A significant proportion (55.8%) of students was classified as having poor sleep quality. After controlling for important demographic, behavioral and psychological factors, females, second year and third year students had significantly higher odds of poor sleep quality. Perceived stress level and symptoms of depression were strongly associated with poor sleep quality. Additionally, moderate to extremely severe levels of anxiety were related with poor sleep quality.
The proportion of students with poor sleep quality in our study population generally consistent with reports from previous reports [7, 8, 10, 14]. The significant sex difference in sleep quality was also consistent with other studies conducted among students [9, 13–15]. This difference by sex could be explained by the significantly higher proportion of female students reported shorter sleep duration and bad subjective sleep quality in this study. Other studies have also reported that female students were more likely to report longer sleep latency [15, 34], sleep disturbance  and lower rating of their sleep quality .
We found that increasing year of study in university was associated with reduced odds of poor sleep quality. This observation is inconsistent with reports from the study of university students in Hong Kong . Another study found higher proportion of short sleep duration among freshman students and longer sleep latency among seniors; however the second study did not find any difference on overall sleep quality by year of study in university . Variations in results across studies may be explained by differences in social, academic demands across universities. Within our own study population, sleep quality varied according to university, where students from the University of Gondar had higher odds of poor sleep quality than other students.
Consistent with previous studies [19, 22, 23, 26], we noted that student mental health status variables were associated with poor sleep quality. For example, perceived stress level and depression were strongly associated with sleep quality. The consistency of our findings with those of the published literature underscore the public health importance and implications of more thoroughly investigating links between sleep habits and sleep problems with mental health and wellness among young adults.
The findings of our study should be interpreted in light of some study limitations. First, given the cross-sectional nature of our study, it is difficult to determine whether poor sleep quality is a result of mood, anxiety and stress symptoms, or whether these psychological symptoms contributed to poor sleep. Second, our use of a self-administered survey that relied on subjective measures of sleep quality and other covariates may have introduced some degree of error in reporting behavioral covariates. However, we believe that these issues are in part reduced by our use of anonymous questionnaire and validated instruments. Furthermore, a thorough pre-testing of our questionnaire likely reduced the possible risk of ambiguity of the questions. Finally, non-response by approximately 9% of enrolled students, may have contributed to some selection bias in our study. However, our evaluation of available data on responders and non-responders suggest similarity across the two populations, thus reducing the level of concern about selection bias.