Study Population and Data Collection
The study population consisted of a cohort of young adults (Figure 1), 20–24 years old (corresponding to the United Nations' definition of young adults). Ten thousand men and 10 000 women, born between 1983 and 1987, were randomly selected from the general population (from a registry held by the Swedish Tax Agency), 50% living in the County of Västra Götaland, Sweden, and 50% in the rest of the country. In October 2007, a questionnaire containing questions about health, work, and leisure-related exposure factors, background factors, and psychosocial factors was sent by post to the selected population. Besides returning the postal questionnaire it was also possible to respond to the questionnaire via the web if desired. A lottery ticket (value approx. 1 Euro) was attached to the cover letter and could be used whether the recipient participated in the study or not. Two reminders were sent by post. The response rate at baseline was 36% (n = 7125). One year later, those respondents who had indicated that they would accept to be offered to participate in further studies (n = 5734) were invited to respond to an identical questionnaire, this time administered via the web. The data collection process was otherwise similar to that at baseline, but with the addition of a third reminder offering a paper version of the questionnaire and two cinema tickets to respondents. The response rate at follow-up was 73% (n = 4163). After excluding those who failed to respond to both questions concerning frequency of mobile phone and SMS use at baseline, 4156 remained in the study group. All in all, non-participation and dropout from the study was 79% (Figure 1).
Mobile Phone Exposure Variables
Information about mobile phone exposure was collected from the baseline questionnaire. This included the average number of mobile phone calls made and received, and of SMS messages sent and received, per day, but also more qualitative aspects of mobile phone use, including how often the respondent was awakened at night by the mobile phone, how he or she perceived demands on availability, and whether he or she perceived the accessibility via mobile phones to be stressful, as well as perceptions regarding personal overuse of the mobile phone. Responses were divided into high, medium, and low categories, based on the frequency distribution of responses, except for overuse which was categorized according to number of items confirmed. A combined quantitative mobile phone use variable was constructed by merging the variables frequency of calls and frequency of SMS messages (Spearman correlation r = 0.35, p < 0.0001). The mobile phone use variable correlated well with the original calls and SMS variables (r = 0.73, p <.0001, and r = 0.84, p <. 0001, respectively).
Mobile phone variables, questionnaire items, response categories, and response classifications are presented in Table 1.
Mental Health Outcome Variables
Information about mental health symptoms was collected from the cohort study questionnaire at baseline and at follow-up.
The outcome variable Current stress was constituted by a validated single-item stress-indicator: Stress means a situation when a person feels tense, restless, nervous, or anxious or is unable to sleep at night because his/her mind is troubled all the time. Are you currently experiencing this kind of stress? Response categories were: a = not at all, b = just a little, c = to some extent, d = rather much, e = very much. The responses were divided into Yes (responses d-e) and No (responses a-c), based on frequency distribution while yet taking content of response categories into account.
The Sleep disturbances variable was constructed by including the most common sleep disorders (insomnia, fragmented sleep and premature awakening) into a single-item, adapted from the The Karolinska Sleep Questionnaire: How often have you had problems with your sleep these past 30 days (e.g., difficulties falling asleep, repeated awakenings, waking up too early)? Response categories were: a = never, b = a few times per month, c = several times per week, and d = every day. The responses were divided into Yes (responses c-d) and No (responses a-b), based on clinical significance.
Symptoms of depression (one item) and symptoms of depression (two items) were made up by the two depressive items from the Prime-MD screening form:During the past month, have you often been bothered by: (a) little interest or pleasure in doing things? (b) feeling down, depressed, or hopeless? Response categories were Yes and No. It is proposed that it is sufficient if one of the two items is confirmed in screening to go forward with clinical assessment of mood disorder. This procedure has high sensitivity for major depression diagnosis in primary care populations.[32,33] In our cohort study group, approximately 50% of the men and almost 65% of the women confirmed at least one of the two depressive items, which indicates that the instrument is probably very sensitive but has low specificity in our study group. Therefore, we constructed two outcomes: Symptoms of depression (one item), in which the Yes category contained those who confirmed only one of the depressive items, and Symptoms of depression (two items), in which the Yes category contained those who confirmed both depressive items. The No category in both outcomes contained those who disclaimed the two depressive items.
Background Factors and Social Support
Background factors were collected to describe the study group and to adjust for in the multivariate analysis, including: relationship status: single or in a relationship; highest completed educational level: elementary school (basic schooling for 6–16-year-olds), upper secondary school, or college or university studies; and occupation: working, studying, or other (other included being on long-term sick leave, or on parental or other leave, or being unemployed). The variable social support was based on the item: When I have problems in my private life I have access to support and help, a one-item adaptation of the social support scale in the Karasek-Theorell job content questionnaire, here relating to private life (rather than work life). Response categories were: a = applies very poorly; b = applies rather poorly; c = applies rather well; d = applies very well. The responses were categorized as low (response categories a and b), medium (response category c), and high (response category d).
All analyses were performed using the statistical software package SAS, version 9.2 (SAS Institute, Cary, NC, USA). Spearman correlation analysis was used to examine associations between the mobile phone exposure variables, and between mobile phone use and social support. The Cox proportional hazard model (PHREG proc with time set to 1) was used to calculate prevalence ratios (PRs) with a 95% confidence interval (CI) for multivariate analysis of cross-sectional and prospective associations between exposure variables and mental health outcomes. The robust variance option (COVS) was used in the cross-sectional analysis to produce adequate CIs.[35,36] The low category in each exposure variable was used as reference level. The PRs were adjusted for background factors including relationship status, educational level, and occupation at baseline. Missing values (non-responses to items) were excluded from the analyses, which means that the n varied in the analyses. Prevalence ratios with a CI not including 1.00 (before round-off) were considered statistically significant. In the prospective analysis, subjects who reported symptoms at baseline were excluded from the analysis of the mental health outcome variable concerned. All analyses were done separately for the men and women.
The study was approved by the Regional Ethics Review Board in Gothenburg, Sweden (Reg. no. 191-05).
BMC Public Health. 2011;11 © 2011 BioMed Central, Ltd.
© 1999-2006 BioMed Central Ltd
Cite this: Mobile Phone Use and Stress, Sleep Disturbances, and Symptoms of Depression among Young Adults - A Prospective Cohort Study - Medscape - Jan 01, 2011.