As expected, ANOVA results confirmed significant changes in sleep pattern with chronological age: a reduction of sleep time (p<.0001); a tendency to go to bed later (p<.0001) and less regularly with age (p<.0001), boys more often than girls (p<.0001). Parental involvement, both at bedtime and risetime, decreased as adolescents got older (bedtime from 3% at 15 yrs to 0.9% at 18 yrs; risetime from 43% at 15 yrs to 32% at 18 yrs). This is different from the report of Carskadon which found that parental influence diminishes at bedtime but becomes more important on waking in the morning. Difficulties in falling asleep were reported in 21% of students, without gender differences, but girls suffered more from nighttime wakings (11,5% vs 8,1% p<.001). As a measure of sleep schedule regularity we computed the difference between weekend and school nights bedtime (BDS). An irregular sleep schedules (BDS higher than 3 hours) was found in 25% of students, boys more often than in girls (14,7% vs 11%).

Regarding sociodemographic factors while socio-economic (SES) level of parents was not correlated to sleep problems, we found that adolescents from non-intact families (8.6%) had more sleep irregularity (BDS 131 vs 115 mins; p<.001). Small differences in some sleep habits among geographical region were found. Students living in the South went to bed significantly later (20 mins) but their sleep schedules are more regular compared to those living in the North and Center. Students living in urban areas showed more sleep irregular sleep and tend to sleep less either in school or in weekend nights.

Moreover, we examined the relationship between subjective evaluation of poor sleep quality, their sleep habits and symptoms of sleep disturbances. The "Do you consider yourself to be a poor sleeper?" question was used to define sleep quality. Of 6632 students, 19.6% considered themselves as poor sleeper, significantly more girls than boys (21% vs 17%; p<.001). Self-perception of poor sleep tended to increase with age in both sexes (14-15 yrs 15%, 15-16 yrs 18%, 16-17 yrs 19%, 17-18 yrs 20% >18 yrs 22.3%; Pearson chi-square 25.5; df=4; p<.001). To better identify variables related to a self-evaluation of poor sleep quality, logistic regression analysis was performed, which showed the following independent variables (female sex, sleep onset insomnia, nightwakings either brief and multiple or prolonged, depression, anxiety and evening phase preference) in association to self-perception of poor sleep. Parasomnias and erratic sleep schedules, although more frequent in poor sleepers, were not statistically significant in relation to self-reported poor sleep quality Sleep length did not contribute to self-perception of poor sleep.

We also investigated the use of sleeping pills in Italian adolescents to determine the prevalence and to identify associated factors. Results showed that 4% of students had used medication to help themselves to sleep at least once during the last six months, but only 1.3% had regularly taken sleeping pills. Girls were more prevalent users than males (5.3% vs 2.04%; p<.001). Consumption was related and tended to increase with age (Pearson chi-square df=4; p<.001). In general the increase was considerably larger for girls than boys and this sex difference was found in all age groups. Ansiolitic benzodiazepines were used by 44.2% of students, while 10% of them had used aspirin or NSAID and 33% had taken herbs, to improve sleep. Finally, homeopathic remedies were used in 2.6% of students while 6.5% of them did not specify the type of substance used. Professional help for sleeping problems had been sought by 2.5% of all the sample. Ansiolitic benzodiazepines were prescribed by general practitioners in 58.6% of cases, while NSAID were suggested by parents in 17% of them of self-prescribed in 25% of students. In order to determine which independent variables were significantly associated with the condition of using substance to promote sleep, logistic regression analysis was performed. Seven independent variables (female sex, living in urban area, multiple midnight awakenings, headache related to sleep, depression, anxiety and consumption of psychoactive substances) were statistically significant in relation to the use of sleeping pills. Sleep onset insomnia, irregular sleep patterns as well as partial arousal disorders were not statistically significant in relation to use of sleeping pills.

Moreover ANOVA results showed an association between poor self-reported school achievement and increased complain of daytime sleepiness (p<.001), greater use of caffeine, alcohol, and tobacco (p<.001), sleep problems (p<.001), evening phase delay preference (p<.001), anxiety (p<.001) and depressive mood (p<.01). We found that, of the small percentage of students who went to bed later on school nights than on weekends, most of them across all age groups (57% at 14 yrs to 37% at 18 yrs) had this late bedtime because of the habit of doing homework in the evening. Furthermore, students who referred attention problems at school slept less either in school (462 vs 477 mins; p<.001) or in weekend nights (526 vs 543 mins; p<.001), had more irregular bedtime (140 vs 109 mins; p<.001) and significant higher scores in all scales but M/E scale where they obtained lower scores. In order to investigate the influence of school starting time on sleep and daytime behavior, we compared sleep pattern for late vs early starting schools. Despite the small difference in school start time in Italy (7.45 vs 8.45 am) across all age groups, we found that students who attended the school with earliest start time had more irregular sleep schedules (p<.001) complained of daytime sleepiness (55% vs 44%, p<.01), tended to fall asleep at school more frequently (9.5% vs 6%; p<.05) were at risk of increased use of stimulants (p<.001) and reported more frequently poor academic performance (7% vs 4%; p<.05).

Of the whole sample, 88% across all ages and without gender difference, reported having had some accidents (at home, at school, driving). The increased vulnerability to accidents showed a positive correlation to daytime sleepiness (r=.15 p<.0001) increased use of stimulants and tobacco (r=.18 p<.001) and mainly to sleep problems (r=.25 p<.00001) and to anxiety (r=.18 p<.001).

Our data confirmed the presence also in Italian adolescents of unhealthy sleeping habits and difficulties more prevalent in late adolescence. According with other studies, adolescent poor sleepers showed significant differences in daytime feelings and mood, complaining also poor school achievement. Poor sleep self-perception occurs with sufficient frequency in healthy adolescent, and is significantly more prevalent in girls. However most of them do not seek help for their sleep. In contrast they were more prone to use substance to help themselves to sleep. Although International comparisons are difficult to establish because of differences in legislation, our results are similar to other adolescent studies. Despite the use of medications to enhance sleep were not very common in Italian adolescents, the tendency, that increases with age, to use a chemical solution may persist in adult life and become a risk factor for habitual use. Since all these topics, with its possible public health significance, has largely been ignored in Italy, much more attention must be paid to this problem in order to prevent the negative impact of poor sleep habits on daytime functioning also in our country.