Occupational status as a determinant of mental health inequities in French young people: is fairness needed? Results of a cross-sectional multicentre observational survey

In France, 447 Local Social Centres (LSC), created in 1982 (Order n°82–273, March 1982), provide social, health and education support to young people without qualification, to be integrated into the workplace. Since its creation, 1,400,000 of young people were taken in each year. There also are 112 preventive Health Examination Centres (HEC), created in 1945 (Article L321–3 Health insurance Code). They provide preventive medical consultations to beneficiaries of the national health insurance for salaried workers and their family. Patients do not pay and can consult without being referred. In 1992, a ministerial order designated who should have priority for these consultations, such as retirees, job seekers, young people just entering the job market, the homeless and people exposed to risk factors for health [18]. These people account for a third of all those annually cared for by the HECs.

A cross-sectional multicentre observational survey was performed in June and July 2010 in 115 French LSC and 74 HEC. LSC and HEC were free to participate.

The survey was based on an anonymous self-administrated questionnaire delivered by social workers or healthcare professionals working in the LSC or HEC. Professionals explained to young people the aim of the survey and were available to answer questions young people could have. The survey was carried out during 2 weeks per LSC or HEC.

The questionnaire was elaborated by a working group between 2007 and 2008. The working group was composed by the CNML, the French national health insurance for salaried workers (Caisse Nationale d’Assurance Maladie-travailleurs salariés- CNAMts), Health Directorate (Direction Générale de la Santé-DGS), Social Welfare Directorate (Direction Générale de l’Action Sociale- DGAS), the CETAF and representatives of LSC and HEC. A first version of 68 items has been elaborated based on questionnaires previously used in the European School Survey Project on Alcohol and Other Drugs (ESPAD), the French Health Barometer and the survey of the French Monitoring Centre for Drugs and Drug Addiction (Observatoire Français des drogues et des Toxicomanies-OFDT). The first version was tested in 24 LCS and 14 HEC on 1342 questionnaires. The second version was then elaborated and used in the present work.

People included were young people age from 16 to 25, coming from LSC or HEC. There were no exclusion criteria. The first 15 to 30 people who were taken in charge by social workers in LSC and health care professionals in HEC were asked to participate in the study. All participants gave their informed consent to be enrolled.

The questionnaire was composed by 54 items on socio-demographic status, deprivation, addictions, psychological disorders, health state, violent behaviour and discrimination and the need of medical, social or family support. The socio-demographic status included 21 items: age, sex, home location, own and parent’s current occupational status, level of education, social support, medical insurance, and serious childhood life events. Details on young people’ occupational status, are displayed in the Additional file 1. Deprivation was measured by the EPICES score, a validated individual score of individual social deprivation [19, 20]. Addictions included items about alcohol; consumption in the last 30 days and drunkenness in the last 12 months, current tobacco consumption and marijuana consumption in the last 30 days and feeling of need to smoke marijuana to feel health since the morning. Life events during childhood were measured by 4 items: have lived family events as break-up, disease, depression, alcoholism etc.; have had accident or serious illness; have been involved in an educative or judicial intervention; how was the childhood and adolescence. Psychological disorders were measured by two validated questionnaires, the Mental Health Index (MHI-5) and the Adolescent Depression Rating Scale (ADRS), and by items about suicide [21]. Health state was also measured by the scale of self-perceived health. Three kinds of violence behaviour, psychological, physical and sexual, as victim or perpetrator were assessed. Discrimination in different place (employment, accommodation, leisure, education and transports) was also integrated into the questionnaire. Then, people were questioned about needs for medical, social or family support they had had during the past 12 months (Additional file 2).

Nine health outcomes were measured: self-perceived health, mental health, addiction and to be victim of violence. Self-perceived health was a validated and relevant measurement of health predicting mortality [22, 23]. It was measured through a numeric scale from 0 for the poorest self-perceived health to 10 for the best self-perceived health. The threshold of 7 was used with poor self-perceived health under 7 [24, 25]. Mental health was assessed by the Mental Health Inventory- 5, a dimension of the well-known and validated SF-36, with psychological distress under 56 and a question about suicide attempt. Addiction to three substances was measured: current tobacco use, alcohol misuse (when young people reported being drunk at least three times in the last year) and marijuana use. Exposition to three kind of violence was explored, psychological, physical and sexual violence.

The main explanatory variable assessed, as in occupational status of French young people, was composed by 13 categories: permanent job, temporary job, interim, specific employment contract, to be at school or secondary school, to study at university or out of university, training school, to be in block release training school, trainee in education, in integrating into the work place, job seeking or unemployed. The other explanatory variables were level of education (low when <2nd school diploma and high when >2nd school diploma), major event in childhood (meaning yes if the person reported at least one kind of major event during his childhood), deprivation (through the calculation of EPICES score), discrimination (i.e.-if the person declared being victim of at least one kind of discrimination), financial social support (i.e.-if the person received at least one kind of financial support), social worker support (if the person met a social worker during the last 12 months), social network support (if the person met family, friends or other relations in the last 12 months), and health care consumption (meaning the person met a health care professional in the last 12 months).

Descriptive analysis was performed to assess the sample characteristics by using percentage for qualitative variables and means with standard deviation for quantitative variables. Multivariate analysis was performed by using stepwise forward logistic regression adjusted on age and gender to assess health outcomes through three steps. The first step integrated the occupational status as explanatory variables adjusted on age and gender. The second step added all the other explanatory variables, excepting individual social deprivation measured by the EPICES score. The third step added all the explanatory variables statistically significant at the second step and the EPICES score, to identify information this explanatory variable specifically brought into the model. Logistic regression models identified meaningful explanatory variables based on the calculation of the adjusted Odds Ratio (OR) and its 95% Confidence Interval [95%CI]. The R² of Nagelkerke describing the appropriateness of fit of the model was presented. A meaningful threshold of 5% was chosen for all the statistical analyses. Statistical analysis was performed on SPSS V 16.0 software.

Occupational status of French young people was a determinant of poorer self-perceived health, addictions and mental health. Young French people who were either not at work, at work without a permanent status or studying in training school or block release training school, were more exposed to poor mental health outcomes.

It is also interesting to point out that to add the measurement of individual social deprivation through the EPICES score into the multivariate analysis undermined occupational status and especially the non-working categories. Non-working status categories therefore seemed measuring individual social deprivation too. On the other hand, greater odds ratio was observed for the category “training school” when EPICES score was added, for consumption of tobacco and marijuana.

Concerning addiction, the protective association of school, secondary school and university, was another point of interest. Indeed, to study at university was not protective for alcohol consumption revealing the impact of binge drinking. Non-working occupational status was not associated with higher alcohol misuse but on the contrary to lower alcohol misuse, job seeking in particular, which means that alcohol misuse seemed to be more significant into social groups.

Non-working status categories were identified as a variable measuring individual social deprivation. This result was support by Benach et al. who explained in their work: “precarious employment is clearly related to absolute and relative social deprivation” [16].

Virtanen et al. revealed that unemployed included three categories; compensation-income, subsidy or low-income, expressed systematically poor self-rate health irrespective of the gender, which mirrors our findings [26]. This conclusion was also drawn for atypical employees defining by all form of work excepted permanent employees and fixed-term employees, as in our survey for young people being in interim or with a specific employment contract. Virtanen et al. deepened their results in a review, revealing an association between temporary employment and increased psychological morbidity [9]. László et al. showed an association between job insecurity and health in European countries, even though it was not significant in France [7]. Bambra et al. provided further evidence of such an association in the 2010 European Working Conditions Survey [8]. Employees having a temporary contract reported significant poor self-perceived health (OR = 1.24; 95%CI [1.13–1.37]). Two recent Italian studies also identified that people in unstable working conditions declared poorer self-perceived health than permanent workers [6, 27]. One of the studies revealed that this association existed among young people in particular [27]. Three previous French studies showed association between employment characteristics, full time vs. part time and permanent job vs. non-permanent job and adverse health outcomes like self-perceived health [28‐30]. When comparing genders, studies identified that both men and women expressed more damage on heath when they are in unstable working conditions. Nonetheless, while one study argues that men are more unhealthy than women, most of the studies showed higher vulnerability in women as revealed in our study [12, 27, 31].

To tackle health inequities among young people belonging to a non-working status category was complicated. Indeed, as previously demonstrated by the WHO into the micro-conceptual framework and Benach et al. through their conceptual model putting precarious employment as a central issue, several factors should be considered simultaneously by policies, which makes these policies harder to be effective [3, 16].

On the other hand, to base strategies on the reinforcement of the protective effect of education appeared to be relevant. It was especially the case for alcohol consumption for which the protective effect seemed to be lower at university. Prevention policies should also take into consideration that young people in training school did not have the same protective effect as secondary school and university, for tobacco and marijuana consumption, in particular.

Individual life trajectories, including major event during childhood seemed to be a significant point to consider.

The present work was a multicentre survey (115 LSC and 74 HEC) with a large sample (4282) of French young people. To our knowledge, this study was the sole which targeted young people in LSC. Occupational status was detailed in 13 categories but working conditions were not explored preventing us to assess deeply potential association between health outcomes and job characteristics such as number of hours worked per day, flexible employment, and if non-permanent job corresponds in personal choice or not. Individual employment trajectories, such as work linked to study field, type of work performed, were also not assessed accurately. Individual life trajectories were approached by the measurement of event during childhood. Furthermore, as young people who participated were volunteers, those in the worst life and health conditions could be not included.