Mental health among displaced and non-displaced populations in Valledupar, Colombia: do inequalities continue?

ABSTRACT During the long-lasting civil war in Colombia, thousands of people were displaced mainly from rural to urban areas, causing social disruption and prolonged poverty. This study aimed at analyzing the traumatic experience many years ago on the current psycho-emotional status of displaced families as well as the ongoing inequalities regarding displaced and non-displaced communities in one of the most affected areas by the armed conflict. An interview survey was conducted among 211 displaced families and 181 non-displaced families in 2 adjacent compounds in Valledupar, Colombia. The questionnaire used questions from the validated national survey and was revised and applied by staff members of the departmental secretary of health who conducted additional in-depth interviews. The study showed that the living conditions of the displaced community were precarious. The past traumatic events many years ago and the current difficult living conditions are associated with psychological problems being more frequent among the displaced people. The displaced people had experienced more violent acts and subsequently had a larger number of emotional symptoms (fright, headache, nervousness, depression, and sleeplessness). Other stress factors like economic problems, severe disease or death of family members and unemployment prevailed among displaced persons. The non-displaced lived in a more protected environment with less exposure to violence and stress, although belonging to a similarly low socio-economic stratum. It is recommended to take measures for a better protection of the displaced community, improve their access to the job market, offer different leisure activities and facilitate public transport.


The international context
Worldwide there are over 50.8 million people living in internal displacement for 2019 [1], most of them have been displaced due to conflict, violence and disasters. Forced displacement associated with conflicts seems to have a profound health and social impact on internally displaced persons (IDPs) [2,3], including acute and long-term effects on mental health [4]. Many forced IDPs lack legal status and psychosocial protection as they are not covered by international refugee laws. They are displaced within the nation's boundaries where they are commonly marginalized [5]. It has been shown that IDPs are more likely to experience poorer mental health indices than refugees or asylum seekers [6][7][8][9]. However, little attention has been paid to their mental health and well-being, and their needs are poorly supported by relief agencies [10].
Research suggests that IDPs are a high-risk group for mental health disorders [11,12]. Studies among IDPs report widely varying estimates of mental health disordersin particular, depression, post-traumatic stress disorder (PTSD) and other anxiety disorders linked to conflict [13][14][15][16][17]. Many displaced people have undergone traumatic episodes before and during their displacement and had to cope with long displacement periods with social and economic adversities [18], making them vulnerable to psychological disorders, particularly PTSD [19]. Porter & Haslam (2005) suggest that mental disorders may persist in IDPs for many years and may even increase over time [8]. A recent study of a population living in conflictaffected areas found a high prevalence of any mental health disorder (58.8%), a decade after the war ended [20]. This increased risk may also be related -among others-to post-migration socio-economic factors [8,21]. Risk factors include female gender, household roles, older age, food insecurity, continued trauma exposure, and poor economic situation [8,22,23]. However, limited evidence is available on how mental health disorders vary within a post-conflict setting following prolonged periods of displacement experiences. Thus, understanding the unique mental health status and possible risk factors for specific populations with long-term displacement is crucial for providing adequate care.

The Colombian context
Civil war is a sad reality of many countries in our contemporary world and Colombia is a prominent example among them. Colombia with a total population of roughly 50 million inhabitants has suffered an armed conflict among different interest groups for several decades, which partially ended in 2016 with the peace treaty of the government with FARC (Fuerzas Armadas Revolucionarias de Colombia, Colombian Revolutionary Armed Forces), one of the major guerrilla groups. More than five million Colombian citizens have suffered violence and internal displacement during this time affecting their mental health and well-being and creating high levels of headache, sadness, nervousness, suicidal attempts and depression [24]. The situational analysis of the health situation in Colombia has shown that the internal displacement has generated vulnerability and affected the physical and even more the mental health of the families involved [25]. According to the National Programme for Attending Psychosocial and Integral Health among Victims (PAPSIVI) the armed conflict has generated a deterioration of living conditions among these groups creating distrust, insecurity and fear [26]. Research also demonstrated that Colombian IDPs present high prevalence of depression, anxiety, PTSD, and substance use disorders [27][28][29], especially women [27,30]. Although there is an increasing research interest in the health of conflict-affected populations in Colombia, there is still a scarcity of information on mental health issues in IDPs many years after the forced displacement.
According to the Central Register of Victims, the State of Cesar is one of the most affected States (departamentos) by the armed conflict in the country [31]. According to this source, the State has more than 300,000 victims, 85% of them being displaced. In Valledupar, the capital of the State, 132,294 victims were registered in July 2018 mainly living in neighborhoods with high insecurity levels and crime. In spite of these problems, no detailed information is available that could help plan a better future for these communities and the information from the National Mental Health Survey is too general to provide concrete guidance for what needs to be done. Health and social workers operating in these neighborhoods are aware of problems but do not have concrete information about the real dimension of human suffering, which can directly lead to action.
In order to understand the magnitude (prevalence) of the psycho-social challenges among displaced families or those subject to other forms of violence by the armed conflict, the here presented study was set up jointly by the State Health Services in Cesar and universities in Colombia (Instituto de Salud Pública de la Pontificia Universidad Javeriana) and Germany (Freiburg University, Center for Medicine and Society) supported by the German Academic Exchange Service (DAAD). The objectives were (i) to quantify the prevalence of mental health issues (mainly emotional distress) among displaced and victimized persons in comparison with non-displaced persons of a similar socio-economic status living in the same environment and ii) to analyze if the cause of the expected increased distress among the displaced is more due to the traumatic experience many years ago or to the current precarious living conditions of the displaced. The overall aim was to provide recommendations for designing programmes for improving the situation.

Study communities
Both displaced and non-displaced study communities were located next to each other in a suburb of Valledupar. The displaced families were all registered as victims of the armed conflict with RUV.
Displaced community: The neighborhood of Hernando Marín was constructed in the outskirts of Valledupar by the National Government as part of a social programme to provide shelter to poor displaced families (belonging to stratum 1 to 2 in the Colombian socio-economic classification system where stratum 5 represents the highest level). Six blocks of five-storey buildings with a total of 1900 departments were built, each apartment having a surface area of 45.6 m2 (see Figure 1). The families can live there for free and have only to pay for water, electricity, gas, and administration.
Non-displaced community: The neighborhood Leandro Diaz was part of a governmental housing programme for employees of the lower socioeconomic level (stratum 2). It has three blocks, each with six five-storey buildings, in total 1300 apartments (the security staff mentioned 1400 apartments). A fence secures the buildings and a security guard controls the entrance (see Figure 1). The families must pay a subsidized rent plus expenses for basic services (water, electricity, gas, administration).

Methodological approach
A mixed method approach was applied with quantitative elements (questionnaire survey) and qualitative elements (in-depth interviews). This corresponds to a 'concurrent embedded approach' where quantitative and qualitative data were collected simultaneously, but the quantitative study (survey) was the primary approach and the qualitative study provided supplementary information [32].

Questionnaire survey
The interview survey was conducted among 211 displaced families and 181 non-displaced families living in two adjacent compounds. The questionnaire was only applied to adults, mostly women (81.1% of the 392 respondents were women and 18.9% were men), who provided information related to the purpose of this study. No child or adolescent below the age of 18 was interviewed in this study. Special care was devoted in the interviewer training to the interview setting. Our 15 interviewers (14 of them women) were familiar with social and mental health challenges of poor families. The way of showing compassion and understanding was discussed in the training and practices. The interviews should be conducted in a quiet place for which reason afternoon hours were preferred for the interviews when the male heads of household were usually absent. In some cases, interviewers had to come back the following day when the levels of noise were lower. The questionnaire was an abbreviated form of the national survey (ENSM 2015). The first draft questionnaire was discussed in the team, then with 15 staff members of the local health services and finally with the community leaders. Based on these discussions and recommendations, some sensitive questions on alcohol abuse were deleted, but the questions on intra-familiar violence were maintained because of public discussions on these issues. The questions included were either directed to the entire family or in some cases only to the respondents (mainly housewives). A selection of the most relevant questions for our study objectives were taken from the national survey on mental health [24], except for the introductory questions for the characterization of the study populations. The ENSM included nationally and internationally validated instruments such as la Gran Encuesta Integrada de Hogares 2013-the Great Integrated Household Survey 2013 [33]; the Self Report Questionnaire (SRQ) [34,35] and the Family APGAR [36] of which only the SRQ was relevant for our study. The Spanish version of the SRQ questionnaire has been in use for quite a while in Latin America and has been validated in previous Colombian studies [37,38]. Respondents were asked if they currently suffered from the following 22 different physical and emotional distress: frequent headaches, loss of appetite, sleeplessness, fright, trembling of hands, tension/ nervousness, mal-digestion, lack of clear thinking, sadness, does not enjoy daily activities, difficulty to take decisions, difficulty to work, incapable of doing something useful, loss of interest, feels to be useless, suicidal thoughts, feels tired all the time, has unpleasant feelings in the stomach, gets easily tired, feels that somebody has tried to wound her, and feels that she is more important than other people think.
The other selected groups of questions were directly related to our specific study topic such as personal and family experience with violence, mental health symptoms and disorders and mental health services. Our own validity testing before starting the survey was to apply the interview to 10 families (women) and then to check in a long conversation if responses were the same as in the formal interview. Discrepancies were minor so that the original questions from the ENSM were used (see Fischer et al. 2019) [35]. 15 staff members of the health and social services of the State Secretary of Health were trained by our research team in a 3-days workshop, which included theoretical and practical components as well as a field visit.

Sample size calculation and sampling procedure
The sample size was calculated using data from the national mental health survey on the difference in proportions of major mental health issues in displaced families compared to the non-displaced. With a confidence level of 95%, a power of 80% and a cumulative prevalence of major emotional problems among the displaced versus non-displaced populations of 55% versus 30%, respectively (taken from the national survey [24]), a minimum sample size of 170 families per study arm was calculated; the sample size was increased to 230 families among the displaced and to 193 among the non-displaced families, but finally 211 displaced and 181 non-displaced families were interviewed due to non-response (see next section). Using systematic sampling of apartments (households where almost exclusively nuclear families) with a random start (using a die), the sampling interval in the displaced families was eight (every eight family to be interviewed; calculated as 1900 families/230 interviews) and in non-displaced it was seven (every seventh family to be interviewed; calculated as 1400 families/193 interviews).
We did not opt for a two-stage sampling procedure (where within-family sampling is the second stage) but for one-stage sampling where apartments were the sampling unit; our respondents were mostly women (mothers or caregivers), which means that the sample is representative for the reference population only regarding household questions while the responses to individual questions reflected more the experience and perceptions of women.

Survey procedures and non-response
After pre-testing the questionnaire, the leaders of both communities were informed and consulted in separate meetings and their consent was obtained as well as their willingness to inform the community. Then, the 15 interviewers (the same staff members as mentioned above) used mainly afternoon hours and sometimes their free weekends to conduct the interviews in the assigned buildings. In these hours mainly housewives were available, rarely with the participation of their husbands. Women are supposed to know best about health issues of their families [39,40] but would have difficulties to talk about sensitive issues like alcohol abuse or sexual violence [41]. This is in line with other studies, which observed that men are often less motivated to participate in surveys on health issues [42,43].
The team leader conducted several control interviews and checked the sampling interval. The nonresponse rate (rejection to participate or absence) was 6% among the displaced sample and 8% in the non-displaced sample, which is low compared to household surveys in high-income countries (nonresponse rate up to 45%) [39] and LMICs (up to 20%) [41]. The typical refusal was that family members opened the door, were informed about the purpose of the visit (NB they had already been collectively informed by the community leaders) but quickly closed the door saying that they were not willing to respond. It would be unethical to insist and try to get their motives.

Qualitative study with in-depth interviews
After the formal interviews, additional in-depth interviews were conducted with 25 interview partners in each group; the interviewers followed a list of guiding questions ('Are you happy living in your current environment? What are the challenges? Do you feel comfortable with your neighbors? Do you often think of the past events? What could be improved in your living conditions?'). The responses were recorded in the field diary. Later, they were ordered in thematic topics and analyzed according to the topic (structural analysis). In the analysis, common statements and experiences were extracted. When reading the interviewer field diaries (raw data) and the summaries of common concerns and complaints (aggregated information) and also later on when listening to the interviewers' passionate discussions regarding the translation of their research findings into policy and practice, we as a research team were confident that we had selected the right interviewers who had been able to get reliable information from our study populations. The qualitative study had put the quantitative results in a broader context and helped to understand and interpret the findings.

Data management and analysis
Interview answers were entered into the computer and analyzed using Excel and SPSS software. The results were tabulated, confidence intervals (95% CI) were calculated, and different bar charts were developed. A comparative analysis examining the sample characteristics and mental health issues of the two study communities was conducted using chi-square test and double-sided t-test for statistical significance testing, which was seen to be appropriate for this kind of study.

The study populations
In the displaced community, 211 families (845 individuals) were interviewed and in the non-displaced population 181 families (732 individuals), showing four persons per family in both groups. The direct respondents to the interview were in both communities mostly women (81.1%) at the age of 30-50 years, almost all of them with school education. However, in the displaced community, 9.6% (81) of families did not comply with the definition of displacement (i.e. they were not registered with RUV) and in the non-displaced community, 12.4% (91) of families were in fact displaced according to the registration with RUV (Central Colombian Register for Victims).
The demographic characteristics of households in both communities were similar (Table 1): slight excess of females (female to male ratio 1.04 in the displaced community and 1.1 in the non-displaced) and slight excess of young adults among the displaced compared to the non-displaced (38.7% of all ages versus 31.6%, p = 0.018). The proportion of children <15 years old was higher among the displaced compared to the nondisplaced. With respect to ethnicity, in both groups, the mestizo-population (ethnic mixture) was predominant (53.3% of all families) while the afro-descendants and indigenous people had suffered more frequently from displacement: 28.0% afro-descendants among the displaced families and 20.4% among the nondisplaced, 5.7% of the indigenous people among the displaced and 2.2% among non-displaced (p = 0.052). The educational level of persons above 12 years of age including qualified technicians was clearly higher in the non-displaced group (35.6% of all educational levels) compared to the displaced group (only 15.7% of all; p = 0.000).

History of displacement
The peak of the violence in Colombia occurred from 2002 to 2004, which means about 18 years before our survey when our mostly female respondents were about 20 years old. Therefore, 55.5% of the displaced families reported that the event of displacement has happened between 16 and 20 years ago. Only a smaller proportion (19.0%) had been displaced 11 to 15 years ago and only 9.5% during the preceding 10 years. However, the in-depth interviews showed that the traumatic experience of displacement and victimization still had a profound emotional effect on the displaced in present times (tears, profound breathing, lifting the voice) and that the fear of going back and receiving bad news from their place of origin kept the memory of those days of life.

People's perception of the current situation (qualitative study)
In the displaced community, the government subsidizes the apartments and the inhabitants did not have to pay a rent. However, they expressed a feeling of insecurity -mentioned by most respondents -due to the open access to the premises; they mentioned the noise that was always present, either loud music or intra-familiar fights or aggression between families; all of them mentioned the existence of gangs of young adults as well as of drug users. The interviewers could witness this; particularly when they arrived in the late afternoon, they saw groups of youngsters and perceived the smell of marijuana. The elected community leaders did not seem to have much authority, as they did not want to accompany the interviewers (mostly females) or introduce them to the families. Taxi drivers could hardly be convinced to enter the compound and bring the interviewers to the buildings because of the bad access road and the feeling of insecurity.
In the adjacent buildings of the non-displaced community, the situation was different. The flats were partially subsidized by the government, but the residents had to pay a rent. The residential area was closed for outsiders by a fence, and special permission was required to enter. There was an administrator of the compound who looked over the security and order and controlled entering unknown people. The respondents mentioned the risk of leaving the compound and being attacked on the street. Although the families belonged to the second lowest socio-economic stratum (similar to the displaced families), they felt superior to the adjacent displaced community.

Experience with violence
When asked in the formal interview survey for the exposure of family members to violence, significantly more displaced families had experienced such events (n = 64.0%) compared to non-displaced families (n = 25.4%; p = 0.000) (Figure 2, Table 2). The responses show that in all categories (intended murder, theft, threat, murder of a family member, kidnapping and sexual violence), the displaced families had to suffer much more frequently than the non-displaced ones. Also, the non-displaced had suffered, although to a less extent, particularly of theft and to less extent of intended murder. Once a violent event has happened, both groups suffered equally of emotional disturbance.
In addition, the intra-familiar violence was clearly more frequent in the displaced community (victims of insults, physical aggression, and armed attacks) compared to the non-displaced community (p = 0.001, 0.035, and 0.001, respectively; Figure 3, Table 3). Other causes of human suffering were important life or vital events such as economic problems (affecting 72.5% of the displaced families compared to 43.1% of the non-displaced; p = 0.000) showing that although both groups belong to the lowest socio-economic class in Colombia (stratum 1 and 2), there are important differences regarding income and employment. Also, severe disease or death of a family member as well as unemployment and severe health problems were more frequent among the displaced families compared to the non-displaced ones ( Figure 4, Table 4).
Current psycho-emotional status of displaced and non-displaced respondents (Mothers/caregivers; Self Reporting Questionnaire, SRQ) 46.4% of the displaced interviewees and only 20.4% of the non-displaced suffered seven or more symptoms (p = 0.000). Only 6.6% of the displaced had no symptoms at all, while among the non-displaced interviewees this was 20.4% (p = 0.000, Figure 5, Table 5). The mean number of symptoms or conditions was 8.2 (SD 5.7) per respondent among displaced; this was significantly higher than among non-displaced respondents (4.7 symptoms, SD 4.1; p = 0.000).
More than half of the 211 displaced respondents felt symptoms like 'get easily frightened' (62.1%) and 'frequent headache' (59.7%), 'felt nervous or tense' (54.0%), all of them being much rarer than among the 181 non-displaced respondents (35.9%, 43.6% and 30.4% respectively; p = 0.000, 0.001 and 0.000 respectively). Other frequent emotion related 64 Figure 2. Victims of violence in displaced and non-displaced families. There were three additional questions on psychotic symptoms ('hears strange voices', 'has strange interferences in her thoughts', and 'has epileptic seizures'), but these were rare in both groups although marginally more frequent among the displaced respondents.

Current psycho-emotional status of displaced respondents living in the non-displaced community
This analysis was done to better understand if the emotional problems among displaced respondents are more due to the displacement or due to the current living conditions, 40.9% of the 91 respondents belonging to the 'displaced living among the non-displaced' reported more than seven symptoms in the SRQ and 16 Figure 3. Intra-family violence among displaced and non-displaced families.  showed a pattern of emotional distress similar to the displaced group rather than to the non-displaced suggesting that the traumatic experience of the armed conflict with displacement and victimization long ago still impacts on the emotional status of displaced individuals.

Use of mental health services
Mental health services in local hospitals and ambulatory care were equally available for the displaced and nondisplaced persons. However, they were rarely used during the preceding 12 months, both by displaced families (16.6%) and even less by non-displaced families (6.1%; p = 0.000). Hospitalizations for mental health problems 'once in their lives' were rare exceptions (4.7% among the displaced and 0.6% among the non-displaced; p = 0.012) keeping in mind that the relative frequency of mental health issues was higher among the displaced compared to the non-displaced populations. The main reasons for not using mental health services during the last 12 months were 'was not necessary' (39.1% among the displaced and 71.6% among the non-displaced; p = 0.000). Other major reasons for not attending mental health services particularly among the displaced families were service far away, no time, no money, unawareness of my rights, bureaucratic hurdles, and distrust. Most patients attending mental health services went there 3 times or less (12.3% among displaced persons and 5.0% among non-displaced; p = 0.001). Patients in both groups went for treatment mainly to hospitals or clinics (56.3% both groups), but some of them to pharmacies (mainly the non-displaced) or to friends and family members including alternative medicine. Additional consultations for mental health problems were mainly sought by non-displaced patients (28.9%) and less by the displaced (17.8%; p = 0.157). The treatment costs for displaced patients were mainly borne by the subsidized insurance companies for the displaced patients (48.9%), while the treatment of non-displaced patients was mainly financed by the prepaid insurance companies (21.1%) and often through out-of-pocket payment.

Discussion
Mental health, as defined by WHO, refers to the capacity of thought, emotion and behavior that enables every individual to realize their own potential in relation to  Figure 5. Number of emotional symptoms reported in the SRQ by formerly displaced and non-displaced respondents. their development stage, to cope with the normal stresses of life, to study or work productively and fruitfully, and to make a contribution to their community' [44]. In this sense, our study has shown the precarious mental health conditions of displaced population groups in comparison with non-displaced groups of a similar socio-economic background. Both their history of violent displacement and current poor living conditions are determinants for inequalities between the two groups. In detail: The civil war in Colombia -as a prominent example of similar conflicts in our current world -with its peak of violence in the early years of our century is officially over with the peace treaty signed by the government and the most important rebel group FARC in 2016. At the climax of the conflict around 2002, many particularly rural families had to leave their homes and farms due to the threat by major parties of the armed conflict, mainly paramilitary groups in Cesar State. They generally fled to the capital of their State (department) where they finally found a place to stay, often in a multi-family building provided by the government.
The national survey on mental health in Colombia [24] had shown an overall picture of the mental health status of the nation, distinguishing in some parts of the report between the general and the displaced population. Some but in general not very strong differences have been elucidated between the two groups but naturally at a 'macro' level and without adjusting for socio-economic differences. Our study was able to give more detailed insights into the mental health status of two populations with a similar socio-economic status living close to each other but with the distinctive characteristic to belong to the displaced community as against the nondisplaced community. Although the event of displacement has happened almost two decades ago, there were still important differences regarding mental health status and associated factors. These will be summarized in the following:

Context and circumstances
Displaced and non-displaced belonged to the lower socio-economic population strata so that in our comparative analysis, 'socio-economic status' may not be an important confounding variable (although persons older than 12 years had in the non-displaced population 2.3 times more frequently a secondary education compared with the displaced one).
The main traumatizing event (displacement or another violent act) occurred 16-20 years before the study but had, according to the in-depth interviews (qualitative study) among the displaced a long-lasting effect on the mental health status of victims (quantitative study) fueled by the impossibility to return to their rural homes because of the continuing insecurity. This is in line with the literature on post-traumatic stress disorder (PTSD) in displaced people by war [45,46]. The proportionally most affected groups were Afro-descendants and indigenous people (1.4 times and 3 times higher proportion Afro-Descendant and Indigenous descendants in the displaced compared to the non-displaced) highlighting the socio-cultural disparity in the society. Researchers have concluded that ethnicity may influence the prevalence of mental disorders [47], resulting in more vulnerable sub-groups at elevated risk for mental illnesses.

Mental health status of mothers
As the respondent related questions were mainly answered by women (mothers), they reflect the perceptions and feelings of this group, while the familyrelated questions provide information about all family members. Significant differences between displaced and non-displaced mainly female respondents were: Currently, more frequent physical and emotionrelated symptoms among displaced persons: higher number of symptoms, particularly frequent being fear, headache, and nervousness but also sleeplessness, problems to think with clarity and other depression-like symptoms The use of mental health services (mainly hospitals and clinics) was slightly higher by displaced than by non-displaced respondents (16.6% versus 6.1%; p = 0.01) but did not reflect the much higher disease burden in the displaced community

Determinants
Experience of violence and other stress factors (according to our survey): • More family members in the displaced community had experienced some form of violence (64.0%) compared to the non-displaced community (25.4%) • Different forms of violence were more frequently reported by displaced families such as intended murder, threat, murder of a family member and kidnapping (NB. Questions on alcohol or substance abuse could not be asked, as they were perceived to be insensitive according to the pilot tests). • Threatening life events as stress factors were more frequent in the displaced community: economic problems, severe disease or death of family members, unemployment, or severe diseases Living conditions of the displaced community showed a number of stress factors (according to our qualitative study): • Open access to the premises allowing anybody to enter • Weak leaders lacking authority • Youth gangs often involved in drug trade or drug use • Frequent unemployment • Noisy neighbors (music, aggressions) • Poor access road and poor connection with public transport In contrast, the non-displaced community could live more in peace (qualitative study): • Closed and protected premises • Strong authority controlling the access and order Better maintenance of the buildings • Higher employment rates of both men and women Challenges for the non-displaced community were (qualitative study) as follows: • Poor access road and poor connection with public transport • Mistrust and social distancing from the neighboring displaced community • Fear of being attacked on the access road In view of these determinants, it is not surprising that the displaced community living under precarious circumstances continues to be exposed to elevated mental health risks [18]. Additional factors such as the presence of gestational diabetes mellitus and other chronic diseases may worsen the mental health of those displaced populations as well as the lack of physical activity [48,49]. In contrast, the neighboring non-displaced population can enjoy some basic commodities such as physical protection by a fence, a strong authority that is able to keep order, and better maintenance of the building. However, the previous trauma of displacement seems also to persist among displaced mothers/caregivers living in the more protected environment of the non-displaced neighborhood as 40.9% of the individuals interviewed suffered from more than seven emotional symptoms, which is similar to the displaced respondents living in an unsafe environment (46.4% with more than seven symptoms; see quantitative study). This long-term persistence of mental health issues after displacement has also been observed in Sri Lanka and Germany [45,46]. The low use of mental health services in our study compared with the national average (33%) [50] confirms that vulnerable groups do not use them as much as they could and that cultural barriers are important. The main reason for not using mental health services was 'not necessary' (39.1% among the displaced and 71.6% among the nondisplaced); this is consistent with the national survey where 47% to 56% of the people with mental health issues considered that they did not need any consultation [24]. Attitudinal barriers toward resorting to mental health services have been frequently associated with the stigma generated by mental illness [51,52].

Recommendations
In general terms, displaced families need attention and help for many years, not only the provision of shelter separated from the civil society and being left on their own but also local solutions for major challenges have to be developed. In our case, this was done by the State public health and social services (where our interviewers worked) based on the findings of the study. To these practical recommendations belong: Increase security -Protective fence around the buildings of the displaced community and security guards (watchmen) -and promote a police station in the area.
Reduce violence by offering leisure activities for the adolescents and a social area for women and provide opportunities for sport, artwork, dancing, youth clubs, and women's clubs. This is also recommended by authors who highlight the importance of providing leisure activities to displaced people [53,54]. Indeed, it has been shown that recreation can help reduce youth violence and gang activities by offering elements that might be otherwise missing in the lives of vulnerable young adults, including the learning of new skills [55][56][57].
Promote advanced education and employment: Provide incentives for entering higher education; encourage reading competition for children or storytelling for adults.
Offer training on productive work (urban gardening, preparation of food, bakery, etc.) to complement the family income.
Strengthen primary mental care in the community through psychologists and social workers in order to overcome the traumatic experiences of the past and face the problems of the current conditions (also recommended by Anaya Mercado and Romero Perez (2019) [58].
These recommendations are in line with the mental health promotion suggested by WHO [59] and other recent reviews [60,61] which have highlighted the beneficial effect of social interventions (sports, cultural, and leisure activities) and employment and education opportunities to reduce mental disorders in targeted people.
Other recommendations include: Special effort to involve those committing violent acts in community activities.
Repeat the survey after 3 years to see if changes have taken place.

Limitations of the study
The one-stage sampling procedure of households (no within-family sampling) enabled the analysis of general living conditions of displaced and non-displaced families but had a focus on emotional reactions of the respondents (mostly mothers) thus providing less information on the mental health status of husbands and children.
Mental health questions are sensitive and special care must be applied regarding the wording of the questions, interviewer behavior and interview setting in order to get reliable answers. The 2015 National Mental Health Survey from where the questions were taken has been carefully validated and previously adjusted to current social and economic conditions in the country [62]. In order to get information on the most prevalent psychological problems among our study groups, the Self Report Questionnaire by Beusenberg and Orley (1994) [34] as an internationally renowned instrument was applied to families. Previous studies have confirmed the potential of this instrument for the screening of psychological distress in poor settings [63][64][65].
The majority of the respondents were happy to answer the interview questions, but there were a small number of families who did not want to participate or were absent during the interview. Some response bias was probably due to mistrust and fear. Other contextual factors before and during the displacement processincluding disruption of social networks -could not be evaluated in depth. However, in spite of these limitations, we are confident that our study results reflect the reality of previously traumatized women (mothers/caregivers) living in quite precarious circumstances compared to others who did not have such a traumatic experience and can live in a more peaceful environment.
Limitations regarding our second objective (to better understand if the distress of displaced persons is mainly due to their traumatic experiences in the past or to their current precarious living conditions) were more serious as the multiple factors influencing over the years the displaced families could not be sorted out retrospectively. However, the fact that those displaced respondents now living in the safer environment of the non-displaced community suffered from a similar distress as the displaced living in the insecure environment of the displaced community is a strong indicator reconfirming what has been observed in the in-depth interviews and reported in the literature (see above) that displaced individuals and families continue to suffer from their traumatic past for many years.

Conclusions
Sixteen to eighteen years after the peak of violence and displacement in Colombia, the victims are still suffering mental hardships (compared to non-victimized persons) which adds to their current precarious living conditions and requires new policies and programs to alleviate human suffering of vulnerable population groups.

Acknowledgments
The German Academic Exchange Service (DAAD) in the context of the PAGEL programme facilitated this work. We are grateful to our co-interviewers for their patience and excellent work under difficult circumstances, particularly to Electa de Jesús Arzuaga Villero, Cecilia Flórez Sanchez, Yusleydis Gutiérrez Ospino, Sunilda María López Blanco, Yarima Vanessa Martínez Castilla, Olga Beatriz Ochoa Díaz, Martha Leonor Romero Orozco, Yerlis Saenz Arias, Nora Inmaculada Santos and Zuleima María Billa Jimenez. We are particularly obliged to all our respondents who took the time and had the courage to answer our questions.

Ethics approval and consent to participate
The interviews were initiated by a standardized description of the aims of the study, the institutions involved, the anonymity (no names were recorded), the right to stop at any time or to skip questions without any harm. No child or adolescent below the age of 18 was interviewed in this study. The consent was obtained verbally as under the given circumstances, a written consent was not possible and not recommended by the community leaders. The procedure was checked by the supervisor, and the consent to participate statement and ethical considerations were approved by the ethical committee of Pontificia Javeriana University. The ethical committees of Pontificia Javeriana University and Freiburg University approved the project on 07.03.2018 stating that the approval was based on 'the pertinence of the investigation, the methodological rigor, its scientific quality, the coherence and adequacy of the proposed budget, and the fulfilment of scientific, technical and ethical norms'. The local health authorities accepted the approvals after a detailed presentation of the study.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Author's contributions
SDM, AMV, JEA, YNF, MAC and AK designed the study, did the interviewer training, and drafted the paper. All authors read and approved the study and the procedures. AMV, YNF, EPQ and AFCN prepared the submission to the ethical committee and prepared the logistics. JEA, AMO, ARRV, JCBC, LAZ prepared the field study, contacted the local leaders, organized the logistics. All authors read and commented on the draft paper and approved the final version.