Quality of sleep in the intensive care unit (ICU) can be affected by several factors. These can be categorised into demographic issues and those about the ICU environment. In the literature, patients' age and sex are commonly related to sleep disturbances in ICUs. A study conducted by Madrid-Valero et al (2017) to evaluate the incidence and quality of sleep among the adult population found a negative correlation between age and quality of sleep, particularly for women.
The most common ICU environmental factors that have a negative impact on patients' quality of sleep are high light and noise levels, nursing interventions, painful procedures, treatment side effects and the alarms on mechanical ventilators and monitors (Nicholson, et al, 2001; Weinhouse and Schwab, 2006; Erol and Enç, 2009; Boyko et al, 2012; Kamdar, et al, 2012; Wang and Greenberg, 2013; Delaney et al, 2015; Tainter et al, 2016). Others include bad smells, pain, stress and physical restraints (Delaney et al, 2015). Among all factors studied, a high noise level was the greatest cause of poor sleep in critical care units (Williams et al, 2013).
Patients' comorbid illnesses might also contribute to poor sleep in critical care units. Examples of these include headache, neuropathy, heart failure, chronic obstructive pulmonary disease (COPD), peptic ulcer, hyperthyroidism, diabetes and rheumatoid arthritis (Schutte-Rodin et al, 2008).
This research was conducted because several previous studies recommended examining this topic using larger samples and more diverse ICU populations. The National Sleep Foundation (2010) reported that sleep quality differed between ethnic groups. This was supported by the findings of a study by Patel et al (2010). Given this, ICU patients' quality of sleep in Jordan and the Arab population might be different from that of other ethnic groups.
The results of this study may provide scientific evidence that could motivate policymakers in Jordanian hospitals to develop strategies to overcome factors that might disturb patients' sleep and consequently reduce complications associated with sleep disturbances. This might reduce the need for sleep-inducing drugs, the length of hospital stay, care costs and overall ICU mortality.
Understanding and managing patients' sleep disturbances require a multidisciplinary commitment by physicians, nurses and hospital administration.
Aim
The general purpose of this study was to investigate the demographic and environmental factors that might adversely affect the patients' quality of sleep. The hypotheses were: patients' demographic characteristics affect their perceived quality of sleep; and there is a correlation between the ICU environmental factors and the patients' perceived quality of sleep.
Methods
Design
A cross-sectional, correlational design was used. The dependent variable in this study is the perceived quality of sleep, while the independent variables are the participants' demographic characteristics and ICU environmental factors. The demographic data include: age, marital status, sex, educational level, diagnoses, medical history, previous hospital admission and the length of stay in the ICU at the time of data collection.
ICU environmental factors include: pain, nursing interventions, heart monitor alarms, ventilator alarms, intravenous pump alarms, nebulisers, doctors' and nurses' pagers and phones ringing, television, bedside phones, finger pulse oximetry, noise, light, diagnostic tests, medicines administration, vital signs measurement, blood sampling, and talking.
Instruments
A self-reported questionnaire was used. It had three parts: a demographic survey of the participants, the Freedman Quality of Sleep Scale (FQSS), and the Richards-Campbell Sleep Scale (RCSS).
Freedman Quality of Sleep Scale
The FQSS scale collects data about 17 environmental factors that are known to affect ICU patients' sleep (Freedman et al, 2001).
The participants were asked to rate on a scale of 1–10 the extent to which certain factors disturbed their sleep quality (1 indicates no disruption while 10 indicates significant disruption). The ICU environmental factors known to disturb sleep quality are pain, noise, light, nursing interventions, undergoing diagnostic testing (eg a chest X-ray), vital signs measurement, blood sampling, finger pulse oximetry and administration of medication, heart monitor alarms, mechanical ventilator alarms, talking, intravenous pump alarms, nebulisers, doctors' and nurses' pagers and phones ringing, and bedside phones ringing. Television as an environmental factor was not examined since there were no televisions at the ICUs in Princess Basma Teaching Hospital and Maan Governmental Hospital.
Permissions to use FQSS and RCSS were obtained from the original authors of these scales.
Richards-Campbell Sleep Scale
The Richards-Campbell Sleep Scale was used to measure the patients' quality of sleep. This tool was originally developed to evaluate the quality of sleep among critical care patients (Ritmala-Castren et al, 2014). It has been validated against polysomnography recordings in a medical ICU population.
The tool comprises five items with a 100 mm visual analogue scale. It measures the perceived depth of sleep, sleep latency (time taken to fall asleep), number of awakenings, returning to sleep after awakening and sleep quality. The overall mean score of the five items was considered to represent the patients' perception of the overall quality of sleep.
Setting
This study was conducted in two multidisciplinary ICUs at two large governmental educational hospitals in Jordan. The first setting was the ICU at Princess Basma Teaching Hospital in Irbid city. This hospital is the largest government hospital in the northern region and serves more than one million residents. The bed capacity of this hospital is 230 and the ICU capacity is 12. The second setting is the ICU at Maan Governmental Hospital, which is one of the largest hospitals in the south region of Jordan. The bed capacity of this hospital is 151 and the ICU capacity is 13. The ICUs at both hospitals are multidisciplinary and receive medical and surgical patients who are severely ill.
Participant selection
The target population for this study was all ICU patients; the accessible population was patients in the ICUs at Princess Basma Teaching Hospital and Maan Governmental Hospital in Jordan.
The inclusion criteria were: age ≥18 years, Glasgow Coma Scale score of 13 or more, able to hear, write and speak the Arabic language, haemodynamically stable with normal vital signs, and able to consent to participate in the study.
The exclusion criteria were: taking sleeping pills, mechanically ventilation, brain damage or any other neurological problems, psychiatric disorders, chronic sleep problems, hearing problems requiring the use of hearing aids, and blindness (as patients would not be able see the light level in the ICU).
The total number of the patients who were admitted to the ICU at the selected hospitals during February, March and April 2018 and met the inclusion criteria was 194. From this number, 128 eligible patients agreed to participate in the study. Of these, 103 patients were able to complete their participation in the study successfully.
Ethical considerations
Permission to conduct this study was obtained from the scientific research committee in the School of Nursing at the University of Jordan, the University of Jordan ethical committee and the institutional review board committee at the Ministry of Health. All ethical principles were applied.
Data collection
Data were collected using a demographic data sheet, the FQSS and the RCSS. The researchers checked the ICU each day to see if any more eligible patients had become available and could take part.
At the time of their transfer or discharge from the ICU and after the purpose of the study had been explained, the patients who met the inclusion criteria and agreed to participate in the study were asked to sign the consent form and fill out the research questionnaire.
Data analysis
First research hypothesis
To test the first hypothesis, the independent samples t-test was used for sex, previous hospitalisation and treatment with narcotic analgesics. ANOVA was used for marital status, educational level, medical diagnosis, presence of other diseases and length of stay in the ICU at the time of data collection, while Pearson correlation coefficient (r) was used for age.
Second research hypothesis
To test the second hypothesis, the Pearson correlation coefficient (r) was used.
Before data analysis was conducted, preliminary data screening and cleaning of outliers, missing values and skewness was performed. All assumptions for all statistical analyses were tested, considering the level of measurement for all variables. Descriptive analysis, visual inspection of histograms, scatter plots, skewness and kurtosis were tested to ensure the accuracy and to assess the normality of the continuous variables.
Results
Factors influencing the quality of sleep among Jordanian intensive care patients were examined in this study. Participants were selected based on their availability, eligibility and willingness to participate. A total number of 128 patients were eligible to take part; 25 were excluded, leaving 103 patients who completed the questionnaire.
The independent samples t-test and ANOVA test were used to investigate the effect of demographic factors on the patients' quality of sleep. The results of these tests indicated that there were no significant effects for: age, sex, marital status, level of education, medical diagnosis, length of stay in the hospital, previous hospitalisation, and treatment with narcotic analgesics on the patients' quality of sleep.
The results of the independent samples t-test indicated there was no significant difference in the mean quality of sleep at an α level of 0.05 between men and women (t(101)=−1.342; P=0.183). There were no significant differences in the mean quality of sleep between people who had previously been admitted to hospital and those who had not (t(101)=1.508; P=0.135. Also, there was no significant difference in the mean quality of sleep between the subjects who were treated by narcotic analgesics or who were not treated t(101)=0.599; P=0.551.
The results of ANOVA test indicated that there were no significant differences in the mean quality of sleep at an α level of 0.05 for marital status, educational level, medical diagnosis, presence of other diseases and length of stay in the ICU. The results of the Pearson correlation coefficient were not statistically significant at α level of 0.05 (r(101)=0.070; P=0.48; r2=0.005). Findings are shown in Table 1.
| Factor | r | t | F | P value |
|---|---|---|---|---|
| Sex | −1.342 | 0.183 | ||
| Age | 0.070 | 0.48 | ||
| Marital status | 1.04 | 0.37 | ||
| Educational level | 1.00 | 0.39 | ||
| Medical diagnosis | 0.389 | 0.85 | ||
| Other diseases | 0.465 | 0.761 | ||
| Length of stay in the ICU | 2.23 | 0.089 | ||
| Previous hospitalisation | 1.508 | 0.135 | ||
| Treatment by narcotic analgesics | 0.599 | 0.551 |
t: t-test; F: ANOVA; r: Pearson correlation coefficient
The factors that had a significant correlation with the perceived quality of sleep at an α level of 0.05 were noise, light, nursing interventions, vital signs measurement, administration of medications, talking and phones ringing (Table 2). These factors accounted for the variance of 4.5-21%. That means that about 4.5-21% of the variance in the patients' quality of sleep could be predicted using these factors.
| ICU environmental factors | Minimum | Maximum | Mean (SD) | Skewness | R | P value |
|---|---|---|---|---|---|---|
| Pain | 1 | 10 | 6.55 (2.29) | −0.66 | −0.17 | 0.07 |
| Noise | 2 | 10 | 6.75 (1.77) | −0.52 | −0.36 | <0.001* |
| Light | 1 | 10 | 6.37 (2.19) | −0.74 | −0.46 | <0.001* |
| Nursing interventions | 1 | 10 | 5.88 (1.70) | −0.57 | −0.36 | <0.001* |
| Diagnostic testing | 1 | 10 | 5.45 (1.69) | 0.38 | −0.004 | 0.97 |
| Vital signs measurements | 1 | 10 | 5.60 (1.47) | 0.06 | −0.212 | 0.031* |
| Blood sampling | 3 | 10 | 5.74 (1.86) | 0.53 | 0.141 | 0.156 |
| Administration of medications | 1 | 10 | 5.77 (1.61) | 0.07 | −0.27 | 0.007* |
| Heart monitor alarms | 1 | 10 | 5.34 (1.90) | −0.21 | −0.007 | 0.94 |
| Ventilator alarms | 1 | 10 | 4.71 (2.04) | −0.11 | −0.156 | 0.115 |
| Finger pulse oximetry | 1 | 10 | 4.17 (1.76) | 0.39 | −0.116 | 0.243 |
| Talking | 1 | 10 | 6.37 (2.15) | −0.56 | −0.440 | <0.001* |
| Intravenous pump alarms | 1 | 9 | 3.95 (1.98) | 0.43 | −0.099 | 0.321 |
| Nebulisers | 1 | 8 | 2.98 (1.77) | 0.87 | −0.067 | 0.501 |
| Doctors'/nurses' pagers and phones | 1 | 9 | 5.54 (2.77) | −0.55 | −0.308 | 0.002* |
| Bedside phones | 1 | 10 | 5.68 (2.49) | −0.39 | −0.116 | 0.243 |
SD: standard deviation; r: Pearson correlation coefficient
Cohen (1988) suggested that r≤0.10 (r2<0.01) is small, an r of about 0.30 (r2=0.09) is medium and r>0.50 (r2=0.25) is large. Therefore, light and talking have the greatest impact on the quality of patients' sleep.
The other ICU environmental factors did not show significant correlations with the patients' perceived quality of sleep (Table 2).
Discussion
To the best of the authors' knowledge, this is the first study in Jordan that evaluates the effect of both ICU environmental factors and demographic characteristics on patients' perceived quality of sleep in ICUs. In line with the results of previous studies, the findings show that the patients in the ICU experienced poor-quality sleep.
This study found no significant correlations between any of the patients' demographic data with their perceived quality of sleep. There are inconsistencies in the literature over whether demographic characteristics are associated with patients' quality of sleep. For example, the findings of this study are supported by those of a study conducted in Jordan by Alafafsheh (2015) which, similarly, did not find significant correlations between the patients' demographic data and quality of sleep.
In contrast, Bihari et al (2012) found that older patients slept better in the ICU than younger patients. However, the findings of a study by Beck et al (2010) reported that the duration of sleep decreases and the number of awakenings increases with age. The effect of age might be related to the aging process, with older people at a higher risk of circadian rhythm disruption (Nakamura, et al, 2011). Consequently, older people spend more time in the light stages of sleep and less time in deep stages, so experience light, fragmented sleep (Bliwise, 2011).
The lack of statistically significant correlations between the subjects' demographic data and their sleep quality in this study might be attributed to a variety of environmental and non-environmental factors. Examples include illnesses, such as myocardial infarction, pulmonary oedema and COPD, as well as postoperative care (Parthasarathy and Tobin, 2004) in addition to the effects of the ICU environment such as high nocturnal sound and light levels.
The results of this study revealed that several ICU environmental factors have significant correlations with patients' perceived quality of sleep. The results were consistent with the results of previous studies. In the literature, there is an agreement about the most common ICU environmental factors that might affect patients' quality of sleep. Of all of these factors, the high noise level was cited in the literature as contributing the most to poor sleep in critical care units (Williams et al, 2013). The average acceptable level of noise in a hospital, as recommended by the World Health Organization, should not exceed 40 dB during the night (Stafford et al, 2014). The findings of previous studies indicated that peak noise levels in the ICU exceeded 80 dB (Hu et al, 2015). Between 1960 to 2005, noise levels in ICUs increased annually and steadily by 0.38 dB in the daytime and by 0.42 dB during the night time (Busch-Vishniac, 2005). In a study by Stewart et al (2017) patients reported that the following factors were barriers to good sleep in the ICU: nocturnal noise levels (reported by 53.6%), discomfort (33.9%), pain (32.1%), medical and nursing procedures (32%), being attached to medical devices (28.6%), stress and anxiety (26.8%), and nocturnal light levels (23.2%).
In addition to the ICU environmental factors, the results of this study might be explained by most people sleeping better in their natural environment at home). They are adapted to their room, bed, mattress, pillow, bedclothes and usual sounds from different sources. Furthermore, some people perform sleep rituals to induce sleep such as eating snacks, reading, watching television, bathing, closing doors and windows, and darkening their bedrooms. Any alteration in the usual environment or the activities performed before bedtime during hospitalisation might negatively affect patients' ability to fall and stay asleep.
Implications and recommendations for nursing practice
The implications of this study start with increasing nurses' awareness about the importance of the patients' sleep in the ICU and the devastating effects poor sleep can have on health and recovery. Sleep quality assessment for ICU patients, using valid and reliable tools, could be integrated into the routine work of ICU nurses. From this assessment, nurses will be able to identify sleep problems, identify their possible causes, and plan and implement individualised nursing care to overcome them.
Several strategies and guidelines to improve patients' sleep in the ICU can be implemented. The use of earplugs and eye masks could be the most practical and effective approach, since they are not expensive, are easy to use and contribute to better patients' sleep in the sleep-disturbing ICU environment. ICU nurses should be educated about different types of eye masks and earplugs, the ability of different types to reduce sound and light levels, and the possible effects of using these on perceived quality of sleep. ICU nurses and patients should be trained in using eye masks and earplugs. To gain the maximum benefits from using these devices, those that reduce sound and light levels the most should be used.
Limitations of the study
This study has some limitations. First is the use of a self-reported questionnaire (the RCSS) to measure the patients' perceived quality of sleep. However, polysomnography is the recognised gold standard to measure patients' quality of sleep, and this scale has been validated against polysomnography. Another limitation was that the participants were recruited from only two ICUs in two state hospitals. This was done because data were collected by the researcher himself and because of time limitations.
Conclusion
The findings of this study did not support the first hypothesis; there were no correlations between the participants' demographic data and the perceived quality of sleep. On the other hand, the findings supported the second hypothesis; there were significant correlations between some ICU environmental factors and the perceived quality of sleep.
In addition to high noise and light levels, several factors such as nursing interventions, diagnostic testing, medication administration, talking, and doctors' and nurses' pagers and phones ringing were significantly correlated with patients' poor quality of sleep.
Although the ICU environmental factors are important in causing poor sleep, additional issues such as myocardial infarction, pulmonary oedema, COPD and postoperative care may contribute to this problem. Being aware of these factors is important when creating guidelines to control them and consequently improve patients' quality of sleep.
Improving sleep quality in the ICU has the potential to decrease the need for sleeping pills, boost patients' recovery, decrease length of stay in ICU, and reduce care costs.