Table 4.

Findings from included cohort studies on the association between social connectedness/engagement and health outcomes among older adults living in congregate long term care settings

(Author, Year)	Study purpose	Measure of social connectedness	Measures of health outcomes	Covariates	Major findings
(Chau, 2021)	Examine key risk factors for depression among older adults in long-term care	Multidimensional scale of perceived social support (MSPSP) (Zimet, Dahlem, Zimet, & Farley, 1988; Zimet, Powell, Farley, Werkman, & Berkoff, 1990)	Geriatric Depression Scale (GDS-15) (Yesavage et al., 1982) M.I.N.I. International Neuropsychiatric Interview (D. Sheehan, 2015; D. V. Sheehan et al., 1997)	Age, gender, education, marital status, and other risk factors (pain, sleep, cognitive impairment, physical function, person-environment fit)	As social support score increased depression score decreased.
(Cohen-Mansfield et al., 1999)	Examine factors associated with mortality	Hebrew Home Social Network Rating Scale (Cohen-Mansfield & Marx, 1992)	Date and cause of death abstracted from death certificates	Age, gender, hearing, surgery in past 2 years, number of medical diagnoses, cognitive function, activities of daily living, appetite, falls, incontinence, sleep aggressive behaviors, screaming	Size and quality of social network were not associated with time to death.
(Fehnel et al., 2015)	Identify factors that predict hospital readmission and mortality after ischemic stroke	Minimum Data Set (MDS) Index of Social Engagement (Mor et al., 1995)	Date of death (whether from Medicare files or MDS not specified) Hospital re-admissions from Medicare Part A	Age, race, sex, education level, non-native English speaker, active tobacco use, do not resuscitate order, do not hospitalize order, hospice care, guardianship status, hospital length of stay, bowel and bladder incontinence, urinary catheter, body mass index, marital status, co-morbid conditions, gastric feeding tube, tracheostomy, activities of daily living, cognitive impairment, communication scale, pressure ulcer, CHESS (changes in health, end-stage disease, symptoms, and signs score) index, Elixhauser score.	Relative to residents with a social engagement score of zero (not engaged), each social engagement score from 1 to 6 (very engaged) was inversely associated with 30-day readmission or death.
(Foebel et al., 2015)	Estimate factors associated with new antipsychotic use in residents newly admitted to long term care	Reduced social engagement measure derived from MDS 2.0 items	Initiation of an antipsychotic as documented on the MDS 2.0	Age, cognitive impairment, physical function, communication, delirium, comorbid conditions, hearing, vision, pain, delusions, motor agitation, depressive symptoms, urinary tract infection, # of medications, other psychoactive medications, restraints, hospital stay, emergency room visits, days between assessments Stratified by sex	61.2% had reduced social engagement. In women, reduced social engagement was associated with initiation of an antipsychotic, but not in men.
(Foster & Cataldo, 1993)	Evaluate the extent to which the onset of depression can be predicted in a long-term care facility	‘How many friends do you have here at the hospital? By friend, I mean someone who you talk to regularly, not just a casual acquaintance.’ (self-reported each month)	Monthly evaluations by a research psychiatrist (≥15 on 21-item Hamilton Depression Rating Scale [Hamilton, 1960] and ≥10 on GDS-15 [Yesavage et al., 1982]) or using DSM-III criteria for affective disorders	At admission: coping, GDS-15, Life satisfaction scale, non-affective psychiatric symptoms, activities of daily living Monthly after admission: Health status, Zung Self-Rating Depression Scale, activities of daily living, non-affective psychiatric symptoms	Mean monthly number of friends for non-depressed patients: 2.4±2.5 Mean monthly number of friends for depressed patients: 1.6±1.3 Number of friends included in the final model and considered an important factor in predicting depression onset.
(Freeman et al., 2016)	Examine the role of physical restraint use, use of antipsychotic medications, and engagement in social activities and cognitive decline	MDS 2.0 Index of Social Engagement (Mor et al., 1995)	Cognitive Performance Scale (John N. Morris et al., 1994) (range: 0 [intact] – 6 [severe impairment])	Sex, age group (years), dementia, physical restraint use, prescribed antipsychotic medication	18.2% had no-low social engagement among those with cognitive decline at follow-up. Moderate-high social engagement scores reduced risk of cognitive decline at follow up.
(Grabowski & Mitchell, 2010)	Examine the association between private oversight and the quality of end-of-life care for residents with advanced dementia	# of hours per week residents had visits by their health care proxies (HCP) (no visit, 1-7 hours/week of visits, >7 hours/week of visits)	Pressure ulcer, pain, dyspnea, hospice use, hospital transfer, do-not-hospitalize (DNH) order, satisfaction with care at the end-of-life in dementia, antipsychotic use	Characteristics of residents: Sex, age, ethnicity, general health and disease severity by Bedford Alzheimer Nursing Scale Characteristics of proxy: Sex, child of nursing home resident or not, age (years), education	Mean number of residents without weekly HCP visitation: 0.12±0.32 Residents with >7 hours/week of visits had increased risk of pressure ulcer, pain, dyspnea, and less likely to have a DNH order.
(Hjaltadottir et al., 2011)	Identify factors associated with 3-year mortality	MDS 2.0 Index of Social Engagement (Mor et al., 1995)	National death registry	Age, gender, activities of daily living, changes in health scale, location admitted to nursing home from	Relative to residents with the highest level of social engagement, each other level of social engagement was associated with a faster time to death.
(Inventor et al., 2018)	Examine which factors influence positive and negative behaviors among people with Alzheimer’s disease	Resident activities: solitary, care-related, family/friend visits, structured, unstructured, large and small (2-5 people)	Positive behaviors: smiling, positive physical expressions, engaging in conversation Negative behaviors: grimacing, verbal aggression, physical aggression, physical non-aggression, repetitive behaviors	Age, gender, race, cognitive function, physical function, special care dementia unit, # of psychotropic medications and measure of social environment	Solitary activities were associated with fewer episodes of positive behaviors.
		Derived from 60 direct 5-minute observations of each resident over 12 consecutive weekdays from 9 am to 9 pm
(Kiely et al., 2000)	Examine the effect of social engagement on mortality in long term care	MDS 2.0 Index of Social Engagement (Mor et al., 1995)	Time to death based on facility records or last available MDS assessment	Age, gender, limitations in activities of daily living, body mass index, weight loss, heart failure, shortness of breath	23.1% lacked any social engagement. Residents with higher social engagement had delayed time to death. Residents with no social engagement died faster than residents with maximum social engagement scores.
(Kiely & Flacker, 2003)	Examine the effect of social engagement on mortality in long term care	MDS 2.0 Index of Social Engagement (Mor et al., 1995)	Survival time was time between baseline MDS assessment and date of death (from the National Death Index)	Age, gender, activities of daily living, body mass index, recent weight loss, unstable condition, shortness of breath, asthma/ emphysema/chronic obstructive pulmonary disease, diabetes, heart failure, pressure ulcer, cancer, leaves food uneaten, swallowing problem	Among long-stay residents, 6.6% lacked any social engagement and 9.9% had the maximum score on the social engagement index. Residents with higher social engagement had delayed time to death and those without died faster than those with highest social engagement scores.
(Kloos et al., 2019)	Quantify the association between the satisfaction of autonomy, relatedness, and competence with depression and well-being among nursing home residents	8-item relatedness subscale embedded within the 21-item Basic Need Satisfaction in Life Scale (Gagné, 2003)	Measured 5-8 months after baseline: 8-item Dutch version of the GDS (Jongenelis et al., 2007) Dutch version (Custers, Westerhof, Kuin, & Riksen-Walraven, 2010) of the 5-item Satisfaction With Life Scale (Diener, Emmons, Larsen, & Griffin, 1985)	Autonomy and competence subscales of the Basic Need Satisfaction in Life Scale	The mean relatedness score at baseline was 4.27± 0.58. Relatedness was associated with GDS, but not Satisfaction With Life Scale.
(Kroemeke & Gruszczynska, 2016)	Examine the moderating effect of institutional care on the association between support and subjective well-being	Berlin Social Support Scales (Schulz & Schwarzer, 2003) (Received and provided support)	Center for Epidemiological Studies Depression Scale (CES-D) (Radloff, 1977) 3 items from the PANAS (Watson, Clark, & Tellegen, 1988)	Baseline level of CES-D and positive affect, instrumental activities of daily living, and subjective health	Received support was associated with positive affect. Providing and receiving support was not associated with positive affect.
(Lou et al., 2012)	Examine the association between social engagement and depressive symptoms	MDS 2.0 Index of Social Engagement (Mor et al., 1995)	7-item Depression Rating Scale (J.N. Morris, Murphy, & Nonemaker, 1995)	Age, gender, education, marital status, familial financial support, activities of daily living, cognition, pain	At baseline, the mean index of social engagement was 2.13±1.81. A 1-point increase on the social engagement scale over 1 year resulted in a 0.11-point decrease in the depressive symptoms score with the effect of the quadratic change following a similar pattern.
(Mao & Han, 2018)	Examine the associations between living arrangements and psychological well-being and life satisfaction, and the mediating role of social support	Types of social support (financial, instrumental, emotional) and sources of social support (spouse, children, informal (non-family), formal) Living arrangement	7-item psychological well-being score (higher scores better) Life satisfaction: “How do you rate your life at present?” Range: 1 (very good) to 5 (very bad); categorized as good/very good versus other	Age, gender, education, marital status, number of children, number children living nearby, and physical and cognitive functions Stratified by rural/urban	Urban: Compared with living with children, living in nursing homes had small positive direct (and total) effects on life satisfaction, with higher life satisfaction through formal financial support and lower life satisfaction due to not receiving instrumental support from a spouse. Rural: Compared with living with children, nursing homes residents received more financial, instrumental, and emotional support from formal sources and formal financial support (rather than from children) which was associated with better psychological well-being.
(Maxwell et al., 2015)	Determine factors associated with hospitalization over 1 year for residents with dementia living in long term care or assisted living	Summary score of strength of social relationships (0-2-low/none, 3-5 moderate/high) Average time in activities when awake	Time to first event (i.e., hospitalization, admission to long term care, death, transfer to other setting) or 1-year	Age, gender, cognitive impairment, health instability, fatigue, number of medications, previous hospitalizations <1 year, bowel incontinence, rural/urban	Among those in assisted living, 20.5% scored low/none on social relationship strength. Residents with no social relationships were hospitalized/died more quickly than those with moderate/high social relationships. Among those in long term care, 38.2% scored low/none on strength of social relationships. No estimates of hazard ratios provided.
(E. Neufeld et al., 2019)	Investigate factors associated with symptoms of depression	MDS 2.0 items Social connectedness /support (binary) Involved in social activities (binary)	7-item Depression Rating Scale (≥3 probable depression) (Burrows, Morris, Simon, Hirdes, & Phillips, 2000; Eva Neufeld, Hirdes, Perlman, & Rabinowitz, 2015)	Age, gender, marital status, limitations in activities of daily living, cognitive impairment, pain, falls, baseline depression	23.5% had social activities at baseline. Residents with social connectedness were less likely to develop indicators of probable depression.
(Palese et al., 2019)	Identify predictors of self-feeding dependence	Social Engagement Scale and Relationship with Others Scale (Pascazio, 2009)	Self-feeding dependence progression over time measured on an ordinal scale	Age, sex, activities of daily living, cognitive performance scale, depression, pain, behaviors, clinical instability, comorbidities, pressure sores, time order, and facility factors	Social engagement deprivation increased the risk of self-feeding dependence. Engagement with family, health-care workers, or volunteers was protective against the progression of self-feeding dependence.
(Pastor-Barriuso et al., 2020)	Estimate the effect of social engagement on long-term all-cause mortality	To what extent do you/does the resident interact with other residents in the institution?	Date of death from facilities and the Spanish National Death Index	Age, gender, education, marital status, number of chronic conditions, functional dependency, length of stay, assigned caregiver, frequency of external visits, facility ownership, facility size	36.0% of residents had low/no social engagement. No differences between those with moderate levels of social engagement and low/no social engagement at 2, 5, or 10 years of follow up. Standardized mortality risk differences at 5 years were lower for those with high social engagement, but not at 2 or 10 years.
(Rozzini et al., 1996)	Identify predictors of mood changes over a 12-month period	Nursing home activities scale (scores 0 [none] to 32 [high level]) (McCaffree & Harkins, 1975) based on observed behaviors	30-item GDS (Brink et al., 1982)	Age, gender, baseline GDS, baseline measures and changes in: cognitive function, satisfaction with the nursing home environment, Barthel index, chronic symptoms, pain, diseases, disabling diseases, drugs, and self-evaluation score	The mean activity score was 5.0 ± 5.4. Neither baseline activities score or changes in activities score was associated with changes in GDS score.
(Vetrano et al., 2018)	Investigate the association between determinants of health and survival	InterRAI LTCF Social participation and family visits (never, ≤3 days, >3 days ago)	InterRAI LTCF death from any cause and date of death	Age, sex, cognitive impairment, functional impairment, body mass index, physical activity, vaccines, and number of diseases	Involvement in social activities and family visits were associated with delayed time to death.
(Y. H. Wang et al., 2019)	Evaluate factors associated with resident outcomes	Interview Schedule for Social Interaction (C. Yeh, L, 1999) Frequency of family visits	GDS-15 (0–5 good, 6–9 melancholy, ≥10 depression) (Luo, Kuo, & Tang, 2012) EuroQol-5D self-reported health status	Gender, age, baseline data collection, mobility, cognitive function, comorbidities, intervention type and facilities random effects	The frequency of family visits was high (42%) and social support from facility staff was also high (24.66 ± 4.64). Support from facility staff was associated with an increase in utility of quality of life, an increase in self-reported health, and a decrease in depression.
(Weiner et al., 1998)	Examine factors associated with pain scales	Quality of residents’ proximate social network (Cohen-Mansfield & Marx, 1992)	Pain scale Pain thermometer	Age, gender, race, education, marital status, analgesics, cognitive impairment, comorbidity, depression, physical function, self-rated health	Quality of residents’ proximate social network was not associated with pain measures.
(Yamada et al., 2016)	Examine whether dual sensory impairment is associated with greater cognitive decline	7 MDS items comparable to Index of Social Engagement (Mor et al., 1995)	Cognitive Performance Scale (John N. Morris et al., 1994) Changes in CPS over time (baseline, 6 months, 12 months)	Age, gender, facility, baseline information on primary diagnosis of Alzheimer’s disease, other dementia, and psychiatric diseases, depressive symptoms, cognitive and functional state, communication problem, and use of sensory assistive device	15.7% had no behaviors indicative of social engagement. Fewer numbers of behaviors indicating social engagement observed were associated with a greater decrease in cognitive function. Concurrent vision and hearing impairment was not associated with a greater cognitive decline when > one behavior indicating social engagement was present.
(K.-P. Yeh et al., 2014)	Evaluate factors associated with rapid functional decline and 12-month mortality	MDS 2.0 Index of Social Engagement (Mor et al., 1995)	Recent activities of daily living decline defined as a 1-point increase in RUG-III activities of daily living score in the past 6 months and 1-year mortality	Age, education, body mass index, comorbidities, pain, cognitive impairment, baseline activities of daily living, RAP triggers	Social engagement was not associated with recent functional decline.