Table 2: Characteristics of studies of social network structures and late-life depression in low- and middle-income countries: citations are repeated if study investigated more than one social network structure typologies.

Community Based Cross-Sectional Surveys
References Sample Size Social Network Typology Outcome Measure Measure of Effect Effect Size (95% C.I)a /Standard Errors Effect (YES/NO)
Aung et al. 2016 435 Size
(Social Network Index)
30-item GDSb Beta/Linear regression Larger size,
-0.18 (-0.28, -0.09)
Protective
Singh et al. 2016 630 Composition:
children, relatives, friends, confidant networks
WHOc Composite International Diagnostic Interview O.Rd/Logistic regression Children, 1.32 (0.79–2.21)
Friends, 0.33 (0.18–0.61)
Relatives, 0.70 (0.39–1.27)
Confidant, 0.549 (0.17–1.75)
Protective if composed of friends
Chen et al. 2016 400 Composition:
family, relatives, friends, health care networks
15-Item GDSb Beta/Linear regression Family, -0.46/0.43
Friends, -1.1/0.55
Health and social carers, 1.17/0.51
Protective if friends,
Predictive if carers
Kim et al. 2015 147 Interaction
(Lubben’s Scale)
15-item GDSb Beta/linear regression Greater connectedness,
-0.02/0.28
NO
Caetano et al. 2013 3649 Interaction:
Reciprocal visits
Self-report of clinician diagnoses O.Rd/Logistic regression No visits,
1.38 (0.83 – 2.29)
NO
Tong et al. 2011 228 Interaction:
Meetings/telephone
15-Item GDSb Beta/Linear regression Telephone, −.083/8.30
Meeting with someone, −.027/2.7
NO
Wahlin et al. 2015 471 Interaction:
Satisfaction with contact
Self Reporting Questionnaire Prevalence rate ratio Fully satisfied with contact,
0.61 (0.46–0.80)
Protective
Sicotte et al. 2008 1905 Geographical Proximity:
Co-resident with children or not
15-Item GDSb O.R/Logistic regression Co-resident children
Women, 0.43, (0.25–0.74)
Men, 0.29, (0.07–1.16)
Non co-resident children
Women, 0.70, (0.44–1.11)
Men, 0.94, (0.40–2.20)
Protective in women with co-resident child
Wahlin et al. 2015 471 Geographical proximity: Co-residency Self Reporting Questionnaire Prevalence rate ratio Co-resident child, 0.78 (0.66–0.94)
No co-resident children, 1.28 (1.02–1.60)
Protective: co-resident Predictive: if not.
Monserud et al. 2015 8704 Geographical proximity:
Co-residency with own children/other relative/friends/network lives in the neighborhood
Modified CES-De 20 Beta/linear regression Co-resident children, 0.18/0.21,
Co-resident other relatives, 0.31/0.17,
Children nearby, 0.05/ 0.10,
Relatives nearby, -0.14/0.10,
Friends nearby, -0.04/ 0.11
Predictive if co-resident with other relatives (i.e., Other than their own children)
Webster 2014 195 Geographical proximity:
Network lives far away
CES-De 11 Beta/linear regression Distant family,
0.43/2.24,
NO
Webster 2014 195 Emotional proximity:
Hierarchical mapping
CES-De 11 Beta/linear regression Distant family, 2.57/1.78,
Close family, −2.31/1.61
NO
Cao, et al. 2015 928 Social standing:
Occupational Prestige Scale
30-Item GDSb Beta/linear regression Prestigious,
-0.07/0.34
Protective
Thiyagarajan, et al. 2014 17,031 Diversity:
PANTf assessment
Geriatric Mental State Examination Prevalence ratio Low diversity,
1.20 (1.08 – 1.34)
Predictive
Sicotte et al. 2008 1905 Diversity:
Types of relationship
15-Item GDSb O.R/Logistic regression High diversity
Women,0.34, (0.17–0.69)
Men,0.14, (0.04–0.48)
Protective
Chan et al. 2009 1042 Social Network Scale (aggregate scores) 15-Item GDSb O.R/Logistic regression Lower social network scores,
3.63, (2.35–5.60)
Predictive
COMMUNITY BASED OBSERVATIONAL STUDIES
Monserud et al. 2015 8704 Geographical proximity Modified CES-De 20 Beta/linear regression Co-resident children, 0.20/0.23,
Co-resident others, -0.31/0.18,
Children nearby, -0.01/0.14,
Relatives nearby, 0.25/0.13,
Friends nearby, 0.24/0.13
Symptomatic improvement if co-resident with other relatives
Gureje et al. 2011 2149 Interaction
(frequency of contacts)
WHOc Composite International Diagnostic Interview Hazard ratios/Cox regression No Family contacts
women, 2.4, (0.3–20.3)
men, 2.5, (0.4–16.0)
No friends contact
Women, 2.1 (1.1–3.7)
Men, 1.5 (0.7–3.4)
Predictive of onset in women with no friends contact

aFinal Adjusted Model, bGeriatric Depression Scale, cWorld Health Organization, d Odds Ratios, eCenter for Epidemiological Studies Depression Scale, fPractical Assessment of Network Types.