Complementary Therapies in Medicine (2010) 18, 150159

available at www.sciencedirect.com

journal homepage: www.elsevierhealth.com/journals/ctim

Effects of music on depression and sleep quality in elderly people: A randomised controlled trial
Moon Fai Chan a,, Engle Angela Chan b, Esther Mok b
a

Alice Lee Centre for Nursing Studies, National University of Singapore, Block E3A, Level 3, 7 Engineering Drive 1, Singapore 117574, Singapore b School of Nursing, The Hong Kong Polytechnic University, Hong Kong, China

KEYWORDS
Elderly people; Depression; Sleep quality; Music intervention

Summary Objectives: To determine the effect of music on sleep quality in elderly people. In addition, we aimed to examine if there were effects on vital signs and depression levels. Background: Sleep disturbances and depression are common in older people, and their impacts on this group, along with its conventional treatment, merit our attention as our population ages. Conventional pharmacological methods might result dependence and impairment in psychomotor and cognitive functioning. Listening to music, which is a non-pharmacological method, might reduce depression and promote sleep quality. Design: A randomised controlled study was conducted during the period December 2006 to January 2007. Setting: At participants homes in Hong Kong. Participants: In total, 42 elderly people (21 using music and 21 controls) completed the study after being recruited in one community services centre. Intervention: Participants listened to their choice of music for 30 min per week, for 4 weeks. Main outcome measures: Blood pressure, heart rate, depression levels and sleep quality variables were collected once a week for 4 weeks. Results: In the experimental group, there were statistically signicant reductions in geriatric depression scores and sleep quality at week 4. In the control group, there were no statistically signicant reductions in depression and improvement of sleep quality over the 4 weeks. However, for all the outcome measures, no signicant differences were found between groups over the 4 weeks. Conclusions: The ndings contribute to knowledge about the effectiveness of soft slow music used as an intervention to improve depression and sleep quality in elderly people. Whilst there were no statistical differences between groups, there was some indication that music yielder higher improvement on some of the parameters, which are worthy of further investigation in larger trials. 2010 Elsevier Ltd. All rights reserved.

This study was funded by the School of Nursing of the Hong Kong Polytechnic University (A-PH29). Corresponding author. Tel.: +65 6516 8684. E-mail address: nurcmf@nus.edu.sg (M.F. Chan).

0965-2299/$ see front matter 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.ctim.2010.02.004

Effects of music on depression and sleep quality in elderly people: A randomised controlled trial

151

Contents
Methods.................................................................................................................. Sample size, study design and participants .......................................................................... Selection of music................................................................................................... Measures ............................................................................................................ Data collection procedure ........................................................................................... Ethical considerations ............................................................................................... Data analysis ........................................................................................................ Results ................................................................................................................... Demographic and health history variables ........................................................................... Physiological measures .............................................................................................. Psychological measures.............................................................................................. Discussion ................................................................................................................ Strengths, limitations and future directions ......................................................................... Conclusion ............................................................................................................... Acknowledgements....................................................................................................... References ............................................................................................................... 152 152 152 152 153 154 154 154 154 154 154 155 157 157 157 157

Sleep disturbances and depression are the most common mental disorders reported among the elderly in various countries.14 Complaints of sleep disturbances such as insomnia, sleep fragmentation and daytime sleepiness, are estimated at 35% among the general population,2,3 with an even higher prevalence (50%) in elderly people.4,5 Previous studies have reported that loss of sleep in elderly people is associated with a greater risk of adverse outcomes, including accidents, falls, poor health status and all-cause mortality.68 Depression is another common condition in older people, with 3.815% of older adults having depressive symptoms.9 Studies evaluating subjects with depressive symptoms or depression have suggested that both conditions are associated with poor social skills,10,11 more disability, poorer physical function and greater perception of poor health status. Within the Chinese community, patients typically fail to report a depressed mood to the physician as a symptom.3 The illness beliefs of the depressed Chinese patients in the study reect a focus on their physical symptoms, seldom highlighting their depression as their chief concern. Current management of depression and sleeping mainly focusses on medication. However, medications may have adverse consequences, with physical and psychological effects such as deterioration of emotional and mental condition, and impaired psychomotor abilities and cognitive functioning.8,12,13 Their safety and efcacy for depression and sleep problems in the elderly have not been established. Therefore, non-pharmacological methods that promote a mindbody interaction without side effects should be tested. In addition, listening to music, a less costly but possibly a more feasible intervention, is one of the alternative methods proposed to address this problem.13 Music has many purposes. Its multidimensional nature touches the individuals physical and psychological levels of consciousness.14 Haas et al.15 and Watkins16 suggested that music exerts its effect through the entrainment of body rhythms. Entrainment is dened as the tendency for two oscillating bodies to lock into phase and thus vibrate in harmony.1719 Spintge20 stated that each system in the body has its own preferential rhythm, which synchronises and is superimposed over

the others. Loss of rhythmicity is the characteristic feature of states of high physical or mental strain, such as anxiety or pain. During high states of anxiety, distress and pain, this synchronisation of rhythmicities is lost. Further, adrenaline will be released from the adrenal medulla,21 which increases heart rate, blood pressure and respiratory rate. Campbell17 pointed out that music therapy evokes psychophysiological responses because of its inuence on the limbic system. The fact that perception of music leads to stirring emotional experiences is an indicator that the limbic system is engaged in processing music stimuli22 and that this system is inuenced by music pitch and rhythm. Music, particularly the classical genre, is a complex and dynamic stimulus with aesthetic and emotional meaning and can be useful in reducing anxiety and pain.20 Based on a psychophysiological theory synthesised from the literature, sedative music induces a relaxation and distraction response,23 which reduces activity in the neuroendocrine and sympathetic nervous systems, resulting in decreased pain,13 stress,24 anxiety23,25 and sleep.26 However, some studies found no difference in pain27,28 and anxiety.2931 For vital signs, some studies showed that music can reduce heart rate,18,32 respiratory rate28,33 and blood pressure,32,34 but some studies did not, for example, nding no differences in heart rate,3537 blood pressure28,29,35,37 and respiratory rate.29,36,38 Vink et al.39 provided a systematic review of the effect of music intervention for people with dementia; they concluded that the methodological quality and reporting of the included studies were too poor to draw any useful conclusions. Most recently, Witzke et al.40 provided another review on the effect of music for dementia patients; they concluded that the evidence supports music as a therapeutic nursing intervention that may serve to enhance the quality of life for many clients with Alzheimers dementia. Many studies in the UK4144 have reported similar ndings to those of Witzke et al.40 , and a recent study conduct in the UK by Maratos et al.52 concluded that there was a greater reduction of depression among the elderly in the music group than to those in the standard care group.45 Music, as a vehicle of feeling, can facilitate a nonverbal expression of emotion. It can reach peoples inner feelings without being

152 threatening, and it can be a tool for emotional catharsis. Music as a nursing intervention may be simple to use and is worth trying. Thus, the purpose of this article was to assess the effects of music therapy on sleep quality in elderly people. In addition, we aimed to explore if there were effects on physiological (e.g., blood pressure) and depression levels. The following four null hypotheses were tested: 1. There is no statistically signicant difference on physiological measures between the elderly in the experimental group and those in a control group. 2. There is no statistically signicant difference on psychological measures between the elderly in the experimental group and those in a control group. 3. There are no statistically signicant changes on physiological measures among the four time points for the elderly in each group. 4. There are no statistically signicant changes on psychological measures among the four time points for the elderly in each group.

M.F. Chan et al. included meditative, Chinese classical, western classical and western modern jazz. All were slow and owing pieces, approximately 6080 beats min1 , instrumental, and 30 min in length.14

Measures
The study instrument was divided into three parts: Part 1. Demographic variables: These included age, gender, religion, marital status, educational level, previous experience of listening to music and medical history. Data were collected in week 1 as baseline information for all subjects. Part 2. Physiological parameters: A digital monitor was used to collect systolic blood pressure (SBP), diastolic blood pressure (DBP) and heart rate (HR) for each participant. For subjects in the experimental group, data were recorded before the 30-min music intervention in week 1 and after each intervention in weeks 24. For subjects in the control group, data were recorded once a week for 4 weeks. Part 3. Psychological parameters: The primary outcome was sleep quality, as measured by the Pittsburgh Sleep Quality Index (PQSI).53 The PQSI is a self-rated questionnaire, which assesses sleep quality and disturbances over a period of time. It is divided into several components that assess subjective sleep quality, sleep latency, sleep duration, sleep efciency, sleep disturbances, use of sleeping medication and daytime dysfunction. The sum of scores ranges from 0 to 21, and a score greater than 5 is indicative of poor sleep quality.54 Acceptable measures of internal homogeneity, consistency (testretest reliability) and validity were obtained. A global score greater than 5 yielded a diagnostic sensitivity of 89.6% and specicity of 86.5% (kappa = 0.75, p less than 0.001), with an internal consistency = 0.83 and testretest reliability, r = 0.85, in distinguishing good and poor sleepers. It conveys information about the severity of the subjects problem, and the number of problems present, through a single simple measure. A Chinese version is available and approval was gained.54 For subjects in the experimental group, data were collected at baseline before the music intervention at week 1, and after the music intervention in weeks 24. For subjects in the control group, data were recorded once a week for 1 month, when the researcher(s) visited them. The secondary outcome was depression level, as measured by the Chinese short version Geriatric Depression Score (GDS-15).51 The GDS-15 is one of the most popular measures used in clinical settings, and its focus is on asking how the elderly subjects felt during the previous week. One point is assigned to each question and a summary of all questions yields a total score from 0 to 15, categorised as 02 (normal), 35 (some depressive symptoms) and 615 (depressed). Jongenelis et al.52 have shown that GDS-15 had 92.2% sensitivity, 95.2% specicity and 0.94 internal consistency using Cronbachs alpha coefcient. For subjects in the music group, data were recorded before the 30-min music intervention as baseline data in week 1, and after the music intervention in weeks 24. For subjects in the control group, data were recorded once a week for 1 month.

Methods
Sample size, study design and participants
The design was a randomised controlled study with repeated measures (see Fig. 1), conducted in one community centre in Hong Kong. The chosen centre joined the study voluntarily and shared a similar mission to that of our team, namely that of providing quality services to the elderly in the community. The data were collected between December 2006 and January 2007. Following a prior power calculation, this study was estimated based on the primary outcome measure, the sleep quality scores. A two-sided MannWhitneys U design was chosen for this study to test for a probability of 0.75 to detect a difference between groups, based on the ndings of previous studies.8 The required sample for each group was 21 (the total sample was 42), which could achieve 80% power at a 5% level of signicance.48 Fig. 1 illustrates elderly subject enrolment patterns. According to the information provided by the community centre, there are 78 elderly people who are members. In the study, there were 45 eligible participants, 3 of whom refused to participate and 42 who were randomly assigned to either the control or experimental group using a random number generated by a random digits table.49 In the end, all of them completed the study. The inclusion criteria included both male and female participants at a community services centre in Hong Kong, who were aged 60 or over. Subjects were not eligible if they were deaf; had an altered mental status (e.g., delusions, confusion) or cognitive impairment (inability to understand and follow directions, or inability to read and write); or had had a recent death in the family.

Selection of music
Recent studies have shown that giving participants a choice of music lowered anxiety, promoted relaxation and led to effective treatment.13,31,50 Therefore, the music chosen by the research team, based on several local studies,23,25,31,50

Effects of music on depression and sleep quality in elderly people: A randomised controlled trial

153

Figure 1

Subjects progress through the trial: CONSORT owchart.

Data collection procedure

After the random allocation of participants to groups, they were visited at home by the researcher at week 1 to collect vital signs and psychological outcomes as baseline data. Thereafter, weekly visits were made for 3 weeks to measure

the same outcomes. Three weeks is a recommended period of time for observing sleep patterns,55,56 and the effects of a new intervention on sleep quality.57 Subjects in the experimental group were provided with an MP3 player with earphones to listen to the music of their choice from a selection of soft, slow music without lyrics. Prior to the music

154 listening session, the researchers introduced the different types of music by giving the titles of the music selections and playing a section of that music to the subjects. The subjects were allowed to choose their preferred music at each home visit. The subjects were also taught how to control the music volume. This was done after choosing the preferred music and 5 min before the music intervention. The researchers then asked the subjects to choose the most comfortable place to listen to the music, for example, in their bedroom. Standardised instructions were given to the subjects by the researchers: (1) listen to the music at room temperature; (2) wear comfortable clothes, turn out the lights and close their eyes; (3) sit back or lie in bed, unfold their legs, not think about anything, let their lips go soft and, as they listened, to let the music relax their body from head to toe; (4) play the MP3 at a comfortable volume; (5) do not worry about turning off the music, but just let it play; (6) do not consume any caffeine or sedative medication before the test; (7) do not talk, remain silent whilst listening to the music unless they experienced any discomfort or needed to ask a question, in which case they should raise a hand; (8) avoid watching the remaining time; and (9) avoid any environmental disturbance such as the telephone ringing. The researcher would leave the subjects alone and stay a short distance away so that she or he could be available for any unexpected response. After 45 min, the researcher stopped the music and within 58 min measured the subjects vital signs and psychological data. Participants in the control group were given an uninterrupted rest period, but the same vital signs and psychological data were collected once a week for 4 weeks. All the data collection, including administering the intervention and collecting the data, was carried out by the same researcher.

M.F. Chan et al. rarely distorts the results. By contrast, with non-parametric tests, there is no assumption about the distribution of the data and no transformation is required, so that it retains its original values, thus making interpretation easier. However, the disadvantage of such techniques is their inability to handle multivariate analysis.58 To address hypotheses 1 and 2, the MannWhitney U-test was used to determine whether any statistically signicant differences were found for all outcome variables between groups at each time point. To address hypotheses 3 and 4, the Friedman test was used to test for any statistically signicant changes for each dependent variable among the four time points for each group. Multiple comparisons were performed to compare each pair, for example, baseline vs. week 2, baseline vs. week 3 and baseline vs. week 4, and the level of signicance was set at p < 0.001.

Results
Demographic and health history variables
The demographic and health data for the 42 participants are presented in Table 1. The majority were aged 75 or above (n = 24, 57.1%), 54.8% (n = 23) were female, and 81% (n = 34) were married. Most of the elderly participants had received primary education (n = 31, 73.8%), and more than 60% of them had no religious beliefs (n = 26, 61.9%). More than half of the subjects income came from their children (n = 23, 54.8%). There were 10 subjects (23.8%), who had tried music therapy before. With regard to their health history, about half of them had hypertension (n = 19, 45.2%) and 23.8% had renal disease (n = 10).

Ethical considerations
Approval was obtained from the ethics committees of the university and the study community. A researcher explained the study to potential participants, and written informed consent was obtained beforehand. The subjects personal identity was protected because all data were identied only by case number, and so condentiality was assured. They were given an opportunity to ask questions, and were told that they could withdraw from the study at any point without adverse effects on their subsequent care. All results for this study were reported as aggregates. In addition, if subjects detected any untoward or unanticipated possible unpleasant effects from the music, the intervention was stopped immediately.

Physiological measures
To address hypothesis 1, the MannWhitney U-test was used to determine whether there were statistically signicant differences for all physiological variables between the two groups at each time point. As shown in Table 2, for the baseline and weeks 24, no statistically signicant differences were found for any of the variables between the two groups. To address hypothesis 3, for each group, the Friedman test was used to determine any statistically signicant changes for each physiological variable among the four time points. For the control group, no statistically signicant differences were found for SBP (p = 0.372), DBP (p = 0.073), and HR (p = 0.124). For the experimental group, no statistically signicant reductions were found in SBP (p = 0.165), DBP (p = 0.194) and HR (p = 0.710).

Data analysis Psychological measures

Descriptive statistics were used to describe the groups characteristics. The ShapiroWilk test was used to examine the normality of the physiological and psychological parameters. The results suggested that non-parametric tests were appropriate. In principle, it is believed that parametric tests are more suitable for use with social science data because of their greater power and exibility to handle multivariate questions than non-parametric tests.58 However, it has been shown that the use of such techniques with ordinal data To address hypothesis 2, the MannWhitney U-test was used to determine whether there were statistically signicant differences in subjects depression and sleep quality scores in the two groups at each time point. As shown in Table 2, on both depression and sleep quality, no such differences were found between the experimental and control group at each time point. To address hypothesis 4, for each group, the Friedman test was used to determine any statistically sig-

Effects of music on depression and sleep quality in elderly people: A randomised controlled trial
Table 1 Demographic data of study sample by group. Total (n = 42) Group Control (n = 21) n Age (years) 6064 6569 7074 7579 80+ Gender Male Female Marital status Married Divorced/separated Widow/widower Education level Illiterate Primary Secondary or above Economic status Government support Children support Saving Religious belief No Yes Taoism Christian Buddhist Others Habit of listening music Yes No Diseases Diabetes mellitus (yes) Hypertension (yes) Coronary Disease (yes) Renal disease (yes) 2 8 8 9 15 19 23 34 1 7 11 26 5 18 23 1 26 16 3 5 6 2 10 32 7 19 6 10 (%) (4.9) (19.0) (19.0) (21.4) (35.7) (45.2) (54.8) (81.0) (2.4) (16.6) (26.2) (61.9) (11.9) (42.9) (54.8) (2.3) (61.9) (38.1) (7.1) (11.9) (14.3) (4.8) (23.8) (76.2) (16.7) (45.2) (14.3) (23.8) n 0 6 5 4 6 10 11 18 0 3 6 13 2 11 9 1 13 8 3 2 2 1 3 18 4 8 2 3 (%) (0.0) (28.6) (23.8) (19.0) (28.6) (47.6) (52.4) (85.7) (0.0) (14.3) (28.6) (61.9) (9.5) (52.4) (42.9) (4.8) (61.9) (38.1) (14.3) (9.5) (9.5) (4.8) (14.3) (85.7) (19.1) (38.1) (9.5) (14.3) Experimental (n = 21) n 2 2 3 5 9 9 12 16 1 4 5 13 3 7 14 0 13 8 0 3 4 1 7 14 3 11 4 7 (%) (9.5) (9.5) (14.3) (23.8) (42.9) (42.9) (57.1) (76.2) (4.8) (19.0) (23.8) (61.9) (14.3) (33.3) (66.7) (0.0) (61.9) (38.1) (0.0) (14.3) (19.0) (4.8) (33.3) (66.7) (14.3) (52.4) (19.0) (33.3)

155

Demographic

nicant changes in their depression and sleep quality among the four time points. The control group showed no statistically signicant difference among the four time points for either depression score (p = 0.791) and sleep quality level (p = 0.252). For the experimental group, there was a significant reduction in both psychological outcomes, especially compared with the baseline and week 4, for depression score (p = 0.001) and sleep quality levels (p = 0.001).

Discussion
The ndings contribute to knowledge about the effectiveness of soft slow music used as an intervention to improve

sleep quality and relieve depression for elderly people. Whilst there were no statistical differences between groups, there was some indication that music yielded improvement on some of the parameters, which are worthy of further investigation in larger trials. The benecial effect of music on the psychological aspect for the elderly was demonstrated in the study. Haas et al.15 pointed out that music would evoke a psychophysiological response by affecting the limbic system. In this study, the subjects experiencing lesser depression or having better emotional feelings implied that music stimuli had been processed in the limbic system. Therefore, our ndings support the notion that music stimuli exert a biologically

156
Table 2 Comparison of primary and secondary outcome analyses by group. Group Control (n = 21) Mean (SD) Median [range] Experimental (n = 21) Mean (SD) Median [range] U

M.F. Chan et al.

MannWhitneys U-test

p value

Primary outcome Pittsburgh Sleep Quality Index (PSQI) Baseline 6.1 (3.7) 6.0 [013] Week 2 6.0 (3.0) 6.0 [112] Week 3 5.5 (3.3) 5.0 [112] Week 4 6.0 (3.6) 6.0 [014] 2 Friedman test = 4.091, p = 0.252 Multiple comparisons+ A (p = 0.671), B (p = 0.189), C (p = 0.430) Secondary outcome Geriatric Depression Scale (GDS-15) Baseline 1.8 (1.7) 2.0 [06] Week 2 1.8 (1.8) 1.0 [06] Week 3 1.9 (2.1) 1.0 [06] Week 4 2.0 (2.4) 1.0[07] 2 Friedman test = 1.042, p = 0.791 Multiple comparisons+ A (p = 0.713), B (p = 0.713), C (p = 0.774) Systolic blood pressure (SBP) Baseline 137.7 (15.2) 136.0 [99.0164.0] Week 2 135.9 (15.2) 134.0 [101.0162.0] Week 3 135.0 (14.7) 135.0 [104.0164.0] Week 4 135.0 (13.7) 130.0 [110.0162.0] 2 Friedman test = 3.132, p = 0.372 Multiple comparisons+ A (p = 0.331), B (p = 0.354), C (p = 0.065) Diastolic blood pressure (DBP) Baseline 80.4 (8.3) 81.0 [64.092.0] Week 2 78.5 (8.7) 80.0 [59.090.0] Week 3 76.6 (8.2) 75.0 [63.094.0] Week 4 77.0 (7.6) 77.0[62.088.0] 2 Friedman test = 6.975, p = 0.073 Multiple comparisons+ A (p = 0.239), B (p = 0.013), C (p = 0.040) Heart rate (HR) Baseline 71.0 (8.5) 71.0[54.084.0] Week 2 73.0 (9.1) 74.0 [52.086.0] Week 3 73.2 (8.9) 75.0 [58.088.0] Week 4 74.2 (7.0) 73.0 [59.086.0] 2 Friedman test = 5.762, p = 0.124 Multiple comparisons+ A (p = 0.141), B (p = 0.340), C (p = 0.048)

7.6 (4.0) 6.0 [015] 6.6 (3.2) 6.0 [011] 6.0 (2.9) 5.0 [012] 5.1 (2.6) 5.0 [010] 2 = 27.761, p < 0.001* A (p = 0.031), B (p = 0.040), C (p = 0.001* )

173.00 200.50 194.50 199.00

.230 .612 .510 .582

4.1 (4.0) 3.0 [012] 3.4 (4.2) 2.0 [013] 2.7 (3.6) 1.0 [012] 2.1 (3.0) 1.0 [09] 2 = 27.87, p < 0.001* A (p = 0.022), B (p = 0.002* ), C (p = 0.001* ) 143.6 (18.5) 144.0 [114.0176.0] 143.9 (18.8) 142.0 [108.0182.0] 139.4 (16.7) 132.0 [115.0175.0] 140.4 (18.4) 136.0 [106.0178.0] 2 = 5.099, p = 0.165 A (p = 0.653), B (p < 0.044), C (p = 0.121) 82.1(11.0) 85.0 [60.090.0] 81.4 (10.9) 82.0 [60.0100.0] 78.7 (8.9) 80.0 [67.098.0] 79.4 (6.1) 80.0 [65.090.0] 2 = 4.718, p = 0.194 A (p = 0.519), B (p = 0.096), C (p = 0.223) 78.2 (13.6) 81.0 [53.0101.0] 78.0 (11.3) 80.0 [57.093.0] 75.6 (8.0) 75.0 [59.092.0] 76.6 (9.2) 78.0 [60.091.0] 2 = 1.381, p = 0.710 A (p = 0.837), B (p = 0.334), C (p = 0.304)

146.50 194.00 204.50 210.00

.058 .495 .549 .590

184.50 154.00 201.50 174.50

.365 .094 .632 .246

186.00 180.00 194.50 178.50

.385 .307 .511 .289

137.00 167.50 195.50 184.00

.035 .181 .529 .358

GDS-15, ranging from 0 to 15; the higher the score, the more depressed the subject; +, A: baseline vs. week 2; B: baseline vs. week 3; C: baseline vs. week 4. * Signicant at p < 0.001.

meaningful effect on human behaviour by engaging specic brain functions.17,19,59 Regarding sleep quality, Gerra et al.60 pointed out that sleep quality can be improved by relaxing the body with sedative music, which decreases circulating noradrenaline. Further, Hass et al.15 and Watkins16 indicated that music exerts its effect through the entrainment of body rhythms, but our results show no signicant changes in vital signs in the experimental group. A systematic review conducted by Nilsson et al.30 found that only 6 out of 22 studies (27%) had reported a signicant effect

of music on heart rates and blood pressure; three out of eight studies (38%) reported a decrease in respiratory rate; so our ndings are consistent with other 16 studies. Perhaps music functions more as a driving input, and therefore the effects of music with a slower tempo on elderly peoples vital signs are minor.30 Haas et al.15 pointed out that music evokes psychophysiological responses because of its inuence on the limbic system. The fact that perception of music leads to stirring emotional experiences is an indicator that the limbic system is engaged in processing music stim-

Effects of music on depression and sleep quality in elderly people: A randomised controlled trial uli, and that this system is inuenced by musical pitch and rhythm.59 Thus, peoples emotional reaction to music may occur because the limbic system is the neurophysiological location of emotional states, feelings and sensations. Hoch and Reynolds55 show that music intervention can have an effect on improving elderly peoples sleep quality and reducing depression levels in the community setting.38 However, uncontrolled and unstructured usage of music could create sensory overload and motility rituals in some patients.40 Several contraindications due to obsession with music can be reected as side effects such as like withdrawal and selfisolation. In fact, studies have reported several side effects whilst working with autistic people, nding that music can sometimes hypnotise people with autism into lethargy, making them oblivious of their environment as a result of their inward turning.46,47 What works for one individual may not produce the same effect for others. Therefore, music therapy needs to be customised to the individual undergoing it, as there is no universality of rules.

157

employed to perform the music intervention in a future study. Further, to be aware of the emotional side effects that may occur in some of the elderly after listening to music, having a psychologist work with the researcher to handle this issue is suggested for future studies. Last but not the least, we used non-parametric tests to examine the effect of the music intervention, which did not include the confounding factors in the analysis, for example, the personal background of music or education; hence, we suggest using parametric methods (e.g., analysis of variance (ANOVA) or analysis of covariance (ANCOVA)) to examine the impacts on the outcome measures in future studies.

Conclusion
From the study, it is shown that listening to music may act as an effective intervention to improve sleep quality and allay depression levels in a group of older people. However, due to small sample sizes, we could consider it as a preliminary pilot study, and further study should proceed by recruiting more subjects. However, the current ndings may help health-care professionals build therapeutic relationships with elderly patients, and they are encouraged to use music as part of their holistic caring in a group of elderly people. In practice, health-care professionals can encourage elderly patients to listen to music as an alternative selfcare skill, enabling them to release their negative feelings, improve poor sleep quality and develop a healing process in their daily lives.

Strengths, limitations and future directions

This research had several methodological strengths. The sample was recruited directly from the community. Followup rates were excellent, with 100% of subjects in both the music and control groups, providing data at 1-month intervention. Despite these strengths, the research also had limitations that affected its outcomes. First, our original plan was too optimistic, only requiring 21 samples per group to achieve a signicant large treatment effect or effect size. However, in reality, this study was underpowered and far below our expectation because our actual observed effect size was small. Therefore, this study needed at least 159 samples per group under a two-sided MannWhitneys U design to test for a probability of around .58 of detecting a difference between groups.48 Second, the study could not distinguish whether the reductions in sleep and depression levels in the experimental group were due to the chosen music or the Hawthorne effect. The experimental group subjects may have perceived that listening to the music was a type of therapy. This perception may have contributed to more positive responses reected in both physiological and psychological outcomes. Further, in the Chinese culture, the patients trust and respectful attitudes towards health-care professionals would further increase the magnitude of the Hawthorne effect. Third, the design of the study called for using an MP3 player with earphones for the music intervention. This was to ensure the subjects concentration on listening to the music. However, some of the elderly reported that they did not like the earphones because of discomfort. This may have affected the results, because they were not in the most comfortable condition. Therefore, in future studies, to ensure that they did full the task of listening to the music for the required time or that the participant is listening to music, not leaving the brain blank or thinking, we need to collect their satisfaction level to evaluate this issue. Fourth, our researcher is not a trained music therapist, and some of the elderly reported that our researcher could not fully explain the differences between the different types of music we had provided to them; therefore, we recommend that a music therapist be

Acknowledgements
We thank L.Y.S. Cheung, C.M. Chow, C.Y.T. Chow, J.T.Y. Kwok, W.C. Kwok, S.O. Lai, W.Y. Lai, Y.Y. Lam, J.Y.L. Leung, S.M. Tsang, and J.T.F. Wong for data collection.

References
1. Woo J, Ho SC, Lau J. The prevalence of depressive symptoms and predisposing factors in an elderly Chinese population. Acta Psychiatrica Scandinavia 1994;89(1):813. 2. Wilson RS, Mendes de Leon CF, Bennett DA, Bienias JL, Evans DA. Depressive symptoms and cognitive decline in a community population of older persons. Journal of Neurology, Neurosurgery & Psychiatry 2004;75(1):1269. 3. Yeung A, Chang D, Gresham Jr RL, Nierenberg AA, Fava M. Illness beliefs of depressed Chinese American patients in primary care. Journal of Nervous and Mental Disease 2004;192(4):3247. 4. Mitchell AJ, Subramaniam H. Prognosis of depression in old age compared to middle age: a systematic review of comparative studies. American Journal of Psychiatry 2005;162(9):1588601. 5. Zahran HS, Kobau R, Moriarty DG, Zack MM, Holt J, Donehoo R. MMWR: Morbidity and Mortality Weekly Report 2005;54(SS4):315. 6. Mornhinweg GC, Voignier RR. Music for sleep disturbance in the elderly. Journal of Holistic Nursing 1995;13:24854. 7. Chinese University of Hong Kong [CUHK]. Press release: study shows sleep problems are common in Hong Kong elderly 1999, Hong Kong. 8. Lai HL, Good M. Music improves sleep quality in older adults. Journal of Advanced Nursing 2005;49(3):23444.

158
9. Department of Health and Department of Community Medicine University of Hong Kong [DH & DCMUHK] 20032004. Collaborative project: population health survey, Hong Kong; 2004. 10. Baldwin RC, Graham N, Chiu E. Guidelines on depression in elder people: practicing the evidence. UK: Aartin Dunitz; 2002. 11. Tse WS, Bond AJ. The impact of depression on social skills: a review. Journal of Nervous and Mental Disease 2004;192(4):2608. 12. Chang S, Chen C. Effects of music therapy on womens physiologic measures, anxiety, and satisfaction during cesarean delivery. Research in Nursing & Health 2005;28:45361. 13. McCaffrey R. Music listening, its effects in creating a healing environment. Journal of Psychosocial Nursing 2008;46(10):3944. 14. Nilsson U. The anxiety- and pain-reducing effects of music interventions: a systematic review. Association of Operating Room Nurses Journal 2008;87(4):780807. 15. Haas F, Distenfeld S, Axen K. Effects of perceived musical rhythm on respiratory pattern. Journal of Apply Physiology 1986;6:118591. 16. Watkins GR. Music therapy: proposed physiological mechanisms and clinical implication. Clinical Nurse Specialist 1997;11:4350. 17. Campbell DG. Introduction to the musical brain. 2nd ed. St. Louis, MO: MMB Music; 1986. NYC. 18. Chan MF. Effects of music on patients undergoing C-clamp procedure after percutaneous coronary interventions: a randomized controlled trial. Heart & Lung 2007;36:4319. 19. Thaut MH. Neuropsychological processes in music perception and their relevance in music therapy. In: Unkefer RK, editor. Music therapy in the treatment of adults with mental disorders: theoretical bases and clinical interventions. New York: Schirmer Boods; 1990. p. 322. 20. Spintge R. MusicMedicine (MMB), vol. 2. USA: Music Publishers Inc.; 1996. pp. 313. 21. Smolen D, Topp R, Singer L. The effect of self-selected music during colonoscopy on anxiety, heart rate and blood pressure. Applied Nursing Research 2002;16(2):12636. 22. Goldstein A. Thrills in response to music and other stimuli. Physiological Psychology 1980;8:1269. 23. Arslan S, Ozer N, Ozyurt F. Effect of music on preoperative anxiety in men undergoing urogenital surgery. Australian Journal of Advanced Nursing 2008;26(2):4654. 24. Allen G. Effect of music therapy of stress response to day surgery. Association of Operating Room Nurses Journal 2007;86:6714. 25. Twiss E, Seaver J, McCaffrey R. The effect of music listening on older adults undergoing cardiovascular surgery. Nursing in Critical Care 2006;11(5):22431. 26. Johnson JE. Progressive relaxation and the sleep of older noninstitutionalised women. Applied Nursing Research 1991;4(4):16570. 27. Broscious SK. Music: an intervention for pain during chest tube removal after open heart surgery. American Journal of Critical Care 1999;8(6):4105. 28. Ikonomidou E, Rehnstrm A, Naesh O. Effect of music on vital signs and postoperative pain. Association of Operating Room Nurses Journal 2004;80(2):26978. 29. Ayoub CM, Rizk LB, Yaacoub CI, Gaal D, Kain ZN. Music and ambient operating room noise in patients undergoing spinal anesthesia. Anesthesia Analogy 2005;100(5):13169. 30. Nilsson U, Rawal N, Unosson M. A comparison of intraoperative or postoperative exposure to musica controlled trial of the effects on postoperative pain. Anaesthesia 2003;58(7):699703. 31. Yung PMB, Szeto CK, Lau BWK, Chan MF. The effect of music in managing preoperative stress for Chinese surgical patients

M.F. Chan et al.

in the operating room holding area: a controlled trial. International Journal of Stress Management 2003;10(1):6474. Tse MMY, Chan MF, Benzie IFF. The effect of music therapy on post-operative pain, heart rate, systolic blood pressures and analgesic use following nasal surgery. Journal of Pain & Palliative Care Pharmacotherapy 2005;19(3):219. ONeill O. The efcacy of music therapy on patient recovery in the post-anaesthetic care unit. Journal of Advanced Perioperational Care 2002;1(1):1926. Mok E, Wong KY. Effects of music on patient anxiety. Association of Operating Room Nurses Journal 2003;77(2):396410. Sendelbach SE, Halm MA, Doran KA, Miller EH, Gaillard P. Effects of music therapy on physiological and psychological outcomes for patients undergoing cardiac surgery. Journal of Cardiovascular Nursing 2006;21(3):194200. Masuda T, Miyamoto K, Shimizu K. Effects of music listening on elderly orthopaedic patients during postoperative bed rest. Nordisk Journal of Music Therory 2005;14(1):414. Peng S-M, Koo M, Yu Z-R. Effects of music and essential oil inhalation on cardiac autonomic balance in healthy individuals. The Journal of Alternative and Complementary Medicine 2008;15(1):537. Lee D, Henderson A, Shum D. The effect of music on preprocedure anxiety in Hong Kong Chinese day patients. Journal of Clinical Nursing 2004;13(3):297303. Vink AC, Birks JS, Bruinsma MS, Scholten RJPM. Music therapy for people with dementia (Article No. CD003477). Cochrane Database of Systematic Review 2003;(4). Witzke J, Rhone RA, Backhaus D, Shaver NA. How sweet the sound: research evidence for the use of music in Alzheimers Dementia. Journal of Gerontological Nursing 2008;34(10):4552. Burns S, Harbuz M, Hucklebridge, Bunt L. A pilot study into the therapeutic effects of music therapy at a cancer help center. Alternative Therapies in Health and Medicine 2001;7(1):4856. Bunt L, Burns S, Turton P. Variations on a themethe evolution of a music therapy research programme at the Bristol Cancer Help Centre. British Journal of Music Therapy 2000;14(2):629. Hooper J. Using music to develop peer interaction: an examination of the response of two subjects with a learning disability. British Journal of Learning Disabilities 2002;30: 16670. Koelsch S. A neuroscientic perspective on music therapy. The Neurosciences and Music IIIDisorders and Plasticity: Annual New York Academic Scientic 2009;1169:37484. Maratos A, Gold C, Wang X, Crawford M. Music therapy for depression. Cochrane Database of Systematic Reviews 2008;(1). DOI: 10.1002/14651858.CD004517.pub2. Wigram T, Gold C. Music therapy in the assessment and treatment of autistic spectrum disorder: clinical application and research evidence. Child: Care, Health and Development 2006;32(5):53542. Gomibuchi T, Gomibuchi Km Akiyama T, Tsuda H, Hayakawa T. Obsession of hearing music: from the viewpoint of Morita theory. Psychiatry and Clinical Neurosciences 2000;54: 2036. nQuery Advisior Statistical Solutions, Belfast, North Durbin; 2001. Neave HR. Statistics tables; 1978. London, pp. 6465. Lee OKA, Chung YFL, Chan MF, Chan WM. Effect of music on the physiological responses and anxiety levels of patients receiving mechanical ventilation: a pilot study. Journal of Clinical Nursing 2005;14(5):60920. Sheikh JI, Yesavage JA. Geriatric depression scale (GDS): recent evidence and development of a shorter version. Clinical Gerontology 1986;5:16573.

32.

33.

34. 35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48. 49. 50.

51.

Effects of music on depression and sleep quality in elderly people: A randomised controlled trial
52. Jongenelis K, Pot AM, Eisses AMH, Gerritsen DL, Derksen M, Beekman ATF, et al. Diagnostic accuracy of the original 30item and shortened versions of the Geriatric Depression Scale in nursing home patients. International Journal of Geriatric Psychiatry 2005;20:106774. 53. Daniel JB, Charles FR, Timothy HM, Susan RB, David JK. The Pittsburgh sleep quality index: a new instrument for psychiatric practice and research. Psychiatry Research 1989;28:193213. 54. Chan YK. Sleep behavior among elderly in Hong Kong. The Hong Kong Polytechnic University School of Nursing Master of Science in Nursing, Hong Kong; April 2004. 55. Hoch C, Reynolds C. Sleep disturbances and what to do about them. Geriatric Nursing 1986;7(1):247. 56. Roehrs T, Zorick F, Roth T. Transient and short-term insomnia. In: Dement WC, editor. Principles and practice of sleep medicine. Philadelphia: W.B. Saunders; 1994. p. 48693.

159

57. Wang YW. Effect of acupressure on the sleep disturbance of Taiwanese elderly. Unpublished Doctoral Dissertation, Case Western Reserve University, Cleveland; 1997. 58. Munro BH. Statistical methods for health care research. 5th ed. USA: Lippincott Williams & Wilkins; 2005. Chapter 4, pp. 109135. 59. Guzzetta CE. Music therapy: nursing the music of the soul. In: Campbell D, editor. Music: physician of time to come. Wheaton, IL: Quest Books; 1991. p. 14666. 60. Gerra G, Zaimovic A, Franchini D, Palladino M, Giucastro G, Reali N, et al. Neuroendocrine responses of healthy volunteers to techno-music: relationships with personality traits and emotional state. International Journal of Psychophysiology 1998;28(1):99111.