Development of the Sleep Quality Scale

J. Sleep Res. (2006) 15, 309 316 Development of the Sleep Quality Scale HYERYEON YI 1, KYUNGRIM SHIN 2 and C H O L S H I N 1,3 1 Institute of Human Genomic Study, College of Medicine, Korea University, Seoul, 2 College of Nursing Science, Ewha Womans University, Seoul and 3 Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Korea University Ansan Hospital, Ansan, Republic of Korea Accepted in revised form 26 June 2006; received 22 December 2005 SUMMARY Sleep quality affects health and the overall quality of life. As the factors that influence sleep quality and their relative importance vary among individuals, a self-report method is essential. Although various questionnaires have been used to assess sleep quality, few all-inclusive assessment scales have been developed. Therefore, the purpose of this study was to develop an instrument for measuring sleep quality and to study its validity and reliability. A Sleep Quality Scale (SQS) was developed using item analysis and factor analysis on items with content validity. SQS, composed of 28 items and six factors, accounted for 62.6% of the total variance. The difference of SQS score between insomniacs and normal subjects confirmed the construct validity (t ¼ )13.8, P ¼ 0.000). Concurrent validity was identified by the significant correlation of SQS with the Pittsburgh Sleep Quality Index (r ¼ 0.72, P ¼ 0.000). The Cronbach s alpha coefficient was 0.92 for internal consistency and the correlation coefficient was 0.81 for test retest reliability at a 2-week interval. The developed SQS was therefore confirmed to be a valid and reliable instrument for the comprehensive assessment of sleep quality. keywords reliability, sleep quality scale, validity INTRODUCTION Sleep is essential for restoration of the body and for maintaining energy (Shapiro and Flanigan, 1993; Spensely, 1993). As sleep quality affects health (Edell-Gustafsson et al., 2002; Ersser et al., 1999; Tynjala et al., 1999), daytime function (Alapin et al., 2000; Chesson et al., 2000) and quality of life (Carpenter and Andrykowski, 1998; Iliescu et al., 2003), the measurements of sleep quality are important in both clinical practice and research. Both subjective and objective methods have been used to assess sleep quality. Polysomnography, which is considered the gold standard, provides accurate information on the physiological indices of sleep quality by using electroencephalogram, electrooculogram and electromyelogram. However, it is expensive, needs technical training and also requires much time for testing and the interpretation of data. In addition, it is difficult to assess natural or usual sleep quality as a lot of equipment is Correspondence: Chol Shin, MD, PhD, FCCP, Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Korea University Ansan Hospital, 516, Gojan 1-dong, Danwon-gu, Ansan-si, Gyeonggi-do 425-707, Republic of Korea. Tel.: +82-31-412-5603; fax: +82-31-412-5604; e-mail: shinchol@pol.net required for its operation. Actigraphy is another method for objective sleep monitoring that measures a person s sleep activity round the clock. Although it can measure sleep quality and quantity for a week or more (Sadeh et al., 1994), it is limited because it cannot distinguish between an awakened state without movement and true sleep (Menefee et al., 2000). On the other hand, self-report methods providing an assessment of the sleep quality experienced by the sleeper include the sleep diary, sleep log and sleep questionnaire. Although these methods do not provide information about the sleep structure, they attempt to measure both quantitative and qualitative aspects of sleep (Gentili et al., 1995). These subjective methods are easily administered, inexpensive, and applicable to large surveys. Sleep is a private experience that is influenced by many factors including age, sex, food, and physical and psychological health. Components of sleep quality and their importance differ according to individuals (Buysse et al., 1989), and thus a self-report-based evaluation is essential to measure sleep quality. Although various questionnaires have been used to assess sleep quality, they feature several limitations. First, there are few all-inclusive assessment scales. The previous scales include items that are unrelated to sleep quality, or they exclude essential items. There are also few questionnaires which provide a total score. Ó 2006 European Sleep Research Society 309

310 H. Yi et al. Therefore, we undertook this study to develop an instrument to measure comprehensively the sleep quality of adults, and subsequently to evaluate its validity and reliability. METHOD Instrument development A methodological design was used. A scale was developed according to the following process using classical test theory (Fig. 1). Definition of the concept No scale can be constructed to measure a concept until the nature of the concept has been delineated. Therefore, firstly sleep quality was defined through a literature review. Design of the scale Two procedures were used to generate items to fully reflect the concept of sleep quality. First of all, domains and their contents of sleep quality were identified by a literature review and then their contents were confirmed by an in-depth interview. These interviews were conducted with the consent of subjects visiting a sleep laboratory and a research institute. The subjects (32 males, 18 females) were 50 adults aged 21 59 years. They comprised 17 subjects with insomnia symptoms, 17 with obstructive sleep apnea, six with snoring, two with narcolepsy, and eight good sleepers. The main interview questions were ÔHow is your sleep these days?õ, ÔWhat do you think a good sleep is?õ, and ÔWhat do you think of a poor sleep?õ. The interview took about 30 60 min for each subject. The subjects underwent one or two interviews which were recorded and later transcribed. Content analysis of the interview data was conducted line-by-line and the results were confirmed by two nursing college professors with considerable experience of qualitative research. Review of items The constructed items were reviewed twice at a 2-week interval by six experts with an academic background in sleep research or clinical experience treating patients with sleep problems. The experts evaluated whether items were valid using a fourpoint ordinal rating scale, where 1 denotes an irrelevant item and 4 an extremely relevant item. The index of content validity (CVI), the most widely used qualification of content validity, indicates the proportion of items that received a rating 3 or 4 by the experts (Waltz and Bausell, 1981). Items with a CVI of at least 0.8 were selected according to Lynn s (1986) criteria. The experts also evaluated the itemsõ clarity, conciseness, and redundancy. In addition, they also identified whether essential items of sleep quality were excluded. The Korean language vocabularies of the items were revised to a sixth grade reading level and language usage was confirmed by consultation with a Korean literature scholar and four elementary school teachers. Pilot study A pilot study was conducted twice to evaluate the degrees of understanding concerning the sentences and vocabularies. As a reading level between the fifth and seventh grades is appropriate for most instruments used in the general population (Devellis, 2003), data for the first pilot study were gathered from sixth grade elementary school students, while data for the second pilot study were collected from the adults of the target population. The time required to complete the questionnaire was also surveyed. Field test Figure 1. Process of scale construction. This step requires the collection of data for item analysis and factor analysis. It has been reported that item analysis requires a sample size of about 100 200 respondents (Spector, 1992) while factor analysis requires a sample that is 10 times larger than the number of items (Nunally, 1978). Data on the 46 items with content validity were collected by trained interviewers from community-dwelling adults aged 20 59 years.

Development of the Sleep Quality Scale 311 Selection of the final items The final items of the scale to measure sleep quality were selected by item analysis and factor analysis. Item analysis was conducted to identify the items that form an internally consistent scale. Item analysis shows the correlation between respective items and the summated scale score. The higher the coefficient for each item the more clear that the item belongs to the scale. Generally, a coefficient of less than 0.3 suggests that the item should be dropped from the scale (Anthoney, 1999; De Vaus, 2002; Fisher et al., 2001). Thus the items with an itemtotal correlation of less than 0.3 were deleted from the scale. Factor analysis was also used to reduce the number of items and sort out the meaning. Among the several methods used in the factor analysis, principal axis factoring and oblimin rotation were used to extract the factors. Factors of scale were selected by considering a scree test and the possibility of a theoretical interpretation. Items with a factor loading of ±0.4 or less, or with a difference between absolute values of factor loadings of 0.1 or less, were eliminated. Evaluation of validity and reliability The validity and reliability of a developed scale must be evaluated by the collection of additional data from large samples (Burns and Grove, 2001). Two convenience samples were used to evaluate the validity and reliability. The first sample, comprising 629 community-dwelling adults aged 18 59 years, was used for construct validity, concurrent validity using the Pittsburgh Sleep Quality Index (PSQI; Buysse et al., 1989) and internal consistency reliability study. Data on the Sleep Quality Scale (SQS), PSQI, questions defining insomnia, sleep problems and demographic variables were collected from adults by trained interviewers. After the exclusion of 17 incomplete responses, data from 612 subjects, 47.9% male and 52.1% female, of mean age 36.2 years, were analyzed. Sample 2 comprised 110, 4-year university student subjects, 43.1% male and 56.9% female, with a mean age of 20.6 years. These subjects completed SQS twice at an interval of 2 weeks to confirm test retest reliability. in this sample was 0.76. The correlation coefficient was produced between SQS and PSQI using sample 1. Construct validity The validity of the instrument can be tested by identifying the groups that are expected to have contrasting scores on the instrument. Samples are then selected from at least two groups that are expected to have opposing responses to the items in the instrument (Burns and Grove, 2001). As insomnia is a typical sleep disorder featuring the complaint of insufficient or poor sleep quality, the sleep quality of insomniacs was compared with that of normal subjects using sample 1. The presence and frequency of four insomnia symptoms Ôdifficulty in initiating sleepõ, Ôdifficulty in maintaining sleepõ, Ôearly morning awakeningõ, and Ônon-restorative sleepõ were assessed according to previous studies (Ohayon and Hong, 2002; Ohayon and Partinen, 2002). Insomnia was considered to be present when it had occurred three or more times a week over the previous month. In sample 1, there were 191 subjects with insomnia and 332 subjects without any of the following sleep problems: insomnia, snoring, obstructive sleep apnea, periodic limb movement disorder, and narcolepsy. Reliability The reliability of a measure denotes the consistency of measures obtained in the use of particular instruments and is an indication of the extent of random error in the measurement method (Burns and Grove, 2001). Two methods were used to test the reliability of SQS. Homogeneity testing examines the extent to which all the items in the instrument consistently measure the construct. The Cronbach s alpha coefficients, the statistical procedure used in this process, were computed using sample 1 to identify the internal consistency of SQS. Stability, meanwhile, is concerned with the consistency of repeated measures of the same attribute with the use of the same scale or instrument. The correlation coefficients were computed using sample 2 to identify the stability of SQS. Concurrent validity Concurrent validity can be tested by identifying the correlation between a new instrument and a previously validated instrument for measuring the same concept (Seong, 2002a). Pittsburgh Sleep Quality Index was chosen as the criterion because it is a scale that has been used widely to assess sleep quality and disturbances over a 1-month time interval. It consists of 19 items on seven component scores: subjective sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbance, use of sleep medicine, and daytime dysfunction. The possible score ranges from 0 to 21, with a higher score representing worsened sleep quality. The Cronbach s alpha coefficient of PSQI was reported at its inception as 0.83 (Buysse et al., 1989). The Cronbach s alpha coefficient of PSQI RESULTS Definition of sleep quality Sleep can be defined as a regular, recurrent, easily reversible state of the organism that is characterized by relative quiescence and by a large increase in the threshold of response to external stimuli relative to the waking state (Kaplan and Sadock, 1988). Quality is defined as the degree of excellence of a thing or a general excellence (Oxford English Reference Dictionary, 1995). That is, sleep quality means the degree of excellence in sleep. It has been measured by many researchers on the basis of this definition (Freedman et al., 1999; Hawkins and Shaw, 1992; De Koninck et al., 1983; Mayers et al., 2003; Shaver et al., 1991; Sickel et al., 1999). Based on a literature

312 H. Yi et al. review of previous questionnaires and studies, the domains of sleep quality were identified as sleep initiation, sleep maintenance, depth of sleep, dreams, getting up after sleep, condition after sleep, effect on daily life, sleep amount and satisfaction with sleep. comprising community-dwelling adults aged 20 59 years by convenience sampling for item analysis and factor analysis. The analyses were based on 817 subjects, 49.3% male and 50.7% female, of mean age 38.9 years, after 198 incomplete data were excluded. Design of the scale Seventy-five items were generated in nine domains through the content analysis of interview data and literature review. The number of items in each domain was as follows: four items in sleep initiation, five in sleep maintenance, four in depth of sleep, three in dreams, six in getting up after sleep, 23 in condition after sleep, 23 in effect on daily life, three in sleep amount, and four in satisfaction with sleep. A four-point rating scale was used to avoid a neutral response. Review of items The index of content validity was calculated using the expertsõ ratings. Thirteen items had a CVI of under 0.8; 11 were excluded and two were revised. Five items were also excluded because of redundancy while a further two were added on the basis of the expertsõ opinion. This produced a total of 61 items. After 14 days, the content validity on the 61 items was reevaluated by the same experts, resulting in four items with a CVI of under 0.8 being excluded. A further 11 items with duplicated meanings were excluded after discussion among the experts, leaving 46 items in the preliminary scale. Pilot study The number of subjects in the first pilot study was five, sixth grade school students. It took 10 15 min for them to complete a questionnaire. The five students were asked whether or not they understood the sentences and words in the questionnaire. All five completely understood the meaning of 43 items and four of them understood the meaning of two items. None of the students understood the meaning of the last item: actual sleeping time. Therefore, the following sentence was added to clarify the meaning: Ôactual sleeping timeõ is different from Ôtime spent lying in bedõ. The second pilot study was conducted with the other representative target population of 43 adults who visited an outpatient department of a hospital and a research institute. It also took 10 15 min for them to complete the questionnaire. All of them understood the meaning of all items well. The following instruction was supplemented to ensure that the meaning of the response categories was clear: ÔfewÕ denotes none or 1 3 times a month, ÔsometimesÕ 1 2 times a week, ÔoftenÕ 3 5 times a week, and Ôalmost alwaysõ 6 7 times a week. Field test As the aim of this scale was to measure adult sleep quality, 1015 data sets were collected from the target population Selection of the final items A new item of sleep efficiency was calculated by using three items (time to go to bed, time to get up, actual sleeping time) before item analysis. This produced a total of 44 items. From the results of item analysis of 44 items, four items with an itemtotal correlation coefficient of under 0.3 were excluded to leave 40 items remaining. Factor analysis on these 40 items yielded six factors, including 28 items, which explained 62.6% of the total variance in Table 1. Factor 1 included 12 items and was termed the daytime dysfunction. It was composed of items representing symptoms resulting from poor sleep, which accounted for 30.8% of the variance. Factor 2 included four items which accounted for 10.8% of the variance. It was termed restoration after sleep, which seemed to measure the primary function of sleep. Factor 3 included four items which accounted for 8.1% of the variance. It was termed difficulty in falling asleep and was composed of items about sleep initiation. Factor 4 included three items which accounted for 5.3% of the variance. It was termed difficulty in getting up. Factor 5 included three items which accounted for 3.9% of the variance. It was termed satisfaction with sleep and was associated with the level of overall gratification and sufficient sleeping time. Factor 6 included two items which accounted for 3.7% of the variance. It was associated with awakening during sleep and was termed difficulty in maintaining sleep. The correlation coefficients between these factor scores ranged from 0.15 to 0.63 (Table 2). Evaluation of validity and reliability Concurrent validity Table 3 shows that SQS was strongly correlated with PSQI (r ¼ 0.72, P ¼ 0.000). All the subscales were positively correlated with the PSQI total score. Construct validity Table 4 showed that the total score of the insomniacs was significantly higher than that of the normal subjects (t ¼ )13.89, P ¼ 0.000). All subscales had significant differences between the two groups. Reliability The SQS homogeneity was high with a Cronbach s alpha coefficient of 0.92. The Cronbach s alpha coefficients of the subscales ranged from 0.61 to 0.90: daytime dysfunction, 0.90; restoration after sleep, 0.84; difficulty in falling asleep, 0.80;

Development of the Sleep Quality Scale 313 Table 1 Rotated factor matrix of Sleep Quality Scale items (n ¼ 817) Items F1 F2 F3 F4 F5 F6 Difficulty in thinking due to poor sleep 0.78 )0.04 0.00 )0.03 0.06 )0.10 Difficulty in concentrating due to poor sleep 0.75 )0.01 0.02 0.08 0.00 )0.03 Increase of mistakes due to poor sleep 0.75 0.01 )0.02 )0.06 )0.04 )0.03 Irritated feeling due to poor sleep 0.71 0.03 )0.10 0.08 0.02 0.10 Decrease of interest in work or others due to poor sleep 0.69 )0.03 0.00 )0.03 0.04 )0.03 Getting tired easily at work due to poor sleep 0.69 )0.01 0.01 0.12 0.01 0.05 Sleepiness that interferes with daily life 0.69 0.02 )0.04 0.05 0.01 )0.05 Painful life due to poor sleep 0.67 0.04 )0.06 )0.09 0.03 0.03 Decrease of desire due to poor sleep 0.64 0.00 0.06 0.08 )0.04 0.05 Increase of forgetfulness due to poor sleep 0.59 0.05 0.02 )0.05 )0.03 0.06 Headache due to poor sleep 0.59 )0.03 0.14 0.02 )0.02 0.02 Decrease of appetite due to poor sleep 0.45 0.01 0.10 0.03 )0.02 0.06 Relief of fatigue after sleep 0.01 0.82 0.04 )0.10 0.01 )0.05 Regaining vigor after sleep )0.04 0.81 0.01 )0.03 )0.06 0.05 Clear-headed feeling after sleep 0.02 0.77 )0.02 0.06 0.00 0.03 Refreshed feeling of body after sleep 0.03 0.63 )0.02 0.09 0.11 )0.04 Difficulty in getting back to sleep after nocturnal awakening 0.00 )0.01 0.78 0.01 )0.02 0.05 Never falling asleep after awakening during sleep 0.02 0.02 0.75 )0.10 0.00 )0.03 Difficulty in falling asleep 0.01 0.00 0.58 0.12 0.10 0.08 Tossing and turning sleeplessly 0.05 0.07 0.50 0.12 0.10 0.13 Wish for more sleep after getting up )0.01 )0.01 )0.10 0.84 0.00 0.05 Difficulty in getting up after sleep 0.06 0.04 0.10 0.70 )0.04 )0.04 Feeling unlikely to sleep after sleep 0.15 0.14 0.22 0.48 0.12 )0.03 Satisfaction with sleep )0.04 0.02 0.04 0.06 0.78 )0.04 Deep sleep )0.03 0.01 0.08 )0.16 0.72 0.09 Enough sleep time 0.04 0.14 )0.07 0.27 0.42 0.02 Waking up easily due to noise )0.01 0.00 0.02 )0.03 0.02 0.72 Waking up during sleep 0.05 0.00 0.02 0.01 0.01 0.70 Eigenvalue 8.63 3.01 2.28 1.47 1.08 1.04 Variance 30.79 10.76 8.14 5.26 3.88 3.74 Cumulated variance 30.79 41.55 49.70 54.96 58.84 62.58 Entries in bold indicate where P-value is unnecessary. Table 2 Correlations between factor scores in SQS (n ¼ 817) Daytime dysfunction Restoration after sleep Difficulty in falling asleep Difficulty in getting up Satisfaction with sleep Restoration after sleep 0.35 (P ¼ 0.000) Difficulty in falling asleep 0.43 (P ¼ 0.000) 0.30 (P ¼ 0.000) Difficulty in getting up 0.45 (P ¼ 0.000) 0.40 (P ¼ 0.000) 0.19 (P ¼ 0.000) Satisfaction with sleep 0.30 (P ¼ 0.000) 0.61 (P ¼ 0.000) 0.43 (P ¼ 0.000) 0.32 (P ¼ 0.000) Difficulty in maintaining sleep 0.32 (P ¼ 0.000) 0.25 (P ¼ 0.000) 0.63 (P ¼ 0.000) 0.15 (P ¼ 0.000) 0.37 (P ¼ 0.000) difficulty in getting up, 0.74; satisfaction with sleep, 0.71; and difficulty in maintaining sleep, 0.61. The correlation coefficient for the 2-week interval test retest reliability was 0.81 (P ¼ 0.000) and the correlation coefficients of the subscales were daytime dysfunction, 0.72 (P ¼ 0.000); restoration after sleep, 0.74 (P ¼ 0.000); difficulty in falling asleep, 0.56 (P ¼ 0.000); difficulty in getting up, 0.62 (P ¼ 0.000); satisfaction with sleep, 0.70 (P ¼ 0.000); and difficulty in maintaining sleep, 0.68 (P ¼ 0.000). DISCUSSION The SQS was developed systematically and was confirmed to be a valid and reliable instrument for the comprehensive assessment of sleep quality. It measures sleep quality of the previous month and provides a global score. The scale is composed of 28 items and six factors, including daytime dysfunction, restoration after sleep, difficulty in falling asleep, difficulty in getting up, satisfaction with sleep, and difficulty in maintaining sleep. It is a four-point Likert scale (few ¼ 0, sometimes ¼ 1, often ¼ 2, almost always ¼ 3). Scores on items in factors 2 and 5 have to be reversed before all the scores are summed. The range of score is from 0 to 84, with a higher score indicating a lower sleep quality. Although the names of SQS factors and number of items were a little different from those of previous questionnaires or scales, the SQS items themselves were similar. Compared with PSQI, the present scale included more items about restorative

314 H. Yi et al. Table 3 Correlations between SQS and PSQI (n ¼ 612) PSQI Daytime dysfunction Use of sleeping medication Sleep disturbances Habitual sleep efficiency Subjective sleep quality Sleep latency Sleep duration Total score Total Score 0.72 (P ¼ 0.000) 0.67 (P ¼ 0.000) 0.51 (P ¼ 0.000) 0.33 (P ¼ 0.000) 0.33 (P ¼ 0.000) 0.42 (P ¼ 0.000) 0.22 (P ¼ 0.000) 0.44 (P ¼ 0.000) Daytime dysfunction 0.61 (P ¼ 0.000) 0.56 (P ¼ 0.000) 0.38 (P ¼ 0.000) 0.26 (P ¼ 0.000) 0.23 (P ¼ 0.000) 0.35 (P ¼ 0.000) 0.18 (P ¼ 0.000) 0.49 (P ¼ 0.000) Restoration after sleep 0.48 (P ¼ 0.000) 0.48 (P ¼ 0.000) 0.38 (P ¼ 0.000) 0.23 (P ¼ 0.000) 0.22 (P ¼ 0.000) 0.26 (P ¼ 0.000) 0.16 (P ¼ 0.000) 0.24 (P ¼ 0.000) Difficulty falling asleep 0.59 (P ¼ 0.000) 0.50 (P ¼ 0.000) 0.57 (P ¼ 0.000) 0.22 (P ¼ 0.000) 0.40 (P ¼ 0.000) 0.37 (P ¼ 0.000) 0.24 (P ¼ 0.000) 0.17 (P ¼ 0.000) Difficulty getting up 0.48 (P ¼ 0.000) 0.46 (P ¼ 0.000) 0.27 (P ¼ 0.000) 0.23 (P ¼ 0.000) 0.12 (P ¼ 0.003) 0.19 (P ¼ 0.000) 0.11 (P ¼ 0.007) 0.49 (P ¼ 0.000) Satisfaction with sleep 0.58 (P ¼ 0.000) 0.56 (P ¼ 0.000) 0.39 (P ¼ 0.000) 0.33 (P ¼ 0.000) 0.29 (P ¼ 0.000) 0.30 (P ¼ 0.000) 0.14 (P ¼ 0.000) 0.30 (P ¼ 0.000) Difficulty maintaining sleep 0.39 (P ¼ 0.000) 0.33 (P ¼ 0.000) 0.28 (P ¼ 0.000) 0.17 (P ¼ 0.000) 0.31 (P ¼ 0.000) 0.43 (P ¼ 0.000) 0.13 (P ¼ 0.001) 0.05 (P ¼ 0.190) functions after sleep, difficulty in getting up and various daytime dysfunctions because of poor sleep. However, sleep efficiency, an important concept of sleep quality (A kerstedt et al., 1994a; Bliwise, 1992), was excluded because of an itemtotal correlation of under 0.3. It is likely that sleep efficiency was a less sensitive concept to measure sleep quality over a 1- month period because it was calculated from items about rough sleep time for the previous 1 month. That is, it is considered as an appropriate concept to measure sleep quality only of the previous night. On the other hand, a factor should include at least three items for meaningful interpretation according to Zwick and Velicer (1986). The sixth factor of this scale was composed of only two items relating to awakening during the night. This factor was selected because sleep fragmentation produced disruption of nocturnal sleep (Roehrs et al., 1994) and sleep continuity was closely reflected in sleep quality (A kerstedt et al., 1994b). Thus it will be necessary to construct more items related to awakening during sleep in future study. When a scale is developed, its validity and reliability must be evaluated. The construct validity of several questionnaires has been studied by factor analysis. The Post-sleep Inventory was composed of 29 items originally. Factor analysis yielded seven factors, including only 26 items, with a total variance of 54.5% (Webb et al., 1976). The Sleep Questionnaire consisted of 55 items when it was developed, but factor analysis produced seven factors, comprising 41 items, which explained 71.7% of the total variance (Domino et al., 1984). The Verran Snyder-Halpern Sleep Scale (VSH Sleep Scale) was reduced from eight items to seven items by factor analysis and these items accounted for 60% of the variance (Snyder- Halpern and Verran, 1987). SQS was made up of items with construct validity confirmed by factor analysis. Six factors composed of 28 items of SQS explained 62.6% of the total variance. This surpassed the criterion for total variance of 60% as set by social science (Chai, 2003), and therefore was considered to be sufficient to account for sleep quality. In contrast, PSQI was composed of items that were based on empirical and clinical experiences, rather than statistical analysis, so the present result could not be compared with the total variance of PSQI. PSQI was validated instead using the group differences among good sleepers, poor sleepers with depression, disorders of initiating and maintaining sleep, and disorders of excessive somnolence (Buysse et al., 1989). Other studies have also validated their sleep questionnaire by comparing with different subject populations (Domino et al., 1984; Webb et al., 1976). The construct validity of SQS was also examined using the known-group technique. From the result, the total score (t ¼ )13.89, P ¼ 0.000) and scores of the subscales discriminated between the subjects with insomnia and the normal subjects. Thus construct validity was confirmed in a similar manner to that of the previous scales or questionnaires. In addition, the concurrent validity of SQS was studied by using PSQI as the criterion. The correlation result (r ¼ 0.72, P ¼ 0.000) confirmed the validity of SQS as a scale to

Development of the Sleep Quality Scale 315 Table 4. Comparison of SQS scores between insomniacs and normal subjects (n ¼ 523) Insomniacs (n ¼ 191), Mean ± SD Normal subjects (n ¼ 332), Mean ± SD t-value P-value Total score 31.1 ± 13.61 15.8 ± 9.06 )13.89 0.000 Daytime dysfunction 8.2 ± 6.46 3.2 ± 3.96 )9.69 0.000 Restoration after sleep 7.8 ± 2.75 4.3 ± 3.10 )12.81 0.000 Difficulty in falling asleep 2.9 ± 2.86 1.1 ± 1.51 )8.26 0.000 Difficulty in getting up 4.7 ± 2.19 2.6 ± 1.78 )11.20 0.000 Satisfaction with sleep 5.7 ± 2.44 3.4 ± 2.35 )10.55 0.000 Difficulty in maintaining sleep 1.8 ± 1.69 1.2 ± 1.23 )4.72 0.000 measure sleep quality. This was a similar result to the correlation coefficient of 0.71 between the Korean Sleep Scale A and VSH Sleep Scale (Oh et al., 1998). Although there are no absolute criteria to evaluate validity, the validity of SQS as an instrument to measure sleep quality was confirmed to be high according to the criteria (0.6 0.8) suggested by Seong (2002b). Reliability was tested in two ways. First of all, the Cronbach s alpha coefficient for homogeneity was 0.92, indicating a high degree of internal consistency. Burns and Grove (2001) suggested that a slightly lower coefficient (0.8 0.9) indicated an instrument that would reflect more richly the fine discriminations in the levels of the construct. This result was higher than that of the Sleep Questionnaire (Cronbach s alpha coefficient 0.76) and that of PSQI (Cronbach s alpha coefficient 0.83). Test retest reliability was also examined for consistency of repeated measures. The correlation coefficient of 0.81 confirmed the good stability of the instrument. The correlation coefficient for PSQI after an interval of 28.2 days (range 1 265) in all groups, including the normal subjects, depressives, disorders of initiating and maintaining sleep, and disorders of excessive somnolence, was 0.85 (Buysse et al., 1989). The correlation coefficient for PSQI after an interval of 19 days (range 3 44) in nursing home residents was 0.82 (Gentili et al., 1995). The correlation coefficient for the Sleep Questionnaire after an interval of 10 weeks in the normal subjects was 0.79 (Domino et al., 1984). This result was similar to the above results although the test intervals differed in the previous studies. On the other hand, it was a little lower than the correlation coefficient of 0.90 for PSQI in patients with insomnia at a 2-day interval in the study of Backhaus et al. (2002). This was presumably because of the short interval of the PSQI test. In summary, SQS was developed using item analysis and factor analysis on items with content validity. The construct validity and concurrent validity were confirmed. The internal consistency reliability and test retest reliability were acceptably high. SQS can also be used to assess sleep quality easily and rapidly and can provide assistance in planning interventions for sleep problems. In addition, it can be utilized in sleep research. Future studies should be conducted for further validation and standardization of the scale in various clinical settings, including obstructive sleep apnea and narcolepsy patients, and with a diverse range of age groups, including the elderly. The ability of the scale to discriminate among the levels of sleep quality needs to be confirmed through further research. REFERENCES Åkerstedt, T., Hume, K., Minors, D. and Waterhouse, J. The subjective meaning of good sleep, an intraindividual approach using the Karolinska sleep diary. Percept. Mot. Skills, 1994a, 79: 287 296. Åkerstedt, T., Hume, K., Minors, D. and Waterhouse, J. The meaning of good sleep: a longitudinal study of polysomnography and subjective sleep quality. J. Sleep Res., 1994b, 3: 152 158. Alapin, I., Fichten, C. S., Libman, E., Creti, L., Bailes, S. and Wright, J. How is good and poor sleep in older adults and college students related to daytime sleepiness, fatigue, and ability to concentrate? J. Psychosom. Res., 2000, 49: 381 390. Anthoney, D. Understanding Advanced Statistics: A Guide for Nurses and Health Care Researchers. Churchill Livingstone, London, 1999. Backhaus, J., Junghanns, K., Broocks, A., Riemann, D. and Hohagen, F. Test retest reliability and validity of the Pittsburgh Sleep Quality Index in primary insomnia. J. Psychosom. Res., 2002, 53: 737 740. Bliwise, N. G. Factors related to sleep quality in healthy elderly women. Psychol. Aging, 1992, 7: 83 88. Burns, N. and Grove, S. K. The Practice of Nursing Research. Conduct, Critique & Utilization, 4th edn. W.B. Saunders, Philadelphia, 2001. Buysse, D. J., Reynolds, C. F. 3rd, Monk, T. H., Berman, S. R. and Kupfer, D. J. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res., 1989, 28: 193 213. Carpenter, J. S. and Andrykowski, M. A. Psychometric evaluation of the Pittsburgh Sleep Quality Index. J. Psychosom. Res., 1998, 45: 5 13. Chai, S. I. Methodology of Social Science Survey. Hak Hyun Sa, Seoul, 2003. Chesson, A. Jr, Hartse, K., Anderson, W. M., Davila, D., Johnson, S., Littner, M., Wise, M. and Rafecas, J. Practice parameters for the evaluation of chronic insomnia. An American Academy of Sleep Medicine report. Standards of Practice Committee of the American Academy of Sleep Medicine. Sleep, 2000, 23: 237 241. De Koninck, J., Gagnon, P. and Lallier, S. Sleep positions in the young adult and their relationship with the subjective quality of sleep. Sleep, 1983, 6: 52 59. De Vaus, D. A. Surveys in Social Research, 5th edn. Allen & Unwin, Australia, 2002. Devellis, R. F. Scale Development: Theory and Application, 2nd edn. SAGE Publications, Thousand Oaks, 2003. Domino, G., Blair, G. K. and Bridges, A. Subjective assessment of sleep by sleep questionnaire. Percept. Mot. Skills, 1984, 59: 163 170. Edell-Gustafsson, U. M., Kritz, E. I. K. and Bogren, I. K. Selfreported sleep quality, strain and health in relation to perceived

316 H. Yi et al. working conditions in females. Scand. J. Caring Sci., 2002, 16: 179 187. Ersser, S., Wiles, A., Taylor, H., Wade, S., Walsh, R. and Bentley, T. The sleep of older people in hospital and nursing homes. J. Clin. Nurs., 1999, 8: 360 368. Fisher, M., King, J. and Tague, G. Development of a self-directed learning readiness scale for nursing education. Nurse Educ. Today, 2001, 21: 516 525. Freedman, N. S., Kotzer, N. and Schwab, R. J. Patient perception of sleep quality and etiology of sleep disruption in the intensive care unit. Am. J. Respir. Crit. Care Med., 1999, 159: 1155 1162. Gentili, A., Weiner, D. K., Kuchibhatla, M. and Edinger, J. D. Test retest reliability of the Pittsburgh sleep quality index in nursing home residents. J. Am. Geriatr. Soc., 1995, 43: 1317 1318. Hawkins, J. and Shaw, P. Self-reported sleep quality in college students: a repeated measures approach. Sleep, 1992, 15: 545 549. Iliescu, E. A., Coo, H., McMurry, M. H., Meers, C. L., Quinn, M. M., Singer, M. A. and Hopman, W. M. Quality of sleep and health related quality of life in haemodialysis patients. Nephrol. Dial. Transplant., 2003, 18: 126 132. Kaplan, H. I. and Sadock, B. J. Synopsis of Psychiatry: Behavioral Sciences/Clinical Psychiatry, 5th edn. Williams & Wilkins, Baltimore, 1988. Lynn, M. R. Determination and quantification of content validity. Nurs. Res., 1986, 35: 382 385. Mayers, A. G., Van Hooff, J. C. and Baldwin, D. S. Quantifying subjective assessment of sleep and life-quality in antidepressanttreated depressed patients. Hum. Psychopharmacol., 2003, 18: 21 27. Menefee, L. A., Cohen, M. J., Anderson, W. R., Doghramiji, K., Frank, E. D. and Lee, H. Sleep disturbance and nonmalignant chronic pain: a comprehensive review of literature. Pain Med., 2000, 1: 156 172. Nunally, J. C. Psychometric Theory, 2nd edn. McGraw-Hill, New York, 1978. Oh, J. J., Song, M. S. and Kim, S. M. Development and validation of Korean Sleep Scale A. J. Kor. Acad. Nurs., 1998, 28: 563 572. Ohayon, M. M. and Hong, S. C. Prevalence of insomnia and associated factors in South Korea. J. Psychosom. Res., 2002, 53: 593 600. Ohayon, M. M. and Partinen, M. Insomnia and global sleep dissatisfaction in Finland. J. Sleep Res., 2002, 11: 339 346. Oxford English Reference Dictionary. Oxford University Press, Oxford, 1995. Roehrs, T., Merlotti, L., Petrucelli, N., Stepanski, E. and Roth, T. Experimental sleep fragmentation. Sleep, 1994, 17: 438 443. Sadeh, A., Sharkey, K. M. and Carskadon, M. A. Activity-based sleep-wake identification: an empirical test of methodological issues. Sleep, 1994, 17: 201 207. Seong, T. J. Validity and Reliability. Hakjisa, Seoul, 2002a. Seong, T. J. Educational Evaluation. Hakjisa, Seoul, 2002b. Shapiro, C. and Flanigan, M. Function of sleep. BMJ, 1993, 306: 383 385. Shaver, J. L., Giblin, E. and Paulsen, V. Sleep quality subtypes in midlife women. Sleep, 1991, 14: 18 23. Sickel, A. M., Moore, P. J., Adler, N. E., Williams, D. R. and Jackson, J. S. The differential effects of sleep quality and quantity on the relationship between SES and health. Ann. N Y Acad. Sci., 1999, 896: 431 434. Snyder-Halpern, R. and Verran, J. A. Instrumentation to describe subjective sleep characteristics in healthy subjects. Res. Nurs. Health, 1987, 10: 155 163. Spector, P. E. Summated Rating Scale Construction: An Introduction. SAGE Publications, Newbury Park, 1992. Spensely, S. M. Sleep inquiry: a look with fresh eyes. Image J. Nurs. Sch., 1993, 25: 249 256. Tynjala, J., Kannas, L., Levalahti, E. and Valimaa, R. Perceived sleep quality and its precursors in adolescents. Health Promot. Int., 1999, 14: 155 166. Waltz, C. W. and Bausell, R. B. Nursing Research: Design, Statistics and Computer Analysis. F. A. Davis, Philadelphia, PA, 1981. Webb, W. B., Bonnet, M. and Blume, G. A post-sleep inventory. Percept. Mot. Skills, 1976, 43: 987 993. Zwick, W. R. and Velicer, W. F. Comparison of five rules for determining the number of components retain. Psychol. Bull., 1986, 99: 432 442.