Psychological and Psychopathological Reactions in Honduras Following Hurricane Mitch: Implications for Service Planning

Investigación original / Original research

Psychological and psychopathological reactions in Honduras following Hurricane Mitch: implications for service planning

Robert Kohn,1 Itzhak Levav,2 Irma Donaire,3 Miguel Machuca,4 and Rita Tamashiro5

Suggested citiation Kohn R, Levav I, Donaire I, Machuca M, Tamashiro R. Psychological and psychopathological reactions in Honduras following Hurricane Mitch: implications for service planning. Rev Panam Salud Publica. 2005;18(4/5):287–95.

ABSTRACT Background. Posttraumatic stress disorder (PTSD) and other psychopathological outcomes have not been sufficiently studied in community-based samples in Latin America. This study explored various psychopathological reactions and their respective risk factors two months after Hurricane Mitch struck Honduras in October 1998. Methods. In the Honduran capital of Tegucigalpa, 800 respondents age 15 and older were selected from residential areas of high, middle, or low socioeconomic status that had suffered either high or low impact from the devastating effects of the hurricane. The Composite Inter- national Diagnostic Interview was used to diagnose PTSD. Depression, alcohol misuse, and grief reaction were examined using screening instruments, and the Self–Reporting Question- naire was used to measure demoralization. The Impact of Event Scale was administered to ascertain the severity of the posttraumatic reaction. Results. PTSD was present in 10.6% of the sample. Respondents from the high-impact residential areas were more distressed, had higher scores on the grief inventory, and showed greater severity in PTSD symptoms. The respondents from the high-impact residential areas also had higher prevalence rates of major depression, alcoholism, and prior emotional problems. The best explanatory model for the risk of developing PTSD included the degree of exposure based on reported traumatic events, and associated increased demoralization. Among the persons with PTSD, its severity was predicted by being female and by the degree of exposure to hurricane-related traumatic events. Conclusions. Out of a total population of 3.3 million adults (age 15 and older) in Hon- duras, it is estimated that over 492 000 of them may have developed PTSD due to Hurricane Mitch. Adequate health disaster preparedness and response requires full acknowledgement of the multiple psychological effects that victims experience.

Key words Disasters; mental health; mental disorders; stress disorders, post-traumatic; Honduras.

1 Brown University, Department of Psychiatry and States of America; telephone 401-455-6277; fax: 4 Representation, Pan American Health Organization/ Human Behavior, Providence, Rhode Island, 401-455-6566; e-mail: [email protected] World Health Organization, Quito, Ecuador. United States of America. Send correspondence 2 Israel, Ministry of Health, Jerusalem, Israel. 5 Universidad Nacional Autónoma, Department of to: Robert Kohn, Butler Hospital, 345 Blackstone 3 Honduras, Ministry of Health, Tegucigalpa, Psychology, Tegucigalpa, Honduras. Blvd., Providence, Rhode Island 02906, United Honduras.

Rev Panam Salud Publica/Pan Am J Public Health 18(4/5), 2005 287 Original research Kohn et al. • Psychological and psychopathological reactions in Honduras following Hurricane Mitch

The scientific literature has repeat- 4.5% in the least-damaged area and of occurs, for example, because the au- edly shown that natural and human- 9.0% in the most-damaged area (7). thorities are overwhelmed by the im- made disasters leave in their wake A study examining the effects of Hur- mediate human and material losses, short-, medium-, and long-term psy- ricane Paulina in Mexico utilized the environmental consequences, destruc- chopathological reactions (1, 2). Post- Revised Civilian Mississippi Scale in the tion of health centers, and risk of epi- traumatic stress disorder (PTSD) is the community; however, a quota rather demics. In addition, the mental health most commonly studied disorder in than a random sample was employed impact of disasters may be ignored due mental health disaster outcome re- (8). De la Fuente (9) conducted two to the stigma of psychiatric disorders. search, although it is not the only psy- studies on the impact that victims living The objective of this community- chopathological reaction. The role of in shelters had suffered from an earth- based study was to investigate the risk culture in PTSD is complex, and so are quake in Mexico. These studies used of PTSD and other psychopathologi- the relationships that dose response, a questionnaire probing psychiatric cal reactions, including demoraliza- depression, demoralization, and sub- symptoms based on the Diagnostic and tion, depression, grief, and alcohol stance use have with PTSD (3). Statistical Manual, third edition (DSM- misuse, several months after Hon- Most of the research on PTSD has III), and found rates of PTSD of 32%. duras was struck by Hurricane Mitch been conducted in industrialized West- To date, only three studies following in 1998. The study was done in Tegu- ern nations, with little research coming natural disasters in Latin America have cigalpa, the capital city of the country. from developing regions, including used structured diagnostic interview Several measures of dose response Latin America. However, when nat- schedules to make a diagnosis of PTSD. were employed, and, unlike other ural disasters are measured by eco- The Diagnostic Interview Schedule studies, the sampling design was de- nomic loss or the number of deaths and (DIS) and Structured Clinical Interview veloped to specifically study the role injuries, Asia is the most vulnerable re- for DSM-III were used to make a diag- of socioeconomic status as a risk factor gion in the world, followed by Latin nosis of PTSD among 524 individuals for developing psychopathology. America. Of the 20 countries where the following an earthquake in Mexico. A most deaths from disasters occurred rate of 32% was found; women had between 1966 and 1990, 6 are located in higher rates than did men (10). In METHODS Latin America (4). Puerto Rico the DIS was readministered To date, the mental health conse- to 521 persons previously interviewed The setting quences of seven disasters in Latin in a representative national sample after America have been studied. In deriving a flood. The group who had been ex- Hurricane Mitch hit Central America a diagnosis of PTSD, most of these stud- posed to the flood had a rate of PTSD of on 25 October 1998 with wind speeds ies applied symptoms checklists rather 3.7%, and the unexposed group had a up to 285 km/hr (150 mi/hr). This was than structured diagnostic instruments. rate of 0.7% (11, 12). the most devastating disaster to affect In addition, the community was not Several studies have suggested that Central America in a century. the basis for the sampling framework. Hispanic populations may be more vul- Honduras suffered the brunt of the Investigations following the volcano nerable to PTSD. Using the DIS, Durkin disaster. The Pan American Health Or- eruption in Armero, Colombia, in- examined the rate of PTSD among vic- ganization (PAHO) and the Honduran cluded persons from shelters and pri- tims of earthquakes in the country of Government estimated that 1.5 million mary care clinics (5). Using a PTSD Chile and in the state of California in the of Honduras’ population of 6.4 million symptom checklist and clinical assess- United States (13); the rate in Chile was were displaced by the hurricane, 5 657 ment, those researchers found a PTSD 19%, versus 3% in California. The were killed, and 8 058 more were never rate of 32% in the shelters and of 42% in Chilean sample was based on residents found. Another 112 272 persons were the clinics. One published study on the of a housing complex, while the Califor- reported injured. Approximately 1 375 effects of Hurricane Mitch on Nicaragua nia sample was based on a household shelters had to be established to accom- included an investigation of adolescents survey. In a study of Hurricane Andrew modate 285 000 persons who were evac- from three public schools who had had in the United States, Spanish-speaking uated (15). In the region where Teguci- different degrees of exposure to the Latinos had higher rates of PTSD symp- galpa is located, 404 225 individuals impact of the hurricane on their com- toms than did either Hispanics who were displaced; of those, 22 429 were munity (6). Assessment for PTSD was choose to be interviewed in English or housed in 165 shelters (16). Fifteen of made using the Child Posttraumatic non-Hispanics (14). the 29 hospitals in the country were Stress Disorder Index. A second study Despite the relative scarcity of cur- rendered inoperable by the hurricane. in Nicaragua, which targeted adults, rent research on Latin America, the Thirty percent of the country’s health was based on consecutive patients in available data suggest that disasters infrastructure was damaged. Eighty four primary health care centers (7). lead to considerable mental health percent of the population had difficulty PTSD symptoms were measured using morbidity. However, health officials accessing drinking water. Marked in- the Harvard Trauma Questionnaire. and decisionmakers often overlook the creases in intestinal diseases, cholera, This study found a prevalence rate of mental health impact of disasters. This malaria, and dengue were seen.

288 Rev Panam Salud Publica/Pan Am J Public Health 18(4/5), 2005 Kohn et al. • Psychological and psychopathological reactions in Honduras following Hurricane Mitch Original research

Sample selection individuals living in the sampled shel- psychiatric cases in the general com- ters, all families were included. munity (19). The SRQ schedule also in- The study was conducted in Decem- In a brief interview the initial house- cluded five items on alcohol abuse, ber 1998 and January 1999. A sample hold resident who was contacted pro- which were modified to tap into be- of 800 adults (age 15 and older) strati- vided a description of the exposure of haviors that occurred since the hurri- fied by the socioeconomic status (SES) the hurricane on the home and the cane. This scale had been used glob- of the neighborhood was selected. family. This informant provided a list ally, including in Latin America (20). One-half of the adults were from low- of all the individuals, including the in- To measure losses resulting from the SES neighborhoods, one-fourth from formant, who resided in the home at hurricane, five items from the Texas In- middle-SES neighborhoods, and one- the time of the hurricane, in order for ventory of Grief were used (21). The fourth from high-SES neighborhoods. us to select for a more-extensive in- Texas Inventory of Grief was adminis- The SES groups were stratified by terview one target respondent per tered in only half of the interviews. high-impact and low-impact neigh- household, according to the Kish pro- To evaluate for the presence of borhoods. The selection and determi- cedure (17). Those age 15 and older current major depression a DSM-IV/ nation of which neighborhoods were were eligible. International Classification of Diseases considered to be high and low impact The Secretary of Health of Hon- Tenth Revision symptom checklist in- was done in consultation with the Sec- duras and the Representation of the quiring into the presence of diagnostic retariat of Health in Honduras. A Pan American Health Organization in symptoms in the past two weeks was neighborhood was considered to have Honduras approved both the inter- used. This instrument had been pre- suffered high impact if nearly all the view schedule and the consent forms viously used in a multicenter study homes were destroyed. Individuals in for human rights safeguard. Written of primary care practices through- the low-SES highly impacted strata informed consent was obtained from out Latin America (22). DSM-IV diag- were further subdivided into two the person providing the information nosis of PTSD was obtained using the groups: those remaining in the com- about the residents of the home and Spanish-language version 2.1 of the munity (either still residing in their also from the subject who underwent Composite International Diagnostic original homes or having moved else- the more-extensive interview. Each of Interview Schedule (CIDI) PTSD mod- where in the community) and those the interviewers was provided a letter ule, a diagnostic interview schedule temporarily housed in shelters. of introduction from the Secretary of (23). The module obtained the current For the persons remaining in the Health identifying them as an inter- diagnosis, and each question was community, the housing units were viewer for this study. modified so that it only referred to the selected from predetermined neigh- hurricane and to no other traumatic borhoods based on SES and impact. event. Presence of current PTSD symp- The number of homes in each neigh- Interview schedule toms and their severity in the past borhood was counted regardless of seven days was further evaluated by whether the house was intact or not. The interview of the target respon- using the Impact of Event Scale (IES) The first house selected was at the be- dent included sections on demogra- (24). The total score on the IES is de- ginning of each street. The subsequent phic information, demoralization, de- rived by the sum of positive responses. house selected was based on the num- pression, posttraumatic stress, alcohol The IES consists of two subscales, one ber of homes in the area divided by the abuse, help-seeking, and exposure to for intrusion and one for avoidance. number of interviews needed for the re- trauma. A psychiatrist and a Honduran Again, the items were linked directly spective strata. Homes were selected psychologist translated the scales that to the hurricane rather than being left regardless of whether or not they were did not have an available Spanish ver- open to other traumatic events. destroyed or occupied. If a house was sion. All items were evaluated for cul- A single item was included in the in- unoccupied or destroyed, the residents tural validity by the team and modified terview schedule to control for the of the home prior to it were located for the local usage, and to fit the events presence of any predisaster psycholog- if they resided in Tegucigalpa by ask- related to the hurricane. ical problem: “Did you have problems ing neighbors and reconstruction work- Demoralization was measured using with your nerves before the hurri- ers. If the residents of the designated the Self–Reporting Questionnaire (SRQ) cane?” This was followed by questions house could not be located, the subse- (18). Since the objective of the study evaluating the degree of traumatic ex- quent house was selected. Consecutive was to determine current psychiatric posure that occurred during the hurri- households were selected in the dis- and psychological problems, the SRQ cane (see Table 2 for a list of items). tricts where houses were destroyed and was modified to inquire about the last Positive responses to these items were in the poor neighborhoods where it seven days rather than the past month. summed to form an exposure index was extremely difficult to ascertain the The SRQ had been validated in neigh- (Chronbach’s alpha = 0.58). individual unit, such as with a slum boring Nicaragua, with a cutoff of nine The interview also asked about ex- dwelling where one family unit was positive symptom items being found to posure to violence following the hurri- built next to or on top of another. For be appropriate as a screen for probable cane (witnessing violence, being as-

Rev Panam Salud Publica/Pan Am J Public Health 18(4/5), 2005 289 Original research Kohn et al. • Psychological and psychopathological reactions in Honduras following Hurricane Mitch saulted, someone assaulting your dergo a change in occupation follow- events were: the house was damaged, child, or assaulting someone). The in- ing the hurricane. The demographic witnessed someone being injured, wit- terview concluded with a series of characteristics of the sample are nessed someone die, was injured, lost items about the utilization of any med- shown in Table 1. one’s belongings, and lost a pet. Each ical or mental health assistance after Table 2 reports the proportion of per- event was reported at a statistically the disaster, and the resources that sons who experienced any of six trau- higher rate in the high-impact group. might be used in the future. matic events whose occurrence during Twelve individuals, including two the hurricane we asked about, for all the from the low-impact group, reported respondents, the low-impact stratum, witnessing someone die. The high-im- Analysis and the high-impact stratum. The six pact group had a significantly increased

Risk factors for psychopathology were examined using both univariate and multivariate analyses. Chi-square TABLE 1. Demographic characteristics (%) of the sample in study of psychological and tests and t tests were used in the uni- psychopathological reactions in Honduras following Hurricane Mitch, for all respondents, variate analysis. Logistic regression those in low-impact neighborhoods, and those in high-impact neighborhoods was used to determine the odds ratios Total Low-impact High-impact (ORs) and the 95% confidence inter- sample neighborhoods neighborhoods vals (CIs) for potential risk of PTSD. Variable (n = 800) (n = 400) (n = 400) Chi-square/(df)a P Multivariate regression analysis was Gender 1.36 (1) 0.24 used to examine predictors of severity Female 61.4 63.4 59.4 of PTSD. Backwards regression anal- Male 38.6 36.6 40.6 yses were utilized to determine the Marital status 8.92 (5) 0.11 best-fitting models. Low SES, middle Married 29.3 29.8 30.1 SES, and high SES plus high-impact or Common-law 21.7 20.7 22.7 Never married 38.0 41.2 34.9 low-impact neighborhood were in- Widowed 4.0 4.1 3.9 cluded in each of the regression mod- Divorced/Separated 7.0 4.2 8.4 els since they were basic stratification Education completed 1.06 (3) 0.79 variables in the sample design. Statisti- No school 6.7 6.3 7.0 cal tests were considered to be signif- Primary school 36.1 34.9 37.3 < Secondary school 34.9 35.8 34.1 icant at the P 0.05 level. The statisti- University 22.3 23.0 21.6 cal analysis was conducted using the Current occupation 5.51 (4) 0.24 SUDAAN software program to adjust Employed 53.1 52.4 53.8 the standard errors (SEs) for the strati- Unemployed 22.3 20.9 23.7 fied sample design (25). Housewife 17.2 18.4 15.9 Student 5.1 6.6 3.7 Retired 2.3 1.7 2.9 Occupation prior to hurricane 16.72 (2) 0.0003 RESULTS Same employment 59.2 63.4 54.9 Other employment 17.3 11.8 22.9 Exposure and psychopathology Unemployed 23.5 24.7 22.2 a The degrees of freedom (df) are in parentheses. Eight hundred interviews were completed, stratified as follows: low impact and high SES, 100; low impact and middle SES, 101; low impact and TABLE 2. Proportion (%) of persons who reported exposure to traumatic events during the low SES, 199; high impact and high hurricane, in study of psychological and psychopathological reactions in Honduras following Hurricane Mitch SES, 100; high impact and middle SES, 100; high impact and low SES not in a Total Low-impact High-impact shelter, 98; and high impact and low sample neighborhoods neighborhoods SES in a shelter, 102. Females consti- Event (n = 800) (n = 400) (n = 400) Chi-squarea P tuted 61.4% of the sample. The average House was damaged 29.5 19.7 39.4 44.49 0.0001 age of the subjects was 37.0 years Saw someone injured 4.9 1.6 8.2 16.19 0.0001 (standard deviation, 16.8 years). Resi- Saw someone die 1.6 0.6 2.7 4.83 0.03 dence in a high-impact or low-impact Respondent was injured 5.1 1.6 8.6 19.06 0.0001 neighborhood was not significantly as- Lost belongings 26.6 4.8 48.5 281.63 0.0001 Lost pet 8.9 2.0 15.9 47.79 0.0001 sociated with any of the demographic variables other than having to un- a Degrees of freedom = 1 for the chi-square test.

290 Rev Panam Salud Publica/Pan Am J Public Health 18(4/5), 2005 Kohn et al. • Psychological and psychopathological reactions in Honduras following Hurricane Mitch Original research exposure index (t = –13.79, degrees of TABLE 3. Prevalence rates (%) of psychiatric disorders by impact group, in study of psy- freedom (df) = 793, n = 793, P < 0.0001) chological and psychopathological reactions in Honduras following Hurricane Mitch in contrast to the low-impact group. Total Low-impact High-impact Exposure to violence following the sample neighborhoods neighborhoods hurricane, reported by over one-third Disorder (n = 800) (n = 400) (n = 400) Chi-squarea P of the population, did not differ significantly by impact group. PTSDb 10.6 7.9 13.4 6.02 0.02 Major depression 19.5 14.9 24.4 10.56 0.002 PTSD, as diagnosed by the CIDI, Alcoholism 5.8 3.7 8.0 6.03 0.02 was present in 10.6% (SE = 1.1%) of the sample two months following the hur- a Degrees of freedom = 1 for the chi-square test. b PTSD = posttraumatic stress disorder. ricane. Table 3 shows the prevalence rates by high-impact and low-impact neighborhood, and Table 4 presents the rates by the three levels of SES. Each of the three disorders was signifi- TABLE 4. Prevalence rates (%) of psychiatric disorders by socioeconomic status (SES), in cantly higher among those in the high- study of psychological and psychopathological reactions in Honduras following Hurricane impact group, and that was also true Mitch for those in the low-SES group. The relationship between SES and the three Low Middle High diagnoses was not a result of higher ex- SES SES SES Disorder (n = 399) (n = 201) (n = 200) Chi-squarea P posure to the traumatic events in the low-SES group, given that the interac- PTSDb 15.7 3.7 7.7 26.66 0.0001 tion between SES and the exposure Major depression 25.9 11.0 15.6 22.79 0.0001 index was not statistically significant. Alcoholism 8.0 4.6 2.8 7.64 0.03 The proxy one-item measure of predis- a Degrees of freedom = 2 for the chi-square test. aster psychopathology, “prior nerves,” b PTSD = posttraumatic stress disorder. was also more prevalent among those in the high-exposure group (23.5%) than in the low-exposure group (16.6%) (chi-square = 5.86, df = 1, P < 0.02). The prevalence of “prior nerves” was in- problems, with that being 10.3% along with impact group and SES, versely related to SES (chi-square = among the high-impact group and each of these three variables remained 16.26, df = 2, P < 0.0003). 7.4% among the low-impact group. statistically significant. There were significant differences Exposure to the traumatic events between the two impact groups in the due to the hurricane that were risk fac- degree of demoralization as measured Risk factors for posttraumatic tors for PTSD included: house being by the SRQ. Respondents in the high- stress disorder damaged (OR = 3.4; 95% CI: 2.0–5.7); impact group were more distressed seeing someone injured (OR = 2.9; 95% (t = –4.45, df = 793, n = 772, P < 0.0001) Each logistic regression analysis that CI: 1.1–7.4); losing one’s belongings and had higher scores in the Texas examined risk factors for CIDI-PTSD (OR = 3.0; 95% CI: 1.5–5.9); and loss of Inventory of Grief (t = –4.88, df = 793, included impact group and SES since a pet (OR = 3.6; 95% CI: 1.9–7.0). Being n = 394, P < 0.0001). The IES also was these characteristics were used to injured during the hurricane did not elevated in the high-impact group, stratify the sample. Females were at a reach statistical significance (OR = 2.2; including the intrusion subscale (t = markedly increased risk (OR = 2.1, 95% CI: 0.9–5.1). The exposure index –4.55, df = 793, n = 783, P < 0.0001), the 95% CI: 1.2–3.4). Increasing age ap- was predictive of PTSD (OR = 1.9; 95% avoidance subscale (t = –4.10, df = 793, peared to have a small but statistically CI: 1.5–2.5). n = 786, P < 0.0001), and the total score significant increased risk for PTSD As for violent events following the (t = –4.74, df = 793, n = 773, P < 0.0001). (OR = 1.02; 95% CI: 1.00–1.03). Living hurricane, witnessing violence (OR = About one-fourth of the total sample in a common-law relationship (in con- 3.4; 95% CI 1.8–6.4) increased the risk of sought medical consultation in the two trast to being married) had an in- developing PTSD related to the hurri- months following the hurricane, but creased risk for PTSD (OR = 2.1; 95% cane. However, having been assaulted the 28.9% proportion among the high- CI: 1.1–4.1). Level of education and (OR = 2.2; 95% CI: 0.8–5.9) or having a impact group and the 24.2% among being unemployed (prior to or since child who was assaulted (OR = 1.2; 95% the low-impact group were not signif- the hurricane) were not significantly CI: 0.9–1.7) were not associated with icantly different. Nor were there statis- associated with increased risk for PTSD due to the hurricane. tical differences with regard to those PTSD. When gender, age, and marital Comorbid psychopathological reac- who sought assistance for emotional status were all included in the model tions and conditions were in part asso-

Rev Panam Salud Publica/Pan Am J Public Health 18(4/5), 2005 291 Original research Kohn et al. • Psychological and psychopathological reactions in Honduras following Hurricane Mitch ciated with increased risk for the diag- ance subscale (OR = 1.1; 95% CI: studies used symptoms checklists nosis of PTSD following the hurricane. 1.1–1.1), and the intrusion subscale rather than diagnostic tools and relied Major depression (OR = 11.1; 95% CI: (OR = 1.2; 95% CI: 1.1–1.2). Severity on selected populations. 6.5–18.9) and predisaster problems was examined using the IES total score Latin American studies on PTSD with nerves (OR = 1.4; 95% CI: 1.3–1.6) among those who had PTSD, using have also led to mixed results concern- were associated risk factors, but not al- multiple regression models. The strati- ing whether a dose response is related coholism (OR = 1.9; 95% CI: 0.8–4.3). fication variables, impact group, and to an increased risk of developing Alcoholism was associated with in- SES were not related to severity of PTSD. A study of Guatemalan refu- creased risk for PTSD, controlling for PTSD. Females with PTSD had higher gees living in Mexico found no associ- gender (OR = 3.4; 95% CI: 2.4–8.2). De- scores than did men on the IES (β = ation between the number of war- moralization was also comorbid with 8.37, SE = 4.23, P < 0.05). Age, marital related trauma events reported and PTSD, based on the SRQ (OR = 1.3; status, and unemployment prior to or symptoms of PTSD (26). In contrast, 95% CI: 1.3–1.4). The Texas Inventory after the hurricane were not related to Caldera et al. (7), who conducted a re- of Grief scale was not associated with PTSD severity. Seeing someone being search study on the effects that Hurri- PTSD. A logistic regression model that injured (β = 10.59, SE = 4.21, P < 0.02), cane Mitch had in the neighboring included only the possible comorbid getting injured during the hurricane country of Nicaragua, suggested that conditions, demoralization, major de- (β = 11.31, SE = 3.95, P < 0.0002), and there was a dose response that depression, and “prior nerves,” along losing one’s belongings (β = 10.43, pended on the degree of impact on the with the stratification variables, was SE = 3.20, P < 0.002) were events sig- different communities from which the created. This model resulted in demor- nificantly associated with increased samples were drawn. Two measures alization (OR = 1.3; 95% CI: 1.2–1.4) severity of the disorder. The exposure of impact/exposure were employed in and major depression (OR = 2.1; 95% index also predicted more severe our study of Honduras: (1) impact on CI: 1.0–4.5) remaining significant co- PTSD (β = 4.47, SE =1.17, P < 0.0001). neighborhood, an ecological measure, variates, but not “prior nerves.” Exposure to post-hurricane violence and (2) an exposure index, which re- A backward logistic regression was not related with severity of PTSD. flected the individual’s own experi- analysis was conducted using impact Major depression, alcoholism, and ence based on the number of trauma group, SES, gender, marital status, age, prior problems with nerves did not events. Although both were associated demoralization based on the SRQ, correlate with severity. Increased de- with an increased risk for PTSD, in the major depression, alcoholism, “prior moralization (β = 0.99, SE = 0.33, P < final model only the individual’s own nerves,” and exposure index to deter- 0.004) was significantly associated exposure experience remained signifi- mine the best-fitting model of the stud- with more severe PTSD, but not for the cant. Importantly, the severity of PTSD ied risk factors for PTSD. Only demor- measure for grief. Seeking medical or was associated only with individual alization (OR = 1.3; 95% CI: 1.2–1.4) mental health services was not associ- exposure factors. and the exposure index (OR = 1.6; 95% ated with severity of PTSD. Major depression, demoralization, CI: 1.2–2.1) remained statistically sig- A backward regression analysis was prior emotional disorder, and alco- nificant. Gender, age, “prior nerves,” conducted to determine which factors holism in men were all associated with major depression, and alcoholism did were predictive of severity of PTSD. PTSD. However, following control for not remain in the model; neither SES Gender, demoralization, exposure in- other covariates, only demoralization nor impact group remained statisti- dex, and the two stratification variables remained significant in the analysis. cally significant. were included in the initial model. The Curiously, comorbidity was not predic- Individuals with CIDI-PTSD were significant variables in the final model tive of severity of PTSD. Our measure much more likely to use health services included being female (β = 9.56, SE = of major depression and alcoholism following the hurricane compared to 3.54, P < 0.008) and the exposure index was based on symptom scales and not those without PTSD (48.5% versus (β = 4.66, SE = 1.13, P < 0.0001). on a diagnostic instrument, unlike the 23.5%; OR = 2.8; 95% CI: 1.7–4.5), as procedure that we followed to ascertain well as to consult for psychological is- PTSD. This is a limitation of this study, sues (30.1% versus 6.1%; OR = 5.7; 95% DISCUSSION and conceivably the relationship with CI: 3.1–10.3). comorbidity might have been different. PTSD appeared in 10.6% of the re- Furthermore, our study could not dif- spondents. This rate of PTSD was ferentiate incident cases of major de- Severity of posttraumatic much lower than in most Latin Amer- pression and alcoholism following the stress disorder ican studies examining disasters. This hurricane from those with a preexisting finding may be explained by the uti- condition. We did use a proxy measure An increased score on the IES was lization of a diagnostic instrument that for prior psychopathology, although its moderately associated with increased was clearly linked to the event, and to validity is open to interpretation. CIDI-PTSD risk for the total score the community-based sampling de- Prior research has suggested that so- (OR = 1.1; 95% CI: 1.1–1.1), the avoid- sign that we used. As noted, previous cioeconomic status is a risk factor for

292 Rev Panam Salud Publica/Pan Am J Public Health 18(4/5), 2005 Kohn et al. • Psychological and psychopathological reactions in Honduras following Hurricane Mitch Original research disaster-related PTSD (2), yet SES has FIGURE 1. The nature of the disaster and its impact on the individual been less often investigated than other issues such as dose response or comor- Type and extent of Disaster Impact factors: losses: Scope Cause: Involvement: bidity. Disasters have the potential to Economic Duration Man-made Peripheral Property Intensity affect people differentially. Disasters Nature Central Social network Preventability may have a higher impact the lower the Kinfolk Forewarning social class of the victim is. That is because poorer people have fewer re- Evacuated: Physical injury Shelter sources before a disaster, and are thus more likely to exhaust their assets be- Pre-disaster fore the disaster situation ends. Fur- Cultural factors: Networks: Ethnicity mediating factors Family functioning thermore, the risk of being affected by a Religion Social support disaster is not equal across SES groups; Sociodemographic: Prior life events the lower the socioeconomic level of a Gender neighborhood, the higher the impact. Age Physical health Our Honduras study found mixed re- SES–Education Occupation Psychological predisposition: sults for the role of SES. Those in the Income Personality low-SES group have markedly higher Locus of control Biological predisposition: Resilience rates of PTSD, suggesting an increased Prior psychiatric history Self-esteem risk. Increased risk of exposure itself Family history–Genetic Fatalism Developmental history Sense of coherence did not explain the inverse relationship Endocrine vulnerability between PTSD and SES, suggesting Neurochemical vulnerability that lower SES itself increases the risk for PTSD. SES, however, is not a strong Bereavement-specific Post-disaster predictor for PTSD and was not signifi- mediating factors mediating factors cant in the final model. In addition, SES Change in networks: did not predict severity of PTSD. Family functioning Confirmation of death vs. Social support The magnitude of the mental health presumed dead: View body reactions associated with Hurricane Physical health Mitch was considerable. Based on the Quality of the relationship New life events: census of Honduras, which indicated a with deceased Exposure to violence total population of 3.3 million age 15 Exposure to other trauma and older in 1998, we estimated that Other life events over 492 000 adults may have devel- Change in SES: oped PTSD from Hurricane Mitch Occupation alone. We also noted in our study the Income post-disaster violence that may have been a result of community disorgani- Bereavement outcome Psychological distress zation and deprivation. Our findings outcome do not reflect post-disaster PTSD from Grief Demoralization secondary trauma. TIME In our study a sizable proportion (30%) of those individuals with PTSD Complicated grief Psychological outcome Parental well-being attempted to get assistance for their Quality of life psychiatric problems. Unfortunately, Family functioning Physical health Academic performance this study did not ascertain the type or outcome Other psychosocial outcomes quality of the assistance provided. Al- though not all those affected may re- Psychiatric outcome Disability outcome quire direct intervention, some large- Body Acute stress function scale interventions may be needed. disorder and Such an intervention might reduce the structure PTSD, Other psychi- risk of secondary reactions by turning post atric disorders Participation non-understandable symptoms into traumatic Activity stress understandable ones and by suggest- disorder ing ways of coping with psychiatric symptoms, including seeking help Negative for an adverse outcome, from mental health services. or acute versus chronic outcome

Rev Panam Salud Publica/Pan Am J Public Health 18(4/5), 2005 293 Original research Kohn et al. • Psychological and psychopathological reactions in Honduras following Hurricane Mitch

Mental health disaster response re- perspective, each of these mediating tal health outcomes are diverse, and quires an understanding of the nature factors needs to be taken into account. they may include bereavement, demor- of the disaster and its impact on the in- For example, our study suggested that alization, psychosocial concerns (such dividual (Figure 1). This includes the personal losses, in particular property as parental well-being, quality of life, type and extent of the loss, the cause of losses, have a great impact on acute family functioning, and academic per- the disaster, the involvement of the in- psychopathology. This might suggest formance), physical health problems, dividual (central versus peripheral), that public health action focused on psychiatric disorders, and disability—as impact of the disaster, whether the restoring housing and other essential well as no adverse outcome. Mental person was evacuated, and if physical goods and in supporting the process of health planning should not cease with injury occurred. Taken together, these financial recovery might alone result just the conclusion of the acute event; factors constitute the individual’s dis- in a positive mental health outcome health care providers must be prepared aster experience. Three broad types of for many. Mental health services need to treat those who remain in protracted mediating factors should be accounted to be designed to address the complex- vulnerability and have long-term for in addressing an individual’s ity of the disaster experience, and the chronic conditions as a result of trauma. needs following a disaster (27): those services need to include surveillance that preceded the disaster, those that of the affected population over time. Acknowledgments. We thank our arose after the disaster, and those that Little research is available on the interviewers for their dedication and are related to situations involving be- long-term outcome of disasters, particu- Beth Hott for her assistance with the reavement. From the public health larly for low-income populations. Men- data management and the manuscript.

REFERENCES

1. Logue JN, Melick ME, Hansen H. Research is- 10. Tapia Conyer R, Sepúlveda Amor J, Medina in Nicaragua. Acta Psychiatr Scand. 1990;82: sues and directions in the epidemiology of Mora ME, Caraveo J, de la Fuente JR. Preva- 82–5. health effects of disasters. Epidemiol Rev. lencia del síndrome de estrés postraumático 20. Climent CE, Arango MV. Manual de psiquia- 1981;3:140–62. en la población sobreviviente de un desastre tría para trabajadores de atención primaria. 2. Norris FH, Friedman MJ, Watson PJ, Byrne natural. Salud Publica Mex. 1987;29:406–11. Washington, D.C.: Organización Panameri- CM, Diaz E, Kaniasty K. 60,000 disaster vic- 11. Bravo M, Rubio-Stipec M, Canino GJ, Wood- cana de la Salud; 1983. (Serie PALTEX para tims speak: part I. An empirical review of the bury MA, Ribera JC. The psychological se- técnicos medios y auxiliares, n. 1). empirical literature, 1981–2001. Psychiatry. quelae of disaster stress prospectively and 21. Faschingbauer TR, Devaul RA, Zisook S. De- 2002;65:207–39. retrospectively evaluated. Am J Community velopment of the Texas Inventory of Grief. 3. Kroll J. Posttraumatic symptoms and the Psychol. 1990;18:661–80. Am J Psychiatry. 1977;134:696–8. complexity of responses to trauma. JAMA. 12. Canino GC, Bravo M, Rubio-Stipec M, Wood- 22. Levav I, Kohn R, Montoya I, Palacio C, Rozic 2003;290:667–70. bury M. The impact of disaster on mental P, Salano I, et al. Training Latin American pri- 4. Noji EK. The nature of disasters: general char- health: prospective and retrospective analy- mary care physicians in the WPA module on acteristics and public health effects. In: Noji ses. Int J Ment Health. 1990;19:51–69. depression: results of a multicenter trial. Psy- EK, ed. The public health consequences of dis- 13. Durkin ME. Major depression and post- chol Med. 2005;35(1):35–45. asters. New York: Oxford University Press; traumatic stress disorder following the Coa- 23. World Health Organization. Composite Inter- 1997. Pp. 3–20. linga and Chile earthquakes: a cross-cultural national Diagnostic Interview, version 2.1. 5. Lima BR, Pai S, Santacruz H, Lozano J. Psy- comparison. J Soc Behav Pers. 1993;8:405–20. Geneva: WHO; 1997. chiatric disorders among poor victims follow- 14. Perilla JL, Norris F, Lavizzo EA. Ethnicity, 24. Horowitz M, Wilner N, Alvarez W. Impact of ing a major disaster: Armero, Colombia. J culture, and disaster response: identifying Event Scale: a measure of subjective stress. Nerv Ment Dis. 1991;179:420–7. and explaining ethnic differences in PTSD six Psychosom Med. 1979; 41:209–18. 6. Goenjian AK, Molina L, Steinberg AM, Fair- months after Hurricane Andrew. J Soc Clin 25. Shah BV, Barnwell BG, Bieler GS. SUDAAN banks LA, Alvarez ML, Goenjian HA, et al. Psychol. 2002;21:20–45. user’s manual. Release 7.5. Research Triangle Posttraumatic stress and depressive reactions 15. Honduras, Secretaría de Salud. Huracán Park: Research Triangle Institute; 1997. among Nicaraguan adolescents after hurri- Mitch en Honduras 1998. Tegucigalpa: Secre- 26. Sabin M, Lopes Cardozo B, Nackerud L, cane Mitch. Am J Psychiatry. 2001;158:788–94. taría de Salud; 1999. Kaiser R, Varese L. Factors associated with 7. Caldera T, Palma L, Penayo U, Kullgren G. 16. Honduras, Secretaría Técnica y de Coop- poor mental health among Guatemalan refu- Psychological impact of the hurricane Mitch in eración Internacional. Número de personas gees living in Mexico 20 years after civil con- Nicaragua in a one-year perspective. Soc Psy- damnificadas y en albergues. Tegucigalpa: flict. JAMA. 2003;290:635–42. chiatry Psychiatr Epidemiol. 2001;36: 108–14. SETCO; 1998. 27. Kohn R, Levav I. Bereavement in disaster: an 8. Norris FH, Perilla JL, Murphy AD. Postdisas- 17. Kish L. Survey sampling. New York: John overview of the research. Int J Ment Health. ter stress in the United States and Mexico: a Wiley & Sons; 1965. 1990;19:61–76. cross-cultural test of the multicriterion con- 18. Beusenberg M, Orley J. A user’s guide to the ceptual model of posttraumatic stress disor- Self Reporting Questionnaire. Geneva: World der. J Abnorm Psychol. 2001;110:553–63. Health Organization, Division of Mental 9. De la Fuente R. The mental health conse- Health; 1994. quences of the 1985 earthquakes in Mexico. 19. Penayo U, Kullgren G, Caldera T. Mental dis- Manuscript received 24 August 2004. Accepted for publi- Int J Ment Health. 1990;19:21–9. orders among primary health care patients cation 14 June 2005.

294 Rev Panam Salud Publica/Pan Am J Public Health 18(4/5), 2005 Kohn et al. • Psychological and psychopathological reactions in Honduras following Hurricane Mitch Original research

RESUMEN Objetivos. En América Latina no se han estudiado suficientemente los trastornos por estrés postraumático (TSP) y otras afecciones piscopatológicas mediante muestras basadas en la comunidad. El presente estudio explora varias reacciones psicopatoló- Reacciones psicológicas y gicas y sus respectivos factores de riesgo dos meses después de que el huracán Mitch psicopatológicas en Honduras azotó Honduras en octubre de 1998. después del huracán Mitch: Métodos. En Tegucigalpa, capital de Honduras, se seleccionó a 800 personas de 15 años de edad o más que vivían en áreas residenciales consideradas de estatus socioe- implicaciones para la conómico alto, medio o bajo y que habían sufrido los devastadores efectos del hura- planificación de los servicios cán en mayor o en menor medida. Para diagnosticar los casos de TSP se utilizó la En- trevista Diagnóstica Internacional Compuesta. La depresión, el abuso de alcohol y la reacción de dolor se evaluaron mediante encuestas de tamizaje, mientras que para la desmoralización se empleó el Cuestionario de Síntomas. Se aplicó la Escala de Im- pacto de Eventos para establecer la gravedad de la reacción postraumática. Resultados. Se encontró TSP en 10,6% de la muestra. Los encuestados de las áreas residenciales más afectadas estaban más angustiados, tuvieron una mayor puntua- ción en la evaluación de sufrimiento, y presentaron síntomas de TSP de mayor gravedad. También presentaron mayores tasas de prevalencia de depresión mayor, alco- holismo y problemas emocionales previos. El mejor modelo explicativo del riesgo de sufrir TSP tomó en cuenta el grado de exposición según los eventos traumáticos in- formados y el incremento de la desmoralización asociada con ellos. Entre las personas con TSP, los factores pronóstico de la gravedad de los trastornos fueron el pertenecer al sexo femenino y el grado de exposición a los eventos traumáticos relacionados con el huracán. Conclusiones. Se calcula que del total de 3,3 millones de adultos (de 15 años de edad o más) habitantes de Honduras, más de 492 000 han sufrido TSP debido al huracán Mitch. Para contar con una adecuada preparación sanitaria y ofrecer una apropiada respuesta ante los desastres es necesario reconocer en profundidad los múltiples efectos psicológicos que experimentan las víctimas.

Palabras clave Desastres, salud mental, trastornos mentales, trastornos por estrés postraumá- tico, Honduras.

A veces no nos dan a escoger entre las lágrimas y la risa, sino sólo entre las lágrimas, y entonces hay que saberse decidir por las más hermosas. Maurice Maeterlinck, autor belga (1862–1949)

Rev Panam Salud Publica/Pan Am J Public Health 18(4/5), 2005 295