Open access Research

Prevalence of post-traumatic stress

disorder and associated factors among
Koshe landslide survivors, Addis
Ababa, Ethiopia: a community-based,
cross-sectional study
Sintayehu Asnakew,1 Shegaye Shumet, 2
Worknesh Ginbare,3 Getasew Legas,1
Kalkidan Haile4

To cite: Asnakew S, Shumet S, Abstract

Ginbare W, et al. Prevalence of Strengths and limitations of this study
Objectives To assess the prevalence of post-traumatic
post-traumatic stress disorder stress disorder (PTSD) and associated factors among the
and associated factors among ►► The nature of cross-sectional design, which might
survivors of Koshe landslide, Addis Ababa, Ethiopia, 2018.
Koshe landslide survivors, have only partially accounted for durable temporal
Design Community-based cross-sectional design.
Addis Ababa, Ethiopia: a relationships.
community-based, cross- Setting Koshe landslide, Addis Ababa, Ethiopia.
►► Social and recall biases might have interfered with
sectional study. BMJ Open Participants We recruited 830 participants for interviews
decisions respondents made when completing the
2019;9:e028550. doi:10.1136/ using the simple random sampling technique.
questionnaire.
bmjopen-2018-028550 Measurement We collected data by face-to-face
►► The post-traumatic stress disorder checklist-civilian
interviews. The civilian version of the PTSD checklist
►► Prepublication history for version scale used in the current study can serve as
was used to measure the symptoms of the disorder. The
this paper is available online. a reference in subsequent studies since it has good
To view these files, please visit Perceived Stress Scale and the Oslo-3 social support
internal consistency.
the journal online (http://​dx.​doi.​ instruments were used to assess the factors. Coded
org/​10.​1136/​bmjopen-​2018-​ variables were entered into Epi data V.4.2 and exported to
028550). SPSS V.24 for analysis. Bivariate and multivariate logistic
regressions with OR and 95% CI were employed. fear, helplessness or horror.1 In order to be
Received 17 December 2018 Result A total of 830 participants were interviewed, diagnosed with PTSD, a person must re-ex-
Revised 31 May 2019 with a response rate of 98.2%. The prevalence of perience the trauma, avoidance of trau-
Accepted 3 June 2019 ma-related stimuli, excessive arousal and
PTSD was 37.3% with 95% CI 34.1 to 40.8. In the
multivariate logistic regression, female sex (adjusted negative alterations in cognition and mood
odds ratio (AOR)=1.74, 95% CI 1.21 to 2.50), divorce which occur within a month after the event.2
(AOR=2.08, 95% CI 1.26 to 3.43), sustained physical A disaster is a traumatic event that might
injury (AOR=8.28, 95% CI 5.04 to 13.61), history of mental have been experienced by many people and
illness (AOR=5.55, 95% CI 2.30 to 13.36), family history causes different mental and physical health
of mental illness (AOR=2.82, 95% CI 1.48 to 5.37), poor
consequences.3 A survey study conducted
social support (AOR=3.64, 95% CI 1.99 to 6.69) and high
on US residents, 13% of the participants
perceived stress (AOR=3.08, 95% CI 1.43 to 6.64) were
© Author(s) (or their associated with PTSD. reported a lifetime exposure to natural or
employer(s)) 2019. Re-use
Conclusion and recommendations The prevalence of human-generated disasters.4 Several studies
permitted under CC BY-NC. No
PTSD among the survivors of Koshe landslide was high. conducted among adults showed that there
commercial re-use. See rights
and permissions. Published by We recommend that an early PTSD-focused regular was an increasing psychological distress
BMJ. screening be carried out by trained health professionals; after natural disasters.5 Although the conse-
1
Psychiatry, Debretabor linkage with mental health service providers also needs to quences of a disaster included a wide range
University, Debretabor, Ethiopia be considered. of psychopathology, a review study indicated
2
Psychiatry, University of Gondar, that PTSD is the most commonly investi-
Gondar, Ethiopia
3
Psychiatry, Amanuel Mental
gated and frequently occurring psychopa-
Specialized Hospital, Addis Introduction thology following disaster.6
Ababa, Ethiopia Post-traumatic stress disorder (PTSD) is a About 8 million adults had PTSD during a
4
Psychiatry, Debremarkos mental health problem that occurs following given year worldwide.7 Over 51 million people
hospital, Debremarkos, Ethiopia a traumatic event in which the individual are being forcibly displaced worldwide, of
Correspondence to experience, witnesses, is confronted with whom 16.7 million were displaced outside
Mr. Shegaye Shumet; either actual or imagined loss of life or their home countries, and this may lead to a
​shumetshegaye@​yahoo.​com serious injury which results in a response of stress-related disorders.8 The global disease

Asnakew S, et al. BMJ Open 2019;9:e028550. doi:10.1136/bmjopen-2018-028550 1

Open access

burden report attributed about 14% of the burden to Disasters including man-made mishaps have negative
neuropsychiatric disorders, mostly because of the long- impacts on the mental health of affected individuals.22
term disabling nature of depression and other common PTSD is the most common psychopathology and notable
mental disorders like PTSD.9 According to World Health public health matter that follows trauma/disaster.
Report 2001, ~0.4% of the total years lived with disability Although PTSD is highly prevalent among postdisaster
(YTD) followed PTSD, and the estimated burden settings, no studies have been done on the prevalence
increased to 0.6% YLD globally.10 Data in USA showed the of PTSD among Koshe landslide survivors, Addis Ababa,
lifetime prevalence of PTSD was 8% in the general popu- Ethiopia. So, determining the prevalence of PTSD and
lation. The lifetime prevalence rate was 10% in women associated factors among the survivors is important for
and 4% in men.11 A study on Israelis aged 18 years and early intervention and the reduction of the burden of
above and exposed to terrorism showed that the average PTSD and to improve the victims’ quality of life.
prevalence of PTSD was 9.4%, 16.2% for women and 2.4%
for men.12 The global economic burden of stress-related Objective
mental illness is expected to rise in the coming decade. This study set out to assess the prevalence of PTSD and
The global disease burden study of WHO estimates that associated factors among the survivors of Koshe landslide,
mental illness, including stress-related disorders, will be Addis Ababa, Ethiopia, 2018.
the second leading cause of disability by the year 2020.13
The prevalence rate of PTSD in developing countries
is higher compared with the developed ones. A study Methods and materials
conducted among a geographically diverse sample of Study settings and period
Mexican adults estimated the prevalence of PTSD at A community-based, cross-sectional study was conducted
19%.14 Different studies in Africa showed that PTSD could in May and June 2018. The study was conducted at
still be a public health concern for several years after the Koshe (dirt), a large open landfill with a surface area of
civil conflict and natural disaster. Reviews of communi- 25 hectares which used to receive 300 000 tons of solid
ty-based studies in South Africa showed that trauma expo- waste from Addis Ababa, the capital of Ethiopia, annually
sure was higher in low-income countries than in their as reported by Clague 2017. It was the only dumping site
counterparts.15 Another study in Uganda during an active available for the entire capital city with more than three
conflict showed that PTSD prevalence varied between million inhabitants. It has been located in the south-
18% and 54% in the general population.16 western part of Addis Ababa bounded by Nefas Silk-Lafto
PTSD is a public health issue that contributes to and Kolfie subcities. The area was a dumping ground
poverty, lack of employment, insecure living circum- for Addis Ababa’s rubbish for more than five decades,
stances, change in the social network and is highly asso- hosting hundreds of rubbish pickers who sell materials
ciated with low quality of life.17 18 Factors that contribute recovered from the waste. Some people even live around
to the development of PTSD have been classified into the site permanently.
pre-existing factors like family history of mental illness,
substance history as well as the traumatic event itself, and Study participants and sampling
post-trauma factors, such as lack of social support.19 We used the multistage sampling technique to select 830
There was a devastating garbage landslide in Addis participants. To reach households, the simple random
Ababa, Ethiopia, in the area of Koshe garbage land fill sampling technique (computer-generated random
on 11 March 2017. The catastrophic slope collapse number) was employed. In each of the areas, household
killed >113 people who were living around the land- lists were obtained from the kebeles/wards/offices and
fill and injured several others. The debris stood from a health extension workers. We proportionally allocated
height of 20 m beyond the actual toe line of the landfill, the sample size to Kilinto, Asko, Addis Hiwot and Koshe
destroying a minimum of 50 houses.20 21 The phenom- garbage dumping area, where victims temporarily settled.
enon occurred in the early morning hours of 11 March Members of the selected households were further sorted
and buried a number of makeshift homes under tons for interviews. In case of more than one eligible partic-
of refuse as reported by Eddie Haywood (March 2017). ipant in a household, the lottery method was used to
Koshe landfill is a large man-made mount formed from choose one.
vast rubbish dump on the outskirts of Addis Ababa, the The study included participants aged 15 years and
capital of Ethiopia. Hundreds of people used to attempt above during data collection in the area. There were a
to make a living by collecting refuse at the landfill site total of 5316 people in ~1035 households. Individuals
and selling it. Some people even lived around the rubbish seriously ill and unable to communicate were excluded.
dump permanently. Even though landslides sometimes
happened in Ethiopia, a man-made garbage hill slide like Sample size determination
this was quite. The landslide left a negative sequel on the We determined the sample size by using the single popu-
victims’ socioeconomic and psychological conditions, lation proportion formula with the assumptions of 48%
for example, in terms of housing, job and loss of family prevalence of PTSD from studies conducted in South
members. Sudan,23 0.48 P, 1.96 Z (standard normal distribution),

2 Asnakew S, et al. BMJ Open 2019;9:e028550. doi:10.1136/bmjopen-2018-028550

 Open access

95% CI, ⍺=0.05 and a 10% non-response rate. Accord- and individuals with higher scores indicating higher
ingly, a representative/probabilistic sample was calcu- perceived stress.27
lated to be 423. After considering design effect, the total
sample was 846. Substance use history
To examine substance use history, respondents were
Study variables asked: ‘Have you ever used any substance in the last
The dependent variable was PTSD measured by the 17 3 months or in your lifetime?’ and the responses were
items of the PTSD checklist-civilian version (PCL-C). yes/no.28
We measured PTSD as a dichotomous variable (yes/
no). Independent variables included sociodemographic History of mental illness
factors (age, sex, marital status, ethnicity, religion, educa- To examine history of mental illness, respondents were
tional and occupational status), clinical variables (family asked: ‘Have you ever been diagnosed with mental illness
history of mental illness, previous history of mental illness and treated’ and responses were yes/no.
and childhood trauma), trauma-related factors (trauma
exposure, perceived life threat), substance-related Family history of mental illness
factors (alcohol consumption, cigarette smoking, khat To examine family history of mental illness, respondents
chewing), awhile psychosocial factors embraced social were asked: ‘Do you know a family member who had
support and stressful life events. experienced a mental illness?’

Experiencing childhood trauma

Data sources and measurement
To examine childhood trauma, respondents were asked:
Data were collected by face-to-face interviews using a
‘Have you experienced childhood physical and sexual
semistructured questionnaire by four trained data psychi-
abuse and neglect’ and responses were yes/no.
atry nurses by means of the Amharic version of the tool
for a month. The questionnaire was designed in English Items on sociodemographic factors
and translated to Amharic and back to English to main- Items on sociodemographic factors (age, sex, ethnicity,
tain consistency. Data collectors were trained on how to religion, marital status, educational status and occupa-
interview participants and explain unclear questions and tional status) were adopted from a variety literature.
the purpose of the study. Furthermore, they were made
aware about ethical principles, such as confidentiality/ Data processing and analysis
anonymity/data management, and securing respondents’ All collected data were checked for completeness and
informed consent for participation. consistency and entered into Epi-data V.4.2 and then
PTSD was measured using the PCL-C. The PCL is a exported to SPSS for windows V.24 for analysis.
standardised self-report rating scale for PTSD comprising We computed descriptive, bivariate and multivariate
17 items that correspond to the key Diagnostic and logistic regression analyses to see the frequency distribu-
Statistical Manual of Mental Disorders-IV symptoms of tion and to test the association between independent and
PTSD. A total symptom severity score (range=17–85) dependent variables, respectively. Factors associated with
was obtained by summing the scores from each of the 17 PTSD were selected during the bivariate analysis with a
items. It had a Likert response options ranging from1 ‘not p<0.05 for further analysis in the multivariable logistic
at all’ to5 ‘extremely’ and a cut-off ≥50, that is, garbage regression analysis. In the multivariable logistic regression
landslide victims had PTSD symptoms.24 We adapted analysis, variables with p<0.05 at 95% CI with adjusted OR
this instrument from a study conducted on Somali and were considered as statistically significant.
Oromo Ethiopians in Minnesota.25 It showed a high
internal consistency, reliability and a strong correlation Ethical consideration
with PTSD diagnosis. We conducted a reliability analysis Permission was obtained from the Addis Ababa Admin-
for the PCL-C questionnaire (Amharic version) and that istration Health Bureau Ethical Committee. We received
it a had high score (Cronbach’s α=0.94). written informed consent from study participants and
assent from officially authorised foster parents after
Social support explaining the purpose of the study. Confidentiality was
Social support was measured using the Oslo 3-items maintained by omitting personal identifiers.
social support scale with scores ranging from 3 to 14:
3–8=poor social support; 9–11=intermediate social Patient and public involvement
support; and 12–14=strong social support.26 In the current study, participants were people who
survived the Koshe landfill, Addis, Ababa, Ethiopia;
Individual stress levels patients were excluded; participants were not involved in
Individual stress levels were measured using the Perceived the study design and recruitment. The results of this study
Stress Scale (PSS). The questions in this scale asked about will be disseminated to the Federal Ministry of Health,
feelings and thoughts last month. PSS was measured with Addis Ababa Health Bureau and Kolfe subcity Health
likert type scale ranging from (0) ‘never’ to4 ‘very often’ Department for timely management of survivors.

Asnakew S, et al. BMJ Open 2019;9:e028550. doi:10.1136/bmjopen-2018-028550 3

Open access

Table 1 Sociodemographic characteristics of study Table 2 Distribution of trauma-related factors of the

participants among residents of Koshe area, Addis Ababa, respondents in Koshe, Addis Ababa, Ethiopia, 2018 (n=830)
Ethiopia, 2018 (n=830) Characteristics Frequency Percentage
Characteristics Frequency Percentage Sustaining physical injury 166 20
Age (years) Witnessing the death of families or 526 63.4
 15–40 675 81.3 friends
Witnessing physical injury of families or 569 68.6
 >40 155 18.7
friends
Sex
Property destruction 240 28.9
 Female 491 59.2 Thinking, they may die 546 65.8
 Male 339 40.8 Perceived life threat
Marital status  Low perceived stress 185 22.3
 Married 428 51.6  Moderate perceived stress 581 70
 Single 249 30  High perceived stress 64 7.7
 Divorced 131 15.8
 Others* 22 2.7
abuse and neglect experience and 79 (9.5%) family
Ethnicity
history of mental illness.
 Amhara 404 48.7 Of the respondents, 569 (68.6%) witnessed phys-
 Tigray 138 16.6 ical injury of families or friends, and about 166 (20%)
 Gurage 135 16.3 sustained physical injury and 581 (70 %) moderate
perceived life threat (table 2).
 Oromo 123 14.8
Out of the total 830 participants, nearly half (48%) had
 Others† 30 3.6 poor social support and the majority, 659 (79.4%), expe-
Religion rienced at least one stressful life event (table 3).
 Orthodox 502 60.5 Regarding substance-related factors, nearly three-
 Muslim 195 23.5 fourths, 602 (72.5%) consumed alcohol, and 516
(62.20%) were doing that at the moment; 164 (19.80%)
 Protestant 94 11.3
smoked, and 102 (12.30%) were smoking; 129 (15.5%)
 Catholic 39 4.7 used khat (leaves) and 102 (12.30%) were using it during
Educational status the study.
 Cannot read and write 153 18.4
Prevalence of PTSD
 Primary school 366 44.1
The prevalence of PTSD among participants was 37.3%
 Secondary school 185 22.3 (95% CI 34.1 to 40.8).
 Diploma and above 126 15.2
Occupational status Factors associated with PTSD
To determine the association of independent variables
 House wife 131 15.8
with PTSD, bivariate and multivariate binary logistic
 Employed 472 56.9 regression analyses were carried out. In the bivariate
 Student 110 13.3 analysis, factors including female sex, >60 years of age,
 Jobless 117 14 divorce, history of mental illness, family history of mental
illness, childhood physical trauma and neglect, physical
*Separated, widowed.
†Silte, Hadya.
injury, witnessing the death and physical injury of families
or friends, property destruction, fear of death, poor social

Result
A total of 830 respondents took part with a response rate Table 3 Distribution of psychosocial factors of the study
of 98.2%. The majority of the respondents, 491 (59.2%), participants among residents of Koshe, Addis Ababa,
Ethiopia, 2018 (n=83)
were female. The mean age of the respondents was
33 (SD ±12) years; 675 (81.3%) were in the age range Characteristics Category Frequency Percentage
of 15–40 years; 428 (51.6%) were married; 502 (60.5%) Social support Poor 398 48
were Orthodox Christian and 404 (48.7%) Amhara by Moderate 324 39
ethnicity. Regarding occupation, more than half (56.6%) Strong 108 13
were employed (table 1). Stressful life events Yes 659 79.4
A small number, 55 (6.6%) of the participants had
No 171 20.6
history of mental illness, 202 (24.3%) childhood physical

4 Asnakew S, et al. BMJ Open 2019;9:e028550. doi:10.1136/bmjopen-2018-028550

 Open access

support and high perceived life threat were significantly Lebanon and 9.1% in sothern Brazil.35–38 The possible
associated with PTSD at a p<0.05. These factors were reason for this variation might be difference in instru-
entered into the multivariable logistic regression model ments. That is, the other study used general health
to control confounding effects. questionnaire (GHQ)-12, structured clinical interview,
The result of the multivariate analysis showed that mini-international neuropsychiatric interview (MINI),
female sex, divorce, history of mental illness, family history the modified version of the composite international
of mental illness, physical injury, poor social support and diagnostic interview, while we utilised the PCL-C. The
high perceived life threat were significantly associated with other variation might be due to the methods they used
PTSD at a p<0.05. Female sex was 1.7 times more likely to for data collection (structured telephone interview) and
develop PTSD compared with male sex (AOR=1.7, 95% CI conducting of studies late after the trauma.
1.2 to 2.5). The odds of developing PTSD were 2.1 times Female sex, being divorce, history of mental illness, family
higher among divorcees compared with the married ones history of mental illness, sustained physical injury, poor
(AOR=2.1, 95% CI 1.3 to 3.4). The odds of developing PTSD social support and high perceived life threat were signifi-
were 5.6 times higher among participants who had history of cantly associated with PTSD. The greater likelihood of PTSD
mental illness compared with those who had no such history among women than men in our work was similar to the
(AOR=5.6, 95% CI 2.3 to 13.4). The likelihood of devel- reports of other studies,29 31 32 39–41 possibly because females
oping PTSD was 2.8 times higher among respondents who experience sexual assaults and child sexual abuse more than
had family history of mental illness compared with those who males. Hence, being exposed to such trauma involves more
had no family history of mental illness (AOR=2.8, 95% CI 1.5 risk than other trauma in causing PTSD.42
to 5.4). The odds of developing PTSD were 8.3 times higher Divorcees were more likely to develop PTSD than
among respondents who sustained physical injury than those married respondents. Participants who lost their partners
who had not (AOR=8.3, 95% CI 5.0 to 13.6). Respondents and needed to support families, especially small chil-
who had poor social support were 3.6 times more likely to dren, single handed were more stressed. Our finding was
develop PTSD compared with those who had strong social supported by that of a study in Serbia.36
support (AOR=3.6, 95% CI 2.0 to 6.7). The odds of devel- History of mental illness was also significantly associ-
oping PTSD were 3.1 times higher among respondents who ated with PTSD. Participants with history of mental illness
had high perceived stress than those who had low perceived might have more neurochemical imbalance and neuronal
stress (AOR=3.1, 95% CI 1.4 to 6.6) (table 4). damage compared with those who had no history of
mental illness. As a result, they might be prone to develop
PTSD after the event. This finding was supported by
Discussion results of studies conducted in various countries.31 34 39 43 44
PTSD is the most common psychopathology and The odds of developing PTSD was 2.8 times higher
important public health matter after experiencing among respondents who had family history of mental
trauma/disaster. We found that, for the entire sample, illness than those who had no such illness. The possible
the garbage landslide had a negative impact on exposed explanation might be the inheritance of the serotonin
individuals’ mental health in terms of housing, income, transporter gene as well as genes associated with the
jobs and family problems resulting from the event. This hypothalamic–pituitary–adrenal axis and psychological
study found that a number of people met the criteria factors which make participants more highly predisposed
for post-trauma stress symptomatology. Some 37.3% of to PTSD.42 45 This finding was consistent with the results
people who experienced the incident presented with of studies conducted in South Korea.43 44
PTSD symptoms according to the PCL-C. Our finding Moreover, experiencing physical injury was a stronger
was consistent with reports of studies on people exposed predictor of PTSD compared with those who experienced
to natural disasters, such as 36.3% among earth quack no such injuries during the catastrophe. The finding was
victims in Kerman, 35.4% Syrian refugees in Lebanon, similar to the results of other studies.29 32 41 The possible
34.9% in Turkey and 34.3% among the bombing victims explanation for the similarity could be the presence of
of Oklahoma city, USA.8 29–31 Conversely, this finding was scars, the impaired part may remind the trauma and cause
lower than the 48% noted in South Sudan, 75.6% among reliving it and victims may believe that the traumatic event
the Rana Plaza building collapse victims in Bangladesh, has left its marks behind, and the body could keep clinging
57% in Saudi Arabia, 83.7% in Croatia and Serbia former to unresolved issues. The odds of developing PTSD was
Yugoslavia, Germany and UK, 59.4% in Fukushima 3.6 times higher among individuals who had poor social
nuclear disaster, Japan.18 23 32–34 The possible reason for support than strong social support. The finding is similar
this difference might be the use of different instruments to results of studies conducted in Southern Brazil and
and cut-off points to measure PTSD, exposure to multiple Mexico.38 46 Lack of help to compensate for physical inca-
trauma, study design and the nature and magnitude of pacity, emotional support and someone to talk with about
the accidents covered in the study. the traumatic experience or to turn to for advice could
On the other hand, our estimations are higher than increase the risk of PTSD.47
findings in other countries, for example, 11.8% in Participants who had high perceived stress were more
northern Uganda, 18.8% in Serbia, 29.3% in Southern likely to develop PTSD compared with respondents who

Asnakew S, et al. BMJ Open 2019;9:e028550. doi:10.1136/bmjopen-2018-028550 5

Open access

Table 4 Factors associated with PTSD among residents of Koshe, Addis Ababa, Ethiopia, 2018 (n=830)
PTSD
Variables Category Yes No COR (95% CI) AOR (95% CI)
Sex Male 106 (31.3%) 233 (68.7%) 1 1
Female 204 (41.5%) 287 (58.5 %) 1.6 (1.2 to 2.1)* 1.7 (1.2 to 2.5)*
Age (years) 15–40 241 (35.7%) 434 (64.3%) 1 1
>40 69 (44.5%) 86 (55.5%) 1.5 (1.0 to 2.1) 1.4 (0.9 to 2.1)
Marital status Married 138 (32.2%) 290 (67.7%) 1 1
Single 87 (34.9%) 162 (65.1%) 1.1 (0.8 to 1.6) 1.2 (0.8 to 1.8)
Divorced 77 (58.8%) 54 (41.2%) 3.0 (2.0 to 4.5)* 2.1 (1.3 to 3.4)*
Others 8 (36.4%) 14 (63.6%) 1.2 (0.5 to 2.9) 1.4 (0.5 to 4.2)
History (Hx) of mental illness Yes 46 (83.6%) 9 (16.4%) 9.9 (4.8 to 20.5)* 5.6 (2.3 to 13.4)*
No 264 (34.1%) 511 (65.9%) 1 1
Family Hx of mental illness Yes 55 (69.6%) 24 (30.4%) 4.5 (2.7 to 7.4)* 2.8 (1.5 to 5.4)*
No 255 (34.0%) 496 (66.0%) 1 1
Experiencing childhood trauma Yes 109 (54.0%) 93 (46.6%) 2.5 (1.8 to 3.4) 1.2 (0.7 to 1.9)
No 201 (32.0%) 427 (68.0%) 1 1
Sustaining Yes 135 (81.3%) 31 (18.7%) 12.2 (7.9 to 18.7)* 8.3 (5.0 to 13.6)*
physical trauma No 175 (26.4%) 489 (73.6%) 1 1
Witnessing the death of family yes 223 (42.4%) 303 (57.6%) 1.9 (1.4 to 2.5) 0.8 (0.5 to 1.4)
or friend No 87 (28.6%) 217 (71.4%) 1 1

Witnessing injury of family or Yes 238 (41.8%) 331 (58.2%) 1.9 (1.4 to 2.6) 0.8 (0.5 to 1.4)
friend No 2 (1.1%) 189 (98.9%) 1 1
Property destruction Yes 117 (48.8%) 123 (51.2%) 2.0 (1.4 to 2.7) 1.0 (0.7 to 1.5)
No 193 (32.7%) 397 (67.3%) 1 1
Thought of death Yes 242 (44.3%) 304 (55.7%) 2.5 (1.8 to 3.5) 1.3 (0.7 to 2.0)
No 68 (23.9%) 216 (76.1%) 1 1
Social support poor 209 (52.5%) 189 (47.5%) 4.9 (2.9 to 8.2)* 3.6 (2.0 to 6.7)*
Moderate 81 (25.0%) 243 (75.0%) 1.5 (0.9 to 2.5) 1.4 (0.8 to 2.6)
Strong 20 (18.5%) 88 (81.5%) 1 1
Perceived threat Low 56 (30.3%) 129 (69.7%) 1 1
Moderate 209 (36.0%) 372 (64.0%) 2.9 (1.9 to 4.6) 1.0 (0.7 to 1.6)
High 45 (70.3%) 19 (29.7%) 10.9 (5.6 to 21.4)* 3.1 (1.4 to 6.6)*
*P<0.05, modelfitness=0.114 (Hosmer and Lemshow), 0.000 (Omnibus test), no multicolinearity (tolerance >0.1 and variance inflation factor
(VIF) <2).
COR, crude odd ratio; PTSD, post-traumatic stress disorder.

had low perceived stress. The result is similar with find- landslide. The presence of earlier catastrophic experience
ings from Southern Israel and South Korea.43 48 Negative might have influenced the disorder due to the landslide.
beliefs about the consequence of the ongoing threat as Furthermore, social desirability and recall bias might
damaging implications will precipitate the onset and also be the other limitations. Since the data collection
persistence of PTSD.49 method was a face-to-face interview which might led indi-
viduals to respond in socially acceptable ways during the
Limitation of the study process, especially in cases of substance-related questions.
The cross-sectional design of the study prevented us from Individuals without PTSD symptoms may have less moti-
concluding the casual relationships of the associations we vation to recall earlier exposure than individuals with the
found. symptoms.
In addition, participants might not tell whether or not In addition, we did not consider other mental health
they had other PTSD symptoms before the onset of the problems that can confound outcomes. For instance, the

6 Asnakew S, et al. BMJ Open 2019;9:e028550. doi:10.1136/bmjopen-2018-028550

 Open access

presence and effects of anxiety and depression symptoms, 3. Norris FH, Friedman MJ, Watson PJ, et al. 60,000 Disaster victims
speak: part I. An Empirical Review of the Empirical Literature,
which are commonly associated with PTSD symptoms and 1981–2001. Psychiatry 2002;65:207–39.
the severity of PTSD, duration of mental illness or expo- 4. Burkle FM. Acute-phase mental health consequences of disasters:
sure to other diseases were not covered. implications for triage and emergency medical services. Ann Emerg
Med 1996;28:119–28.
The strength of the study was it used a relatively large 5. Gates MA, Holowka DW, Vasterling JJ, et al. Posttraumatic stress
sample and sampling methods. disorder in veterans and military personnel: epidemiology, screening,
and case recognition. Psychol Serv 2012;9:361–82.
Since we have employed face-to-face interviews, we 6. Galea S, Nandi A, Vlahov D. The epidemiology of post-traumatic
addressed individuals who had PTSD symptoms for stress disorder after disasters. Epidemiol Rev 2005;27:78–91.
further investigation and intervention. 7. Hg L, Heather Graham L. How Common is PTSD? National Center
for PTSD. 2015.:1. National Center for PTSD. 2015:1.
8. Kazour F, Zahreddine NR, Maragel MG, et al. Post-traumatic
stress disorder in a sample of Syrian refugees in Lebanon. Compr
Psychiatry 2017;72:41–7.
Conclusion 9. Prince M, Patel V, Saxena S, et al. No health without mental health.
The prevalence of PTSD was found to be high. This The Lancet 2007;370:859–77.
10. Ayuso-Mateos JL. Global Burden of post-traumatic stress disorder in
study confirmed that the garbage landslide had a nega- the year 2000: version 1 estimates. World Health Organ 2002.
tive impact on the mental health of affected individuals. 11. Kaplan BJ. Kaplan and Sadock’s Synopsis of Psychiatry.
Female sex, divorce, history of mental illness, family Behavioral Sciences/Clinical Psychiatry. Tijdschrift voor Psychiatrie
2016;58:78–9.
history of mental illness, sustained physical injury, 12. Bleich A, Gelkopf M, Solomon Z. Exposure to terrorism, stress-
poor social support and high perceived life threat related mental health symptoms, and coping behaviors among a
nationally representative sample in Israel. JAMA 2003;290:612–20.
were significantly associated with PTSD. Therefore, we 13. Kalia M. Assessing the economic impact of stress--the modern day
recommend a PTSD-focused early regular screening by hidden epidemic. Metabolism 2002;51:49–53.
trained health professionals and linkage with mental 14. Norris FH, Murphy AD, Baker CK, et al. Epidemiology of trauma
and posttraumatic stress disorder in Mexico. J Abnorm Psychol
health service providers. It is necessary to give emphasis 2003;112:646–56.
to individuals with family history of mental illness, 15. Atwoli L, Stein DJ, Koenen KC, et al. Epidemiology of posttraumatic
stress disorder: prevalence, correlates and consequences. Curr Opin
women and history of mental illness of those who expe- Psychiatry 2015;28:307.
rienced physical trauma during the disaster. 16. Ozer EJ, Best SR, Lipsey TL, et al. Predictors of posttraumatic stress
disorder and symptoms in adults: a meta-analysis. Psychol Bull
2003;129:52–73.
Acknowledgements The authors acknowledge the University of Gondar and
17. Akinyemi OO, Owoaje ET, Ige OK, et al. Comparative study of mental
Amanuel Mental Specialized Hospital for funding the study. We extend our gratitude health and quality of life in long-term refugees and host populations
to data collectors, supervisors and study participants for their time and effort. We in Oru-Ijebu, Southwest Nigeria. BMC Res Notes 2012;5:394.
also thank Mr. Demeke Dessu for his great contribution in language editing of the 18. Priebe S, Matanov A, Janković Gavrilović J, et al. Consequences
manuscript. of untreated posttraumatic stress disorder following war in former
Yugoslavia: morbidity, subjective quality of life, and care costs. Croat
Contributors SA developed the proposal, supervised the data collection, analyzed Med J 2009;50:465–75.
the data and wrote the draft manuscript. WG, GL, KH revised the proposal, checked 19. Keane TM, Marshall AD, Taft CT. Posttraumatic stress disorder:
the data analysis. SS, revised the proposal, check data analysis, revised and etiology, epidemiology, and treatment outcome. Annu Rev Clin
approved the manuscript. Psychol 2006;2:161–97.
20. Raviteja K, MunwarBasha B. Probabilistic back analysis of Koshe
Funding The authors research have no a specific grant from any funding agency in landfill slope failure. Indian geotechnical conference. 2017.
the public, commercial or not-for-profit sectors. 21. Organization WH. Weekly Update on Outbreaks and other
Competing interests None declared. Emergencies: Week 11: 11-17 March 2017. Weekly Update on
Outbreaks and other Emergencies. 2017.
Patient consent for publication Not required. 22. Catapano F, Malafronte R, Lepre F, et al. Psychological
consequences of the 1998 landslide in Sarno, Italy: a community
Ethics approval Ethical approval was obtained from Ethical Review Board of
study. Acta Psychiatr Scand 2001;104:438–42.
the University of Gondar. Ethical clearance was obtained from joint ethical review 23. Karunakara UK, Neuner F, Schauer M, et al. Traumatic events and
committees of the University of Gondar and Amanuel mental specialized hospital. symptoms of post-traumatic stress disorder amongst Sudanese
A formal letter of permission obtained and submitted to the respective town nationals, refugees and Ugandans in the West Nile. Afr Health Sci
administration. 2004;4:83–93.
24. Ruggiero KJ, Del Ben K, Scotti JR, et al. Psychometric properties
Provenance and peer review Not commissioned; externally peer reviewed. of the PTSD Checklist-Civilian Version. J Trauma Stress
Data sharing statement No additional data is available for this study; all the data 2003;16:495–502.
are included in the manuscript. 25. Jaranson JM, Butcher J, Halcon L, et al. Somali and Oromo
refugees: correlates of torture and trauma history. Am J Public Health
Open access This is an open access article distributed in accordance with the 2004;94:591–8.
Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which 26. Dalgard OS. The Oslo 3-items social support scale. 20002.
permits others to distribute, remix, adapt, build upon this work non-commercially, 27. COHEN. Perceived Stress Scale. 1983:7–8.
and license their derivative works on different terms, provided the original work is 28. Humeniuk R, Ali R, Babor TF, et al. Validation of the alcohol, smoking
properly cited, appropriate credit is given, any changes made indicated, and the use and substance involvement screening test (ASSIST). Addiction
2008;103:1039–47.
is non-commercial. See: http://​creativecommons.​org/​licenses/​by-​nc/​4.​0/.
29. Parvaresh N, Bahramnezhad A. Post-traumatic stress disorder in
bam-survived students who immigrated to Kerman. four months after
the earthquake 2009.
30. Ataman M. Prevalence of PTSD and related factors in communities
living in conflictual area: Diyarbakir case.. 2008.
References 31. North CS, Nixon SJ, Shariat S, et al. Psychiatric disorders among
1. Heinrichs M, Wagner D, Schoch W, et al. Predicting posttraumatic survivors of the Oklahoma City bombing. JAMA 1999;282:755–62.
stress symptoms from pretraumatic risk factors: a 2-year prospective 32. Fitch T, Villanueva G, Quadir M, et al. Prevalence and risk factors
follow-up study in firefighters. Am J Psychiatry 2005;162:2276–86. for PTSD in injured workers in Bangladesh: a study of surviving
2. Association AP. Diagnostic and statistical manual of mental disorders workers from the Rana Plaza building collapse. Lancet Glob Health
(DSM-5®): American Psychiatric Pub. 2013. 2015;3:S33.

Asnakew S, et al. BMJ Open 2019;9:e028550. doi:10.1136/bmjopen-2018-028550 7

 Open access

33. Alghamdi M, Hunt N, Thomas S. Prevalence rate of PTSD, 41. Zhou X, Kang L, Sun X, et al. Prevalence and risk factors of post-
depression and anxiety symptoms among Saudi firefighters. Journal traumatic stress disorder among adult survivors six months after the
of Traumatic Stress Disorders and Treatment 2016;6:1–6. Wenchuan earthquake. Compr Psychiatry 2013;54:493–9.
34. Tsujiuchi T, Yamaguchi M, Masuda K, et al. High prevalence of post- 42. Heim C, Newport DJ, Heit S, et al. Pituitary-adrenal and autonomic
traumatic stress symptoms in relation to social factors in affected responses to stress in women after sexual and physical abuse in
population one year after the Fukushima nuclear disaster. PLoS One childhood. JAMA 2000;284:592–7.
2016;11:e0151807. 43. Song JY, Jeong KS, Choi KS, et al. Psychological risk factors for
35. Mugisha J, Muyinda H, Wandiembe P, et al. Prevalence and factors posttraumatic stress disorder in workers after toxic chemical spill
associated with Posttraumatic Stress Disorder seven years after the in Gumi, South Korea. Workplace Health Saf
conflict in three districts in northern Uganda (The Wayo-Nero Study). 2018;66:393–402.
BMC Psychiatry 2015;15:170. 44. Brewin CR, Andrews B, Valentine JD. Meta-analysis of risk factors for
36. Lecic-Tosevski D, Pejuskovic B, Miladinovic T, et al. Posttraumatic posttraumatic stress disorder in trauma-exposed adults. J Consult
stress disorder in a Serbian community: seven years after trauma Clin Psychol 2000;68:748–66.
exposure. J Nerv Ment Dis 2013;201:1040–4. 45. Keane T, Fisher L, Krinsley K, et al. Handbook of prescriptive
37. Farhood L, Dimassi H, Lehtinen T. Exposure to war-related traumatic treatments for adults: Springer, 1994.
events, prevalence of PTSD, and general psychiatric morbidity in a civilian 46. Kaniasty K, Norris FH. Longitudinal linkages between perceived
population from Southern Lebanon. J Transcult Nurs 2006;17:333–40. social support and posttraumatic stress symptoms: sequential
38. Brunnet AE, Bolaséll LT, Weber J, et al. Prevalence and factors roles of social causation and social selection. J Trauma Stress
associated with PTSD, anxiety and depression symptoms in Haitian 2008;21:274–81.
migrants in southern Brazil. Int J Soc Psychiatry 2018;64:17–25. 47. Guay S, Billette V, Marchand A. Exploring the links between
39. Jenkins R, Othieno C, Omollo R, et al. Probable post traumatic posttraumatic stress disorder and social support: processes and
stress disorder in kenya and its associated risk factors: a cross- potential research avenues. J Trauma Stress 2006;19:327–38.
sectional household survey. Int J Environ Res Public Health 48. Besser A, Neria Y, Haynes M. Adult attachment, perceived stress,
2015;12:13494–509. and PTSD among civilians exposed to ongoing terrorist attacks in
40. Arnberg FK, Bergh Johannesson K, Michel PO. Prevalence and Southern Israel. Pers Individ Dif 2009;47:851–7.
duration of PTSD in survivors 6 years after a natural disaster. J 49. Taylor S. Clinician's guide to PTSD: A cognitive-behavioral approach:
Anxiety Disord 2013;27:347–52. Guilford Publications, 2017.

8 Asnakew S, et al. BMJ Open 2019;9:e028550. doi:10.1136/bmjopen-2018-028550