The Lower Rio Grande Valley (LRGV) comprises 4 counties in Texas on the Mexico border: Cameron, Hidalgo, Starr, and Willacy. LRGV had 978,369 residents in 2000; residents are predominantly Hispanic (88%), of Mexican American origin (86% of Hispanics), and low income (35% below the poverty threshold in 1999) (16). Brownsville is the county seat of Cameron County; its sister city in Mexico is Matamoros in the state of Tamaulipas. Matamoros is less than 3 miles south of Brownsville and in 2000 had a population of 412,544 (17).

Data for this study came from 2 school-based surveys. The first survey (18) was conducted in English during the 2000-2001 school year in LRGV. This survey was then translated into Spanish, back-translated into English, and cross-culturally adapted for Mexican nationals. This second survey was conducted during the 2002-2003 school year in Matamoros (12). A full description of both surveys, study designs, participants, and samples has been reported previously (12,18). Briefly, ninth-grade students from both sides of the Texas-Mexico border were surveyed to determine risk behaviors that contribute to illness and death among adolescents. Both surveys were self-administered during regularly scheduled classes. Passive parental consent was obtained in LRGV and active parental consent was obtained in Matamoros; the children provided consent in both surveys. Survey protocols were reviewed and approved by the University of Texas Health Science Center at Houston's Committee for the Protection of Human Subjects and by the Secretariat of Education of the state of Tamaulipas.

All ninth-grade students from randomly selected schools in LRGV were invited to participate. The LRGV sample included 4,901 students, all of Hispanic origin, from 13 of 18 high schools; this sample represented 23% of ninth-grade students in LRGV during the 2000-2001 school year. In Matamoros, we randomly selected 8 of 28 high schools. All ninth-grade students from selected schools were invited to participate in the survey. The Matamoros sample included 669 students and represented 12% of ninth-grade students in Matamoros during the 2002-2003 school year. The response rates for the surveys were 67% from the LRGV and 58% from Matamoros. We did not collect data on students who did not participate.

The surveys asked 7 questions about sleep patterns and problems related to sleep. These questions were used previously to operationalize Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (19), symptom criteria for sleep disturbance (5,6). Two questions were related to amount of sleep reported in the past 4 weeks on weekday and weekend nights. Restorative sleep was assessed with the question, "During the past 4 weeks, how often have you felt really rested when waking up in the morning?" Overall quality of sleep was assessed with the question, "During the past 4 weeks, how would you rate the quality of your sleep overall?"

Three questions assessed problems related to sleep disturbance: “During the past 4 weeks, how often would you say you have had any of these problems related to your sleep: 1) trouble falling asleep (difficulty initiating sleep), 2) waking up in the middle of the night and finding it hard to get back to sleep (difficulty maintaining sleep), and 3) waking up very early and not being able to get back to sleep (early morning awakening)?” An adolescent who experienced at least 1 of these problems often or almost every day in the past 4 weeks was defined as having disturbed sleep.

We analyzed demographic covariates as potential risk factors for disturbed sleep among adolescents; these covariates included sex, age, ethnic origin, employment status, and perceived family standard of living (used as a proxy for socioeconomic status). Ethnic origin was assessed in the LRGV cohort with the question, "If you are Hispanic or Latino in origin, are you Mexican American (born in United States), Mexican American (born in Mexico), Mexican national, or other Hispanic or Latino (Cuban, Central American, Puerto Rican, etc)?" Because of regulations, we could not ask about country of origin in Matamoros, and all students were considered to be Mexican nationals. Perceived family standard of living was collapsed into 3 categories: very well off, living comfortably or just getting along, and nearly poor or poor.

As an indicator of mental health, the survey also included questions on substance use. We asked about the frequency of smoking cigarettes, drinking alcohol, and using marijuana in the past 30 days. We also asked about lifetime use of cocaine and steroids. To measure mood disturbance, we asked, "During the past 12 months, did you ever feel so sad or hopeless almost every day for 2 weeks or more in a row that you stopped doing some usual activities?" Attempting suicide was measured with the question, "During the past 12 months, did you ever seriously consider attempting suicide?"

LRGV students self-reported their height and weight, and Matamoros students had their height and weight measured to the nearest 1 mm and 0.1 kg, respectively, while the students were wearing light clothing and no shoes. Body mass index (BMI) was calculated by dividing weight in kilograms by the square of the height in meters. We used BMI-for-age growth charts (http://www.cdc.gov/growthcharts/) to classify students as underweight or normal weight (<85th percentile), overweight (≥85th percentile to <95th percentile), or obese (≥95th percentile).

We assessed violence by asking whether students had been in a physical fight in the past 12 months. We also asked whether they had ever been forced to have sex. We measured 1 unhealthy dietary behavior by asking, "During the past 30 days, did you go without eating for 24 hours or more to lose weight or to keep from gaining weight?"

Both surveys used a multistage stratified cluster design combined with probability proportional to school size sampling. Sampling weights were used in the analysis to account for the multistage stratified sampling design. In Matamoros, nonresponse adjustment and ratio adjustment for population-fixed totals were used from school records to ensure that the sex composition of the sample was the same as that of the total school enrollment. Adjustment was not done in the LRGV survey because of limitations in the sampling frame. Weighted percentages, means, standard errors, and test statistics were calculated with SUDAAN version 9.0 (RTI International, Research Triangle Park, North Carolina). Multiple logistic regression was used to estimate adjusted odds ratios with their corresponding 95% confidence intervals. Differences were considered significant at P < .05.

Our study has several limitations. First, our measure of disturbed sleep was not inclusive. We investigated hours of sleep, restorative sleep, and other symptoms of disturbed sleep, but we did not examine parasomnias or formal diagnostic criteria for insomnia (19). Second, we relied on self-reported data, although these are widely used in community surveys and correlate well with data obtained by objective measures of disturbed sleep (19,25). A third limitation was that our sleep items asked whether participants experienced symptoms of disturbed sleep almost every day for the past 4 weeks. Therefore, we could not differentiate between students with chronic versus acute sleep problems. A fourth limitation was conducting the survey in Matamoros 2 years after the survey was conducted in LRGV; however, results should be comparable because of the lack of interventions during those 2 years. A fifth limitation of our surveys is that we did not measure any chronic diseases, which may influence the sleep process (6).

The challenge of overcoming the high prevalence of disturbed sleep among adolescents on the Texas-Mexico border region is immense. Teaching adolescents about healthy sleep habits may reduce this prevalence. Nevertheless, studies are needed that identify additional risk factors and potential strategies to reduce disturbed sleep among adolescents.