Table 3.

Summary of Meta-analyses on social support and health/disease using risk and odds ratio 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Meta-analysis	Published/Unpublished studies	Participants + countries-continents	Design	Social support measures	Outcome measures	Moderators/Confounders	Heterogeneity/Publication Bias	Effect size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.]
Gilbert et al. (2013)	39 (10 with social support measures)	3,685,137 USA + international	Cross-sectional	Combined: General index of social support	Health: Self-reported health (+)	Level of analysis: (individual, neighborhood, community) Country	Yes/Yes	Odds ratio	1.30 [1.13 150]
Kuiper et al. (2015)	19	24,966 4 continents	Longitudinal	Structural: Frequency of contact Social participation Functional: Loneliness Satisfaction with network	Disease: Incidence of Dementia or Incidence of Alzheimer Disease (AD)	Age, depression, alcohol Baseline cognition Physical activity functional disability, chronic disease Follow-up time, AD vs. dementia	Yes/Yes	Risk ratio	Frequency of contact: 1.57 [1.32 1.85] Social participation: 1.41 [1.13 1.75] Loneliness: 1.58 [1.19 2.09] Satisfaction social network: NS
Valtorta et al. (2016)	19	181,006 4 continents	Longitudinal	Structural: Social isolation Functional: Loneliness	Disease: Incidence of Coronary Heart Disease (CHD) or Stroke	Age, gender, SES Internal validity Small-study effect	Yes/Yes	Risk ratio	CHD: 1.29 [1.04 1.59] Stroke: 1.32 [1.04 1.68]
Penninkilampi et al. (2018)	33	2,370,452 3 continents	Longitudinal + case-control	Combined: Good social engagement Poor social engagement Functional: Loneliness	Disease: Incidence of Dementia	Age, gender, education Depression, physical activity Study quality, follow-up time Country	Yes/Yes	Risk ratio	Good social engagement:: 1.23 [1.14 1.35]* Poor social engagement: 1.41 [1.21 1.65] Loneliness: NS
Heerde and Hemphill (2018)	52	126,939 4 continents	Longitudinal + cross-sectional	Functional: (Actions by others in support of a distressed individual)	Disease: Youth (aged 12–19) Internalizing behaviors Externalizing behaviors Substance use Educational outcome Bullying victimization	Age, gender, source of support Study design, sample size Country	Yes/Yes	Odds ration	Overall: 1.30 [1.20 1.41]* Internalizing: 1.75 [1.47 2.08]* Externalizing: NS Substance use: 1.35 [1.19 1.54]* Educational outcome: 1.42 [1.29 1.57] Bullying: 1.47 [1.32 1.67]* Victimization: 1.82 [1.64 2.0]*
Wen et al. (2020)	7	1,248 Asia: China, India, Nepal Africa: South Africa, Kenia South America: Ecuador, Perú	Group comparison: Support Intervention vs. No intervention	Functional: Material Informational Emotional companionship (integrated measure)	Disease: Drug resistant tuberculosis Treatment success (cured + treat completed) Lost to follow-up (treatment interruption)	Type of outcome Type of social support	Yes/Yes	Odds ratio	Integrated measure: Treatment success: 2.58 [1.80 3.69] Reduction of lost to follow-up: 5.88 [1.82 20]* Without material support: NS
Desta et al. (2021)	13 (4 with social support measures)	1,647 Ethiopia	Cross-sectional	Combined: General index of social support	Disease: Post-partum depression	History of depression Marital status satisfaction Partner violence Substance abuse Quality of study Region (within Ethiopia)	Yes/Yes	Odds ratio	6.27 [4.83 8.13]

^*Inverted ratio; NS, not significant; SES, socio economic status; AD, Alzheimer disease; CHD, coronary heart disease.