Psychological distress (PD) can be defined as a state of poor mental health associated with symptoms from the anxiety-depressive spectrum. Significant levels of psychological distress and other mental health–related problems have been found in college students around the world 1 2 3 4 5 6 , with prevalences often much higher than those found in the general population 1 2 6 . Indeed, autonomy 7 8 , logistical independence 4 9 , financial difficulties 10 and aspects of academic life such as academic overload, pressure to succeed, competition and uncertainty about the future 1 3 10 11 12 might act as specific stressful factors in students and then lead to poor mental health. Although PD does not directly reflect a clinical diagnosis 13 14 and does not necessarily require therapeutic or clinical intervention, PD is nevertheless a key subject for primary preventive policies since it could lead to maladaptive behaviours, especially substance use and abuse or misuse of pharmaceutical drugs, and to psychiatric disorders 15 16 17 .

Prevalence rates of PD in students differ considerably from country to country and according to the tools used, ranging from 25.7% in a French sample of 1,723 first-year university students evaluated using the MH-5 scale 18 , to 67.4% in an Australian cohort of 6,479 students evaluated using the Kessler-10 screening scale 19 and 83% in an American cohort of 1,773 students 20 evaluated using the Dysphoria Domain of the Trauma Symptom Inventory 21 . In research targeting medical students, Dyrbye et al. 22 found that nearly all (82%) students were experiencing at least one form of distress (such as burnout, depression, stress, mental quality of life (QOL), physical QOL or fatigue), with 58% having at least 3 different forms of distress. In addition, some areas of study could be more at risk than others: Verger et al. asserted that medical studies were a specific risk factor for PD 12 .

Besides the expected higher prevalence of PD in women than men 1 2 3 23 , other factors identified as associated with PD in students include psychological factors such as a feeling of lack of mastery 12 , recent life events and a psychiatric history 12 , plus some situational factors such as being a full-time student, being an undergraduate 23 , financial stress 23 , academic factors 6 and lack of social support 24 25 26 .

The general applicability of results concerning students’ mental health is limited. Most studies are based on samples that are not representative of the general student population. Most of the samples studied are confined to a single university or a limited number of universities 3 4 5 , to the first or second years of study 3 4 , to one academic field - particularly medicine 1 3 4 - or are selected using non-probabilistic methods 1 3 4 .

Moreover, most studies have looked into psychological, socioeconomic or sociodemographic factors or life events. Little is known about the impact on student mental heath of more socio-environmental aspects such as academic factors (year of study, area of study) or social participation. To our knowledge, few studies have looked into the impact of academic factors 6 (e.g. year of study, area of study), loneliness 3 or social support, and none has looked into the effects of social participation.

However, a direct link between PD and variables assessing social participation has been shown in the general population. Berry et al. 27 defined as the “big 7” a list of seven factors measuring social participation and social support that reduced the levels of PD in an Australian cohort in New South Wales. The seven factors were: contact with immediate household, extended family, friends, and neighbours, participating in organised community activities, taking an active interest in current affairs, and religious observance. Murray et al. 28 , controlling for age, gender and income, found that social participation was also positively associated with positive affects and satisfaction with life, and negatively associated with negative affects. In a Swedish cohort of 51,414 subjects aged 16–84 years, Ahnquist et al. 29 likewise found a negative link between social participation and PD in men (evaluated using GHQ-12 30 ). A link between social participation and psychological distress has also been shown in specific contexts or populations. Indeed, it seems that social participation can act, directly or as a moderator, on the effects of negative life events on mental health in those with a low socioeconomic status and many stressors and in vulnerable people such as disadvantaged ethnic minorities 24 25 26 .

In the context of college, student life considered as a specific period of vulnerability due to the psychological and environmental stressors mentioned earlier, social participation could act as in other vulnerable populations 24 25 26 , working as a protective factor in the prevention of psychological distress and its negative consequences (increased risk of anxiety, depression, substance use and personality disorders, as well as school failure, work difficulties and adverse social consequences later in life 31 32 ).

Using a representative sample of French college students, this study aims to investigate the prevalence of and factors associated with PD, and to look more closely at the relative impact of academic and social variables and especially social participation. A focus on these variables should contribute to a better understanding of how PD can develop and provide inspiration for developing health promotion interventions that involve acting on students’ academic and social environment.

In the whole sample of participants aged 18–30 years, the PD rate was 13.8% in students. It was 12.6% in working people and 23.4% in unemployed people.

Table 1 shows the bivariate analyses (F statistic and P from the design-based Pearson’s chi-squared test) for social and educational factors, life events and sociodemographic data and levels of PD in students, together with adjusted regression analyses (odd ratios and 95% CI).

*p < .05, **p < .01, ***p < .001.

The results of this study revealed that, within a representative sample of French students, the prevalence of PD is not as high as expected (13.8%) when assessed with the MH-5 scale. Moreover, we did not find a higher prevalence of PD in young or first-year students. This contrasts with several studies which suggested that students were a vulnerable population with high rates of PD (e.g. 22 ), specifically in first-year students 12 . French students are a specific population in terms of socioeconomic characteristics. Public academic studies cost about € 500 a year, so that a high proportion of young adults of the same age can benefit from university curricula without taking out loans, as can be observed in many other national contexts. Indeed, the low proportion of students reporting debts or being in financial difficulty (the latter being significantly more at risk of psychological distress) could partly help reduce the overall prevalence of psychological distress found in our sample of French students.

While we found that, as per the literature, violence in the last year was highly related to psychological distress, lifetime sexual assault was not significantly associated with our dependent variable, although a clear trend was identifiable in the bivariate analyses (22.7% of participants who had been a victim of sexual violence in life reported PD, compared with 13.5% of those who had not been a victim of such assaults; F = 3.7; p = .05); this link became non-significant when controlled for other variables. Co-variation with other independent variables (gender and social isolation) might partially explain this finding. The limited number of variables used to define the process by which lifetime sexual violence affects the expression of distress in students does not allow us to draw any firm conclusions from this study.

Although expected, a link between contacts with friends and family and psychological distress was not observed. Several interpretations might explain this lack of connection. On the one hand we can assume that the measurement was not sufficiently specific (friends and family were combined, as were telephone contacts and physical ones). On the other hand, whereas the quantity of contacts was assessed, no account was taken of their quality. Finally, the geographical coverage of universities in France may be sufficiently dense for students to be able to live closer to their family and social network of friends. However, the limits of our measurement of contacts with friends and family mentioned above do not allow us to draw any firm conclusions from this result.

The higher prevalence notably found by Verger et al. 12 in a French student sample by comparison with our study could be explained mainly by a difference in the assessment of psychological distress and to a lesser extent by a difference in sample structure. Verger et al. used the first version of the MH-5 (cut point at 52 with a 6-point scale) whereas we used, as in the 2010 National Health Barometer secondary analyses, a second version of the MH-5 with a cut point at 56 (and a 5-point scale) 38 . This does not allow us to make any comparison between these two studies. Moreover, the sample used by Verger et al. consisted exclusively of first-year students from only public universities, excluding about 30% of young adults studying in private schools. However, conditions of education, social and professional opportunities are more uncertain in French public universities than in engineering or business schools, where more than 8 out of 10 students find a job within a year after completing their studies. On the basis of our results which show a very low proportion of psychological distress in engineering or business schools, the inclusion of these students in our survey might also have helped to reduce the overall prevalence of psychological distress in our sample.

Consistently with previous research (e.g. 12 23 44 ), we identified that, after controlling for all variables, gender (OR in female students = 2.6), financial stress (OR = 2.4) and nonsexual violence in the past year (OR = 2.8) were associated with PD. The effect of medicine as field of study, which was identified by Verger et al. as a risk factor for PD 12 , was not detected in our study, whereas participants studying law were less at risk of PD (OR = 0.3). This could be at least partly due to the parents’ socioeconomic status which is higher in French law students by comparison with other fields of study. Finally, as expected we found that low participation in social activities (the item was: “In the last week, have you had any group or leisure activities such as meetings, religious activities, club activities, going to the theatre, doing sports, going to parties…?”) was associated with a higher level of PD, starting with students who said they did not participate in any group or leisure activities (OR = 4.1) to students who said they participated a little (OR = 2.3), compared with those who said they participated a lot.

Participation in social activities is thus the biggest predictor of PD that we observed in our study. This finding adds to the results previously published on the general population 29 or on disadvantaged ethnic minorities 24 25 26 the fact that students are also and even particularly likely to benefit from a programme using social participation as a lever of prevention and health promotion. This offers opportunities for developing actions aimed at promoting group or leisure activities as a general preventive instrument in student populations.

Recently, a number of local initiatives have been set up in France to address social isolation in students. Social meetings initiatives, world cafés aimed at providing meetings with mental health practitioners and hence at reducing the reluctance to ask for counselling or help, local training for community leaders in visiting isolated students or door-to-door visiting by psychology students have been initiated (see http://www.fage.org, http://apsytude.e-monsite.com), but have not yet been assessed.

To conclude, our results suggest that social participation should be considered a key factor in further research into the development of preventive interventions for students. Work on social participation seems to be promising when targeted at students who have to spend a lot of time in common and collective spaces (campus, classrooms, halls, cafeterias) that are particularly convenient for the development of participative projects and actions. Combating loneliness and promoting collective initiatives may result in improved mental health within student communities. Students’ social environment (i.e. campus with multiple facilities) should make possible local initiatives that promote social participation.

These local initiatives should pay special attention to women and to students with financial difficulties, both of whom have higher rates of PD. It has been shown that financial stress in students often comes with poorer housing conditions and social isolation. Working on this specific subgroup together with social workers would allow poorer students and need-based scholarship holders to be targeted for specific housing and socially-based preventive services aimed at preventing the mental health consequences of poor living conditions, together with services aiming at promoting better living conditions. As an example, social groceries have been appearing recently in France (see http://www.gaelis.fr) to support students on low income by offering low-cost essential foods and developing social events for every student living on the campus.

As regards strengths and limitations, the main strength of our findings is that they are based on a random sample that is representative of the general French student population, including all years of study and a large variety of academic fields. Several limitations however deserve attention in the interpretation of the findings. First, the response rate obtained in the survey was 61%, which is satisfactory for such health surveys in France, but lower than the rates obtained in other epidemiological mental health surveys such as the NCS-R 45 or the NESARC 46 . Moreover, focusing on current college students, we can underline the fact that PD is a risk factor for dropping out of college 47 48 which may skew our estimates. The limited number of variables used in this study demand the inclusion, in future surveys of the student population, of more control variables (such as depression) and the expansion and refinement of variables assessing social support and social participation.

The authors declare that they have no conflicting interests.

Devised and designed the experiments: TS EDR LV FS FB. Performed the experiments: TS EDR RG JBR SL FB. Analysed the data: TS EDR LV RG JBR SL FS VK FB. Contributed reagents/materials/analysis tools: TS EDR LV RG JBR SL FS VK FB. Wrote the paper: TS EDR RG FB. All authors read and approved the final manuscript.