Table 3.
Study Characteristics and reported utility weights by level of loneliness from extracted studies
| Author, Country | Population characteristics | Study design (Sample size) | Reported health conditions | Utility instrument (tariff used) | Loneliness measure | Overall population mean loneliness score (SD) mean utility (U) score (SD) | Mean utility score (U) for those ‘not lonely’, ‘lonely’ and by levels of loneliness* | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Not lonely | Lonely | Low loneliness | Moderate loneliness | Moderately High loneliness | ||||||||
| Zhu et al., China [32] |
≥ 60 years, mean age (SD): 71 (7.73) Community-dwelling women: 51%, Married: 64%, Empty nester: 46.6% |
Cross-sectional design (n = 732) | Chronic disease 76.4% | EQ-5D-3L (Chinese general population) | 20-item UCLA LSa |
40.73 (8.73) U: 0.93 (0.12) |
20–34: U: 0.97 |
35–49: U: 0.93 |
50–64:b U: 0.86 |
|||
| Ko et al., South Korea [46] |
≥ 60 years, mean age (SD): 77 (5.87) Women: 78% All Living alone Widowed: 76% |
Cross-sectional design (n = 1023) |
Mild to severe cognitive impairment: 57.2% Moderate depression 26% |
EQ-5D-3L (UK general population) |
20-item UCLA LSa |
41.54 (13.12) U: 0.81 (0.18) |
35–49c 41.54 U: 0.81 |
|||||
| Rodriguez-Blazquez et al., Spain [47] |
≥ 60 years older adults Non-institutionalised group mean age (SD): 72 (8.16) Without partner:43% Women: 58% Institutionalised group mean age (SD: 81 (7.06) Without partner:81% Women: 65% |
Cross-sectional design (n = 468) |
Mean number of medical conditions (SD) Non-institutionalised 3.57 (2.62) Institutionalised 6.50 (2.79) |
EQ-5D-3L (not stated) | 6-item De Jong Gierveld loneliness Scale d |
Non- institutionalised 2.02 (1.81) U: 0.81 (0.26) |
Institutionalised 2.54 (1.58) U: 0.57 (0.36) |
2–6:d 2.02 U: 0.81 2.54 U: 0.57 |
||||
| Liira et al., Finland [48] based on Pitkala et al. [67] |
≥ 75 years: mean age (SD): 80 (4) Home dwelling Women: 73% Widowed: 68% |
RCT for psychosocial group intervention to explore HRQoL (n = 235) |
Having subjective feelings of loneliness and numerous health conditions |
15D (not stated) |
Study specific single question f |
NR U: 0.78 (0.12) g |
Yes: U: 0.78 |
|||||
| Taube et al., Sweden [49] |
≥ 65 years, mean age (SD): 82 (6.4) Women: 67% All Home dwelling Not lonely sample (n = 61) mean age (SD): 81 (6.5) Windowed 21.3% Lonely sample (n = 61) mean age (SD): 82 (6.2) Windowed 67.4% |
Cross-sectional design (n = 153) Not Lonely (n = 61) Lonely (n = 92) |
Dependency in daily activities, repeated contacts with the health care services Not lonely: No of health complaints: 9.8 (4.7) Lonely: No of health complaints: 12.1 (4.6) |
EQ-5D-3L (UK general population) | Study specific single question (Yes/No) h |
NR U: 0.59 (0.27) |
No: U: 0.63 |
Yes: U: 0.56 |
||||
| Van Houwelingen et al., Netherland [42] |
≥ 75 years median age 82 All Home dwelling Living alone 55 Widowed 53% Women: 68% |
RCT to assess efficacy of a simple structural monitoring system to detect deterioration in the functional, somatic, mental or social health of individuals (n = 2713) |
No. of chronic diseasesi median (IQR): 4 (3–6) |
EQ-5D-3L (UK general population) |
11-item De Jong Gierveld Loneliness Scale j |
Problems Functional:k 3 U: 0.65 Somatic: 2 U: 0.69 Mental: 3 U: 0.69 Social 4 U: 0.72 |
No problems Functional: 2 U: 0.81 Somatic: 1 U: 0.84 Mental: 1 U: 0.81 Social 1 U: 0.81 |
0–2 j 2 U: 0.81 2 U: 0.69 1 U: 0.84 1 U: 0.81 1 U: 0.81 |
3–8j 3 U: 0.65 3 U: 0.69 4 U:0.72 |
|||
| Gardner et al., Australia [33] |
18–25 years old: Mean age (SD): 21 (2.21) Women: 48% Living with family: 72% |
Observational study (n = 159) | Serious mental illnesse |
AQoL-8D (Australian general population) |
20-item R-UCLA LSc |
52.48 (12.94) U: 0.42 (0.21) |
50–64c: 52.48 U: 0.42 |
|||||
| Mountain et al., UK [39] |
≥ 65 years mean age (SD): 72 (65.92) Community living Women: 68% Normal cognition Living alone 55% |
RCT involving weekly group sessions Intervention (n = 145) Control (n = 143) |
No reported health conditions | EQ-5D-3L (UK general population) | 11-item De Jong Gierveld Loneliness Scale j |
Control Baseline: 4.6 (3.6) U: 0.77 (0.24) 6 months 4.1 (3.4) U: 0.76 (0.23) 24 months: 4.8 (3.6) U: 0.71 (0.28) |
Intervention Baseline: 4.1 (3.5) U: 0.73 (0.25) 6 months: 3.5 (3.2) U: 0.71 (0.25) 24 months 3.7 (3.4) U: 0.73(0.24) |
3–8 j 4.1 m U: 0.73 4.6 U: 0.77 3.5 U: 0.71 4.1 U: 0.76 3.7 U: 0.73 4.8 U: 0.71 |
||||
| Maxwell et al., Canada and USA [44] |
≥ 65 years mean age (SD): 81 (8.4) Home care clients; Women: 72% Widowed 54% Not lonely 77% Self-rated health: Good/Excellent 69% |
Cross-sectional design (n = 514) | More than 3 chronic diseases#: 75.3% | HUI2 (Canadian adult population) | Self-reported loneliness status (Yes/No) |
NR U: 0.49 (0.18) |
No: U: 0.50 |
Yes: U: 0.46 |
||||
| Action et al., UK [37] |
Mean age (SD): 75 (16.21) All Home dwelling |
RCT to examine the effect of a home visit–based visual rehabilitation intervention (n = 67) | Sight loss causing difficulties in carrying out daily tasks | EQ-5D-5L (UK general population) | 20-item UCLA LSn |
Control Baseline: 10.47 (13.81) U: 0.60 (0.3) 6-months: 12.81 (14.71) U: 0.58 (0.29) |
Intervention: Baseline: 13.14 (13.34) U: 0.51 (0.31) 6-months: 14.51 (15.19) U: 0.52 (0.34) |
|||||
| Packer et al., Australia [35] |
Living with diabetes mean age (SD): 60 (11.0) Women: 48% Number of health conditionsk: 2 (1.1) Living with any chronic condition mean age (SD): 70 (10.5) Women: 66% number of health conditions: 2.6 (1.3) |
Quasi-ongoexperimental design to investigate the impact of generic and diabetes-specific self-management programmes (n = 485) |
Living with diabetes (n = 222) living with any chronic condition (n = 236) |
AQoL-8D (Australian general population) |
R-UCLA LS 3 itemo Single question (SQ) with four responsesp |
Living with diabetes: R-UCLA: 4.02 (1.6) U: 0.73 (0.2) Living with any chronic condition R-UCLA: 4.64 (1.7) U: 0.48 (0.2) Living with diabetes: SQ: 3.24 (0.7) U: 0.48 (0.2) Living with any chronic condition SQ: 3.49 (0.7) U: 0.73 (0.2) |
||||||
| Sarant et al., Australia [36] | Mean age (SD) 72 (6.8) | Interventional study investigating the impact of cochlear implants (n = 20) | With severe–profound hearing loss [49] |
HUI3 (Not stated) |
11-item De Jong Gierveld Loneliness Scale j |
Baseline: 2.64 U:0.56 18-months: 1.39 U:0.67 |
0–2 2.64 U:0.56 1.39 U:0.67 |
|||||
| Weiss et al., Netherlands [43] |
Median age: 60 years; IQR (48.3, 68.0) 77% living alone Low socioeconomic status |
RCT to examine the effect of the positive psychology intervention ‘Happiness Route’ [HR] (n = 58) compared to ‘Customised care’ [CC] (n = 50) |
Severely lonely High level of comorbidity (median health conditions = 3) |
EQ-5D (Dutch population) |
11-item De Jong Gierveld Loneliness Scale j |
Control Baseline 9.41 (0.29) U: 0.46 (0.05) 3 months 9.43 (0.39) U: 0.54 (0.06) 9 months 8.85 (0.52) U: 0.5 (0.06) |
Intervention Baseline 9.41 (0.29) U: 0.46 (0.05) 3 months 9.43 (0.39) U: 0.54 (0.06) 9 months 8.85 (0.52) U: 0.5 (0.06) |
8–11 j 9.41 m U: 0.46 9.43 U: 0.54 8.85 U: 0.5 9.41 U: 0.46 9.43 U: 0.54 8.85 U: 0.5 |
||||
| Marianne Harris et al., Canada [45] |
≥ 35 Mean age: 53 years (8.3) |
Cross-sectional study (n = 856) | Diagnosed with HIV for atleast a year | ED-5D-5L (Not stated) | Study specific single questionf |
NR U: 0.70 (0.19)g |
No: U:0.88 |
Yes U:0.70 |
||||
| Weng Yew et al., Singapore [41] |
Mean age: 54.3 years (16.8) Women: 56.2% Community-dwelling adults |
Cross-sectional study (n = 1510) | Skin diseases | EQ-5D-5L | R-UCLA LS 3 itemo |
Having a skin disease 3.5 (1.2) U: 0.89 (0.18) Not having a skin disease 3.3 (0.8) U: 0.95 (0.12) |
3–5 3.5 U: 0.89 3.3 U: 0.95 |
|||||
aUCLA LS 20-item scores range from 20 to 80, with greater scores defining greater degrees of loneliness. Scores range from 20 to 34 depict low loneliness, 35–49 moderate loneliness, 50–64 moderately high loneliness, 65–80 high loneliness
bA significant difference was found for the utility scores for groups with different degrees of loneliness (p < 0.001)
cCategorised by lead author into reported levels of loneliness based on UCLA LS 20 scoring cut-offs provided in Zhu et al. [32]. Scores range from 20–34 low loneliness, 35–49 moderate loneliness, 50–64 moderately high loneliness, 65–80 high loneliness
dThe 6-item De Jong Gierveld scale ranges from 0 to 6, where scores ranging from 0 to 1 denote a state of not being lonely and scores 2–6 denote a state of being lonely
eSerious mental illness included depressive disorders, anxiety disorders, schizophrenia, personality disorder, trauma-related disorder, bipolar related disorder, eating disorder, obsessive–compulsive and related disorders, substance-related and addictive disorders
fSingle question with three responses: “Do you suffer from loneliness” (seldom or never, sometimes, often or always)
gUtility score was provided for only those who reported to be lonely
hSingle-item question with yes/no response: ‘Looking back over the last year, which response alternative corresponds best for you?’ (no = 0; yes = 1)
iIncluding self-reported diabetes, heart failure, malignancy, COPD, incontinence, arthritis, osteoporosis, dizziness, LUTS, depression, anxiety, dementia, vision, deafness, fracture, stroke/TIA, myocardial infarction
jDe Jong Gierveld Loneliness Scale 11-item scale score range 0–11 scale with higher scores indicating more loneliness, categorised as 0–2 not lonely, 3–8 moderately lonely, 9–11 strongly lonely
kThe paper examined loneliness and quality of life by four health indicators (functional, somatic, mental and social); as such six values are reported here, which were categorised by levels of loneliness
lReported as Median [IQR]
mThis RCT collected loneliness and utility data across 3 measurement points (baseline, 6-months, 24-months) for the intervention and the control group
nOriginal 20-item UCLA loneliness scale ranging from 0 to 60; no established cut-off points
oScores of 3–5 denote “not being lonely” and scores 6–9 indicate “being lonely”
pSingle question with four responses: always feel lonely (1); often feel lonely (2), sometimes feel lonely (3),never feel lonely (4)
*Utility scores were obtained from the overall population and by presence/absence and level of loneliness based on categorisation reported by authors or using established cut-off points. Since the 20-item UCLA scale does not have any established cut-offs, the scores were categorised as per the cut-offs reported by Zhu et al. [32]