Precarious employment is a risk factor for poor mental health in young individuals in Sweden: a cohort study with multiple follow-ups
Canivet et al. BMC Public Health
Precarious employment is a risk factor for poor mental health in young individuals in Sweden: a cohort study with multiple follow-ups
Catarina Canivet 0 3
Theo Bodin 1 2
Maria Emmelin 0 3
Susanna Toivanen 4
Mahnaz Moghaddassi 0 3
Per-Olof Östergren 0 3
0 Division of Social Medicine and Global Health, Department of Clinical Sciences Malmö, Lund University , SE-205 02 Malmö , Sweden
1 Institute of Environmental Medicine, Karolinska Institutet , SE-171 77 Stockholm , Sweden
2 Division of Occupational and Environmental Medicine, Lund University , SE-221 85 Lund , Sweden
3 Division of Social Medicine and Global Health, Department of Clinical Sciences Malmö, Lund University , SE-205 02 Malmö , Sweden
4 CHESS (Centre for Health Equity Studies) , Sveavägen 160, SE-106 91 Stockholm , Sweden
Background: The globalisation of the economy and the labour markets has resulted in a growing proportion of individuals who find themselves in a precarious labour market situation, especially among the young. This pertains also to the Nordic countries, despite their characterisation as well developed welfare states with active labour market policies. This should be viewed against the background of a number of studies, which have shown that several aspects of precarious employment are detrimental to mental health. However, longitudinal studies from the Nordic region that examine the impact of precarious labour market conditions on mental health in young individuals are currently lacking. The present study aims to examine this impact in a general cohort of Swedish young people. Methods: Postal questionnaires were sent out in 1999/2000 to a stratified random sample of the Scania population, Sweden; the response rate was 58 %. All of those who responded at baseline were invited to follow-ups after 5 and 10 years. Employment precariousness was determined based on detailed questions about present employment, previous unemployment, and self-rated risk of future unemployment. Mental health was assessed by GHQ-12. For this study individuals in the age range of 18-34 years at baseline, who were active in the labour market (employed or seeking job) and had submitted complete data from 1999/2000, 2005, and 2010 on employment precariousness and mental health status, were selected (N = 1135). Results: Forty-two percent of the participants had a precarious employment situation at baseline. Labour market trajectories that included precarious employment in 1999/2000 or 2005 predicted poor mental health in 2010: the incidence ratio ratio was 1.4 (95 % CI: 1.1-2.0) when excluding all individuals with mental health problems at baseline and adjusting for age, gender, social support, social capital, and economic difficulties in childhood. The population attributable fraction regarding poor mental health in the studied age group was 18 %. Conclusions: This study supported the hypothesis that precarious employment should be regarded as an important social determinant for subsequent development of mental health problems in previously mentally healthy young people.
Longitudinal studies; Mental health; Unemployment; Stress; Psychological; Young adult; Employment
Due to several decades of globalisation combined with
predominantly neoliberal economic policies, long-term
employment contracts and a high level of job security is
no longer the dominant form of labour market relation
in a global perspective. Instead, short-term contracts,
involuntary part time employment, employment through
‘staff-for-hire’ enterprises, and shorter or longer periods
of unemployment have become increasingly common
]. Already in 1998, in a report from the European
Foundation for the Improvement of Working and Living
Conditions (Eurofound; published online in 2012), it was
shown that fixed-term and temporary contracts were
associated with low incomes, short working hours, and
poor working conditions [
]. The temporary
employment rate in the 27 EU countries has since climbed from
11.2 % in 2001 to 12.8 % in 2012. The corresponding
figures for Sweden are 13.8 % and 14.7 % [
Several aspects of precarious employment have been
shown to be detrimental to mental health in a number
of population surveys. A recent review by Benach et al.
found studies supporting a linkage between mental
health problems and major organisational restructuring
and downsizing, perceived job insecurity, and temporary
]. New research indicates that the latest
financial crisis of 2008 led to an increase in poor mental
health among the European population [
]. Since there
is a strong likelihood of bidirectional causality between
labour market factors on the one hand and mental
health on the other [
], longitudinal studies are of
paramount importance in order to disentangle causal
The impact on health of labour market relations
among young individuals has been studied
longitudinally in the Northern Swedish Cohort, which was
established in 1981 [
]. Unemployment was associated with
subsequent mental health problems [
], and similar
results have been found in studies on young cohorts in
] and US [
]. In another longitudinal study,
from New Zeeland, the association between
unemployment and poor mental health disappeared after
adjustment for confounders and reverse causality [
]. It is
sometimes assumed that persons early in their working
life may be able, later on, to overcome economic and
health setbacks associated with unemployment. This
was one of the working hypotheses in a recent large
meta-analysis covering studies on unemployment and
mental health. However, contrary to expectations, the
effect of unemployment was larger among persons of
young (and old) age versus those in middle-age [
In line with this finding are studies demonstrating a
‘scarring’ effect of unemployment in youth, i.e. that the
difference in mental health between those who have
been exposed to labour market insecurity, versus those
who have not, tends to become more pronounced over
It is often argued that well developed welfare states
with active labour market policies can buffer the
negative effects of a precarious labour market [
]. In a
recent Swedish study, it was found that programs directed
at young persons at risk of labour market exclusion led
to a positive mental health development later in life,
compared to those who experienced unemployment
during the same time period [
Young persons are overrepresented both among those
unemployed and among those in temporary employment
]. In Sweden, the proportion of persons aged 20–
24 who were not in education, employment, or training
(NEET) was 9.9 % as of 2014 . In 2015, temporary
contracts were held by 33% of all women aged 20–34
active in the labour market, and by 27% of the men, while
the corresponding figures for the whole labour force
were 16% and 14% [
]. Moreover, prevalence figures do
not reveal the true proportion of the young population
that is exposed to unemployment or a precarious work
situation over time, and the number affected over a
longer period of time could be expected to be considerably
There are few longitudinal studies from the Nordic
countries, as well as from an international perspective,
that examine the situation of young people entering the
labour market and that use data sets collected after year
2000. The aim of the present study was thus to
investigate the associations between precarious employment
situations and mental health later in life among young
adults aged 18–34 in a Swedish setting, using a
longitudinal study design.
The Scania Public Health Cohort, Sweden, was
established in 1999/2000; for details see Carlsson et al. [
In short, 58 % (N = 13604) of a stratified random sample
of the Scania population, 18–80 years of age, responded
to a postal questionnaire, which was sent out again in
2005 and 2010 to all responders. Out of those who were
18–34 years old at baseline (N = 3420), 1802 participated
also in the two follow-up surveys. The cohort selected
for the present study consists of those (N = 1135) who
had complete data from 1999/2000, 2005, and 2010,
both on employment precariousness and the outcome
measure, i.e. mental health status, and who, according to
their inquiry responses, were considered to be part of, or
close to the labour market (see below).
Outcome variable – poor mental health
Mental health was measured at all three time points
using the 12-item version of the General Health
Questionnaire (GHQ-12). We used the 0-0-1-1 scoring
method recommended by the creators of the instrument
(range 0–12), with poor mental health or
‘GHQ-caseness’ defined as a scoring of 2 or higher [
At the time of the baseline survey, a validated measure of
employment precariousness was not available. However,
the health effects of precarious employment have been
described using several research approaches, included
focusing on a) temporary vs. permanent employment, b)
perceived job insecurity, and c) major organisational
restructuring and downsizing [
]. The construct of work
history has also been proposed to constitute a useful
element in defining precariousness [
], and unemployment at
a young age was recently shown to predict later
unemployment during 15 years of follow-up [
]. We chose
to create a dichotomous variable – non-precarious (NP)
vs. precarious employment (PE) – based on a combination
of data on present unemployment, previous
unemployment, temporary vs. permanent employment, and
perceived job insecurity. The questions asked were as
Which of the following applies best to you at the
present moment (response alternatives: working for
a living, student, no work outside home, disability
pension/long term sick leave, unemployed)?
If employed, which are the terms of your contract?
Response alternatives were ‘permanent’, ‘substitute’,
‘fixed term’, and ‘on demand’. Several alternatives
could be chosen, and the responses were dichotomised
into ‘permanent (only)’ and ‘contingent’.
Have you been involuntarily unemployed at some
point during the past three years (no, yes)?
How do you rate your own risk of involuntary
unemployment within the coming year? (high,
moderate, low, none, ‘I do not wish to work a year
Those on disability pension/long tem sick leave were
excluded from the cohort, as well as those who answered
that they did not wish to work a year from now, since
they were not considered to be relevant for the target
group in focus of this research, i.e. young persons active
in the labour market.
Those categorised as NP were those with permanent
work and no or low self-rated risk of future
unemployment and those with contingent work and no self-rated
risk of future unemployment and no previous
unemployment. Those categorised as PE were others with
contingent work, others with previous unemployment, all
those with moderate to high self-rated risk of future
unemployment, and all presently unemployed.
The following measures were also assessed (for details
about inquiry questions and response alternatives, see
Additional file 1): Age (continuous variable), country of
origin, marital status, education level at baseline, economic
difficulties, alcohol consumption [
] physical activity [
emotional and instrumental support, social participation
], and social anchorage with neighbourhood.
We used economic difficulties in the family while
growing up as a proxy for socioeconomic position. The
response alternatives to this question were ‘no, none
worth mentioning’, ‘yes, slight or relatively short periods
with economic difficulties’, and ‘yes, severe and/or longer
periods with economic difficulties’. The dichotomisation
was performed at the level of ‘no’ versus the other two.
The study was approved by the Research Ethics
Committee of Lund University (1999–99; 2005–471, and
The relationships between background factors and poor
mental health in 2010 are presented as percentages and
age-adjusted incidence rate ratios (IRR) which is a good
estimate of relative risks, using a modified Poisson
regression model with robust standard errors [
In an attempt to clarify possible causal mechanisms,
several analyses were performed in which all participants
with poor mental health at baseline (N = 349) were
excluded. In Table 4, PE in 1999/2000 or 2005 is thus
tested against poor mental health in 2010 with the
stepwise addition of possible confounders, measured at
baseline. Since emotional and instrumental support at
baseline correlated strongly (r = 0.5 in this group), a
composite variable was used (no exposure versus
exposed to either or both) in the multivariate analyses. The
population attributable fraction (i.e. the proportion of
disease cases over a specified time that would be
prevented following elimination of the exposures, assuming
the exposures are causal) [
] for PE in poor mental
health was calculated using the formula: PAF = pd × (RR
− 1)/RR [
], where pd was the proportion of cases
exposed to PE, and the RR was the IRR for poor mental
health, after full adjustment with the same covariates as in
Table 3. Tests for effect modification were performed for
economic difficulties in childhood, gender, education level,
and age groups. This was done by creating dummy
variables and synergy indexes, as proposed by Rothman [
Two standard statistical analysis programs were used,
i.e. SPSS version 22.0 and Stata version 12.
In the original study population (N = 3420) of young
persons participating in the baseline inquiry in 1999/2000,
there were 1802 (53 %) who also took part in both of the
surveys in 2005 and 2010. Those who did not take part
in both surveys (N = 1618, 47 %) were to a higher degree
younger (mean age 26 vs. 27), of male gender (50 vs
41 %), of low education level (65 vs. 51 %), born outside
Sweden (13 vs 9 %), in economic difficulties (38 vs.
27 %), affirming poor self-rated health (27 vs 23 %), and
in a PE situation (52 vs 43 %). However, there was no
difference in mental health status; thus, 33 vs. 31 % had
a GHQ-score indicating poor mental health (Pearson
chi-square 1.4; p = 0.24). (The numbers missing in the
above-mentioned analyses varied from 0 to 81, except
for PE: missing = 781, whereof 410 from the group of
Both among respondents and non-respondents, the
baseline cross-sectional association between PE and
poor mental health was strong, with a Pearson chi
square value in respondents and non-respondents of
56.7 and 32.7, respectively (both p < 0.001).
Thereafter, in the group of participants in all three
surveys (N = 1802), those constituting the final study cohort
(N = 1135, with complete data, see Methods) were
compared to those with lacking data (N = 667). There were
no differences regarding gender, country of birth,
economic difficulties, self-rated health, or mental health.
However, a low education level was more common
among those with lacking data (56 vs. 49 %).
Data on employment precariousness is presented in
Table 1. In this cohort, the proportion of participants in
a PE situation decreased during the 10-year study
period, from 42.1 % in 1999/2000 to 22.9 % in 2010. The
individual trajectories are seen in Table 2. The most
common trajectory (40.2 %) was having a NP situation
at all measurement points. The next most common
trajectory was PE – NP – NP (17.9 %), whereas 11.0 % had
a PE situation at all measurement points. At baseline, 62
persons were unemployed (5.5 %) and at follow-up 53
persons, but only seven had been unemployed at both
time points (not shown in Table).
The proportion of individuals with poor mental health
was 31 % in 1999/2000, 31 % in 2005, and 26.5 % in
Table 3 shows that PE in 1999/2000 or 2005 was
associated with poor mental health at follow-up in 2010. The
age-adjusted IRR was 2.0 (95 % CI (confidence interval):
1.6–2.6) for those exposed to PE in both 1999/2000 and
As can be seen in Table 4, excluding all persons with
poor mental health at baseline, PE in 1999/2000 or 2005
was associated with poor mental health at follow-up also
in the fully adjusted model; the IRR was 1.4 (95 % CI:
Among those mentally healthy at baseline, and after
adjustment with the same variables as in Table 4, the
population attributable fraction for PE in 1999/2000 or
2005 and poor mental health at follow-up was 18 %.
No effect modification on the relationship between PE
and mental health was found for economic difficulties in
childhood, age groups, gender, or education level (results
This study found that among previously mentally
healthy young adults in southern Sweden, labour market
trajectories including a precarious employment situation
in 1999/2000 or in 2005 was a predictor for poor mental
health in 2010. The IRR was 1.4 (95 % CI: 1.1–2.0) after
adjustment for age, gender, emotional and instrumental
support, social participation and neighbourhood
anchorage, and economic difficulties in childhood. The
calculated population attributable fraction for a precarious
employment situation regarding poor mental health in
this group was 18 %.
Findings in relation to other studies and possible mechanisms
Mental health problems have become increasingly
common among young people during the last decades, and
particularly so in Sweden. In 2011, seven percent of men
and 10 % of women aged 16–24 years old had been in
touch with a psychiatric clinic or used psychotropic
]. It has been suggested that recent changes in
the labour market, increasing the risk of experiencing
precariousness, have contributed to the rapidly
deteriorating trend in mental health among young people [
In a meta-analysis from 2002, covering international
studies on job insecurity and its consequences, a
relatively strong effect size for mental health was found
(mean correlation: −0.237) [
]. Several recent
] and longitudinal [
] studies also
support the relationship found between job insecurity
and mental health problems. Benach et al. suggest that
one pathway from job insecurity to adverse health
outcome may consist of stress response to sustained
uncertainty, unpredictability, and lack of control over the
The current study primarily focuses on aspects of
employment insecurity in the definition of PE used here.
Another link between PE and poor mental health may
consist of poorer working conditions for persons with
PE. These may include poorer supervision, inadequate
training, exposure to higher risk tasks, lack of workplace
voice, economic and reward pressures, disorganisation at
the workplace, and regulatory failure [
In one study, insecure employment negatively affected
the likelihood of getting married and having children,
which could be a mediating factor for poorer mental
health in young people [
Fear of the stigmatisation connected to a precarious
labour market may also play a part. In a recent
qualitative study, indications of negative stereotyping about
‘unemployed persons’ were found among nurses working
in a healthcare program for job seekers [
]. On the
other hand, it could be argued that since precarious
employment is steadily increasing, and in particular among
young people, the potential stigmatisation of not having
a stable employment may become less pronounced with
This study was performed in Sweden. The degree of
health impairment of individuals with precarious
employment was least in Scandinavian settings in one
comparative study [
]. Scandinavian societies have been
characterised as egalitarian welfare states, with effective
collective bargaining institutions, lifelong job training,
and generous unemployment schemes, all of which may
contribute to a buffering effect. Therefore, it could be
hypothesised that a similar study performed elsewhere
might have shown even stronger associations between
precarious employment and poor mental health.
However, a study performed in Sweden today might
yield greater associations compared with previous
decades, since ‘Scandinavian welfare’ is not a constant. A
recent Swedish study showed that mental distress among
women increased between 2006 and 2010, and more so
among groups outside the labour market [
authors suggested that one of the reasons might be the
considerable modifications, e.g. stricter eligibility criteria
and lower benefit levels, which have been implemented
by the Swedish social insurance system during the last
Moreover, the global trend towards precarious labour
market relations seems different than previous cyclic
unemployment situations. Not only have the past decades
of neoliberal politics, with general deregulation and
privatisations, led to a shift in power relations characterised
by a markedly increased influence of employers vis-à-vis
], but also, a broad range of changes have
occurred regarding the individual’s relation to society
and the capacity of the welfare state to buffer the
negative impacts among those exposed to this situation. The
British sociologist Guy Standing has developed the
concept of precarity, and he states: ‘This is not just a matter
of having insecure employment, of being in jobs of limited
duration and with minimal labour protection, although
all this is widespread. It is being in a status that offers
no sense of career, no sense of secure occupational
identity and few, if any, entitlements to the state and
enterprise benefits that several generations of those who saw
themselves as belonging to the industrial proletariat or
the salariat (non-manual employees with secure
employment) had come to expect as their due.’  p. 24.
Standing also suggests that it is of major public health
0.8 – 1.4
1.3 – 1.9
1.4 – 2.0
1.2 – 1.8
interest to assess the potential negative health impacts of
such living conditions for the increasing numbers of
persons living in this ‘class-in-the-making’ [
The study was designed to investigate a causal relation
between a PE situation and mental health problems in
young persons. We thus used a model excluding all
persons with poor mental health at baseline and followed
the trajectory over two subsequent follow-up occasions,
which reduces the risk of reversed causality.
Other strengths of our study include the adjustment
for a substantial number of potential confounders and
the recruitment from a large random general population
sample. However, a first selection took place in the very
establishment of The Scania Public Health Cohort, since
the response rate was 58 % for women 18–30 years old
and 45 % for men in the same age range [
our study group (participating in all three surveys)
differed from the original cohort regarding several
characteristics (education level, country of origin, economic
difficulties, and self-rated health), which makes it
reasonable to assume [
] that PE may have been more
common among non-respondents. This was indeed the case
for those non-respondents who supplied this
information at baseline (1208 out of 1618), where 52 % reported
PE. Since the baseline association between exposure and
outcome was similar in both groups, we conclude that
selection bias in our final study sample may have
resulted in an underestimation of the association at
follow-up between PE and poor mental health in the
general population of young people.
Our measure of PE is based on extensive information
on relevant indicators but has not been validated
previously. It is highly probable that a multifaceted and
continuous measure of precarity such as the comprehensive
EPRES (The Employment Precariousness Scale) [
would have resulted in a more accurate description and
conceivably a discernible dose–response association
between exposure and outcome.
Moreover, the ‘forced’ dichotomisation of our
instrument may have led to some misclassified cases. For
instance, we chose to classify those with ‘contingent work
but no previous unemployment and no self-rated risk of
future unemployment’ as NP cases. However, and perhaps
particularly in this population of young people, it may be
argued that they are correctly classified. As stated in a
recent EU-commission report on new forms of employment,
aspects of the ‘flexible’ labour market may be utilised as a
positive choice by some individuals who have competitive
characteristics regarding the labour market, and who do
not risk unemployment despite the lack of a long-term
]. This phenomenon is discussed also in Guy
Standing’s seminal book ‘The precariat – the new
dangerous class’ [
In this context it should be noted that we chose to
categorise unemployment as one form of PE. This is in
contrast to the EPRES, where persons without employment
contracts were excluded [
], and also in contrast to the
‘peripheral employment score’ [
]. In the latter study,
exposure to peripheral employment was positively related to
psychological distress, but adjustment for unemployment
attenuated the association. Our model, in which
unemployment is seen as an extreme form of precarious
employment, could thus be debated. However, the fact that
only 7 out of 62 unemployed persons remained so after
ten years could indicate that there is a considerable
mobility in and out of this category, and excluding these very
persons, as was done in the EPRES study, would restrict
the study findings to those concerning persons with a less
Changes in precarious employment status may have
taken place between the measurement of exposure in
1999/2000 to 2005 and the measure of the outcome in
2010. In particular, the economic recession in 2008 may
have led to a number of persons being misclassified as
NP, which could have biased the results towards the null,
and thus led to an underestimation of the associations.
On the other hand, only 23 % of the participants were in
a PE situation at follow-up, versus 42 % in 1999/2000,
which partly may be explained by the fact that the
participants had had 10 years to establish themselves on the
The determination of a cut-off point for the GHQ-12
test is a trade-off between sensitivity and specificity.
Since the mean values in the population were below
1.85 (1.22 in 1999/2000 and 1.01 in 2010), we chose
the threshold of 1/2, as advocated by the creator of
the test [
There were several reasons for choosing financial
difficulties during childhood as a proxy for socioeconomic
status in the multivariate analyses. A large proportion
(N = 186) of the participants were students, i.e. thus with
uncertain present and future socioeconomic position.
Furthermore, since the focus of the study was on
precarious employment, we firstly considered it essential to
identify and categorise all unemployed persons, and
secondly, we expected that being unemployed would
influence and overrule any socioeconomic position to the
point where a classification of socioeconomic status
according to job description would become meaningless.
Lastly, socioeconomic position according to job
description showed no correlation with the outcome. On the
other hand, it is reasonable to assume that parental
social position in the form of background economic and
cultural resources must have influenced these young
persons’ present circumstances, both in terms of risk for
a precarious employment situation and for poor mental
In Sweden, precarious employment has increased since
the Nineties and accelerated after 2008. Cross-sectional
studies suggesting a relationship between precarious
employment and poor mental health are abundant, but the
causal direction could be debatable in this type of
studies. The present investigation is one of the first cohort
studies to show that a precarious employment situation
is an important risk factor for subsequent development
of mental health problems among previously mentally
healthy young adults. Moreover, due to the probable
influence of selection bias in our final study sample, the
current association found between precarious
employment and poor mental health should be regarded as an
underestimate. Therefore, further research is needed to
clarify the nature of the association and the underlying
mechanisms. The estimated population attributable
fraction of 18 % in the group of mentally healthy young
people in this study should serve as a wake-up call for
politicians and policy makers.
Additional file 1: Other variables and corresponding questions in the
questionnaire, i.e. a list with those variables that were used in the
analyses, and mentioned, but not described in detail, in the manuscript.
(DOCX 16 kb)
CI, confidence interval; EPRES, the employment precariousness scale;
GHQ, general health questionnaire; IRR, incidence rate ratio; NP, non-precarious
employment; PAF, population attributable fraction; PE, precarious employment
Preliminary results from this study were presented by Per-Olof Östergren at
the 7th European Public Health Conference in Glasgow, 19–22 November
2014. We wish to thank Dr. Elizabeth Cantor-Graae for helpful comments and
English language revision.
This work was funded by FORTE (Swedish Research Council for Health,
Working Life and Welfare; Grant Number 2013–1269) and the Medical
Faculty at Lund University. The funding institutions had no role in the design
of the study, data collection, analysis and interpretation of data, or in writing
of the manuscript.
Availability of data and materials
The data can be obtained by contacting the main author,
P-OÖ participated in the conception of the study, acquired the data,
participated in analysis and interpretation of the data, and the drafting of the
manuscript. CC participated in the conception of the study, the analysis and
interpretation of the data, and in the drafting of the manuscript. TB participated
in the conception of the study and the drafting of the manuscript. ME
participated in the conception of the study. ST participated in the conception
of the study and the drafting of the manuscript. MM participated in the analysis
and the interpretation of the data. All authors read and approved the final
The authors declare that they have no competing interests.
Consent for publication
Ethics approval and consent to participate
The study was approved by the Research Ethics Committee of Lund
University (1999–99; 2005–471, and 2010–392). The participants were
informed about the study in a cover letter to the postal questionnaire, which
also stated that participation in the study was voluntary.
Submit your next manuscript to BioMed Central
and we will help you at every step:
1. Puig-Barrachina V , Vanroelen C , Vives A , Martinez JM , Muntaner C , Levecque K , et al. Measuring employment precariousness in the European Working Conditions Survey: the social distribution in Europe . Work. 2014 ; 49 : 143 - 61 .
2. Eurofound: Precarious employment and working conditions in Europe. 2012 . https://www.eurofound.europa.eu/sv/publications/report-summary/ 1998 / other/precarious-employment-and-working-conditions-in-the-europeanunion-summary . Accessed 12 Apr 2016 .
3. Eurofound: Recent developments in temporary employment: Employment growth, wages, and transitions . Publications Office of the European Union, Luxembourg . 2015 . http://www.eurofound.europa.eu/publications/report/ 2015/labour-market/ recent-developments-in-temporary-employmentemployment-growth-wages-and-transitions . Accessed 12 Apr 2016 .
4. Benach J , Vives A , Amable M , Vanroelen C , Tarafa G , Muntaner C . Precarious employment: understanding an emerging social determinant of health . Annu Rev Public Health . 2014 ; 35 : 229 - 53 .
5. Karanikolos M , Mladovsky P , Cylus J , Thomson S , Basu S , Stuckler D , et al. Financial crisis, austerity, and health in Europe . Lancet . 2013 ; 381 : 1323 - 31 .
6. Fergusson DM , Horwood LJ , Woodward LJ . Unemployment and psychosocial adjustment in young adults: causation or selection? Soc Sci Med . 2001 ; 53 : 305 - 20 .
7. Cornwell K , Forbes C , Inder B , Meadows G . Mental illness and its effects on labour market outcomes . J Ment Health Policy Econ . 2009 ; 12 : 107 - 18 .
8. Hammarström A , Janlert U , Theorell T. Youth unemployment and ill health: results from a 2-year follow-up study . Soc Sci Med . 1988 ; 26 : 1025 - 33 .
9. Hammarström A , Gustafsson PE , Strandh M , Virtanen P , Janlert U . It's no surprise! Men are not hit more than women by the health consequences of unemployment in the Northern Swedish Cohort . Scand J Public Health . 2011 ; 39 : 187 - 93 .
10. Goldsmith A , Veum J , Darity WJ . The psychological impact of unemployment and joblessness . J Socio Econ . 1996 ; 25 : 333 - 58 .
11. Mossakowski KN . The influence of past unemployment duration on symptoms of depression among young women and men in the United States . Am J Public Health . 2009 ; 99 : 1826 - 32 .
12. Paul KI , Moser K. Unemployment impairs mental health: Meta-analyses . J Vocat Behav . 2009 ; 74 : 264 - 82 .
13. Daly M , Delaney L . The scarring effect of unemployment throughout adulthood on psychological distress at age 50: estimates controlling for early adulthood distress and childhood psychological factors . Soc Sci Med . 2013 ; 80 : 19 - 23 .
14. Strandh M , Winefield A , Nilsson K , Hammarström A . Unemployment and mental health scarring during the life course . Eur J Public Health . 2014 ; 24 : 440 - 5 .
15. Kim IH , Muntaner C , Vahid Shahidi F , Vives A , Vanroelen C , Benach J . Welfare states, flexible employment, and health: a critical review . Health Policy . 2012 ; 104 : 99 - 127 .
16. Strandh M , Nilsson K , Nordlund M , Hammarström A . Do open youth unemployment and youth programs leave the same mental health scars? - Evidence from a Swedish 27-year cohort study . BMC Public Health . 2015 ; 15 : 1151 .
17. Vives A , Amable M , Ferrer M , Moncada S , Llorens C , Muntaner C , Benavides FG , Benach J . Employment precariousness and poor mental health: evidence from Spain on a new social determinant of health . J Environ Public Health . 2013 ; 3 : 978656 .
18. Tsurugano S , Inoue M , Yano E. Precarious employment and health: analysis of the Comprehensive National Survey in Japan . Ind Health . 2012 ; 50 : 223 - 35 .
19. Statistics Sweden: Labour Force Surveys (LFS) . http://www.scb.se/en_/ Finding-statistics/Statistics-by -subject-area/Labour-market/Labour-forcesurveys/Labour-Force-Survey-LFS . Accessed 12 Apr 2016 .
20. Statistics Sweden: [Sysselsatta 20 -64 år efter ålder och anknytningsgrad till arbetsmarknaden ; in Swedish] http://www.scb.se/sv_/Hitta-statistik/ Temaomraden/Jamstalldhet/Fordjupningar/I-och-utanfor-arbetskraften/ Anknytningsgrad/Sysselsatta-2064 - ar -efter-alder-och-anknytningsgrad-tillarbetsmarknaden- 2015 /. Accessed 12 Apr 2016 .
21. Carlsson F , Merlo J , Lindstrom M , Östergren PO , Lithman T. Representativity of a postal public health questionnaire survey in Sweden, with special reference to ethnic differences in participation . Scand J Public Health . 2006 ; 34 : 132 - 9 .
22. Goldberg DP , Oldehinkel T , Ormel J . Why GHQ threshold varies from one place to another . Psychol Med . 1998 ; 28 : 915 - 21 .
23. Sirvio A , Ek E , Jokelainen J , Koiranen M , Jarvikoski T , Taanila A . Precariousness and discontinuous work history in association with health . Scand J Public Health . 2012 ; 40 : 360 - 7 .
24. Helgesson M , Johansson B , Nordqvist T , Lundberg I , Vingård E. Unemployment at a young age and later unemployment in native Swedish and immigrant young adults . Mod Econ . 2014 ; 5 : 24 - 31 .
25. Isacsson SO , Hanson BS , Janzon L , Lindell SE , Steen B . Methods to assess alcohol consumption in 68-year-old men: results from the population study 'Men born in 1914' Malmö . Sweden Br J Addict . 1987 ; 82 : 1235 - 44 .
26. Annerstedt M , Östergren PO , Björk J , Grahn P , Skärbäck E , Währborg P . Green qualities in the neighbourhood and mental health - results from a longitudinal cohort study in Southern Sweden . BMC Public Health . 2012 ; 12 : 337 .
27. Hanson BS , Östergren PO . Different social network and social support characteristics, nervous problems and insomnia: theoretical and methodological aspects on some results from the population study 'men born in 1914', Malmö . Sweden Soc Sci Med . 1987 ; 25 : 849 - 59 .
28. Zou G. A modified poisson regression approach to prospective studies with binary data . Am J Epidemiol . 2004 ; 159 : 702 - 6 .
29. Rockhill B , Newman B , Weinberg C . Use and misuse of population attributable fractions . Am J Public Health . 1998 ; 88 : 15 - 9 .
30. Rothman K. Modern epidemiology . Boston: Little, Brown & Co; 1986 .
31. Socialstyrelsen ( Swedish National Board of Health and Welfare). [Psykisk ohälsa bland unga ; in Swedish]. 2013 . http://www.socialstyrelsen.se/Lists/ Artikelkatalog/Attachments/19109/2013-5-43.pdf. Accessed 12 Apr 2016 .
32. Lager AC , Bremberg SG . Association between labour market trends and trends in young people's mental health in ten European countries 1983- 2005 . BMC Public Health. 2009 ; 9 : 325 .
33. Sverke M , Hellgren J , Naswall K. No security: a meta-analysis and review of job insecurity and its consequences . J Occup Health Psychol . 2002 ; 7 : 242 - 64 .
34. Vives A , Vanroelen C , Amable M , Ferrer M , Moncada S , Llorens C , Muntaner C , Benavides FG , Benach J . Employment precariousness in Spain: prevalence, social distribution, and population-attributable risk percent of poor mental health . Int J Health Serv . 2011 ; 41 : 625 - 46 .
35. Virtanen P , Janlert U , Hammarström A . Exposure to temporary employment and job insecurity: a longitudinal study of the health effects . Occup Environ Med . 2011 ; 68 : 570 - 4 .
36. Burgard SA , Brand JE , House JS . Perceived job insecurity and worker health in the United States . Soc Sci Med . 2009 ; 69 : 777 - 85 .
37. Wyn J , Andres L . Navigating complex lives: a longitudinal, comparative perspective on young people's trajectories . Early Interv Psychiatry . 2011 ; 5 Suppl 1 : 17 - 21 .
38. Underhill E , Quinlan M. How Precarious Employment Affects Health and Safety at Work: The Case of Temporary Agency Workers . Ind Relat . 2011 ; 66 : 397 - 421 .
39. Artazcoz L , Benach J , Borrell C , Cortes I. Social inequalities in the impact of flexible employment on different domains of psychosocial health . J Epidemiol Community Health . 2005 ; 59 : 761 - 7 .
40. Romppainen K , Jahi R , Saloniemi A , Virtanen P . Encounters with unemployment in occupational health care: Nurses' constructions of clients without work . Soc Sci Med . 2010 ; 70 : 605 - 8 .
41. Blomqvist S , Burström B , Backhans MC . Increasing health inequalities between women in and out of work-the impact of recession or policy change? A repeated cross-sectional study in Stockholm county, 2006 and 2010 . Int J Equity Health . 2014 ; 13 : 51 .
42. Fuller S , Stecy-Hildebrandt N . Career pathways for temporary workers: exploring heterogeneous mobility dynamics with sequence analysis . Soc Sci Res . 2015 ; 50 : 76 - 99 .
43. Scott HK . Reconceptualizing the nature and health consequences of workrelated insecurity for the new economy: the decline of workers' power in the flexibility regime . Int J Health Serv . 2004 ; 34 : 143 - 53 .
44. Standing G. The precariat : the new dangerous class . London: Bloomsbury Academic; 2011 .
45. Region Skåne , Sweden: [Hälsoförhållanden i Skåne. Folkhälsoenkät 2000 ; in Swedish]. 2000 . http://utveckling.skane.se/siteassets/publikationer_ dokument/fh_2000_rapport. pdf?id=5258. Accessed 12 Apr 2016 .
46. Vives A , Amable M , Ferrer M , Moncada S , Llorens C , Muntaner C , Benavides FG , Benach J . The Employment Precariousness Scale (EPRES): psychometric properties of a new tool for epidemiological studies among waged and salaried workers . Occup Environ Med . 2010 ; 67 : 548 - 55 .
47. Eurofound : New forms of employment . Publications Office of the European Union, Luxembourg . 2015 . http://www.eurofound.europa.eu/sv/ publications/report/2015/working-conditions -labour-market/new-forms-ofemployment . Accessed 12 Apr 2016 .
48. Waenerlund AK , Gustafsson PE , Virtanen P , Hammarström A . Is the coreperiphery labour market structure related to perceived health? Findings of the Northern Swedish Cohort . BMC Public Health . 2011 ; 11 : 956 .