive care unit. Intensive Care Med 1998;24:845-7.
9Kahn DM, Cook TE, Carlisle GC, Nelson DL, Kramer NR, Millman RP. Identification and modification of environmental noise in an ICU setting. Chest 1998;114:535-40.
10Ryherd EE, Waye KP, Ljungkvist L. Characterizing noise and perceived work environment in a neurological intensive care unit. J Acoust Soc Am 2008;123:747-56.
11MacKenzie DJ, Galbrun L. Noise levels and noise sources in acute care hospital wards. Build Serv Engg Res Technol 2007;28:117-31.
12Xie H, Kang J. The evolving perception of staff on the acoustic environment of Critical Care Units. Inter Noise 2010. Lisbon; 2010.
13Xie H, Kang J. Sound field of typical single-bed hospital wards. Appl Acoust 2012;73:884-92.
14Xie H, Kang J, Mills GH. The sound power levels of typical medical equipment in intensive care units. Acta Acustica United with Acustica. 2012;98:651-8.
15Department of Health. Health Building Notes 4 In-patient Accommodation: Options for Choice. Leeds: NHS Estates, Knowledge and Information Portal; 1997.
16International Organization for Standardization. ISO 18233:2006, Acoustics Application of new measurement methods in building and room acoustics American National Standards Institute; 2007.
17Xie H, Kang J. On the Relationships Between Environmental Noise and Socio-Economic Factors in Greater London. Acta Acustica United with Acustica. 2010;96:472-81.
18Tsara V, Nena E, Serasli E, Vasileiadis V, Matamis D, Christaki P. Noise levels in Greek hospitals. Noise Health 2008;10:110-2.
19Vinodhkumaradithyaa A, Srinivasan M, Ananthalakshmi I, Kumar DP, Rajasekhar RV, Daniel T, et al. Noise levels in a tertiary care hospital. Noise Health 2008;10:113.
20Freedman NS, Gazendam J, Levan L, Pack AI, Schwab RJ. Abnormal sleep/wake cycles and the effect of environmental noise on sleep disruption in the intensive cave unit. Am J Respir Crit Care Med 2001;163:451-7.
21Gabor JY, Cooper AB, Crombach SA, Lee B, Kadikar N, Bettger HE, et al. Contribution of the intensive care unit environment to sleep disruption in mechanically ventilated patients and healthy subjects. Am J Respir Crit Med 2003;167:708-15.
22Padhye NS, Hamlin S, Brazdeikis A, Hanneman SK. Cardiovascular impact of manual and automated turns in ICU. Conference Proceedings of the International Conference of IEEE Engineering in Medicine and Biology Society; 2009; Minneapolis, MN 2009.