In our study sample, which gathered the highest number of civilian adults meeting criterion A of the DSM-5 definition for PTSD [1] of all French studies on the impact of terrorist attacks to date, 66% of all respondents, and 53% of those suffering from probable partial or full PTSD or depression, did not initiate MSPC after the attacks. The most frequent reason given for this was a perceived lack of need. This finding corroborates results from studies conducted with people with severe mental health problems both in everyday life [34] and in the aftermath of a terror attack [19]. The psychological reasons for not initiating a MSPC reflected those often cited in literature [19, 35, 36]. In terms of prevention, this finding highlights the need to help persons to balance these feelings with the perceived benefits of engaging in a MSPC. Another reason frequently mentioned for not initiating a MSPC was that “it was not the right time to talk about what happened”. This underlines the importance of providing victims with sustainable comprehensive information and treatment options, which can be adapted to changing perceived care needs over time [37]. Indeed, some victims put on the back burner the need for psychological treatment of trauma as long as the needs for information and reorganisation of daily life disrupted by exposure to the event are not met. Moreover, post-traumatic stress disorders may appear late in some people and it is a chronic condition that may last for several months or years. The financial aspect in the decision to initiate MSPC was mentioned quite frequently (17%) by those who had PTSD related to the attacks. This result justifies the initiative of the French Ministry of Health, which offered free-of-charge consultations with a specialist to persons registered as victims. Finally, difficulty of access to information and to care were also cited reasons for no MSPC initiation, but to a lesser extent.

Among those suffering from probable partial or full PTSD, or probable depressive disorder, our study also showed an association between MSPC initiation and modalities of exposure to the attacks. This is in line with the association between contact with a mental health service and exposure found in other studies on terrorist attacks [6, 21]. In our study, witnesses with full/partial probable PTSD and/or depression were less likely to initiate MSPC than persons who were threatened and had one or more of these disorders. This may be the result of perceived legitimacy to receive care on the part of the witnesses, as well as health care policy on the part of authorities, which focused on more direct victims.

We found that the initiation of a MSPC was associated with somatic problems. In a study by Holman et al. [38], the stress caused by exposure to the 9/11 attacks increased the frequency of somatic disorders by 18% and led to greater utilisation of mental health care. Stuber et al. highlighted that people with physical health problems after the 9/11 attacks were more likely to have sought mental health services [19]. After the attack on Utoya Island, the cumulative somatic problems score predicted the use of specialist mental health care [5]. Somatic problems may strengthen links with the health care network and therefore increase the likelihood of seeking mental health care specialist [5]. Somatic problems contribute also to diminished functioning and could maintain mental illness at long-term [5]. They deserve a special attention as an alert for psychological problems.

A history of treatment for depression or stress prior to the November 2015 attacks was associated with the initiation of a MSPC in our study. Stuber et al. found that six months after the 9/11 attacks [19], the use of mental health services by New Yorkers was clearly linked to prior contact with the mental health care system and with knowledge of how it works.

In the literature, woman gender is associated with a greater risk of PTSD after exposure to terrorist attacks [39]. The influence of gender on the use of mental health care is less clear. Although some post-attack studies showed no differences in gender by mental health care uptake [6, 19, 20, 40], others showed greater use by women [5, 21, 37, 41]. However, in multivariate models, this difference remained significant in only one study [41]. Among the relatives of the Utoya island victims, analysis of health care registries showed an increase in the use of specialist mental health care only in women, while visiting GPs increased in both genders [42]. In our multivariate models, gender was significant in terms of MSPC initiation, which suggests that among persons with probable partial or full PTSD or probable depressive disorders, an increased probability of initiation of a MSPC after terrorist attacks is associated with being a woman.

Having a GP was associated with accessing mental health care in a general population study after the 9/11 attacks [38]. GPs are considered the ‘expected’ care route to consultation with a specialist [7], although their role may vary according to the health system and the specific context [1]. In our study, visiting a GP was associated with MSPC initiation in univariate analysis but not in multivariate analysis; this suggests that it is not a strong determinant of MSPC initiation, given other characteristics of the population studied. On the contrary, visits to a specialist were strongly associated with the initiation of a MSPC. This may reflect a tendency for victims to use a specialist for their psychological problems rather than a general practitioner. This propensity may have been reinforced by a free-of-charge specialist consultation policy for victims of terrorist attacks in France. It may also reflect that specialists recognize more easily the indications for psychological care than GPs. Nevertheless, it could also be that the respondent declared their first session of a MSPC as a “visit to a specialist”.

The fact that psychological care initiation was significantly associated with OPS and contact with associations for victims after the multivariable adjustments in model 2 and 3, suggests that the field-based consultations and healthcare information provided by OPS as well as the collective support framework provided by associations for victims may facilitate access to psychological care in case of needs. For example, the members of the CUMPs who provided consultations in the support and information centers set up near the sites of the attacks, informed patients about the symptoms they might be experiencing as a result of the attacks and distributed a document containing the addresses of places where they could obtain psychological support.

In our models, a feeling of social isolation was not associated with MSPC initiation. In other studies, while the perceived quality of social support was inversely associated with the intensity of psychological symptoms experienced [23] and negatively correlated with feelings that care needs were unmet [23, 29], the influence of social support on seeking care was less clear. On the one hand, this support can reduce the intensity of symptoms and the need to seek care [43], in turn reducing the propensity of initiating a MSPC. On the other hand, it can facilitate the sharing of information and access to care [43]. The association that we found between being in a relationship and initiating a MSPC may be a marker for the role that a partner plays in supporting their mate to initiate mental health care.

Some post-attack studies have shown that people under 65 years of age were more likely to use mental health care [41], and that low level of education was associated with less use of care in times of need [21]. The lack of influence of age and education on MSPC initiation in our study may be due to the homogeneity of the Parisian population exposed to the attacks: a majority were middle-aged, well educated, and with an intermediate to high professional status. Moreover, access to healthcare in Paris is better than in other areas of France [1]. This could mitigate the influence of the above-mentioned social demographic variables on access to care.

Eight to 12 months after the November 2015 terror attacks in Paris and its suburbs, 34% of study’s participants who didn’t have already a MSPC at the time of the attacks, had initiated a MSPC. During the first month, the Nice and TENTS guidelines [17, 44,45,46] for post-disaster psychosocial care underscore the importance of first and foremost promoting social support, and reserving mental health care interventions to prevent PTSD for specific clinical indications (e.g., acute stress disorder) for which there is already evidence of intervention effectiveness [44]. In our study, 50% of MSPC initiation occurred within the first month. This result suggests that an important proportion of the people exposed felt a need for psychological care with a follow-up before PTSD or depression had time to appear.

Our results should be interpreted keeping in mind the specificities of the survey methodology used.

We did not have access to the Ministry of Justice’s list of victims (itself non-exhaustive) [47] in order to compare it with our list of participants, or to calculate the participation rate. The eligible persons were of course free of responding or no to the survey. Therefore, there may have been differences between terror-exposed individuals who participated in the study and those who did not, which may have introduced selection bias. Given that the study took place just a few months after the attacks, those suffering the most may have felt that it was too difficult to participate [36]. On the other hand, those suffering less may have felt less motivated, or that their participation would be less legitimate. Furthermore, our web-based survey excluded people who had no internet access and those most socially disadvantaged [36]. Nevertheless, the Parisian population affected by the attacks was essentially middle-aged, active, educated and with access to the Internet.

Our web-based questionnaire may have encouraged higher response rates to questions on sensitive topics compared to face-to-face interviews [48]. We used validated scales for screening the main mental disorders examined, with expected good sensitivity and specificity [30, 32]. But as it was web-based, there was no clinical examination, which is the reference diagnostic method. This may have resulted in inaccuracy in our diagnoses. The HAD_d scale measures depressive symptom scores but does not diagnose directly any specific depressive disorder that would warrant MSPC initiation. Although, studies on various populations have established associations between these scores and the likelihood of presenting depressive disorders, results of evaluations of the HAD_d for screening for possible depressive disorders among civilians exposed to terrorist attacks are particularly relevant. The study on the psychological impact of the January 2015 terrorist attacks in the Paris Region [49] which used simultaneously the HAD_d and face-to-face interview with the Mini International Neuropsychiatric Interview (Mini) underlined satisfactory performances of the HAD_d to screen for depression disorder among the exposed civilians with an optimal threshold score of 7.5 [50]. That being said, although these scales measure the intensity of symptoms, they may not sufficiently assess the presence of functional impairment, which is the main trigger for subsequent care [51]. For this reason our study considered the negative impact of PTSD symptoms on everyday life (criterion G “functional significance”), in addition to DSM-5 criteria A, B, C, D, E and F.

As data in our study were collected from participants’ self-reports, our results are subject to recall bias about care consumption [36]. However, this bias is more important concerning the number of visits rather than the declaration of using or not the different types of medico-psychological cares available [52].

The socio-demographic and health care network specificities of the Paris region may make it difficult to extrapolate the results to the whole of France in terms of MSPC initiation. The external validity of the findings to other countries may also depend on their health systems and plans for post-disaster psychosocial care.

Finally, the cross-sectional nature of the study prevented us from assessing whether the observed correlations might be causal.

Our method of collecting information gives only an instantaneous view of the person’s state of health at the time of responding and our analysis compared it to the initiation of a psychological care at any time after the exposure. Yet, while PTSD tends to be chronic in the absence of treatment, it is typical for individuals to experience fluctuating symptoms, including remission and reappearance of symptoms over time [53]. It is known that a proportion of people with PTSD will recover from the disorder within a few months, that another proportion will not recover without treatment, and that a small proportion may not even develop the disorder until several months after the exposure [53]. The time elapsed between exposure and participation in the study may have influenced the prevalence of the disorders studied. However, events considered as intentional (like terrorist attacks) are associated with greater persistence of PTSD symptoms than when the event is unintentional [54]. Studying the impact of the attack 8 to 11 months after the exposure is therefore not too late to measure its impact.

The psychological consequences of trauma exposure tend to be more severe and disabling when they result from interpersonal violence [53]. The high level of disability, mental and physical co-morbidity and loss of quality of life experienced by these people, and the significant personal and collective costs resulting from the social consequences of the disability (housing problems, absenteeism and unemployment) [55], argue for the development of collective strategies to identify people with disabling mental health disorders at an early stage, so that they can be offered access to appropriate care if they so wish. A great deal of research remains to be done to gain a better understanding of the factors that predict different trajectories, and to develop more effective strategies for screening and providing psychological care to people exposed to attacks at the right time [16]. Addressing all aspects of this field requires numerous studies. In the opinion of the psychologists and first-aid psychiatrists who helped us construct the questionnaire, it was unrealistic to expect all patients to identify and be able to name precisely the type of psychotherapy they received. However, it was reasonable to expect respondents to identify whether or not they had undergone a multiple-session psychological treatment, as opposed to nothing or a one-time consultation. Despite its limitations, this information contributes to building knowledge about to what extent psychological care was provided to those with probable mental health problems in the wake of the terrorist attacks.