Table 2.
Summary of Meta-analyses on social support and mortality using risk, odds, and hazard ratios as effect size measure.
| Meta-analysis | Published/Unpublished studies | Participants + countries-continents | Design | Social support measures | Outcome measures | Moderators/Confounders | Heterogeneity/Publication bias | Effect size measure | Summary effect size [95% C.I.] |
|---|---|---|---|---|---|---|---|---|---|
| Pinquart and Duberstein (2010) | 87 | 10,795,137 4 continents |
Longitudinal | Functional: Perceived Structural: Network size, marital status |
Mortality: Cancer patients |
Age, gender, SES Comorbidity, smoking, alcohol Cancer site, cancer stages Time of measures, length of study |
Yes/Yes | Risk ratio | Perceived support: 1.22 [1.12 1.33]* Network size: 1.25 [1.14 1.39]* Marital status: 1.15 [1.11 1.19]* |
| Holt-Lunstad et al. (2010) | 148 | 308,849 4 continents |
Longitudinal + cross-sectional | Functional: received, perceived, loneliness Structural: marital status, social network, integration, living alone Isolation, complex measure Combined |
Mortality: All causes |
Age, gender, health status Type of support, cause of mortality Length follow-up, country |
Yes/Yes | Odds ratio | 1.50 [1.42 1.59] Functional: Odds ratio range: 1.22–1.45 Structural: Odds ratio range: 1.19–1.91 Combined: 1.47 [1.34 1.60] |
| Barth et al. (2010) | 20 | 34,292 5 countries 2 continents |
Longitudinal | Functional: Perceived Structural: Living alone |
Mortality: Coronary Heart Disease (CHD) + all causes |
Age, SES, smoking Metabolic factors Myocardial function factor |
Yes/Yes | Risk ratio or hazard ratio | Functional: 1.59 [1.21 2.08] CHD + all causes Structural: 1.41 [1.17 1.70] all causes Structural: 1.56 [0.94 2.58] NS CHD |
| Sbarra et al. (2011) | 32 | 6.5 million 3 continents |
Longitudinal | Structural: Divorced-separated |
Mortality: All causes |
Age, gender Length follow-up Number of covariates Country |
Yes/Yes | Risk ratio or hazard ratio | 1.23 [1.17 1.30] Men: 1.31 [1.23 1.39], Women: 1.18 [1.10 1.27] Age: Younger than 65: 1.31 [1.23 1.39] Older than 65: 1.13 [1.05 1.23] |
| Roelfs et al. (2011) | 94/1 | 500 million 4 continents |
Longitudinal + cross-sectional | Structural: Singles (never married) |
Mortality: All causes |
Age, gender Baseline at start time Sample size Study quality Country |
Yes/Yes | Hazard ratio | 1.30 [1.24 1.37] Men: 1.32 [1.23 1.41] Women: 1.23 [1.14 1.32] Gender difference diminishes with time Age: Ratio decreasing at older age from 2.28 to 1.22 |
| Roelfs et al. (2012) | 120/4 | 500 million 4 continents |
Longitudinal + cross-sectional | Structural: Widowed |
Mortality: All causes |
Age, gender, health problems Time elapse from widowed Sample size, study quality Country |
Yes/Yes | Hazard ratio | 1.20 [1.16 1.25] Men: 1.27 [1.19 1.35], Women: 1.15 [1.08 1.22] Age: ratio decreasing at older age from 1.95 to 1.19 Gender difference diminishes with age (faster for men) |
| Shor et al. (2012) | 102/2 | 600 million 5 continents |
Longitudinal + cross-sectional | Structural: Divorced-separated |
Mortality: All causes |
Age, gender, SES, health status Life style, life stress, design Length follow-up, Recency div-sep Study quality, country |
Yes/Yes | Hazard ratio | 1.30 [1.23 1.37] Men: 1.37 [1.27 1.49], Women: 1.22 [1.13 1.32] Age: ratio decreasing at older age from 1.55 to 1.22 Gender difference diminishes with age (faster for men) |
| Shor et al. (2013) | 50 | 100,000 4 continents |
Longitudinal + cross-sectional | Functional: Perceived strength of family ties |
Mortality: All causes |
Age, gender, health status Source of support (family, friends, others) Degree of support, study quality Country |
Yes/Yes | Hazard ratio | 1.11 [1.05 1.17] Men: 1.09 [1.00 1.19], Women: 1.13 [1.02 1.25] Family: 1.15 [1.04 1.27], Friends: 0.99 [0.84 1.17] NS Others: 1.15 [1.07 1.24] |
| Shor and Roelfs (2015) | 91 | 400,000 3 continents |
Longitudinal + cross-sectional | Structural: Social contact frequency |
Mortality: All causes |
Age, gender, health status Source of contact Degree of support, study quality |
Yes/Yes | Hazard ratio | 1.13 [1.09 1.17] Women: 1.14 [1.04 1.25], Men: NS Non-family members: 1.07 [1.01 1.13] Family members: NS |
| Holt-Lunstad et al. (2015) | 70 | 3,407,134 4 continents |
Longitudinal | Structural: Isolation Living alone Functional: loneliness |
Mortality: All causes |
Age, gender, SES, health status Depression, physical activity, smoking Cause of mortality, year initial data Length follow-up |
Yes/Yes | Odds ratio | 1.30 [1.16 1.46] Isolation: 1.29 [1.06 1.56] Living alone: 1.32 [1.14 1.53] Loneliness: 1.26 [1.04 1.53] Age:: <65 OR = 1.57, 65–75 OR = 1.25, >75 OR = 1.14 |
*
Inverted ratio; SES, socio economic status; NS, not significant.