Abstract
Study Objectives:
To estimate the prevalence of, and to identify correlates of clinically significant sleep problems in adult Israeli citizens exposed to chronic terrorism and war trauma or threat thereof.
Methods:
A population-based, cross-sectional study of 1001 adult Israeli citizens interviewed by phone between July 15 and August 26, 2008. The phone survey was conducted in Hebrew and assessed demographics, trauma/stressor exposure, probable posttraumatic stress disorder (PTSD), probable depression, and sleep problems. Probable PTSD and depression were assessed with the PTSD Symptom Scale (PSS) and Patient Health Questionnaire (PHQ-9), respectively, following DSM-IV diagnostic criteria. Sleep problems in the past month were assessed with the Pittsburgh Sleep Quality Index (PSQI), on which a global composite score ≥ 6 indicates a clinical-level sleep problem.
Results:
Prevalence of probable PTSD and depression was 5.5% and 5.8%, respectively. Prevalence of clinically significant sleep problems was 37.4% overall, but was significantly higher for probable PTSD (81.8%) and probable depression (79.3%) subgroups. Independent correlates of poor sleep included being female, older, less educated, experiencing major life stressors, and experiencing psychosocial resource loss. Psychosocial resource loss due to terrorist attacks emerged as the strongest potentially modifiable risk factor for sleep problems.
Conclusions:
Sleep problems are common among Israeli adults living under chronic traumatic threat and trauma exposure. Given the continuing threat of war, interventions that bolster psychosocial resources may play an important role in preventing or alleviating sleep problems in this population.
Citation:
Palmieri PA; Chipman KJ; Canetti D; Johnson RJ; Hobfoll SE. Prevalence and correlates of sleep problems in adult Israeli Jews exposed to actual or threatened terrorist or rocket attacks. J Clin Sleep Med 2010;6(6):557-564.
Keywords: Sleep problems, trauma, war, terrorism, PTSD, depression, resource loss
Trouble sleeping is a common and costly phenomenon. It is estimated that over 40 million people in the United States have chronic or intermittent sleep problems1,2 and it is thought that most cases may be undiagnosed.2 The Institute of Medicine has deemed sleep problems to be an “unmet public health problem.”3 Poor sleep has been associated with negative psychological (e.g., depression, anxiety) and physical health (e.g., cardiovascular disease, motor vehicle accident injuries) outcomes and with poor occupational functioning (e.g., poorer job performance, more workplace accidents, higher absenteeism). Sleep problems have been estimated to have direct costs of $15 billion and indirect costs of $50-150 billion annually.4
Sleep problems are prominent aspects of psychiatric disorders. For example, insomnia (difficulty falling asleep, difficulty staying asleep, difficulty falling back asleep after awakening during the night) and recurring nightmares are two symptoms of posttraumatic stress disorder (PTSD). Sleep disturbance has been referred to as the hallmark of PTSD.5 Sleep symptoms are among the most commonly reported PTSD symptoms, estimated as high as 70% among PTSD cases,6 are associated with more severe PTSD,7 and tend to persist even following otherwise successful PTSD treatment and increase risk of relapse.8 There also is evidence that disturbed sleep predicts development of PTSD in trauma survivors.9,10 Despite all this, research that focuses specifically on sleep problems in traumatized populations and that assesses such problems with more than an item or two is scant. A notable exception is a study of Israeli adult community members that was done approximately half-way through the 41-day 1991 Gulf War, which found that 38% of the respondents suffered from acute insomnia as measured by 5 yes/no sleep-wake quality items.11 To our knowledge, however, no studies have focused on sleep problems in civilians living under persistent war and terrorist attacks or threat of such attacks.
BRIEF SUMMARY
Current Knowledge/Study Rationale: Sleep problems are prevalent and costly in the general population and are prominent aspects of psychiatric disorders such as posttraumatic stress disorder. Despite this, research that thoroughly assesses sleep problems and examines their prevalence and correlates in traumatized populations is scant, and none have focused on civilians living under chronic war/terrorism trauma or threat thereof.
Study Impact: This study reveals the extent of sleep problems in civilians living under persistent attack or threat of attacks, and identifies several correlates of these problems, including psychosocial resource loss, a strong and potentially modifiable risk factor that heretofore has not received attention in the sleep literature. Interventions that bolster psychosocial resources thus may be particularly useful in preventing or alleviating sleep problems and related psychopathology in this population.
The towns of Sderot, Ashkelon, and other Gaza vicinity communities in Israel have sustained thousands of rocket attacks by Hamas and Islamic Jihad over the past several years, leading to many casualties and hundreds of injuries.12 After years of attacks, the citizens of these areas near the Gaza Strip have suffered a substantial amount of direct trauma exposure. Yet, citizens of other areas of Israel with no or low levels of such attacks also have been affected by them, sustaining significant indirect exposure levels as media information and the threat to family and friends in high risk regions reach all areas of the country. Thus, the aims of the current study were to estimate the prevalence of sleep problems in adult Israeli citizens with direct or indirect exposure to terrorist or rocket attacks in a context of long-term chronic traumatic threat and trauma exposure (i.e., following the Al Aqsa Intifada), and to identify correlates of clinical levels of global sleep problems in this population.
Given that we focused on a trauma population and that sleep problems are such a prominent component of PTSD, we expected that variables associated with PTSD might also be predictive of sleep problems. Thus, we included several demographic (e.g., female gender, lower education, lower socioeconomic status, moderate/traditional religiosity) and exposure/stressor (e.g., amount of trauma exposure, other major life stressors, economic loss, and psychosocial resource loss) variables that have been shown in several studies of Israeli adults to be associated with PTSD.13–15
METHODS
Sampling
We randomly sampled 1001 Hebrew speaking Israeli civilian adults (age 18 and over) from two regions: 500 from Western Negev communities with documented high exposure to rocket/missile attacks, and 501 from settlements throughout Israel that had no specific history of rocket attacks or that had relatively lower levels of such exposure. Prospective participants from each region were randomly selected from a comprehensive database of Israeli landline phone numbers. Samples from each region were stratified by gender, based on census data indicating 52% female.
A total of 5913 phone numbers were called in the high exposure region. We excluded numbers that were disconnected (885, 15.0%); business/fax (91; 1.5%); unanswered (1991, 33.7%), busy (121, 2.0%), or re-directed to voicemail (152, 2.6%) on each of seven call attempts; and that connected to a household where language barriers precluded participation (1105, 18.7%) or where there otherwise were no eligible respondents (613, 10.4%). Of the remaining 955 eligible connections, 276 (28.9%) refused to participate, 154 (16.1%) asked to be called back later, 23 (2.4%) partially completed the survey, and 502 (52.6%) fully completed it. A total of 7651 phone numbers were called in the lower exposure region. We excluded numbers that were disconnected (902, 11.8%); business/fax (173, 2.3%); unanswered (2276, 29.7%), busy (390, 5.1%), or re-directed to voicemail (360, 4.7%) on each of seven call attempts; and that connected to a household where language barriers precluded participation (1215, 15.9%) or where there otherwise were no eligible respondents (1322, 17.3%). Of the remaining 1013 eligible connections, 354 (34.9%) refused to participate, 136 (13.4%) asked to be called back later, 22 (2.2%) partially completed the survey, and 501 (49.5%) fully completed it. Proportions across sample region were significantly different for 9 of the 11 outcomes of phone call attempts mentioned above (p values < 0.05); rates of refusal and of partial survey completion were not significantly different. Though significantly different, the proportions between sample regions were generally similar; the only phone outcome with a difference in proportions greater than 5% was no eligible person at the household (17.3% for the high exposure group versus 10.4% for the lower exposure group). The final sample of 1001 respondents from both geographical regions who fully completed the survey reflects a 51.0% response rate among the 1968 eligible individuals that were contacted.
Data Collection
Telephone interviews were conducted between July 15 and August 26, 2008 by native or fluent Hebrew speakers. Informed consent was obtained at the beginning of each phone call. Institutional review boards of Kent State University and the University of Haifa approved the study.
Instruments
Measures included in the survey instrument were translated from English to Hebrew. Translations of most measures in the present study have been used in previous studies with similar populations and have demonstrated good reliability and validity.13
Sample Characteristics
We assessed the following demographic variables: sex, age, religiosity, income, education, and marital status.
Exposure and major life stress indicators
Direct trauma exposure was assessed by three items asking how many times participants had experienced (a) death of a family member or close friend as a result of rocket or terror attacks; (b) an injury to oneself, a family member, or close friend as a result of rocket or terror attacks; and (c) witnessing rocket or terror attacks or being present where there were injuries or fatalities. Data were trichotomized to indicate that 0, 1 (single), or 2+ (multiple) types of traumatic experiences were endorsed. Internal reliability was not calculated because each trauma type is discrete and is not necessarily expected to predict other types of trauma exposure. A single item inquired about how many major stressful life events (e.g., death, disease or injury of a close person, divorce, or loss of work) were experienced in the past year that were not related to rocket or terror attacks. This measure (major life stressors) was used as a control variable when examining the influence of trauma exposure on probable PTSD and probable major depressive disorder (MDD).
Recent economic loss was measured with one item (yes, no) asking if participants suffered financial loss or property damage as a result of rocket or terror attacks during the past year. Psychosocial resource loss was assessed with four items from the Conservation of Resources Evaluation,16 a comprehensive measure of loss or threat of loss. Our abbreviated measure asked respondents to rate on a 4-point scale (coefficient α = 0.74) the extent to which they have experienced loss of the following in the past year as a result of rocket or terror attacks: “hope,” “you feel that you have control of your own life,” “closeness to one or more of your family members,” and “the feeling that you are a person of great value to other people”. Item scores were summed to yield a scale score ranging from 0 to 12, with higher scores indicating more loss. The scale variable was trichotomized to reflect no loss (scale score = 0), a small degree of loss (scale score = 1 to 2), and a higher degree of loss (scale score > 2).
Probable PTSD
Probable PTSD was measured with the PTSD Symptom Scale (PSS)17 and an item assessing impaired functioning. All items were phrased to assess symptoms directly related to rocket or terror attacks, as opposed to other potentially traumatic life events. The PSS contains 17 items that correspond with the DSM-IV18 PTSD symptom criteria, is a widely used measure to assess probable PTSD, and has demonstrated strong psychometric properties in a variety of trauma populations, including those similar to that of the current study.15 Because a separate purpose of the overall project was to pilot an expanded version of the PSS that separated two of the avoidance items (items 6 and 7) into 5 distinct avoidance items, half of the respondents in each of the samples received the original PSS and half received the augmented 20-item version. We combined these back into the original 2 items for scoring purposes. Each item was rated 0 to 3 for frequency of experiencing each symptom over the past month (coefficient α = 0.88 for both versions). Items rated 2 (moderately severe) or higher were considered clinically significant and were counted as endorsed symptoms. We used DSM-IV scoring criteria (at least one re-experiencing, at least 3 avoidance, and at least two hyperarousal symptoms) to determine whether symptom criteria were met. The impairment item used a 0 to 3 scale to assess the extent to which individuals felt that feelings and thoughts about rocket or terror attacks interfered with routine functioning. Respondents that met the full symptom criteria and answered 2 (moderately) or higher on the impairment item were considered to have probable PTSD.
Probable Depression
Probable depression was measured with the Patient Health Questionnaire-9 (PHQ-9),19 the 9-item depression module of the full PHQ that corresponds to the DSM-IV symptoms of a major depressive episode. This measure has demonstrated good reliability and validity19 and has been used in similar trauma populations to that of the current study.13 Items were rated from 0 to 3 for frequency of experiencing each symptom over the past 2 weeks (coefficient α = 0.79). Items were considered clinically significant at 2 (moderately severe). Individuals with ≥ 5 clinically significant symptoms, including depressed mood and/or anhedonia, were considered probable depression cases.
Sleep Problems
The Pittsburgh Sleep Quality Index (PSQI)20 was used to assess sleep quality and disturbances over the past month. It includes 18 items that measure 7 sleep components (sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbances, use of sleeping medication, daytime dysfunction) on a 0 (no difficulty) to 3 (severe difficulty) scale. These component scores are summed into a global score ranging from 0 to 21. Global composite scores ≥ 6 are considered to reflect clinical levels of sleep disturbance.20 The PSQI has demonstrated good psychometric properties in diverse populations, including substantial correlations with polysomnography and sleep diary data.20,21
Statistical Analyses
Descriptive statistics were computed for study demographic variables (sample, sex, age, religiosity, income, education, marital status), exposure indicators (trauma exposure, other major life stressors, financial loss, psychosocial resource loss), sleep problems, probable PTSD, and probable MDD. Independent samples t-tests (2-tailed) and χ2 tests at the 0.05 α level were conducted to test for group differences across sample, PTSD status, and depression status. Bivariate associations were computed between probable clinical sleep problems and the demographic and stressor variables, utilizing Pearson χ2 tests and Fisher exact tests (2-tailed) at the 0.05 α level. Variables that were significant at the p < 0.10 α level were included in a multivariate logistic regression model with probable sleep problems as the dependent variable to enable us to evaluate correlates after controlling for all other variables.
RESULTS
Sample Characteristics
Sample demographics are provided in Table 1. Consistent with census data, just over half the sample was female. Mean age was 43.2 years, (SD, 17.7; range = 18–95). With respect to religiosity, the majority self-identified as nonsecular. There was a broad distribution of income levels reported from below average to above average. The sample was highly educated, with only a small minority completing less than high school education. With respect to marital status, the majority were married or cohabiting.
Table 1.
Full sample characteristics (N = 1001)
| Variable | Full Sample N (%) or M(SD) | High Exposure N (%) or M(SD) | Lower Exposure N (%) or M(SD) | P-Value for χ2 test, Fisher exact test, or t-test |
|---|---|---|---|---|
| Demographic indicators | ||||
| Exposure group | ||||
| High exposure region | 500 (50.0) | |||
| Lower exposure region | 501 (50.0) | |||
| Sex | ns | |||
| Male | 481 (48.1) | 240 (48.0) | 241 (48.1) | |
| Female | 520 (51.9) | 260 (52.0) | 260 (51.9) | |
| Age | ns | |||
| 18–25 | 218 (21.8) | 112 (22.4) | 106 (21.2) | |
| 26–35 | 164 (16.4) | 81 (16.2) | 83 (16.6) | |
| 36–50 | 245 (24.5) | 118 (23.6) | 127 (25.3) | |
| 51–65 | 239 (23.9) | 122 (24.4) | 117 (23.4) | |
| 66+ | 125 (12.5) | 62 (12.4) | 63 (12.6) | |
| Religiosity | 0.003 | |||
| Secular | 367 (36.7) | 166 (33.2) | 201 (40.1) | |
| Traditional | 400 (40.0) | 216 (43.2) | 184 (36.7) | |
| Religious | 127 (12.7) | 74 (14.8) | 53 (10.6) | |
| Ultra-orthodox | 84 (8.4) | 32 (6.4) | 52 (10.4) | |
| Income | ns | |||
| Low | 337 (33.7) | 175 (35.0) | 162 (32.3) | |
| Medium | 253 (25.3) | 123 (24.6) | 130 (25.9) | |
| High | 263 (26.3) | 130 (26.0) | 133 (26.5) | |
| Education | 0.006 | |||
| Less than high school | 119 (11.9) | 68 (13.6) | 51 (10.2) | |
| High school | 372 (37.2) | 201 (40.2) | 171 (34.1) | |
| More than high school | 262 (26.2) | 127 (25.4) | 135 (26.9) | |
| College | 227 (22.7) | 93 (18.6) | 134 (26.7) | |
| Marital Status | ns | |||
| Single/Divorced/Separated/Widowed | 377 (37.7) | 187 (37.4) | 190 (37.9) | |
| Married/Cohabiting | 614 (61.3) | 306 (61.2) | 308 (61.5) | |
| Exposure/Stressor indicators | ||||
| Trauma Exposure | < 0.001 | |||
| No | 531 (53.0) | 217 (43.4) | 314 (62.7) | |
| Yes, one type | 285 (28.5) | 165 (33.0) | 120 (24.0) | |
| Yes, 2+ types | 185 (18.5) | 118 (23.6) | 67 (13.4) | |
| Death of family member or close friend | ns | |||
| No | 804 (80.3) | 388 (77.6) | 416 (83.0) | |
| Yes, 1 | 127 (12.7) | 71 (14.2) | 56 (11.2) | |
| Yes, 2 or more | 63 (6.3) | 36 (7.2) | 27 (5.4) | |
| Injury to self, family, or close friend | 0.010 | |||
| No | 831 (83.0) | 399 (79.8) | 432 (86.2) | |
| Yes, 1 | 105 (10.5) | 66 (13.2) | 39 (7.8) | |
| Yes, 2 or more | 56 (5.6) | 31 (6.2) | 25 (5.0) | |
| Witnessed injuries or fatalities | < 0.001 | |||
| No | 653 (65.2) | 276 (55.2) | 377 (75.2) | |
| Yes, 1 | 166 (16.6) | 106 (21.2) | 60 (12.0) | |
| Yes, 2 or more | 161 (16.1) | 103 (20.6) | 58 (11.6) | |
| Economic loss | < 0.001 | |||
| No | 892 (89.1) | 421 (84.2) | 471 (94.0) | |
| Yes | 109 (10.9) | 79 (15.8) | 30 (6.0) | |
| Psychosocial resource loss | 0.026 | |||
| None | 466 (46.6) | 217 (43.4) | 249 (49.7) | |
| Slight degree | 253 (25.3) | 123 (24.6) | 130 (25.9) | |
| Higher degree | 277 (27.7) | 157 (31.4) | 120 (24.0) | |
| Major life stressors | ns | |||
| No | 699 (69.8) | 348 (69.6) | 351 (70.1) | |
| Yes | 292 (29.2) | 144 (28.8) | 148 (29.5) | |
| Outcomes | ||||
| Probable PTSD | ns | |||
| No | 946 (94.5) | 472 (94.4) | 474 (94.6) | |
| Yes | 55 (5.5) | 28 (5.6) | 27 (5.4) | |
| PSS | 9.09 (8.58) | 10.01 (8.90) | 8.16 (8.15) | 0.001 |
| Probable MDD | ns | |||
| No | 933 (93.2) | 465 (93.0) | 468 (93.4) | |
| Yes | 58 (5.8) | 30 (6.0) | 28 (5.6) | |
| PHQ-9 | 4.43 (4.62) | 4.67 (4.73) | 4.18 (4.49) | ns |
| Probable sleep problems | ns | |||
| No | 588 (58.7) | 281 (56.2) | 307 (61.3) | |
| Yes | 374 (37.4) | 198 (39.6) | 176 (35.1) | |
| PSQI | 5.51 (3.48) | 5.61 (3.43) | 5.42 (3.54) | ns |
Numbers within categories may not add up to 1001 due to missing data.
PTSD, posttraumatic stress disorder; PSS, PTSD Symptom Scale; MDD, major depressive disorder; PHQ-9, Patient Health Questionnaire-9; PSQI, Pittsburgh Sleep Quality Index.
The lower exposure group was somewhat more educated than the high exposure group and was more likely to self-identify as secular. The groups did not differ on any other demographic variables.
Exposure to Trauma and Other Major Life Stressors
Trauma exposure and other stressors are reported in Table 1. The most frequently endorsed exposure item was witnessing attacks or being present where there were injuries or fatalities following attacks, endorsed by approximately one-third of the sample. Reporting having a family member or close friend die as a result of attacks was reported in 19% of the sample, while having been personally injured or having a close friend or family member injured in attacks was reported by 16% of the overall sample. Those in the high exposure group reported slightly higher levels of psychosocial resource loss than those in the lower exposure group. The samples did not differ on exposure to major life stressors in the past year.
Probable PTSD and Depression
Probable PTSD was diagnosed in 5.6% of the high exposure group and 5.4% of the lower exposure group. Probable depression was diagnosed in 6.0% of those in the high exposure group and 5.6% of those in the lower exposure group. Neither probable PTSD rate nor probable depression rate differed significantly by exposure level.
Sleep Problems
Descriptive statistics for sleep variables are reported in Tables 1 and 2. Clinical levels of sleep difficulty can be identified in individuals who score ≥ 6 on the PSQI. Clinical sleep problems were diagnosed in 37.4% of the respondents and did not differ across exposure groups. Respondents with probable PTSD and probable depression had strikingly high rates of concurrent clinical sleep difficulty. Considering the individual sleep component scores and the global composite score, the probable PTSD and probable MDD groups had similarly severe sleep difficulties that were significantly more severe than in the no PTSD and no MDD groups (t range from −7.67 to −2.33; p range from < 0.001 to 0.024), with the exception of sleep efficiency (i.e., lower percentage of total time in bed when trying to sleep is actually spent sleeping) for PTSD versus No PTSD, which was marginally significant in this direction. Specifically, they reported poorer sleep quality, longer sleep onset latency after going to bed for the night, shorter total duration of nighttime sleep, less efficient sleep (for the probable depression group comparison only), more sleep disturbances (e.g., due to feeling too hot or too cold, being in pain, having bad dreams, needing to go to the bathroom), more frequent use of medications to facilitate falling asleep, and more daytime fatigue. The above sleep results change negligibly when sleep related symptom criteria are excluded from the PTSD and MDD diagnostic classifications. In addition, no significant differences were found on any sleep variables between the high exposure and lower exposure samples.
Table 2.
Descriptive statistics and tests of group differences for sleep variables in the full sample and by diagnostic groups
| Full Sample N = 1001 | No PTSD n = 946 | PTSD n = 55 | No MDD n = 933 | MDD n = 58 | |
|---|---|---|---|---|---|
| % (N) with clinical levels of sleep problems (PSQI ≥ 6) | 37.4 (374) | 34.8 (329) | 81.8 (45)*** | 35.0 (327) | 79.3 (46)*** |
| Global PSQI Score, M (SD) | 5.51 (3.48) | 5.29 (3.32) | 9.38 (3.92)*** | 5.26 (3.25) | 9.61 (4.50)*** |
| Sleep Quality, M (SD) | 0.96 (0.81) | 0.91 (0.78) | 1.74 (0.94)*** | 0.91 (0.78) | 1.70 (0.94)*** |
| Sleep Latency, M (SD) | 0.85 (0.91) | 0.81 (0.90) | 1.58 (1.03)*** | 0.80 (0.89) | 1.66 (0.84)*** |
| Sleep Duration, M (SD) | 1.19 (0.88) | 1.17 (0.86) | 1.58 (1.03)** | 1.17 (0.86) | 1.57 (1.06)** |
| Sleep Efficiency, M (SD) | 0.51 (0.91) | 0.49 (0.90) | 0.72 (1.05)† | 0.49 (0.89) | 0.84 (1.17)* |
| Sleep Disturbance, M (SD) | 1.08 (0.58) | 1.04 (0.55) | 1.71 (0.73)*** | 1.05 (0.56) | 1.62 (0.67)*** |
| Sleep Medication Use, M (SD) | 0.23 (0.72) | 0.21 (0.68) | 0.62 (1.10)* | 0.19 (0.65) | 0.83 (1.26)** |
| Daytime Dysfunction, M (SD) | 0.69 (0.78) | 0.64 (0.75) | 1.52 (0.90)*** | 0.64 (0.75) | 1.47 (0.96)*** |
PTSD, posttraumatic stress disorder; MDD, major depressive disorder; PSQI, Pittsburgh Sleep Quality Index.
The percentages in the first row of the table change negligibly if PTSD and MDD classifications are made after excluding the sleep related symptom criteria for these disorders: 35.0% for No PTSD and 82.0% for PTSD, Fisher exact test, p < 0.001; 36.3% for No MDD; and 77.1% for MDD, Fisher exact test, p < 0.001. Descriptive statistics in the remaining rows also are negligibly changed if sleep symptoms are excluded from the PTSD and MDD prevalence calculations. T-tests and χ2 tests were conducted for lower exposure versus higher exposure, no PTSD versus PTSD, and no MDD versus MDD. There were no significant differences by trauma exposure level. Significant test results are indicated in the PTSD and MDD columns.
p < 0.10
p < 0.05
p < 0.01
p < 0.001
Bivariate Analyses
Table 3 displays results of bivariate analyses of probable clinical-level sleep difficulty and other study variables. Demographic variables significantly associated with probable clinical sleep difficulty included sex, age, religiosity, income, and education. Significant trauma and other stressor variables associated with clinical sleep difficulty included financial loss, psychosocial resource loss, and major life stressors.
Table 3.
Bivariate associations between respondent characteristics and clinical levels of sleep problems (N = 1001)
| Variable | % Sleep Difficulty (N) | p-Value |
|---|---|---|
| Demographic indicators | ||
| Exposure group | 0.13 | |
| High exposure region | 39.6 (198) | |
| Lower exposure region | 35.1 (176) | |
| Sex | < 0.001 | |
| Male | 31.4 (151) | |
| Female | 42.9 (223) | |
| Age | < 0.001 | |
| 18 – 25 | 16.6 (62) | |
| 26 – 35 | 14.2 (53) | |
| 36 – 50 | 23.5 (88) | |
| 51 – 65 | 27.3 (102) | |
| 66+ | 17.4 (65) | |
| Religiosity | 0.02 | |
| Secular | 31.9 (117) | |
| Traditional | 42.5 (170) | |
| Religious | 38.6 (49) | |
| Ultra-orthodox | 33.3 (28) | |
| Income | < 0.001 | |
| Low | 44.8 (151) | |
| Medium | 41.5 (105) | |
| High | 28.5 (75) | |
| Education | < 0.001 | |
| Less than high school | 16.8 (63) | |
| High school | 38.5 (144) | |
| More than high school | 23.8 (89) | |
| College | 18.4 (69) | |
| Marital Status | 0.12 | |
| Single/Divorced/Separated/Widowed | 34.2 (129) | |
| Married/Cohabiting | 39.3 (241) | |
| Exposure/Stressor indicators | ||
| Trauma Exposure | 0.40 | |
| No | 35.6 (189) | |
| Yes, one type | 39.3 (112) | |
| Yes, 2+ types | 39.5 (73) | |
| Economic loss | 0.003 | |
| No | 36.0 (321) | |
| Yes | 48.6 (53) | |
| Psychosocial Loss | < 0.001 | |
| None | 34.0 (127) | |
| Slight degree | 25.9 (97) | |
| Higher degree | 40.1 (150) | |
| Major life stressors | < 0.001 | |
| No | 33.3 (233) | |
| Yes | 46.9 (137) |
Numbers within categories may not add up to 1001 due to missing data. P-values are from χ2 tests and Fisher exact tests (all 2-tailed).
Multivariate Logistic Regression Analysis
Variables found to be associated with probable clinical sleep difficulty at the p < 0.10 level were included in a simultaneous logistic regression model, in order to determine the relative importance of each independent variable after controlling for all other variables. Results are provided in Table 4. Significant demographic correlates of probable clinical sleep difficulty included gender, age, and education. Neither religiosity nor income was significantly associated with probable sleep difficulty. Significant exposure/stressor variables included experiencing psychosocial resource loss and experiencing major life stressors in the past year. Incurring financial loss did not significantly predict sleep problems when other variables were simultaneously entered in the model.
Table 4.
Multivariate associations (logistic regression odds ratios) between respondent characteristics and clinically significant sleep problems (N = 809)
| Predictors | OR (95% CI) |
|---|---|
| Demographic Indicators | |
| Sex | |
| Male | 1.00 |
| Female | 1.46 (1.06 – 2.00)* |
| Age | |
| 18–25 | 1.00 |
| 26–35 | 1.59 (0.94 – 2.70)† |
| 36–50 | 1.52 (0.94 – 2.46)† |
| 51–65 | 2.06 (1.27 – 3.33)** |
| 66+ | 4.47 (2.46 – 8.11)*** |
| Religiosity | |
| Secular | 1.00 |
| Traditional | 1.34 (0.93 – 1.91) |
| Religious | 1.23 (0.75 – 2.00) |
| Ultra-orthodox | 1.44 (0.80 – 2.61) |
| Income | |
| Low | 1.00 |
| Medium | 1.10 (0.76 – 1.59) |
| High | 0.71 (0.48 – 1.05)† |
| Education | |
| Less than high school | 1.00 |
| High school | 0.58 (0.35 – 0.95)* |
| More than high school | 0.50 (0.30 – 0.86)* |
| College | 0.48 (0.27 – 0.83)** |
| Exposure/Stressor indicators | |
| Economic loss | |
| No | 1.00 |
| Yes | 1.48 (0.91 – 2.43) |
| Psychosocial Loss | |
| None | 1.00 |
| Slight degree | 1.51 (1.03 – 2.20)* |
| Higher degree | 2.91 (2.00 – 4.25)*** |
| Major life stressors | |
| No | 1.00 |
| Yes | 1.53 (1.10 – 2.13)* |
Sample size for the logistic regression model was lower due to list-wise deletion. The 192 respondents excluded from this analysis did not differ from the rest of the sample on any of the model variables and differed on only one of the study variables (marital status); they were more likely to be single/divorced/separated/widowed (Fisher exact test, p = 0.009, 2-tailed).
OR, odds ratio; CI, confidence interval.
p < 0.10
p < 0.05
p < 0.01
p < 0.001
DISCUSSION
We surveyed a large sample of adult Israelis from regions sustaining high and relatively lower levels of terrorist or rocket attacks in an effort to assess the prevalence and correlates of civilian sleep problems in war zones. Trauma exposure and other major life stressors were high overall. Rates of probable PTSD and MDD were 5.5% and 5.8%, respectively, lower than found in some studies,15,22 yet similar for depression and higher for PTSD in another study.23 These rates did not differ across the high and lower exposure regions, perhaps due to the ongoing conflict in the region and the resulting indirect exposure through relatives and friends, migration throughout the country, media exposure, and the pervasive trauma history across generations of people in our sample. It also speaks to the harmful effects of ongoing conflict—that direct attacks may have minimal additional impact on mental health functioning. Further, as this sample is random and non-treatment seeking, these rates are especially important from a public health perspective. Using recent census data,24 our observed rates of probable PTSD and depression suggest that approximately 250,000 adult Israelis are suffering from each of these disorders.
More than a third of the sample (37.4%) was classified as having clinical-level sleep problems, a result similar to the 38% found during a much briefer period of threat,11 and substantially higher than many general population rates elsewhere in the world.25,26 Using census data again,24 the 37.4% rate suggests that over 1.7 million adult Israelis are experiencing clinical-level sleep problems. Furthermore, approximately 80% (i.e., 200,000) of the individuals with probable PTSD and approximately 80% (200,000) of the individuals with probable MDD meet criteria for overall sleep problems. From a public health perspective, these rates and the direct and indirect costs associated with sleep problems are staggering.
Also notable is that rates of sleep problems did not differ across exposure regions, suggesting that it is not only those who directly experience attacks that develop sleep problems; rather, being under threat of attack, or knowing someone who might be in harm's way, seems to be sufficient to elicit sleep problems. This may be an additional indication that exposure is not limited to direct exposure, but extends to a wealth of indirect and historic trauma that may have a significant impact on mental and physical health. Further, we cannot rule out that persistent threat of attack may be a cause of worry and poor sleep that are felt generally by the population due to the constant media exposure.
Several study variables independently correlated with poor sleep. In terms of demographics, being female and being older were each associated with a higher likelihood of sleep problems, a finding consistent with research focusing on both healthy and psychiatric samples.27–29 Being more highly educated was associated with a lower likelihood of sleep problems. Religiosity and income were not significantly associated with poor sleep. With respect to exposure/stressor indicators, more psychosocial resource loss due to attacks and having major stressful life events each were associated with poorer sleep; whereas sustaining financial loss due to attacks was not an independent risk factor for disturbed sleep. Psychosocial resource loss, in particular, emerged as one of the strongest correlates of sleep problems and was the strongest risk factor that is potentially modifiable through intervention. Unexpectedly, trauma exposure was not even significantly associated with sleep problems bivariately, perhaps due to the ubiquitous threat of war and terrorist attack.
Conservation of resources theory (COR)30 predicts that resource loss is the major predictor of distress and related outcomes. This is the first study to show that resource loss is one of the best of several candidate variables associated with sleep disturbance in the context of chronic traumatic threat and trauma exposure. There are a number of possible mechanisms to explain this relationship. These losses may be sources of worry that interfere with going to sleep. They may also affect dream content and may cause anxiety that interferes with falling back asleep upon nighttime awakenings. As poor sleep impairs concentration and increases irritability, it is known to interfere with performance at work and to interfere with interpersonal relationships,3,26,31,32 which in turn result in further losses. In general, resource loss as measured in this study may signify to individuals the degree to which their lives have been negatively affected by traumatic events such as terrorism and war, and this appears to be more important than the degree of exposure itself.
Clinical Implications
Clinically, it is important to address sleep issues directly in members of this population with PTSD, since it is nearly a given (approximately 80%) that they will have sleep problems. It is possible that standard PTSD treatments will not suffice, as presence of sleep problems has been linked to more severe PTSD,7 and some sleep symptoms seem to persist after PTSD treatment.8 Thus, adjunct treatment targeting sleep problems specifically may be beneficial. Cognitive-behavioral interventions exist for sleep problems generally33 as well as for trauma-related sleep difficulty.34 Sleep problems also are important to address regardless of PTSD status, given the risk of PTSD in this highly trauma-exposed population and how sleep problems are predictive of subsequent PTSD9 and may play an etiological role in the development of PTSD. Of course, prevention and treatment of PTSD and sleep problems are complicated by ongoing war and terrorist attacks in this population. Though we cannot make causal inferences based on the current data, it is plausible that the loss of coping resources such as hope and personal control may make symptoms less bearable and thus potentially more interfering with sleep functioning. This highlights the importance of considering psychosocial resource loss as a possible critical point of intervention, as although we cannot necessarily stop terrorism and war, we may be able to stem, slow, or reverse ensuing resource losses that may contribute to or exacerbate sleep problems. Future research should continue to investigate the role of psychosocial losses in the development and maintenance of sleep disturbances.
Limitations/Future Research
Several limitations of the study need to be considered when interpreting the results. First, the onset of sleep problems, PTSD, and depression is not known. Second, although phone interviews often yield similar estimates of clinical symptoms as do in-person interviews, ours were not conducted by clinically trained professionals. Third, medical problems, home sleeping environment, and other factors that might affect sleep functioning were not assessed and need to be included in future studies. Fourth, we estimated sleep parameters through self-report only. Given that subjective reports can underestimate or overestimate certain sleep variables,35 research using more objective assessment methods would help build on the current findings. Fifth, if we assume that some of the phone calls that were unanswered, busy, or directed to voicemail were from households with eligible respondents, our actual response rate is somewhat lower than 51%. Finally, our cross-sectional data do not allow for causal inferences to be made about the role of sleep in subsequent outcomes. Follow-up data, especially for those individuals reporting current sleep problems but not meeting criteria for PTSD or depression, will help elucidate the role of sleep problems in the development of these disorders. It also is likely that sleep problems and PTSD/depression symptoms exacerbate each other in a bi-directional or spiraling manner, something that can be examined more fully with longitudinal data.
Summary
This is the first investigation focusing on sleep problems in civilians exposed to long-term war and terrorism trauma. Sleep problems are highly prevalent overall and extremely prevalent in civilians with probable PTSD or depression. Given that trouble sleeping has previously been found to predict subsequent PTSD and other negative health outcomes, the high rate of sleep problems in the current study does not bode well for long term functioning. As poor sleep may interfere with one's natural resilience and recovery abilities, sleep functioning should be assessed routinely and interventions that address disturbed sleep and its modifiable risk factors, especially psychosocial resource loss, may help alleviate or prevent psychopathology in this highly traumatized population.
DISCLOSURE STATEMENT
This was not an industry supported study. The authors have indicated no financial conflicts of interest.
ACKNOWLEDGMENTS
This study was made possible by a grant from the National Institute of Mental Health (R01 MH073687).
REFERENCES
- 1.National Heart, Lung, and Blood Institute. National sleep disorders research plan. Washington, DC: National Institute of Health; 2003. [Google Scholar]
- 2.Sexton-Radek K, Graci G. Combating sleep disorders. Westport, CT: Praeger Publishers; 2008. [Google Scholar]
- 3.Institute of Medicine Brief Report. Sleep disorders and sleep deprivation: An unmet public health problem. 2006. [PubMed] [Google Scholar]
- 4.Stoller MK. Economic effects of insomnia. Clin Ther. 1994;16:263–87. [PubMed] [Google Scholar]
- 5.Ross RJ, Ball WA, Sulllivan KA, Caroff SN. Sleep disturbance as the hallmark of posttraumatic stress disorder. Am J Psychiatry. 1989;146:697–707. doi: 10.1176/ajp.146.6.697. [DOI] [PubMed] [Google Scholar]
- 6.Ohayon MM, Shapiro CM. Sleep disturbances and psychiatric disorders associated with posttraumatic stress disorder in the general population. Compr Psychiatry. 2000;41:469–78. doi: 10.1053/comp.2000.16568. [DOI] [PubMed] [Google Scholar]
- 7.Germain A, Buysse DJ, Shear MK, Fayyad R, Austin C. Clinical correlates of sleep disturbance severity in posttraumatic stress disorder. J Trauma Stress. 2004;17:477–84. doi: 10.1007/s10960-004-5796-6. [DOI] [PubMed] [Google Scholar]
- 8.DeViva JC, Zayfert C, Pigeon WR, Mellman TA. Treatment of residual insomnia after CBT for PTSD: Case studies. J Trauma Stress. 2005;18:155–9. doi: 10.1002/jts.20015. [DOI] [PubMed] [Google Scholar]
- 9.Koren D, Arnon I, Lavie P, Klein E. Sleep complaints as early predictors of posttraumatic stress disorder: A 1-year prospective study of injured survivors of motor vehicle accidents. Am J Psychiatry. 2002;159:855–7. doi: 10.1176/appi.ajp.159.5.855. [DOI] [PubMed] [Google Scholar]
- 10.Mellman TA, Bustamante V, Fins AI, Pigeon WR, Nolan B. REM sleep and the early development of posttraumatic stress disorder. Am J Psychiatry. 2002;159:1696–701. doi: 10.1176/appi.ajp.159.10.1696. [DOI] [PubMed] [Google Scholar]
- 11.Askenasy JJM, Lewin I. The impact of missile warfare on self-reported sleep quality. Part 1. Sleep. 1996;19:47–51. doi: 10.1093/sleep/19.1.47. [DOI] [PubMed] [Google Scholar]
- 12.Besser A, Neria Y. PTSD symptoms, satisfaction with life, and prejudicial attitudes toward the adversary among Israeli civilians exposed to ongoing missile attacks. J Trauma Stress. 2009;22:268–75. doi: 10.1002/jts.20420. [DOI] [PubMed] [Google Scholar]
- 13.Hobfoll SE, Canetti-Nisim D, Johnson R. Exposure to terrorism, stress-related mental health symptoms, and defensive coping among Jews and Arabs in Israel. J Consult Clin Psychol. 2006;74:207–18. doi: 10.1037/0022-006X.74.2.207. [DOI] [PubMed] [Google Scholar]
- 14.Hobfoll SE, Canetti-Nisim D, Johnson RJ, Varley J, Palmieri PA, Galea S. The association of exposure, risk and resiliency factors with PTSD among Jews and Arabs exposed to repeated acts of terrorism in Israel. J Trauma Stress. 2008;21:9–21. doi: 10.1002/jts.20307. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Palmieri PA, Canetti-Nisim D, Galea S, Johnson RJ, Hobfoll SE. The psychological impact of the Israel–Hezbollah War on Jews and Arabs in Israel: The impact of risk and resilience factors. Soc Sci Med. 2008;67:1208–16. doi: 10.1016/j.socscimed.2008.06.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Hobfoll SE, Lilly RS. Resource conservation as a strategy for community psychology. J Community Psychol. 1993;21:128–48. [Google Scholar]
- 17.Foa EB, Riggs DS, Dancu CV, Rothbaum BO. Reliability and validity of a brief instrument for assessing post-traumatic stress disorder. J Trauma Stress. 1993;6:459–73. [Google Scholar]
- 18.American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington, DC: American Psychiatric Association; 1994. [Google Scholar]
- 19.Kroenke K, Spitzer RL, Williams JBW. The PHQ-9: Validity of a brief depression severity measure. J Gen Intern Med. 2001;16:606–13. doi: 10.1046/j.1525-1497.2001.016009606.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Buysse DJ, Reynolds CF, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index (PSQI): A new instrument for psychiatric research and practice. Psychiatry Res. 1989;28:193–213. doi: 10.1016/0165-1781(89)90047-4. [DOI] [PubMed] [Google Scholar]
- 21.Backhaus J, Junghanns K, Broocks A, Riemann D, Hohagen F. Test-retest reliability and validity of the Pittsburgh Sleep Quality Index in primary insomnia. J Psychosom Res. 2002;53:737–40. doi: 10.1016/s0022-3999(02)00330-6. [DOI] [PubMed] [Google Scholar]
- 22.Bleich A, Gelkopf M, Solomon Z. Exposure to terrorism, stress-related mental health symptoms, and coping behaviors among a nationally representative sample in Israel. JAMA. 2003;290:612–20. doi: 10.1001/jama.290.5.612. [DOI] [PubMed] [Google Scholar]
- 23.Levinson D, Zilber N, Lerner Y, Grinshpoon A, Levav I. Prevalence of mood and anxiety disorders in the community: Results from the Israel National Health Survey. Israel J Psychiatry. 2007;44:94–103. [PubMed] [Google Scholar]
- 24.Central Bureau of Statistics. Israel in figures 2007. Jerusalem, Israel: Central Bureau of Statistics; 2007. [Google Scholar]
- 25.Ford DE, Kamerow DB. Epidemiologic study of sleep disturbance and psychiatric disorders: An opportunity for prevention? JAMA. 1989;262:1479–84. doi: 10.1001/jama.262.11.1479. [DOI] [PubMed] [Google Scholar]
- 26.Kim K, Uchiyama M, Okawa M, Okawa M, Liu X, Ogihara R. An epidemiological study of insomnia among the Japanese general population. Sleep. 2000;23:41–7. [PubMed] [Google Scholar]
- 27.Bixler EO, Vgontzas AN, Lin HM, Vela-Bueno A, Kales A. Insomnia in central Pennsylvania. J Psychosom Res. 2002;53:589–92. doi: 10.1016/s0022-3999(02)00450-6. [DOI] [PubMed] [Google Scholar]
- 28.Buysse DJ, Reynolds CF, Monk, TH, Hoch CC. Quantification of subjective sleep quality in healthy elderly men and women using the Pittsburgh Sleep Quality Index (PSQI) Sleep. 1991;14:331–8. [PubMed] [Google Scholar]
- 29.Doi Y, Minowa M, Uchiyama M, Okawa M. Subjective sleep quality and sleep problems in the general Japanese adult population. Psychiatry Clin Neurosci. 2001;55:213–5. doi: 10.1046/j.1440-1819.2001.00830.x. [DOI] [PubMed] [Google Scholar]
- 30.Hobfoll SE. Conservation of resources: A new attempt at conceptualizing stress. Am Psychol. 1989;44:513–24. doi: 10.1037//0003-066x.44.3.513. [DOI] [PubMed] [Google Scholar]
- 31.Sexton-Radek K, Graci G. Combating sleep disorders. Westport, CT: Praeger Publishers; 2008. [Google Scholar]
- 32.World Health Organization Report. WHO technical meeting on sleep and health. Copenhagen, Denmark: World Health Organization; 2004. [Google Scholar]
- 33.Perlis ML, Benson-Junquist CB, Smith MT, Posner DA. Cognitive behavioral treatment of insomnia: A session-by-session guide. Springer; 2005. [Google Scholar]
- 34.Krakow B, Hollifield M, Johnston L, et al. Imagery rehearsal therapy for chronic nightmares in sexual assault survivors with posttraumatic stress disorder. JAMA. 2001;286:537–45. doi: 10.1001/jama.286.5.537. [DOI] [PubMed] [Google Scholar]
- 35.Carskadon MA, Dement WC, Mitler MM, Guilleminault C, Zarcone VP, Spiegel R. Self-reports versus sleep laboratory findings in 122 drug-free subjects with complaints of chronic insomnia. Am J Psychiatry. 1976;133:1382–8. doi: 10.1176/ajp.133.12.1382. [DOI] [PubMed] [Google Scholar]