EDS is a typical symptom of sleep disorders such as OSA, narcolepsy, and idiopathic hypersomnia. Studies have reported that 2% of women and 4% of men in the general middle-aged population suffer from OSA
[29, 30]. ESS has proven to be a useful tool in the assessment of EDS in English and some non-English-speaking countries. Our study reported the response rate, reliability, and validity of the Mandarin version of the ESS when applying it to the general population in five cities of China.
We found that 90.42% of the respondents answered every ESS item, indicating that the Mandarin version of the ESS was acceptable to the majority of the population. In addition, the different response rates among some associate factors were explicable: Response rates were higher for urban than rural areas because, as a whole, the urban respondents were more educated than rural ones. Additionally, rural respondents had fewer opportunities to encounter some items in their daily life such as Items 4 and 8. For the elderly, due to their physical and intellectual limitations, it was difficult to complete the questionnaires without others’ help. Consequently, the response rate of the elderly was lower than that of the young. The fact that more educated respondents had a higher response rate was quite natural, as they had greater knowledge and more social responsibilities. Office workers engaged more often in mental work, and likely had better comprehension and more energy to complete the questionnaires than manual workers. Thus, the response rate of office workers was higher than that of manual workers. Finally, it is understandable that respondents with chronic diseases had more missing data, as their physical condition might limit their ability to complete the scale.
The response rate of each item was mainly determined by how frequently the situation was encountered in daily life. Sitting and talking was very common every day, so Item 6 had the highest response rate. In contrast, sitting and reading was seldom encountered in some populations like the elderly, rural residents, the less educated, and manual workers, so Item 1 had the lowest response rate.
The split-half reliability coefficient was 0.81. The internal consistency of the questionnaire was fair, with a Cronbach’s α coefficient of 0.80, which was similar to what has been reported previously (0.74–0.90)
[1, 15, 19, 21, 31, 32]. This consistency did not increase when any of the items were excluded. All of these values indicated excellent reliability and internal consistency
. Each item score significantly correlated with total score and all other item scores. In addition, the results of the factor analysis showed that only one factor was extracted with an eigenvalue greater than one. Hence, this questionnaire was thought to have good construct validity. All of these results suggest that the ESS is suitable for the Chinese population despite the differences in language and cultural background.
The mean (SD) of the ESS scores among the 2906 respondents who answered all eight ESS items was 6.75 (5.11) in our study. In the Turkish version of the ESS, the mean (SD) ESS score was 3.6 (3) among 60 healthy controls and 12.6 (6) among 60 respondents with sleep-disordered breathing
; in the Korean version, the mean (SD) ESS score was 5.07 (2.93) among 60 healthy controls and 8.21 (4.23) among 213 patients
. In most previous studies, the ESS scores were observed from the healthy controls vs. patients with sleep-disordered diseases, using a small population sample, while in our study, the ESS scores were observed from a large general population sample. We thought that it was also important to apply the ESS to the general population. Our results indicated that 22.16% of respondents had an ESS score above 10. According to the original version of the questionnaire, a score of 10 is considered the cut-off point that distinguishes normal respondents from patients with EDS
[1, 21]. Nearly a quarter of respondents in the selected sample had subjective EDS according to this standard. Previous studies conducted in Japan showed that 2.5% of the Japanese general population and 15% of the Japanese general adult population had EDS, as assessed by other questionnaires
[5, 33]. Another study reported that the prevalence of EDS as assessed by the ESS (ESS scores > 10) was 12.2% among the adult population of Korea
. From these and our findings, we could say that EDS is more common in China than in these two East Asian countries. However, the basis of this upper limit of normal sleepiness needs to be further investigated. The more widespread prevalence of EDS in China may be due to dramatic increases in obesity and overweight in Mainland China, especially in metropolises such as Beijing and Shanghai
. The extensive use of electronic products, such as mobile phones, computers, and television, has also likely contributed to the high prevalence of EDS
. Moreover, there is a cultural peculiarity of China in that many Chinese people tend to take afternoon naps, which might make respondents give higher ESS scores. In addition, we found an interesting correlation between education and ESS scores. People with a low level or a high level of education had higher ESS scores, compared with those with a medium level of education. We propose the following explanations for this phenomenon: First, a higher proportion of people with a low level of education are unemployed; second, people with a high level of education are mainly brainworkers and insomnia is one of the diseases exceedingly common among this group. Thus, people with lower or higher levels of education are more prone to sleep during the daytime. We also found that EDS was more common among respondents with obesity and other chronic diseases, a finding consistent with those of previous studies
Item analysis of the Mandarin version of the ESS revealed results that were very close to those reported for the original version
[15, 39]. Item 5 was the most soporific and got the highest score, while the scores for Items 6 and 8 were very low. This means when “lying down to rest in the afternoon when circumstances permit,” respondents were most likely to fall asleep, while when “sitting and talking to someone” or “in a car while stopped for a few minutes in the traffic,” they were less likely to sleep.
Two other Chinese versions of the ESS have been developed, in Hong Kong
 and Taiwan
. However, these are in traditional Chinese, while the current version is in simplified Chinese. Furthermore, the lifestyles and economic backgrounds of the populations in these two regions differ from those in Mainland China. Zhang et al.
 and Peng et al.
 have also developed two simplified Chinese versions of the ESS. In the former study, the researchers created a modified ESS by adding two backup items to the original ESS and forming a ten-item Sleepiness Questionnaire; however, this practice might have changed the original meaning of the scale. Zhang et al. found that the modified ESS was not reliable for normal respondents, though they mainly applied it among patients with suspected sleep-disordered breathing. In Peng et al.’s study, the translations of the items were not the same as those in our study. For example, in Item 8, “in a car” was translated as in their study as “when driving,” and “while stopped for a few minutes in the traffic” was translated as “while waiting for the traffic lights or when stopped in traffic for a few minutes.” Obviously, this would lead the respondents to give a lower item score, because people are less likely to sleep when they are a driver rather than a passenger, and when they are waiting for the traffic lights to change, which is a rather short time, instead of waiting in a traffic jam. In addition, the respondents in their study were patients with suspected obstructive sleep apnea hypopnea syndrome. The strength of our study is that we made the translation strictly in accordance with the meaning of the original version and then applied it to the general population.
The correlations between ESS scores and the scores of the SF-36 dimensions indicated that the eight dimensions of the SF-36 negatively correlated with ESS scores, even after controlling for the sociodemographic variables. SF-36 scores significantly decreased as ESS scores increased. The scores of the GH and VT scales decreased the most. EDS influences more or less all parts of life to such an extent that people with the condition perceive themselves as being generally more limited by their health than those without it. They are more likely to worry about their future and whether the symptoms will improve, leading to increasing dependency on family and friends
. The VT scale was the second most affected dimension in this study. People with EDS experience ineluctable sleep episodes. These symptoms may lead to the feeling of physical tiredness, causing reduced vitality and energy
. These results were consistent with those reported in other countries
Therefore, we concluded that the Mandarin version of the ESS was acceptable and applicable among the general population of five cities in China. As measured by the ESS, subjective EDS is common in China. Special care should be taken for those people who suffer from EDS, as it can disrupt a person’s social life and threaten public health and safety
[47–49]. We propose that the Mandarin version of the ESS can be used as an auxiliary judgment of EDS, and that early detection and treatment after the onset of symptoms might be important in reducing the negative effects on the quality of life of the patients.
There were four limitations in our study. First, we did not contact the original developer of the ESS, which might lead to a lack of conceptual equivalence with the original version. Second, this was only a general population-based study, and the Mandarin version of the ESS was not assessed using patients diagnosed with EDS; thus, the response rate, reliability, and validity of the Mandarin version among the patient population are uncertain. Furthermore, some items, which can be used as the criteria for known-groups validity such as sleep-wake patterns, symptoms and history of sleep disorders, using sleep pills, sleep quality (using the Pittsburgh Sleep Quality Index and/or other validated questionnaires), and so on, were not measured. Third, the ESS is only a subjective method that can be used as an auxiliary judgment of EDS diagnosis, but cannot act as a gold-standard test, so the prevalence of EDS found in this study might be overestimated. The diagnosis of EDS should be further validated by objective methods. Finally, the sample population was selected from five economically developed regions in China. Therefore, the situation in less developed regions was not investigated. To address these limitations, further investigation should be undertaken.