Table 2.
Detailed results on vulnerability factors
| Author, date, (reference), [quality score*] | Context (definition of older age) | Outcome of interest/study design (analysis) | Comparison group | Vulnerability factors |
|---|---|---|---|---|
| Mental Health Outcomes | ||||
| Ardalan et al., 2010. [25] [9/10] | Iran/earthquake (60+) |
Older people’s perceptions of needs post disaster/qualitative - focus-group and interviews (content analysis). | NA (qualitative) | Feelings of insecurity: Not being able to protect oneself or belongings from thieves due to frail physical state and fear of being targeted for this reason. Emotional distress: many still very emotionally upset due to trauma during earthquake even years later. Difficult to adjust to new life after earthquake. Causes of distress: losing children, close relatives and friends, experiencing hopeless days, witnessing the destruction of historic parts of the city and losing valued documents and memorabilia. |
| Ardalan et al., 2011[a] [26] [7/10] | Iran/earthquake (60+) | Psychological state/quality of Life (WHOQOL-BREF)/quantitative - cross-sectional (multivariate) | Earthquake affected vs. non-earthquake affected |
Within earthquake-affected sample (regression coefficients): higher age (−0.113, p = 0.003); female gender (−1.169, p = 0.016); urban residence (−1.043, p = 0.044); being unmarried (−1.144, p = 0.018); history of earthquake related injury (−1.542, p = 0.028); dissatisfaction with quality of current living place (−2.718, p < 0.001); functional dependence (ADL) (1.151, p = 0.004). Comparison with non-earthquake-affected populations: survivors had lower psych. State scores (mean 11.88 12.80, p = 0.03) |
| Cao et al., 2014 [29] [5/10] | China Earthquake (60+) | Psychological distress (SRQ-20/quantitative survey - cross-sectional (bivariate) | None |
Older men: loss of family members (OR 1.32–31.53, p = 0.02); displacement from residence (OR 1.08–32.33, p = 0.04). Older women: higher educational level (OR 0.09–0.77 p = 0.02); chronic illness (OR 1.11–13.78 p = 0.03); loss of family members (OR 2.87–76.51 p = 0.00); displacement from residence (OR 3.37–37.18 p = 0.00). |
| Chaaya et al., 2007 [30] [7/10] | Palestian.refugee, Lebanon/(60+) | Depression (GDS)/quantitative - cross-sectional (multivariate) | None | Regular religious attendance (OR 0.41, p = 0.041); sufficient income (OR 0.42, p = 0.003); ADL difficulties (OR 2.05, p = 0.015); Illness during last year OR 2.89, p < 0.001). |
| Chan et al., 2009[a] [31] [4/10] | Pakistan Earthquake (45+) | Psychosocial (SRQ)/quant. Cross-sectional (bivariate) | Rural vs. urban | Rural more likely to experience feeling depressed/helpless (72% vs. 44%, p < 0.001). Rural more likely to experience sleeplessness (65% vs. 45%, p < 0.001). |
| Chen et al., 2012 [33] [8/10] | China Earthquake (60+) | PTSD (CAPS for DSM)/quantitative - cross-sectional (multivariate) | None | Female gender (OR 1.592 [95% CI 1.236–2.057]); aged 81 years or older (OR 1.557 [95% CI 1.057–2.292]); widowed (OR 1.464 [95% CI 1.281–1.660]); low education level (OR 1.395 [95% CI 1.073–1.804]); low monthly income (OR 1.670 [95% CI 1.401–1.992]); suffering bodily injury (OR 2.468 [95% CI 1.863–3.246]); bereavement OR 2.064 [95% CI 1.363–3.994]); low social support (OR 1.826 [95% CI 1.054–3.162]. |
| Goenjian et al., 1994 [36] [4/10] | Armenia Earthquake (59+) | PTSD (PTSD Reaction Index) /quantitative - cross-sectional (bivariate) | Older vs. younger |
Within older population: living in high impact zone: higher PTSD scores (p < 0.05) Comparison with younger: older people had lower PTSD score re-experiencing (2.1 vs. 2.5, p < 0.05); older people had higher PTSD score arousal (2.7 vs. 2.4, p < 0.05). |
| Handicap Int. & HelpAge Int, 2014 [37] [4/10] | Syrian refugees in Lebanon & Jordan/war (60+) | Psychological distress(SRQ)/ quantitative survey - cross-sectional (bivariate) | Older vs. younger | Older age populations 3 times more likely than non-elderly to show signs of psychological distress. (no p-value or exact figures reported) |
| Havelka et al., 1995 [38] [3/10] | Croatia/war (60+) | Psychosomatic Disorders (SSPD)/quantitative survey- cross-sectional (bivariate) | Older vs. younger |
Older age a risk factor for often experiencing the following outcomes related to psychosomatic feelings (older vs. non-elderly):
Persistent memory of stressful event: 81.5% vs. 55.2%, p 0.001; depression 80.6% vs. 52.9%, p = 0.001; insomnia 53.4% vs. 34.3%,p = 0.005; nightmares 33.0% vs. 13.8%, p = 0.001; distraction 32.0% vs. 14.6%, p = 0.001; forgetfulness: 30.1% vs. 11.3%, p = 0.001; emotional numbness 12.6% vs. 6.7%, p = 0.039; fear: 22.3% vs. 10.5%, p = 0.010) Being older protective against often experiencing the following outcomes (older vs. non-elderly): Bitterness and resentment towards others 18.4% vs. 36.4%, p = 0.001; aggressive behaviour 6.8% vs. 11.7%, p = 0.014. |
| Jia et al., 2010 [39] [8/10] |
ChinaEarthquake (60+) | PTSD (PCL-C), psychiatric morbidity (GHQ-12)/quant. - cross-sectional (multivariate) | Older vs. younger |
PTSD: older age versus non-elderly (OR 3.56, p = 0.002). General psychiatric morbidity: older age vs. non-elderly (OR 2.14, p = 0.005). |
| Johns Hopkins & Policy Studies, 2012 [40] [6/10] | IDPs in Georgia/war (60+) | Depression (GDS), anxiety (GAI)/quantitative - cross-sectional (bivariate) | Long-term(20 years) IDP vs. short-term (4 years) IDP |
Depression: Females higher depression scores than males (p < 0.01) (no mean depression score reported). Anxiety: short-term IDP higher prevalence than long-term IDP (76% vs. 70.3%, p < 0.02). Females higher scores than males (p < 0.01) (no mean scores reported). |
| Kohn et al., 2005 [41] [7/10] | Hondurashurricane (60+) | Psych. distress (SRQ); alcohol misuse (SRQ); depression (DSM-IV/ICD-10); PTSD (CIDI; IES)/quantitative - cross-sectional (multivariate) | Older age vs. younger |
Risk factors for all psychopathology (except severity of PTSD avoidance subscale) in older age: Exposure inventory (exposure to hurricane); prior “nerves” Risk factors for PTSD and psychological distress: living in high impact area Older age vs. non-elderly: none (no p-values or effect measures available). |
| Li et al., 2011 [42] [7/10] | China, earthquake (55+) | Depressive symptoms, stress reaction (Impact of Event Scale), sense of community (SoC Index)/quantitative – cross-sectional (multivariate) | None | High event impact, a reduced sense of community, and social support were associated (P < 0.05) with higher rates of depression. |
| Nomura et al., 2010 [44] [7/10] | Sri Lankatsunami (60+) | PTSD (IES-R)/quantitative - cross-sectional (multivariate*) | None | Increasing age by 10-year interval (coef. -0.27, p = 0.04) Loss of or injury to family members due to the tsunami (coef. 6.12, p < 0.001). |
| Prueksaritanond et al. 2007 [48] [2/10] | Thailand, tsunami (60+) | Depression (Zung Self-Rating Depression Scale)/Quantitative – cross-sectional /quantitative (descriptive) | None | Factors associated with increased symptoms of depression were female (odd ratio [OR] 2.81; 95% confidence interval [CI] 0.73–10.77, p = 0.12), aged of 65 years old and over (OR 2.0; 95% CI 0.52–7.7, p = 0.25), living alone such as single, divorce, or separation (OR 1.47; 95% CI 0.35–6.13, p = 0.44), no income was generated after the tsunami (OR 1.26; 95% CI 0.34–4.75, p = 0.5), hypertension (OR 1.25; 95% CI 0.34–4.59, p = 0.5) and loss of family members (OR 1.14; 95% CI 0.31–4.20, p = 0.56). |
| Strong et al., 2015 [52] [5/10] | Syrian & Palestinian refugees Lebanon/war (60+) | Negative emotions (SRQ)/ quantitative - cross-sectional (bivariate). | Palestinian (longer displaced) vs. Syrian refugees (shorter displaced)** | Palestinians (i.e. longer-term displaced): higher prevalence of depression (40% vs. 25%, p = 0.050)Palestinians (i.e. longer-term displaced): higher prevalence of feeling scared (33% vs. 18%, p = 0.036) Among entire sample (Palestinian and Syrian refugees combined): Older age (p = 0.017) and higher education (p = 0.023) with feeling depressed. Lower social support (p = 0.006) with anxiety. |
| Viswanath et al., 2012 [53] [5/10] | Indiatsunami (60+) | Psych. morbidity, adjustment disorder, PTSD, depressive episode, panic disorder, alcohol dependence, phobic disorder, anxiety (ICD-10 criteria)/ quantitative - cross-sectional (bivariate) | Older age vs. youngerDisplaced vs. non-displaced | Total sample: (older age vs. non-elderly): Older people more likely to suffer from adjustment disorder (50% vs. 37%, p = 0.030).Displaced (older age vs. non-elderly): Older people less likely to suffer from depressive episodes (6% vs. 20%, p = 0.019). Non-displaced: (older age vs. non-elderly): Older people less likely to suffer depressive episode (7% vs. 27%, p = 0.002). Elderly more likely to suffer PTSD (18% vs. 8%, p = 0.036). Within older age sample (older age vs. non-displaced): non-displaced more likely to suffer adjustment disorder (61% vs. 17%, p = 0.001). Displaced more likely to suffer depressive episode (44% vs. 7%, p < 0.001) and unspecified anxiety disorder (22% vs. 4%, p < 0.011). |
| Wu et al., 2015 [56] [5/10] |
China, flooding (60+) | ‘Health related quality of life’ (HRQoL), incl. Role limitations due to emotional problems, mental health/ quantitative – cross-sectional (multivariate) | Pre-flood rural older people (from National Health Services Survey 2008) | Self-reported HRQoL lower in those aged 80–99 (vs. 60–79), lower in those who are single (vs. married), lower in those with poor sleep patterns, lower in those with pre-existing chronic diseases, lower if hospitalised within the last year, lower if living alone (vs. with spouse), |
| Zhang et al., 2012[b] [58] [9/10] | China earthquake (60+) | PTSD (PCL-C); anxiety/dep. (HSCL-25)/ quantitative - cross-sectional (multivariate). | None |
PTSD: loss of livelihood (OR 3.06 [95% CI 1.30–7.21]); initial fear (OR 1.74 [95% CI 1.16–2.54]). Anxiety: female (OR 2.03 [95% CI 1.09–3.39]); bereavement (OR 2.59 [95% CI 1.17–5.77]); injury (OR 2.03 [95% CI 1.03–4.11]). Depression: Initial fear (OR 1.44 [95% CI 1.03–2.01]). |
| Zhang et al., 2012(c) (5/10] [59] | China earthquake (60+) | Quality of Life (QoL) score/quantitative – cross-sectional (bivariate) | National average | Lower QoL scores are associated (p < 0.5) with: female gender; age over 70; single; lower income; non-smoker; disability in self or family member, poor family relationship |
| Physical health outcomes | ||||
| Andre et al., 2013 [24] [3/10] | Rural Democratic Republic of Congo/war (65+) | Nutritional status (MNA-SF/LF)/quantitative - cross-sectional (bivariate) | None | Differences in nutritional status (normal vs. malnourished): mean age (years) 68.4 (+ − 4.0) vs. 74(+ − 6.7) (p < 0.001); BMI <18.5 15.7% vs. 81% (p < 0.001); smoking 31.4% vs. 2.9% (p < 0.001); physical exercise (1–5/week) 100% vs. 2.9% (p < 0.001); >3 prescription drugs/day 19.6% vs. 68.6 (p < 0.001); ADL limitation 50.9% vs. 87.6% (P < 0.001); IADL limitation 11.8% vs. 94.3% (P < 0.001); history of falls: 35.8% vs. 61% (p = 0.003) |
| Ardalan et al. 2011[a] [26] [7/10] | Iran earthquake (60+) | Physical Quality of life (QoL) (WHOQOL-BREF)/ quantitative - cross-sectional (multivariate) | Earthquake affected vs. non-earthquake affected | Within earthquake affected sample (regression coefficients): Higher age (−0.113) (p = 0.001); being female (−1.320) (p = 0.017)Being injured due to earthquake (−2.370) (p = 0.006); dissatisfaction with quality of current living place (−2.411) (p < 0.001); functional (ADL) dependence (−1.963) (p = 0.001) |
| Ardalen et al., 2011[b] [27] [7/10] | Iran earthquake (60+) | Functioning (ADL and IADL, 2 months, 2 years and 5 years after event)/quantitative - cross-sectional (multivariate) | None |
Determinants of functional capacity (regression coefficients, p < 0.05):
Model 1 (controlled for all eligible factors except ADL and IADL scores at preceding time period):1. ADL scores at 2 months after the earthquake: age (−0.60); living with others (−.68); and chronic diseases (−0.66). 2. ADL scores at 5 years after the earthquake: age (−0.64); gender (0.41); living with others (−0.86); and chronic diseases (−0.40) 3. IADL scores at 2 months and 5 years after the earthquake: age (−1.64 & -1.61); education (1.44 and −1.47); study area (1.21 and 1.12); living with others (−1.78 & -1.91). Model 2 (controlled for all eligible factors including ADL and IADL scores at preceding time period): 1.ADL score at two months after the earthquake: age (−0.42), living with others (−0.49), and ADL before the earthquake (0.81). 2. ADL score at 5 years after earthquake: gender (0.16), living with others (0.12), and ADL at 2 months after earthquake (1.03). 3. IADL score at 2 months after earthquake: age (−0.85); living with others (−0.80); and IADL before the earthquake (0.68). 4. IADL score at 5 years after earthquake: IADL at 2 months after the earthquake (0.99). |
| Arlappa et al., 2009 [28] [4/10] | Rural India/drought (60+) | Chronic Energy Deficiency (CED) and BMI/quantitative – cross-sectional (descriptive) | None |
Age (70+ vs. 60–69): higher CED in both genders (males: 59.2% vs. 47.5%; females: 56.6% vs. 45.8%, P < 0.001). Age (18–59 vs.60+): higher CED among older adults (males: 51.8% vs.38.1%; females: 48.5% vs. 40.5%,,p < 0.001). Socio-economic factors: Caste (scheduled caste and scheduled tribe 57% vs. Backward caste and others: 44.2%, (p < 0.001); pension (availing: 55.1%, not availing: 49.7% and not required: 45.8%, p < 0.05); occupation (non-agricultural: 53.4%, Agricultural: 52.1% and others 46.1%, (p < 0.001); total land acres (none: 49.3%, 0.01–2.5: 53.7%, 2.5–5: 52.3% and >5: 43.1%, p < 0.001). |
| Chan et al., 2009[a] [31] [4/10] | Pakistan earthquake (45+) | Dental, visual, eating, hearing, headaches,dizziness,muscle/ joint pain (all SRQ); health seeking behaviour/health access/quantitative - cross-sectional (bivariate). | Rural vs. urban |
Rural prevalence (%) compared to older people in urban areas: β Dental (100 vs. 25) (p < 0.0001); visual (75 vs. 38) (p < 0.0001); weight loss (75 vs. 50, p = 0.001); eating problem (87 vs. 50) (p = 0.002); hearing (54 vs. 40) (p = 0.043); headache (40 vs. 23, p = 0.043); having known medical problem for which never having had treatment (65 vs. 30, p < 0.001); having known medical problem with treatment discontinued (80 vs. 40, p < 0.001) Urban – higher prevalence (%) compared to older age in rural areas: β a known underlying medical problem (38 vs. 25, p = 0.02). |
| Godfrey & Kalache, 1989, [35] [3/10] | Ethiopian refugees in Sudan/war and famine (45+) | Mortality rates; prevalence of disability (SRQ)/quantitative - cross-sectional (descriptive) | None |
Age-specific mortality rates estimated since arriving in Sudan 1 year (using population estimate as denominator):
45–49 years: 5/1000 per year (N = 1); 50–59 years: 35/1000 per year (N = 5); 60+ years: 273/1000 per year (N = 3). Age-specific mortality rates for 2-year period prior to migration (denominator all those reported in Tigray households during this period): 45–49 years: 14/1000 per year (N = 5); 50–59 years: 41/1000 per year (N = 4); 60+ years: 91/1000 per year (N = 8). |
| Pieterse et al., 1998 [45] [6/10] | Rwandan refugees in Tanzania/war (60+) | BMI, AMA, AFA, MUAC/quantitative - cross-sectional (bivariate) | Older vs. younger |
Older higher prevalence of malnutrition (BMI < 18.5): Men (23.2% vs. 15.0%, p < 0.05); women (15.1% vs. 10.9%, p < 0.05) Older lower mean AMA (important in relation to ability to remain active and independent): Men 50–59, 60–69 and 70+ (34.7, 32.3, 30.9, respectively, p < 0.05); women 50–59, 60–69 and 70+ (35.1, 33.0, 31.5, respectively, P < 0.05). |
| Pieterse et al., 2002 [46] [8/10] | Rwandan refugees in Tanzania/ war (50+) | Handgrip strength/quantitative - cross-sectional (multivariate) | Older vs. younger | Men: BMI contributes 5.7% to variation in Handgrip strength. (coef 0.262, p < 0.001); AMA contributes 10.2% to variation in handgrip strength (coef 0.303, p < 0.001).Women: BMI contributes 3.5% to variation in handgrip strength (coef 0.188, p < 0.001); AMA contributes 2.8% to variation in handgrip strength (coef 0.153, p < 0.001). |
| Pieterse & Ismail, 2003 [47] [4/10] | Rwandan refugees in Tanzania/war (50+) | Perceptions of nutritional risk factors by older persons/ qualitative interviews (ranking methodology) | None (qualitative) | Older people’s perceptions of main problems of the less well-off were: physical impairment; no purchasing power, income, tools and utensils; no people to provide assistance and moral support, social isolation. Older people’s perceptions of who were the most vulnerable: widows and widowers; physically impaired and disabled; those living alone, have no children living nearby, have care-giving responsibilities (for example for young children or old spouse). |
| Ramji &Thoner, 1991 [49] [2/10] | Displaced in Moz- mbique/Zimbab-we/war (45+) | BMI/quantitative - cross-sectional (descriptive) | None | Older women in Mozambique (displaced on average 6 months) had a mean BMI significantly lower than older Mozambique women displaced to Zimbabwe (displaced on average 2 years). BMI 17.3 vs. 21.1 (p < 0.001) |
| Sibai et al., 2001 [50] [9/9] | Lebanon, war (50+) | Mortality/quantitative – cohort (multivariate) | Participants in a 1983 community-based health survey | Women exposed to human losses had a significant excess risk of both CVD and total mortality (RR 3.37 and RR 2.28 respectively). Exposure to property losses carried a greater mortality risk for men. Positive trend in the rate ratios for mortality endpoints with an increase in the intensity of exposure to a cumulative number of war events. |
| Sibai et al., 2007 [51] [8/9] | Lebanon/war (50+) | All-cause mortality, cardiovascular mortality (ICD-9)/ quantitative - cohort (multivariate) | None |
Cardiovascular mortality: Men: unmarried (RR 2.50 [95% CI 1.28–4.89]); living with ≥3 generations (RR 1.99 [95% CI 1.32–3.00]); living with married child (RR 1.63 [95% CI 1.03–2.57]). Women: none All-cause mortality: Men: widowed/divorced/separated (RR 1.63 [95% CI 1.06–2.52]); living with ≥3 generations (RR 1.56 [95% CI 1.12–2.15]); living with married child (RR 1.70 [95%CI 1.19–2.43]).Women: living with married child (RR1.55[95%CI1.04–2.32]). |
| Strong et al., 2015 [52] [5/10] | Syrian & Palestinian refugees inLebanon/war (60+) | Negative emotions (using SRQ), and functional status (Katz Index of Independence in Activities of Daily Living/ quantitative - cross-sectional (bivariate). | Palestinian (i.e. longer-term displaced) vs. Syrian refugees (i.e. shorter-term displaced)** | Palestinians (i.e. longer-term displaced) higher prevalence of: hypertension (86% vs. 53%, p < 0.001); diabetes (81% vs. 38%, p < 0.001); eye disease (28% vs.16%,p 0.002); lung disease (44% vs.11%,p < 0.001); digestive tract disease (23% vs. 9%, p0.010); difficulty walking (65% vs. 39%,p 0.002); impaired vision (70% vs. 13%, p < 0.001); impaired hearing (49% vs. 8%, p < 0.001) Palestinians (i.e. longer-term displaced) lower prevalence of Arthritis, injury or back pain (7% vs. 31% p 0.007); Among entire sample (Palestinian and Syrian refugees combined): Older age (p = 0.002) and larger household size (p = 0.003) with worse functional status. Older age and lower educational status with worse self0reported health status. β |
| Wen et al., 2010 [54] [7/10] | China, earthquake (65+) | Mortality, physical injury/quantitative – cross-sectional (descriptive) | All earthquake-affected patients | Extremities the most common location of trauma in older patient admitted to hospital. Mortality significantly higher in this age group - secondary to e.g. thoracic visceral and craniocerebral injuries. Admission of older age patients peaking on the third, fifth, and eighth days. β |
| Wong et al., 2015 [55] [5/9] | 21 crisis-affected countries (conflict & natural disasters) | Intra-operative mortality & surgical procedure types – retrospective cohort of routine data from 93,385 operative cases (11% were older people) at MSF facilities, June 2008 to Dec 2012 (descriptive) | Younger populations (<50) from same crisis-affected populations | Intra-operative mortality increased with each age stratum from 60 years onwards |
| Wu et al., 2015 [56] [5/10] | China, flooding (60+) | HRQoL: physical functioning, role limitations due to physical illness, bodily pain, general health perceptions, vitality, social functioning/quantitative – cross-sectional (multivariate). | Pre-flood rural older age (National Health Services Survey 2008) | Self-reported physical health lower in those aged 80–99 (vs. 60–79), lower in those who are single (vs. married), lower in those with poor sleep patterns, lower in those with pre-existing chronic diseases, lower if hospitalised within the last year, lower if living alone (vs. with spouse), lower in those with illness in the last two weeks, lower in females |
| Zhang et al., 2012[a] [57] [7/10] | China earthquake (65+) | Clinical features and outcomes of crush patients with acute kidney injury, mortality rate/ quantitative - cross-sectional using medical records (multivariate) | Older age vs. younger |
Clinical and lab findings: (older age vs. younger): higher systolic pressure (131.9 vs. 115.2, p = 0.001); lower incidence of oliguria (13.2% vs. 41.0%, p = 0.001) ; lower creatinine (220.4 vs. 352.6, p = 0.001); lower potassium (4.1 vs. 5.3, (p < 0.001); lower serum phosphorus (1.2 vs. 1.9, (p < 0.001); lower creatinine (8173.6 vs. 57,423.0, p = 0.001). Trauma events and medical complications (older age vs. younger): lower % of extremity crush injury (71.1% vs. 88.6%, p = 0.004); higher proportion of thoracic trauma (35.6% vs. 18.7%, p = 0.016); higher proportion of extremity fracture (42.2% vs. 19.9%, p = 0.002); higher proportion of rib fractures (26.7% vs. 7.2%, p = 0.002); higher proportion of vertebral fractures (17.8% vs. 6.7%, p = 0.020); higher proportion of pneumonia (42.2% vs. 25.9%, p = 0.035) Risk factors for death: older people receiving dialysis had higher mortality rate compared to younger adults (62.5% vs. 10.5%, p < 0.001). Risk factors for death in older people (controlling for BP, no of injuries, ISS, thoracic trauma, ARDS, sepsis/or dialysis): dialysis (OR 15.14, p = 0.011); sepsis (OR 13.24, p = 0.030). |
For studies using both bivariate and multivariate analysis only multivariate factors were extracted. Only significant associations were extracted (p < 0.05) for studies that conducted statistical tests
β Did not include tests for statistical significance
*For detailed results on quality assessment, please email corresponding author
**Length of displacement is an assumption by review authors based on history of Palestinian and Syrian displacement and not explicitly reported by study authors (Strong et al., 2015)