A cross-sectional, register-based study design was used comparing refugees with non-refugees with the same origin. As Sweden rarely accepts labour migrants from the same places of origin as current large refugee populations group, the non-refugees are people who have been granted residence permit in Sweden for reasons of family reunion with a refugee.
To be included, the person had to fulfil the following criteria according to the Swedish Population Registration System and the Swedish Board of Migration:
Aged 18 to 65 years and identified during the year 2006.
Born abroad and settled in Sweden.
Reside in Sweden for less than ten years.
Hold a resident permit either after being granted asylum in Sweden (refugees) or for reason of family reunion with a refugee (the non-refugee group).
Not hold a residence permit on reason for particularly distressing circumstances (formally called humanitarian grounds).
The reason for excluding persons admitted to Sweden on reason for particularly distressing circumstances are that these permits are given to persons in exceptional cases, often suffering from life-threatening diseases.
The study size was determined by the number of persons in Sweden fulfilling the inclusion criteria at the time of the study. Persons who can be assumed to have left the country without informing Swedish tax authorities were excluded with methods described by Weitoft. Of the original study population, 1.77% were excluded according to this criterion.
Register based studies use data in official registers collected for generic purposes. As a means of identification, all Swedish citizens, or people living in Sweden with a permanent residence permit, are assigned a personal identity number in the Population Registration System. After ethical approval and permission, it is possible to link registers for research purposes, and the data is made anonymous after collection. This study was performed with data from Statistics Sweden (http://www.scb.se/), the National Board of Health and Welfare (http://www.socialstyrelsen.se/) and the Swedish Board of Migration (http://www.migrationsverket.se/) administered by Statistics Sweden.
Mental ill health, as measured with the proxy variable psychotropic drugs purchased, was used as the outcome. The prescriptions include those from GPs and specialists but exclude those to inpatients (0.2% of the population). The relevant agents include antidepressants (ATC-code N06A), tranquilizers (ATC-code N05B) and sedatives (ATC-code N05C) but exclude antipsychotic agents (ATC-code N05A). The reasons for excluding antipsychotic agents are that antipsychotic agents are given for severe mental illnesses known to have a much more complicated aetiology than moderate mental ill health. There is also a need to separate the limited group with disabling psychiatric illnesses from those with severe psychological reactions to trauma and the majority who adapts once peace and order are restored among trauma affected populations.
The outcome variable is turned into a binary variable (has filled the prescription or not). Note that the outcome is not a proxy for any specific mental health state or diagnosis (such as PTSD or depression), but merely a measure of the physician's decisions.
Sweden grants asylum to refugees according to the Geneva Convention and to those with a well-grounded fear of suffering the death penalty, torture or who need protection due to an internal or external armed conflict or an environmental disaster in their native country, as mentioned above.
The non-refugees are all family members of those granted asylum in Sweden, such as spouses, children under 18 years and, in exceptional cases, relatives over 18, who are granted residence permit in Sweden for reasons of family reunion.
Gender and the country or area of origin; including Afghanistan, Iraq, Iran, the Middle East (except Iran and Iraq), Somalia and former Yugoslavia. The countries or areas of origin represent origins with large refugee groups in Sweden during the year 2006. The Iraqi group was chosen as a reference category. A low-income country is here defined as a country not classified as a high-income country by the World Bank at the time the immigrants arrived in Sweden.
Age, marital status, education and duration of stay in Sweden.
Age was coded into five dummy age groups: 18-24, 25-34, 35-44, 45-55 and 55-64 years. The 18-24 age-group was chosen as a reference category.
Marital status was divided into single, married, divorced and widow(-er). The marital status group, single, was chosen as a reference group.
Education was coded in the following dummy groups: less than 9 years, 9 years, 12 years, more than 12 years of schooling and unknown length of schooling. Statistics Sweden converted education completed outside Sweden into equivalent levels of schooling in Sweden. The education group, less than 9 years of schooling, was chosen as a reference category.
Duration of stay was coded as ten dummy groups by year.
There is always a risk of misclassification when studying specific diagnoses of mental ill health among immigrants due to inter-cultural diversity. We therefore combined antidepressants, tranquilizers and sedatives into a single dichotomous outcome variable to minimise the impact of differential diagnoses in groups from different countries or areas of origin. However, some patients are prescribed psychotropic drugs without having mental ill health, such as sedatives prescribed to patients with rheumatic pain. In order to test if there were differences in the patterns of purchase of antidepressants, tranquilizers or sedatives we made a sensitivity test of the outcome. There were no different patterns for any of the outcomes.
Demographic variables were analysed using chi square tests. The association between the outcome: psychotropic drugs purchased, and the covariates, and the outcome and the potential confounders were analysed using logistic regression. Adjusted multiple logistic regression analysis was used for the association between the outcome, exposure and potential confounders. For the adjusted model, several effect modifiers (indicated with a star * in the text) were tested with stepwise logistic regression. The -2Log Likelihood value was used to assess what model had the best fit. Results are presented as odds ratios (OR) with 95% confidence limits.
This study was approved by Stockholm Regional Ethical Review Board (2008/732-31).