اکتیگرافیک ضعیف و الگوهای خواب خود گزارش شده بزهکاری و اختلال در طول روز در میان نوجوانان در معرض خطر را پیش بینی می کند

Abstract Objective To evaluate associations between actigraphic sleep patterns, subjective sleep quality, and daytime functioning (ie, sleepiness, symptoms of depression, and delinquency and other conduct problems) in at-risk adolescents. Design Prospective, observational cohort study. Setting Providence, RI, predominantly home and school and 2 visits to the Brown Center for the Study of Children at Risk. Participants A diverse group of low-income 13-year-olds (n = 49) with and without prenatal drug exposure. Interventions None. Measurements Actigraphy, sleep diaries, and sleep and health questionnaires. Results Above and beyond the effects of prenatal drug exposure and postnatal adversity, actigraphic daytime sleep was a significant predictor of daytime sleepiness and delinquency. Subjective sleep quality was a significant predictor of daytime sleepiness, delinquency, and depressive symptoms. Later bed times predicted increased delinquency. Conclusions There was a unique effect of actigraphic daytime sleep duration, subjective nighttime sleep quality, and bedtime on daytime functioning (ie, sleepiness, symptoms of depression, and delinquency and other conduct problems) of at-risk adolescents. In these vulnerable youth, these problematic sleep patterns may contribute to feeling and behaving poorly. Intervention studies with at-risk teens should be conducted to further explore the role of these sleep parameters on daytime functioning.

Results Descriptive results There were 2 significant differences in demographic characteristics between the Providence cohort MLS subjects who chose not to participate in this sleep study and those subjects who did choose to participate: Those who chose not to participate in the sleep study were exposed to significantly more community violence and were less likely to be low SES (See Table 1). Although we do not have sleep data on the Providence cohort MLS subjects who chose not to participate in this sleep study, we measured mean differences between them and our sleep participants on our other outcome variables: delinquency, depressive symptoms, and conduct disorder symptoms. The subjects who chose not to participate in this sleep study had significantly more conduct disorder symptoms (M, 3.23 [SD, 3.5]; t = − 2.41 (142.3), P = .017) than our sleep participants (M, 2.12 [SD, 2.13]). There were no significant differences in the 2 groups on delinquency or depressive symptoms (M, 4.1 [SD, 3.82] vs 3.78 [SD, 3.64], and M, 4.8 [SD, 3.17] vs 4.53 [SD, 3.27], respectively). Data from all 49 completers are presented here. See Table 2 for sleep characteristics of the sample and Table 3 for sample means and SDs for the measure of each construct of interest (eg, adversity index, sleep problem index, sleepiness scale, delinquency scale, conduct disorders symptom count, and depressive disorder symptom count). See Table 4 for correlations among the 6 sleep parameters composing the sleep problem index. Table 2. Sleep characteristics of the sample (n = 49). Actigraphy, M (SD) Nighttime sleep duration 494.21 min (53.16) Sleep start (lights out) time 11:19 pm (1.07 h) Week sleep schedule variabilitya 1.99 (2.03) Daytime sleep duration (nap length among nappersb) 21.61 min (24.04) Sleep environment 1.84 (1.4) Subjective sleep quality 2.36 (.81) Final wake time 7:35 am (1.01 h) Nighttime sleep (min) 449 min (46.01) Nighttime wake (min) 45.25 min (36.56) a Sleep schedule variability for the school week was calculated using the summed SDs of the sleep start times and sleep end times. b Twenty-three participants (46%) took at least 1 nap during the assessment week. Table options Table 3. Measure descriptives of the study sample. M SD Min Max Adversity index 3.8 2.83 0 10 Sleep problem index 2.87 1.17 0.38 5.38 Sleepiness scale 7.4 4.7 1 18 Delinquency scale 3.85 3.64 0 12 Conduct disorder symptom count 2.16 2.13 0 7.5 Depressive disorder symptom count 4.54 3.3 0 16.5 Table options Table 4. Correlations among 6 sleep parameters. Nighttime sleep duration Sleep start time Week sleep schedule variability Daytime sleep duration Sleep environment Subjective sleep quality Nighttime sleep duration 1.000 Sleep start time − 0.561⁎⁎ 1.000 Week sleep schedule variability − 0.403⁎⁎ 0.504⁎⁎ 1.000 Daytime sleep duration − 0.432⁎⁎ 0.234 0.484⁎⁎ 1.000 Sleep environment − 0.330⁎ 0.456⁎⁎ 0.161 0.320⁎ 1.000 Subjective sleep quality 0.024 0.032 0.135 0.298⁎ 0.220 1.000 ⁎⁎ Significant at the 0.01 level (2-tailed). ⁎ Significant at the 0.05 level (2-tailed). Table options For the main sleep interval, the correlations between actigraphic and sleep diary lights out time was r = 0.472, P = .001, and between actigraphic and sleep diary final wake time was r = 0.745, P < .001. We expected these sleep parameters to be significantly correlated given that we interviewed participants about their actigraphic data using procedures developed by Acebo et al, 17 in which actigraphic sleep start and end times are corroborated with subjective (ie, sleep diary or interview) report. In addition, actigraphy and sleep diary estimates of nighttime wake minutes were significantly correlated (r = 0.5; P < .001) as were actigraphy and sleep diary estimates of nap length (r = 0.88; P < .001). See Table 5 for HLR unstandardized coefficients of the explanatory variables on daytime sleepiness, delinquency, conduct problems, and depression. Table 5. Unstandardized regression coefficients of explanatory variables on daytime sleepiness, delinquency, conduct disorder, and depression. Daytime sleepiness Delinquency Conduct disorder Depression B P B P B P B P Prenatal drug exposure 0.098 .868 − 0.259 .588 − 0.037 .892 0.531 .213 Adversity indexa 0.21 .936 0.6 .002⁎⁎ 0.332 .003⁎⁎ 0.235 .192 Sleep problem indexb 1.459 .027⁎ 0.991 .052 − 0.277 .322 0.082 .862 Daytime sleep durationc 0.106 .001⁎⁎ 0.034 .214 0.013 .455 Sleep environmentc Subjective sleep qualityc 1.871 .012⁎ 1.58 .01⁎⁎ 1.414 .016⁎ Bedtimec 0.636 .195 a With prenatal drug exposure in the model. b With prenatal drug exposure and the adversity index in the model. c With the adversity index in the model (except daytime sleep duration and subjective sleep quality on daytime sleepiness). ⁎⁎ Significant at the 0.01 level (2-tailed). ⁎ Significant at the 0.05 level (2-tailed). Table options Daytime sleepiness Sleep problem index In an HLR (test 1), prenatal drug exposure was entered in step 1; prenatal drug exposure and the adversity index, respectively, were entered in step 2; and prenatal drug exposure, the adversity index, and the sleep problem index, respectively, were entered in step 3, to determine the unique effect of the sleep problem index on daytime sleepiness adjusting for covariates. Neither prenatal drug exposure nor the adversity index was a significant predictor of daytime sleepiness. The sleep problem index was a significant predictor of daytime sleepiness, explaining 10.8% of the variance of daytime sleepiness ([1,42], P = .027) adjusting for prenatal drug exposure and postnatal adversity. Specific sleep predictors To determine which components of the sleep problem index predicted daytime sleepiness, the 6 components (nighttime sleep duration, bedtime, sleep schedule regularity, sleep environment, subjective sleep quality, and daytime sleep duration) were entered into a stepwise regression (test 2): daytime sleep duration (ie, nap duration) and subjective sleep quality were retained as significant predictors of daytime sleepiness (R2 = 0.28 [1,47], P < .001, and R2 = 0.09 [1,46], P = .012, respectively). Delinquency Sleep problem index The same initial HLR (test 1) was run to determine the unique effect of the sleep problem index on delinquency adjusting for prenatal drug exposure and the adversity index. Prenatal drug exposure was not a significant predictor of delinquency. The adversity index explained 20.4% of the variance of delinquency ([1,42], P = .002), and the sleep problem index explained an additional 7% of the variance of delinquency, which did not reach significance ([1,41], P = .052). Specific sleep predictors To determine which components of the sleep problem index predicted delinquency, the 6 sleep variables were entered into a stepwise regression (test 2): daytime sleep duration (ie, nap duration), subjective sleep quality, and bedtime (ie, sleep start time) were retained as significant predictors of delinquency (R2 = 0.2 [1,46], P = .002; R2 = 0.088 [1,45], P = .023; R2 = 0.067 [1,44], P = .039, respectively). These 3 sleep values remained significant (R2∆ = 0.025 [3,41], P = .008) above and beyond the adversity index (test 3). Conduct disorder symptoms Sleep problem index The initial HLR (test 1) was run to determine the unique effect of the sleep problem index on conduct disorder symptoms adjusting for prenatal drug exposure and the adversity index. Prenatal drug exposure was not a significant predictor of conduct disorder symptoms. The adversity index explained 18% of the variance of conduct disorder symptoms ([1,43], P = .003). The sleep problem index was not significant above and beyond the adversity index. Specific sleep predictors To determine if any components of the sleep problem index predicted conduct disorder symptoms, the 6 sleep variables were entered into a stepwise regression (test 2). Daytime sleep duration (ie, nap duration) was the only sleep value retained as a significant predictor of conduct disorder symptoms (R2 = 0.091 [1,47], P = .035). Daytime sleep duration was not significant, however, above and beyond the adversity index (test 3). Depressive symptoms Sleep problem index Among prenatal drug exposure, the adversity index, and the sleep problem index (test 1), there were no significant predictors of depressive symptoms. Specific sleep predictors To determine independent contributions of any component of the sleep problem index on depressive symptoms, the 6 sleep variables were entered into a stepwise regression (test 2). Subjective sleep quality was the only sleep value retained as a significant predictor of depressive symptoms. Sleep quality explained 14.2% ([1,47], P = .008) of the variance of depressive symptoms and remained significant (test 3) above and beyond the adversity index.