Study design and setting
This cross-sectional study is based on data from the national health survey “Health and Wellbeing of Icelanders” gathered from October 2017 until February 2018 by the Directorate of Health in Iceland19.
Participants
The data included 6,776 of the 9,887 eligible participants, Fig. 1 demonstrates their response rate.
The Icelandic Directorate of Health attained a random population sample from Statistics Iceland, including all Icelanders older than 18 living in Iceland at the time. The response rate was 68.5% [19],a total of 6,776 respondents, and the sample was stratified to ensure sufficient participation of all age groups and from all geographical areas. We only analysed those who completed the Depression Anxiety and Stress – 21 scale (DASS-21-scale) questionnaire.
Outcome measures and co-variables
The primary outcomes of interest were scores on the shorter version of the DASS [20] scale called DASS-21 [21], and includes 21 of the 42 questions of the DASS scale. There are seven questions for each part (depression, anxiety, and stress) designed to measure the severity of symptoms common to depression, anxiety, and stress during the previous week. Participants were asked to rate the extent to which they had experienced each symptom on a 4-point Likert scale with possible scores for each answer ranging from 0 (did not apply to me at all) to 3 points (applied to me very much/most of the time), the higher a score, the worse the symptoms. The total scores for each subgroup are divided into five categories: normal (0–9 for depression, 0–7 for anxiety and 0–14 points for stress), mild (10–13 for depression, 8–9 for anxiety and 15–18 points for stress), moderate (14–20 for depression, 10–14 for anxiety and 19–25 points for stress) severe (21–27 for depression, 15–19 for anxiety and 26–33 points for stress) and extremely severe (all scores above). The score attained from the DASS-21 are doubled to yield a score, comparable to the DASS scale.
The DASS-21 scale is reliable and has been validated in clinical [20] and non-clinical samples [22]. The scale has been validated in Icelandic and is deemed to have the same correlation between DASS and DASS-21 as in other languages and sufficient validity (Ingimarsson, B. The psychometric testimonials on the DASS Self-Assessment Scale. Depression, anxiety and stress. Unpublished cand. psych. Dissertation).
Co-variables
Sex was defined as male or female. Age was obtained by participants’ year of birth as documented in the survey and divided into 10-year ranges with the last age range, 70 + years, being open ended. Education was the level of attained education, divided into three: basic, middle, and university, with basic indicating compulsory education from 6 to 16 years of age, middle representing secondary school and technical or vocational training, and university representing degrees starting with the bachelor level.
Marital status contained five categories: married/cohabiting, dating, divorced, single, and widowed. Financial difficulty was assessed by the question: “How easy or difficult has it been for you and your family to make ends meet over the past 12 months?”. Possible answers: very easy, rather easy, neither, difficult, and very difficult. Residency was defined as either urban or rural, with the former group living in the greater Reykjavik area.
Unemployment was dichotomized into unemployed or not. Information on current and previous medical treatment for mental disorders was obtained from three questions “Have you taken prescribed medication for depression/anxiety/other mental disorders?”, with possible answers (a) “yes, in the past two weeks” (currently), (b) “yes, more than two weeks ago” (previously) and (c) “no, never” (never).
Statistical methods
Data were analysed using StataIC (version 15). We calculated the scores on DASS-21 for all participants and divided into three groups: Normal, mild/moderate (called medium), or severe/extremely severe (called high) scores for each subgroup (depression, anxiety, and stress).
Scoring ranges are presented as follow: normal (0–9 for depression, 0–7 for anxiety and 0–14 points for stress), medium (10–20 for depression, 8–14 for anxiety and 15–25 points for stress) and high (> 20 for depression, > 14 for anxiety and > 25 points for stress).
(The distribution of the full DASS-21 scores by severity groups and by gender are displayed in Additional file 1, supplementary tables 4 and 5).
The data were analysed by multinomial logistic regression adjusting for co-variates: age, sex, education, marital status, financial difficulty, living area, unemployment, and current and previous use of psychiatric medication for mental disorders.
This approach allowed us to estimate the relative risk ratio (RRR) within each group of participants, comparing the risk of scoring medium or high to scoring normal and controlling for the confounding variables.
To study effect modification by medication, and sex we performed sub-group regression analysis and evaluated the significance of interaction effects in the regression analysis, adjusting for other covariates.
Multinomial logistic regression was performed to look at dissimilarities in education between the sexes, since studies have demonstrated that men with lower educational levels seek less medical help for mental health problems [23] than women. A likelihood-radio test was performed additionally to see if antidepressant medication intake was an effect modifier of the gender differences in depression scores. (Additional file 1, supplementary tables 1,2 and 3).