Journal of Psychiatric Research 180 (2024) 482–488

Contents lists available at ScienceDirect

Journal of Psychiatric Research

journal homepage: www.elsevier.com/locate/jpsychires

Clinical and lifestyle predictors of loneliness: A two-year longitudinal study

Thyago Antonelli-Salgado a,b,c,* , Bruno Braga Montezano a,b,c , Thiago Henrique Roza a,i,
Vitória Bouvier c , Aline Zimerman a,b, Lucas Tavares Noronha h , Grasiela Marcon a,b,d ,
Maurício Scopel Hoffmann c,e,f, André Russowsky Brunoni g , Ives Cavalcante Passos a,b,c
a
Laboratory of Molecular Psychiatry, Centro de Pesquisa Experimental (CPE) and Centro de Pesquisa Clínica (CPC), Hospital de Clínicas de Porto Alegre (HCPA), Porto
Alegre, RS, Brazil
b
Instituto Nacional de Ciência e Tecnologia Translacional em Medicina (INCT-TM), Porto Alegre, RS, Brazil
c
Universidade Federal do Rio Grande do Sul, School of Medicine, Graduate Program in Psychiatry and Behavioral Sciences, Department of Psychiatry, Porto Alegre, RS,
Brazil
d
Universidade Federal da Fronteira Sul, School of Medicine, Department of Psychiatry, Chapecó, SC, Brazil
e
Universidade Federal de Santa Maria (UFSM), Avenida Roraima 1000, building 26, office 1446, Santa Maria, 97105-900, Brazil
f
Care Policy and Evaluation Centre, London School of Economics and Political Science, Houghton Street, London, WC2A 2AE, United Kingdom
g
Departments of Internal Medicine and Psychiatry, Faculdade de Medicina da Universidade de São Paulo (USP), São Paulo, Brazil
h
Pontifícia Universidade Católica do Rio Grande do Sul, School of Medicine, Porto Alegre, RS, Brazil
i
Department of Psychiatry, Universidade Federal do Paraná (UFPR), Curitiba, PR, Brazil

A R T I C L E I N F O A B S T R A C T

Keywords: Background: While loneliness is a global public health problem, the literature lacks studies assessing loneliness
Loneliness predictors in low- and middle-income countries. Therefore, we aimed to analyze clinical and lifestyle predictors
Depression of loneliness.
Anxiety
Methods: We conducted a 2-year longitudinal study in Brazil based on a snowball sample and online surveys
Lifestyle
Physical activity
(baseline: May 6 to June 6, 2020). We assessed clinical and lifestyle predictors of loneliness using multiple
Social relationships regression models. The analyses were adjusted for several sociodemographic variables and weighted for attrition
and sampling procedures.
Results: The study included a nationwide sample of 473 participants (18–75 years; 87.1% females). After
adjusting for sociodemographic factors, we identified as risk factors: depressive symptoms (RR: 1.214; 95%CI:
1.08–1.36; p = 0.001), anxiety symptoms (RR:1.191; 95%CI: 1.04–1.35; p = 0.007), alcohol abuse (RR: 1.579;
95%CI: 1.32–1.88; p < 0.001), and cannabis use (RR: 1.750; 95%CI: 1.25–2.39; p < 0.001). More than 150 min/
week of physical activity (RR: 0.177; 95%CI: 0.07–0.34; p < 0.001) and good/excellent quality of family re­
lationships (RR: 0.73; 95%CI: 0.60–0.87; p < 0.001) and sleep (RR: 0.483; 95%CI: 0.39–0.59; p < 0.001) were
protective factors.
Conclusion: Several clinical factors (depression, anxiety, alcohol, and cannabis) have been identified as risk
factors for loneliness, while lifestyle factors (physical activity, better quality of sleep, and family relationships)
have been associated with a lower incidence of loneliness. Addressing clinical and lifestyle factors may therefore
be essential to preventing loneliness.

1. Introduction systematic review and meta-analysis (Surkalim et al., 2022) with

pre-pandemic data from 113 countries or territories showed that lone­
Loneliness is highly prevalent and rising worldwide. A recent liness prevalence in adults ranged from 2.9% to 24.2%, depending on

* Corresponding author. Laboratory of Molecular Psychiatry, Centro de Pesquisa Experimental (CPE) and Centro de Pesquisa Clínica (CPC), Hospital de Clínicas de
Porto Alegre (HCPA), Porto Alegre, RS, Brazil.
E-mail addresses: [email protected] (T. Antonelli-Salgado), [email protected], [email protected] (B.B. Montezano), roza.h.
[email protected] (T.H. Roza), [email protected] (V. Bouvier), [email protected] (A. Zimerman), [email protected]
(L.T. Noronha), [email protected] (G. Marcon), [email protected] (M.S. Hoffmann), [email protected], [email protected] (A.R. Brunoni),
[email protected] (I.C. Passos).

https://doi.org/10.1016/j.jpsychires.2024.11.025
Received 25 July 2024; Received in revised form 31 October 2024; Accepted 8 November 2024
Available online 10 November 2024
0022-3956/© 2024 Published by Elsevier Ltd.
T. Antonelli-Salgado et al. Journal of Psychiatric Research 180 (2024) 482–488

the region or age group. Another systematic review with meta-analysis This study report follows the STROBE guidelines (von Elm et al.,
(Ernst et al., 2022) that included longitudinal studies with data before 2007). Table S1 in the supplementary material presents the STROBE
and during the COVID-19 pandemic showed increased loneliness scores checklist.
and prevalence rates.
These findings reinforce the urgency of addressing loneliness as a 2.2. Measures
public health concern, considering that there is a robust association
between loneliness and various negative health outcomes. For instance, 2.2.1. Survey instruments
loneliness can induce several cardiometabolic changes, such as an The online questionnaires used during data collection included
increased coronary heart disease risk, increased activation of the validated scales and tools to assess loneliness and specific clinical vari­
hypothalamic-pituitary-adrenal axis (Hodgson et al., 2020), high blood ables. In addition, the questionnaires covered lifestyle and sociodemo­
pressure, and increased cholesterol levels (Leigh-Hunt et al., 2017; Lim graphic variables.
et al., 2020). Moreover, it is associated with neurological diseases such
as stroke (Valtorta et al., 2016), Parkinson’s disease (Terracciano et al., 2.2.2. Loneliness
2023), and dementia (Salinas et al., 2022). In addition, loneliness can Loneliness was measured with the 3-item short version of the R-
increase overall mortality comparable with well-established risk factors UCLA (Hughes et al., 2004). The scale asks “How often do you feel you
for mortality (Holt-Lunstad et al., 2010, 2015). In psychiatry, we have lack companionship?”, “How often do you feel left out?”, and “How
several studies showing loneliness as a predictor of depressive symp­ often do you feel isolated from others?”. Response options for each item
toms, anxiety, and suicidality (Antonelli-Salgado et al., 2021; Mann are “hardly ever or never”, “some of the time”, or “often” (equating to
et al., 2022). scores of 1,2, and 3, respectively). Total scores range from 3 to 9, and
Despite the increase in studies of loneliness as a predictor of these higher scores indicate greater loneliness (Hughes et al., 2004), with
various clinical outcomes, the literature still lacks longitudinal studies scores ≥6 indicating important loneliness (Steptoe et al., 2013). We used
on predictors of loneliness, especially in low- and middle-income Cronbach’s α and McDonald’s ω to test internal reliability (Lucke, 2005;
countries and collectivist societies (Luhmann et al., 2023; Surkalim Raykov, 1997), and the analyses and results are described in a previous
et al., 2022). Collectivist societies are those in which the group’s needs article (Antonelli-Salgado et al., 2021).
are valued over the individuals. Brazil scored 36 on Hofstede’s Indi­
vidualism scale (0–100). The score shows a mix of individualistic and 2.2.3. Mental health
collectivist characteristics but is predominately collectivist (Hofstede, Depressive symptoms were measured with the PHQ-9 (Kroenke
2024). et al., 2001). A score equal to or greater than 9 is considered a positive
Therefore, the primary aim of this study is to carry out a longitudinal screening result for depression in the Brazilian population (Matias et al.,
assessment of clinical factors for the development of loneliness in a 2016; Santos et al., 2013). Anxiety symptoms were measured using the
sample of adults in Brazil. The secondary objective is to analyze lifestyle GAD-7 (Moreno et al., 2016). A positive indicator of signs and symptoms
factors as a predictor of loneliness. of anxiety disorders is a value equal to or greater than 10. The AUDIT-C
was used to evaluate alcohol use (Bush et al., 1998). In men, a score of 4
2. Methods or more is considered positive; in women, a score of 3 or more is
considered positive. The questions about cannabis use were based on the
2.1. Design, participants and setting ASSIST (Alcohol, Smoking, and Substance Involvement Screening Test)
instrument (WHO ASSIST Working Group, 2002).
The present study is part of a larger cohort with four temporal waves
(baseline, 1, 6, and 24 months). This larger cohort aimed to investigate 2.2.4. Lifestyle/social
the impact of the COVID-19 pandemic on specific mental health vari­ Lifestyle measures were assessed in Likert scale format but analyzed
ables within the Brazilian population across various pandemic phases. dichotomously. Physical activity was separated into two groups, using
Data collection was performed using online questionnaires due to the the cutoff point of 150 min/week, based on recommendations from the
need to comply with recommendations on social distancing. The inclu­ World Health Organization (WHO, 2022). The variables Sleep quality,
sion criteria were being at least 18 years old, living in Brazil at the time Family relationship, and Friendship relationship were divided into
of data collection, and having access to the Internet. For inclusion in the Poor/Average or Good/Excellent.
subsequent waves of data collection, participants needed to provide
their email addresses at the end of the baseline questionnaire and agree 2.2.5. Covariates
to receive an email with the specific questionnaires regarding that wave. To adjust our results, we chose several sociodemographic factors that
All online questionnaires were provided to eligible participants using an were already related to loneliness in previous studies, such as age,
online platform (Survey Monkey) (“Survey Monkey,” 2024), and the gender, sexual orientation, income, education, and race (Buczak-Stec
baseline questionnaire was advertised through social media (Facebook, et al., 2023; Lim et al., 2020; Lin, 2023; Luhmann et al., 2023; Solmi
Instagram, and WhatsApp) to reach participants. et al., 2020). The supplemental material includes a Portuguese (original
This study is based on this cohort’s first and fourth waves of data version) and an English-translated version of the questions used in this
collection. All predictor variables and covariates were obtained from the study (see Method S1).
first wave (baseline). The outcome (loneliness) was obtained from the
fourth wave. We distributed the Wave 1 (W1) questionnaire between 2.3. Statistical analysis
May 6 and June 6, 2020, and the Wave 4 (W4) two years after W1.
Considering that our study began approximately two months after the We only analyzed participants who completed all questions related
confirmation of the first case of COVID-19 in Brazil (Brazilian Ministry to the variables of interest in this study. Descriptive analyses were re­
of Health, 2022), our baseline refers to an early stage of the pandemic. ported as means or absolute and relative frequencies. We filtered the
The research ethics committee of Hospital de Clínicas de Porto Alegre data with subjects that presented valid values for UCLA (UCLA Loneli­
(Rio Grande do Sul, Brazil) approved the study (CAAE: ness Scale 3-item version) in the first and fourth waves and evaluated
30222820.4.0000.5327). All participants electronically signed the just those that did not have loneliness (UCLA score <6) at baseline (W1).
informed consent form before answering the online questionnaires. Then, we divided participants into two groups based on UCLA’s total
After completing each questionnaire, participants were provided infor­ score at W4: absent and incident cases of loneliness. In the fourth-wave
mation about mental health support centers in Brazil. assessment, UCLA scores greater or equal to 6 were considered incident

483
T. Antonelli-Salgado et al. Journal of Psychiatric Research 180 (2024) 482–488

cases (positives), and scores less than 6 were coded as absent cases weighting (IPW) (Seaman and White, 2013). Age, region, gender, skin
(negatives). color, income, education, marital status, living alone, unemployment,
First, we used chi-squared with Rao & Scott’s second-order correc­ physical activity, history of psychiatric diagnosis, anxiety symptoms
tion or Mann–Whitney U test to analyze demographic and clinical var­ (GAD-7), and AUDIT-C score were included in the regression model to
iables between these two groups. Afterward, we used hierarchical multi- predict attrition. IPW was trimmed to the 5th and 95th percentiles. We
predictor Poisson regression analysis and allocated the variables in three also calculated survey weights to account for demographic representa­
different levels: (i) Sociodemographic; (ii) Lifestyle; and (iii) Clinical. tion at the national level. This procedure applies iterative
We then calculated risk ratios (RR) and respective confidence intervals post-stratification to match population margins to the survey sample
(95% CI) for each domain using bivariate analysis, considering p-values proportions, which can approximate the demographic characteristics of
below 0.20 statistically significant. The RR is the ratio of the cumulative the sample to the Brazilian population. We weighted our sample using
incidences in the exposed and unexposed groups, and provides a useful Brazilian population margins regarding sex at birth, age groups, the
measure for comparing the risks for each group. The variables within the region of residency, race/ethnicity, and household income according to
threshold were allowed to integrate the sequence of multi-predictor the last Brazilian census (Instituto Brasileiro de Geografia e Estatística,
analysis (Poisson regression), starting from the most distal level to the 2010). The survey weight was trimmed up to 20. The final weights were
most proximal ones (sociodemographic → lifestyle → clinical). The calculated as the product of IPW and survey weights and were used in all
variable was retained if the subsequent level analysis calculated a p- regressions to minimize bias due to attrition and sample representation.
value <0.10. At each of the following levels, the Poisson regression The generalized linear models were fitted within the glm R function
model was adjusted with the corresponding level variables and the from stats package. Survey weights were calculated using the packages
previous groups’ statistically significant variables. The final model was survey in R (Lumley, 2020) (rake function). All analyses were performed
composed of the variables of the level that presented a p-value <0.05 using the R programming language (version 4.3.2). The following
and the other variables of the previous levels in the same condition. additional R packages were used: dplyr (version 1.1.4), huxtable (version
Hierarchical regression is a type of regression model in which the in­ 5.5.3), tidyr (version 1.3.0) and gtsummary (version 1.7.2). All statistical
dependent variables are entered in blocks. Each block represents one estimates were performed using survey weight.
step. The order (or which independent variable is entered into which
block) was determined by the researchers based on theory.
Missing data at follow-up were addressed using inverse probability

Table 1
Demographic and personal characteristics of participants.
Not lonely group Lonely group

Characteristics n = 348 %(n) %(D) %(A) %(C) n = 125 %(n) %(D) %(A) %(C) p valueX

Sex ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 0.079
Male 45 12.9 43.7 18.9 58.7 16 12.8 23.8 18.5 33.3 ​
Female 303 87.1 56.3 81.1 41.3 109 87.2 76.2 81.5 66.7 ​
Sexual orientation (heterosexual) 309 88.8 79.1 87.9 77.2 105 84.0 77.9 81.2 70.7 0.7
Skin Color ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 0.9
Non-White 136 39.1 48.6 44.2 57.0 59 47.2 45.9 55.6 54.7 ​
White 212 60.9 51.4 55.8 43.0 66 52.8 54.1 44.4 45.3 ​
Education ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 0.6
Up to high-school 20 5.7 12.8 8.1 18.0 15 12.0 29.4 17.6 31.1 ​
Some college education 135 38.8 42.6 42.2 40.8 54 43.2 32.6 43.2 32.9 ​
Master or doctorate 193 55.5 44.6 49.7 41.1 56 44.8 38.0 39.3 36.0 ​
Socioeconomic status ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 0.8
Lower 93 26.7 66.6 32.2 72.5 42 33.6 71.0 39.9 72.3 ​
Middle 75 21.6 10.7 21.8 8.6 33 26.4 13.9 24.0 11.6 ​
Upper 180 51.7 22.6 46.0 18.9 50 40.0 15.1 36.0 16.1 ​
Unemployed 34 9.8 29.5 10.4 34.1 23 18.4 38.1 22.4 42.3 0.7
Social distancing 327 94.0 95.4 93.5 95.9 119 95.2 97.5 95.7 97.7 0.5
Physical activity ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ <0.001
Less than 150 min/week 305 87.6 88.8 89.9 92.7 119 95.2 98.4 96.6 99.0 ​
More than 150 min/week 43 12.4 11.2 10.1 7.3 6 4.8 1.6 3.4 1.0 ​
Sleep quality ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 0.004
Better 163 46.8 51.2 45.7 50.4 41 32.8 18.9 29.6 16.7 ​
Worse 185 53.2 48.8 54.3 49.6 84 67.2 81.1 70.4 83.3 ​
Marital status (grouped) ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 0.5
Without Partner 146 42.0 56.8 39.2 56.0 68 54.4 72.2 51.8 65.5 ​
With partner 202 58.0 43.2 60.8 44.0 57 45.6 27.8 48.2 34.5 ​
Friendship relationship quality ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 0.6
Better 231 66.4 66.2 65.2 62.2 72 57.6 55.1 56.4 52.9 ​
Worse 117 33.6 33.8 34.8 37.8 53 42.4 44.9 43.6 47.1 ​
Family relationship quality ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 0.13
Better 263 75.6 78.3 74.0 80.7 79 63.2 65.9 61.0 64.0 ​
Worse 85 24.4 21.7 26.0 19.3 46 36.8 34.1 39.0 36.0 ​
Religion 165 47.4 53.9 47.7 56.4 64 51.2 65.1 51.4 63.0 0.7
Alcohol risk zone ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 0.040
Low risk 253 72.7 82.9 70.6 83.3 92 73.6 59.2 71.5 57.5 ​
High risk 95 27.3 17.1 29.4 16.7 33 26.4 40.8 28.5 42.5 ​
Cannabis use 14 4.0 5.4 4.0 4.2 5 4.0 5.2 5.3 6.3 0.7

Weighted characteristics of participants stratified by loneliness incidence. D. Weighted by demographics. A.Weighted by attrition. C. Weighted by demographics and
attrition, combined. X. Chi-squared test with Rao & Scott’s second-order correction; Wilcoxon rank-sum test for complex survey samples applied to the weighted by
demographic and attrition data.

484
T. Antonelli-Salgado et al. Journal of Psychiatric Research 180 (2024) 482–488

3. Results same model also identified higher education as a protective factor, with
significance observed for college education (RR: 0.684; 95%CI:
3.1. Sample size, descriptive results and bivariate analysis 0.551–0.851; p < 0.001) and master’s or PhD education (RR: 0.705;
95%CI: 0.556–0.893; p = 0.004). Furthermore, a quadratic terms
We included 473 participants in our sample. The participant selec­ analysis revealed that both extremes of age were associated with an
tion flowchart and reasons for dropouts in all waves are shown in Fig. S1 increased risk of loneliness (RR: 2.849; 95%CI: 1.881–4.287; p < 0.001).
(supplementary material). The mean age of the participants was 38.7 All these results from the multi-predictor poison analysis, including
(SD: 12.1) years and 87.1% identified as females. Based on data from the those that were not statistically significant, are shown in Table 3. A
last national census, our weighted sample closely represents the Bra­ comparison of the effect size of the predictors of loneliness is shown in
zilian population. The incidence of loneliness at W4 was 26.42%, cor­ Fig. S2 (supplementary material).
responding to 125 out of 473 responses. Table 1 presents the weighted
demographic and personal characteristics of our included sample and 4. Discussion
the comparison between participants with and without loneliness
through the combination of all weightings. Tables S3, S4, and S5 present This is the first longitudinal study to evaluate clinical and lifestyle
in the supplementary material provide this comparison with each type of predictors of loneliness in adults from a low- or middle-income country
weight used. The results of the bivariate Poison regression are presented with collectivist society characteristics. After adjustment for multiple
in Table 2. confounding variables, clinical factors such as depressive symptoms,
anxious symptoms, alcohol abuse, and cannabis use were risk factors for
loneliness.
3.2. Multi-predictor analysis Our findings are in line with the scientific literature that points to a
bidirectional relationship between loneliness and symptoms of depres­
The hierarchical Poisson regression model identified several clinical sion and anxiety; that is, depressive and anxious symptoms can increase
variables as significant risk factors associated with the incidence of the risk of developing loneliness, but also have a greater chance of
loneliness. Specifically, depressive symptoms (RR: 1.21; 95%CI: developing among individuals with loneliness. A recent study (Chen
1.08–1.366; p = 0.001), anxiety symptoms (RR: 1.19; 95%CI: et al., 2023) conducted a meta-analysis (37 studies and 39,511 partici­
1.049–1.35; p = 0.007), severe risk for alcohol abuse (RR: 1.579; 95%CI: pants) using a cross-lagged approach to examine the reciprocal relations
1.32–1.88; p < 0.001) and cannabis use (RR: 1.75; 95%CI: 1.25–2.40; p between loneliness and depressive symptoms. They showed that lone­
< 0.001) were considered risk factors. liness and depressive symptoms reciprocally predicted each other over
Regarding lifestyle variables, the model identified protective factors time with similar effect sizes (β = 0.18). Depression can alter social
for the incidence of loneliness. More than 150 min/week of physical decision-making capacity, leading to greater social avoidance.
activity (RR: 0.177; 95%CI: 0.078–0.348; p < 0.001) and good/excellent Depressed people may have significantly less positive (happiness) and
quality of family relationships (RR: 0.730; 95%CI: 0.609–0.875; p < more negative (shame, guilt, disappointment) feelings about social ac­
0.001) and sleep (RR: 0.483; 95%CI: 0.390–0.594; p < 0.001) were tivities (Fernández-Theoduloz et al., 2019).
protective factors. The number of people living in the same household The relationship between loneliness and depressive symptoms is
also has a protective tendency, though statistical significance was not complex and some studies point to another possibility. Loneliness as part
sustained after adjusting for all factors (RR: 0.968; 95%CI: 0.913–1.024; of the symptoms of a depressive disorder (Gijzen et al., 2021; Manfro
p = 0.262). et al., 2023; Mullarkey et al., 2019, 2021). In this sense, the loneliness
Lastly, when analyzing the association between sociodemographic that precedes depressive disorder could not be a risk factor but rather a
variables and the incidence of loneliness, several risk factors were prodrome. In this same understanding, loneliness after depressive dis­
recognized: female gender (RR: 1.719; 95%CI: 1.393–2.119; p < 0.001), order would be a residual symptom of this disorder.
non-heterosexual orientation (RR: 1.518; 95%CI: 1.227–1.872; p < Another recent meta-analysis (Gabarrell-Pascuet et al., 2023) of
0.001), middle (RR: 1.388; 95%CI: 1.094–1.743; p = 0.006) and upper cross-sectional data that included 73 studies (1,020,461 participants)
(RR: 1.361; 95%CI: 1.062–1.735; p = 0.014) socioeconomic status. The

Table 2
Bivariate analysis of predictors of the sociodemographic, lifestyle, social and clinical domains.
Variable Category β SE RR RR (lower) RR (upper) p-value

Age – − 0.089 0.033 0.914 0.856 0.976 0.007

Age (quadratic term) – − 0.053 0.031 0.948 0.892 1.007 0.087
Age (cubic term) – − 0.028 0.028 0.973 0.920 1.027 0.328
Skin color White 0.072 0.071 1.074 0.934 1.234 0.314
Education Some college education − 0.552 0.088 0.576 0.484 0.685 <0.001
Master or Doctorate − 0.489 0.087 0.614 0.484 0.685 <0.001
Gender at birth Female 0.804 0.075 2.234 1.930 2.591 <0.001
Heterosexual No 0.253 0.078 1.288 1.105 1.498 0.001
Household income Middle 0.220 0.112 1.246 0.995 1.543 0.049
Upper − 0.122 0.097 0.885 0.995 1.543 0.209
Unemployed No − 0.263 0.072 0.769 0.669 0.885 0.000
Family relationship Good/excelent − 0.618 0.074 0.539 0.467 0.623 <0.001
Friend relationship Good/excelent − 0.287 0.071 0.750 0.653 0.862 <0.001
Physical activity More than 150 min/week − 1.771 0.349 0.170 0.079 0.315 <0.001
Sleep quality Good/excelent − 1.301 0.095 0.272 0.225 0.327 <0.001
Number of people living in your house – − 0.146 0.024 0.864 0.824 0.906 <0.001
Anxiety symptoms (z-score) – 0.531 0.034 1.70 1.591 1.818 <0.001
Depressive symptoms (z-score) – 0.513 0.029 1.67 1.577 1.769 <0.001
Alcohol risk zone High risk 0.916 0.072 2.500 2.172 2.875 <0.001
Cannabis use Yes 0.301 0.146 1.351 1.003 1.778 0.039

The bivariate analyses were performed in order to pre-select the variables that will be used in the multi-predictor Poisson regression. The cut-off point we will use is
based on p < 0.2. The lower and upper bounds (RR lower and RR upper) were calculated based on a 95% confidence interval (CI).

485
T. Antonelli-Salgado et al. Journal of Psychiatric Research 180 (2024) 482–488

Table 3 loneliness (OR = 1.60), which can predict loneliness (OR = 1.24). This
Multi-predictor Poisson regression analysis to evaluate factors associated with increased risk of loneliness due to anxiety reinforces previous findings
loneliness incidence. The table presents the risk ratios and the corresponding that show anxiety and depression are associated with worse psychoso­
95% CIs with p-values. The columns illustrate the successive introduction of cial functioning (social connectedness, quality of social relationships)
variables in the Poison regression, with each block of variables represented in a (Cowden et al., 2021).
separate column. The final column presents the result obtained with the full set
The association between loneliness and alcohol or cannabis misuse
of variables.
has been reported previously in the literature. According to a cross-
Block 1: Block 2: Block 3: sectional population-based study from Canada (n = 3772) that investi­
Sociodemographic Lifestyle and Clinical
gated factors associated with loneliness during the COVID-19 pandemic,
social
severe loneliness (according to the UCLA-3) was significantly associated
(Intercept) 0.258 [0.202, 0.448 [0.325, 0.240 [0.169,
with past month cannabis use (OR of 1.47) and past-month binge
0.329], p<0.001 0.616], 0.341],
p<0.001 p<0.001 drinking (OR of 1.39 for binge drinking at least once a month, and OR of
Age (z-score) 0.386 [0.276, 0.693 [0.471, 0.343 [0.224, 1.70 for binge drinking at least once a week) (Lin, 2023). In addition,
0.547], p<0.001 1.029], p = 0.529], according to a longitudinal study from the US (n = 210), with data
0.066 p<0.001 collected before the COVID-19 pandemic, participants with moderate
Age (quadratic term) 2.501 [1.790, 1.475 [1.011, 2.849 [1.881,
and severe loneliness at baseline presented a significantly higher fre­
3.464], p<0.001 2.133], 4.287],
p¼0.041 p<0.001 quency of cannabis and alcohol use at follow-up in adjusted models
Education (college 0.423 [0.348, 0.567 [0.465, 0.684 [0.551, (Gutkind et al., 2022). Even though substance misuse may be a potential
education) 0.514], p<0.001 0.691], 0.851], consequence of loneliness (representing a potential maladaptive coping
p<0.001 p<0.001
mechanism) (Gutkind et al., 2022; Lin, 2023), our study also suggests it
Education (master or 0.458 [0.372, 0.458 [0.368, 0.705 [0.556,
PhD) 0.563], p<0.001 0.568], 0.893],
as a potential risk factor, which may be contributing to the worsening of
p<0.001 p¼0.004 feelings of loneliness.
Sex (female) 2.672 [2.265, 2.317 [1.926, 1.719 [1.393, As described in the literature, our study also found physical activity
3.160], p<0.001 2.795], 2.119], as a protective factor for loneliness (Vancampfort et al., 2019). Links
p<0.001 p<0.001
between social relationships and physical activity may be particularly
Sexual orientation 2.203 [1.809, 1.976 [1.612, 1.518 [1.227,
(non-heterosexual) 2.675], p<0.001 2.414], 1.872], important, since sustained physical activity is associated with a range of
p<0.001 p<0.001 beneficial outcomes, including reduced stress and anxiety symptoms,
Socioeconomic 1.380 [1.088, 1.497 [1.185, 1.388 [1.094, improved sleep quality, reduced levels of depressive symptoms, and
status (middle) 1.732], p¼0.007 1.871], 1.743],
prevention and reduced mortality from chronic diseases such as high
p<0.001 p¼0.006
Socioeconomic 1.311 [1.030, 1.432 [1.139, 1.361 [1.062,
blood pressure and diabetes, as well as providing socialization and
status (upper) 1.660], p¼0.026 1.790], 1.735], quality of life (Schrempft et al., 2019). A study carried out during the
p¼0.002 p¼0.014 covid 19 pandemic among 7203 young university students demon­
Number of people 0.891 [0.845, 0.902 [0.852, 0.968 [0.913, strated that less physically active students were more likely to experi­
living in your house 0.939], p<0.001 0.953], 1.024], p =
ence loneliness than students who were more physically active (Wenig
p<0.001 0.262
Marital status (with 0.924 [0.782, ​ ​ et al., 2023). We did not find studies in the literature that relate the
partner) 1.089], p = 0.349 amount of time spent doing physical activity to loneliness. Our study
Positive family ​ 0.689 [0.574, 0.730 [0.609, found that above 150 min would be a protective factor. We hypothesize
relationship 0.828], 0.875], that longer time is related to more frequent activities, consequently
p<0.001 p<0.001
Positive friendship ​ 0.970 [0.817, ​
bringing greater benefits in terms of physical health and socialization.
relationship 1.151], p = Concerning the connection between loneliness and sleep, our study
0.725 found an association between loneliness and quality of sleep, similar to
Physical activity ​ 0.161 [0.074, 0.177 [0.078, results found in other studies (Guerra-Balic et al., 2023). A meta-analysis
(more than 150 min/ 0.304], 0.348],
correlated loneliness with higher self-reported sleep disturbance,
week) p<0.001 p<0.001
Good sleep quality ​ 0.340 [0.278, 0.483 [0.390, defined as impaired sleep quality and insomnia symptoms. Loneliness
0.414], 0.594], was also associated with sleep inadequacy and dissatisfaction, but not
p<0.001 p<0.001 sleep duration (Griffin et al., 2020). McLay et al. (2021), in a large
Depressive ​ ​ 1.214 [1.080, sample of 95,045 older adults (>65 years of age), showed that loneli­
symptoms (PHQ-9) 1.366],
p¼0.001
ness, social isolation, and health concerns are related and may produce
Anxiety symptoms ​ ​ 1.191 [1.049, sleep disorders, especially in women (McLay et al., 2021). A
(GAD-7) 1.352], propensity-score-matched case-control study with 11,696 participants
p¼0.007 found that loneliness was associated with extended sleep latency,
Alcohol risk zone 1.579 [1.322,
​ ​
increased nocturnal awakenings, and reduced subjective sleep quality
(moderate, high or 1.882],
severe) p<0.001 and daytime function but was not associated with sleep behavior,
Cannabis use (yes) ​ ​ 1.750 [1.250, including sleep onset and offset timings (Peng et al., 2021). As our study
2.399], only addressed general sleep quality, we are unable to differentiate these
p<0.001 issues. Note that the majority of the studies present in the literature
n 473 473 473
demonstrate loneliness as a risk factor for sleep quality and not the
-2log likelihood − 1793.697 − 1678.612 ¡1621.319 opposite relationship. We have few studies (Christiansen et al., 2016;
AIC 3609.395 3385.223 3276.637
Segrin and Domschke, 2011) examining sleep disturbance and loneli­
BIC 3655.145 3443.450 3347.342
ness, being cross-sectional studies. Longitudinal research identified
sleep disturbance as a partial mediator of the connection between
during the COVID-19 Pandemic showed the association of loneliness loneliness and self-reported health, thereby uncovering sleep distur­
with symptoms of depression (r = 0.49), but also symptoms of anxiety (r bance as a treatment target to mitigate the potential effect of loneliness
= 0.40). A two-year follow-up study (Domènech-Abella et al., 2019) on health (Griffin et al., 2021).
including data from 5066 adults aged 50 years and older in Ireland, Our study showed that better quality of family relationships and
showed that loneliness and anxiety are bidirectional. Anxiety predicts higher education was a protective factor. The following covariates:

486
T. Antonelli-Salgado et al. Journal of Psychiatric Research 180 (2024) 482–488

female gender, non-heterosexual orientation, higher socioeconomic Conceptualization. Grasiela Marcon: Writing – review & editing,
status, and extremes of age (in a U-shape way) were associated with an Visualization, Validation, Resources, Methodology. Maurício Scopel
increased risk of loneliness. An umbrella review (Solmi et al., 2020) of Hoffmann: Writing – review & editing, Visualization, Validation,
observational studies (795 studies and 746,706 participants) about Methodology. André Russowsky Brunoni: Writing – review & editing,
loneliness-related factors supported these findings. According to this Writing – original draft, Visualization, Validation, Supervision, Meth­
review, factors cross-sectionally associated with loneliness included odology, Investigation, Data curation, Conceptualization. Ives Cav­
poor quality of social contacts, female sex, and age (in a U-shape way). alcante Passos: Writing – review & editing, Writing – original draft,
However, contrary to our results, loneliness was associated with low Visualization, Validation, Supervision, Resources, Project administra­
socioeconomic status. We believe that our findings reflect the nuances of tion, Methodology, Investigation, Funding acquisition, Formal analysis,
a country with collectivist characteristics. Usually, people from birth are Data curation, Conceptualization.
integrated into cohesive groups (especially represented by the extended
family; including uncles, aunts, grandparents, and cousins) (Hofstede, Data availability statement
2024). People with low socioeconomic status commonly live in com­
munities with a large number of people living in small spaces. Brazil has Data statement is available upon request: Dr. Thyago Antonelli Sal­
more than 10 thousand favelas and urban communities, where 16.6 gado - [email protected]:antonelli.thyago@gmail.
million people live (8% of the Brazilian population) (IBGE, 2022). In this com
context, greater social dependence exists to meet basic survival needs.
This can encourage greater and better social connection and reduce the Role of the funding source
feeling of loneliness. Regarding sexual orientation, several other studies
have also shown that non-heterosexuals have a higher risk of loneliness The Fundo de Incentivo à Pesquisa e Eventos (FIPE) do HCPA (Brasil)
(Buczak-Stec et al., 2023; Luhmann et al., 2023). Finally, a higher level helped with the costs of maintaining the online data collection platform.
of education is also a protective factor in other studies in the literature The support was essential to make this longitudinal study viable.
(Lim et al., 2020; Lin, 2023).
Even though our study presents limitations, we implemented some
Declaration of competing interest
strategies to mitigate them. First, we conducted an online self-report
survey and there may be variations in how participants interpret some
ICP receives authorship royalties from Springer Nature and ArtMed.
questions. To reduce this risk, we used scales validated for our country
ICP has served as a consultant, advisor, or CME speaker for the following
or carried out psychometric tests. Second, we have a sampling bias when
entities: Janssen, LundBeck, Libbs, Daiichi Sankyo, EMS, and Pfizer.
inviting participants via social media, and there was a significant loss of
participants over the two-year follow-up. However, we used weights
Acknowledgments
that brought our sampling closer to the characteristics of the Brazilian
demographic census and weighted for loss to follow-up. Even with the
Dr. Hoffmann is supported by the United States National Institutes of
adjustments, we need to be cautious when interpreting the rates as a
Health grant R01MH120482 under his post-doctoral fellowship at
representative prevalence of the Brazilian population. Third, we use a
UFRGS and by the Wellcome Mental Health Data Prize, granted by the
reduced version of the UCLA scale; because of this, we are unable to
Wellcome Trust (grant number 84494R). At the time of the beginning of
analyze different categories of loneliness, such as emotional and social.
this study, Dr. Hoffmann was supported by the UK Academy of Medical
In addition to some strengths already presented to minimize the
Sciences (Newton International Fellowship NIF\R5A\0028). This study
limitations highlighted above, this study has other strengths to high­
was financed in part by the Coordenação de Aperfeiçoamento de Pessoal
light. First, the study has a longitudinal design that allows for a better
de Nível Superior - Brasil (CAPES) - Finance Code 001. This work
understanding of the cause-and-effect relationship between predictors
received financial support from Conselho Nacional de Desenvolvimento
and outcomes. Second, our national sample recruited individuals
Científico e Tecnológico (CNPq), Brazil. ICP is a CNPq research fellow.
covering all 27 Brazilian federative units. Third, the sample size allowed
This work received financial support from Fundo de Incentivo à Pesquisa
us to analyze a wide range of variables and thus corrected our results for
e Eventos (FIPE) do HCPA, Brasil.
several confounding factors.
In summary, our study showed that after two years of follow-up and
Appendix A. Supplementary data
adjustment for multiple confounding variables, clinical factors such as
depressive symptoms, anxiety symptoms, alcohol abuse, and cannabis
Supplementary data to this article can be found online at https://doi.
use were risk factors for loneliness. Among lifestyle factors, more than
org/10.1016/j.jpsychires.2024.11.025.
150 min/week of physical activity, good/excellent family relationship
quality, and good/excellent sleep quality were associated with a lower
References
incidence of loneliness. Therefore, addressing such predictors may be
essential for preventing loneliness. Antonelli-Salgado, T., Monteiro, G.M.C., Marcon, G., Roza, T.H., Zimerman, A.,
Hoffmann, M.S., Cao, B., Hauck, S., Brunoni, A.R., Passos, I.C., 2021. Loneliness, but
CRediT authorship contribution statement not social distancing, is associated with the incidence of suicidal ideation during the
COVID-19 outbreak: a longitudinal study. J. Affect. Disord. 290, 52–60. https://doi.
org/10.1016/j.jad.2021.04.044.
Thyago Antonelli-Salgado: Writing – review & editing, Writing – Brazilian Ministry of Health, 2022. Painel coronavírus [WWW Document]. URL. https://
original draft, Visualization, Validation, Supervision, Software, Re­ covid.saude.gov.br/, 1.20.21.
Buczak-Stec, E., König, H.-H., Hajek, A., 2023. Sexual orientation and psychosocial
sources, Project administration, Methodology, Investigation, Funding factors in terms of loneliness and subjective well-being in later life. Gerontol. 63,
acquisition, Formal analysis, Data curation, Conceptualization. Bruno 338–349. https://doi.org/10.1093/geront/gnac088.
Braga Montezano: Writing – review & editing, Visualization, Method­ Bush, K., Kivlahan, D.R., McDonell, M.B., Fihn, S.D., Bradley, K.A., 1998. The AUDIT
alcohol consumption questions (AUDIT-C): an effective brief screening test for
ology, Formal analysis, Data curation. Thiago Henrique Roza: Writing
problem drinking. Ambulatory Care Quality Improvement Project (ACQUIP).
– review & editing, Validation, Methodology, Data curation, Concep­ Alcohol Use Disorders Identification Test. Arch. Intern. Med. 158, 1789–1795.
tualization. Vitória Bouvier: Writing – review & editing, Visualization, https://doi.org/10.1001/archinte.158.16.1789.
Methodology, Conceptualization. Aline Zimerman: Writing – review & Chen, Z., Song, X., Lee, T.M.C., Zhang, R., 2023. The robust reciprocal relationship
between loneliness and depressive symptoms among the general population:
editing, Visualization, Validation, Methodology. Lucas Tavares Noro­ evidence from a quantitative analysis of 37 studies. J. Affect. Disord. 343, 119–128.
nha: Writing – review & editing, Visualization, Methodology, https://doi.org/10.1016/j.jad.2023.09.035.

487
T. Antonelli-Salgado et al. Journal of Psychiatric Research 180 (2024) 482–488

Christiansen, J., Larsen, F.B., Lasgaard, M., 2016. Do stress, health behavior, and sleep general population. Soc. Psychiatr. Psychiatr. Epidemiol. 57, 2161–2178. https://
mediate the association between loneliness and adverse health conditions among doi.org/10.1007/s00127-022-02261-7.
older people? Soc. Sci. Med. 152, 80–86. https://doi.org/10.1016/j. Matias, A.G.C., Fonsêca, M. de A., Gomes, M. de L. de F., Matos, M.A.A., 2016. Indicators
socscimed.2016.01.020. of depression in elderly and different screening methods. Einstein 14, 6–11. https://
Cowden, R.G., Chapman, I., Houghtaling, A., 2021. Additive, curvilinear, and interactive doi.org/10.1590/S1679-45082016AO3447.
relations of anxiety and depression with indicators of psychosocial functioning. McLay, L., Jamieson, H.A, France, K.G., Schluter, P.J., 2021. Loneliness and social
Psychol. Rep. 124, 627–650. https://doi.org/10.1177/0033294120918809. isolation is associated with sleep problems among older community dwelling women
Domènech-Abella, J., Mundó, J., Haro, J.M., Rubio-Valera, M., 2019. Anxiety, and men with complex needs. Sci. Rep. 11, 4877. https://doi.org/10.1038/s41598-
depression, loneliness and social network in the elderly: longitudinal associations 021-83778-w.
from the Irish Longitudinal Study on Ageing (TILDA). J. Affect. Disord. 246, 82–88. Moreno, A.L., DeSousa, D.A., Souza, A.M.F.L.P., Manfro, G.G., Salum, G.A., Koller, S.H.,
https://doi.org/10.1016/j.jad.2018.12.043. Osório, F. de L., Crippa, J.A. de S., Others, 2016. Factor structure, reliability, and
Ernst, M., Niederer, D., Werner, A.M., Czaja, S.J., Mikton, C., Ong, A.D., Rosen, T., item parameters of the brazilian-Portuguese version of the GAD-7 questionnaire.
Brähler, E., Beutel, M.E., 2022. Loneliness before and during the COVID-19 Temas em Psicologia 24, 367–376.
pandemic: a systematic review with meta-analysis. Am. Psychol. 77, 660–677. Mullarkey, M.C., Marchetti, I., Beevers, C.G., 2019. Using network analysis to identify
https://doi.org/10.1037/amp0001005. central symptoms of adolescent depression. J. Clin. Child Adolesc. Psychol. 48,
Fernández-Theoduloz, G., Paz, V., Nicolaisen-Sobesky, E., Pérez, A., Buunk, A.P., 656–668. https://doi.org/10.1080/15374416.2018.1437735.
Cabana, Á., Gradin, V.B., 2019. Social avoidance in depression: a study using a social Mullarkey, M.C., Marchetti, I., Bluth, K., Carlson, C.L., Shumake, J., Beevers, C.G., 2021.
decision-making task. J. Abnorm. Psychol. 128, 234–244. https://doi.org/10.1037/ Symptom centrality and infrequency of endorsement identify adolescent depression
abn0000415. symptoms more strongly associated with life satisfaction. J. Affect. Disord. 289,
Gabarrell-Pascuet, A., García-Mieres, H., Giné-Vázquez, I., Moneta, M.V., Koyanagi, A., 90–97. https://doi.org/10.1016/j.jad.2021.02.064.
Haro, J.M., Domènech-Abella, J., 2023. The association of social support and Peng, A., Tang, Y., He, S., Ji, S., Dong, B., Chen, L., 2021. Association between loneliness,
loneliness with symptoms of depression, anxiety, and posttraumatic stress during the sleep behavior and quality: a propensity-score-matched case–control study. Sleep
COVID-19 pandemic: a meta-analysis. Int. J. Environ. Res. Publ. Health 20. https:// Med. 86, 19–24. https://doi.org/10.1016/j.sleep.2021.08.008.
doi.org/10.3390/ijerph20042765. Raykov, T., 1997. Scale reliability, Cronbach’s coefficient alpha, and violations of
Gijzen, M.W.M., Rasing, S.P.A., Creemers, D.H.M., Smit, F., Engels, R.C.M.E., De essential tau-equivalence with fixed congeneric components. Multivariate Behav.
Beurs, D., 2021. Suicide ideation as a symptom of adolescent depression. a network Res. 32, 329–353. https://doi.org/10.1207/s15327906mbr3204_2.
analysis. J. Affect. Disord. 278, 68–77. https://doi.org/10.1016/j.jad.2020.09.029. Salinas, J., Beiser, A.S., Samra, J.K., O’Donnell, A., DeCarli, C.S., Gonzales, M.M.,
Griffin, S.C., Mladen, S.N., Williams, A.B., Dautovich, N.D., Lageman, S.K., Aparicio, H.J., Seshadri, S., 2022. Association of loneliness with 10-year dementia
Dzierzewski, J.M., Perrin, P.B., Rybarczyk, B.D., 2021. Sleep disturbance mediates risk and early markers of vulnerability for neurocognitive decline. Neurology 98,
the association between loneliness and health in older Americans. Int. J. Behav. Med. e1337–e1348. https://doi.org/10.1212/WNL.0000000000200039.
28, 64–72. https://doi.org/10.1007/s12529-020-09897-2. Santos, I.S., Tavares, B.F., Munhoz, T.N., de Almeida, L.S.P., da Silva, N.T.B., Tams, B.D.,
Griffin, S.C., Williams, A.B., Ravyts, S.G., Mladen, S.N., Rybarczyk, B.D., 2020. Patella, A.M., Matijasevich, A., 2013. Sensibilidade e especificidade do Patient
Loneliness and sleep: a systematic review and meta-analysis. Health Psychol. Open 7, Health Questionnaire-9 (PHQ-9) entre adultos da população geral. Cad. Saúde
2055102920913235. https://doi.org/10.1177/2055102920913235. Pública. https://doi.org/10.1590/s0102-311x2013001200006.
Guerra-Balic, M., González-González, C.S., Sansano-Nadal, O., López-Dóriga, A., Schrempft, S., Jackowska, M., Hamer, M., Steptoe, A., 2019. Associations between social
Chin, M.-K., Ding, K., Yang, J., Durstine, J.L., 2023. Impact of COVID-19 lockdown isolation, loneliness, and objective physical activity in older men and women. BMC
on physical activity, insomnia, and loneliness among Spanish women and men. Sci. Publ. Health 19, 74. https://doi.org/10.1186/s12889-019-6424-y.
Rep. 13, 2912. https://doi.org/10.1038/s41598-023-30173-2. Seaman, S.R., White, I.R., 2013. Review of inverse probability weighting for dealing with
Gutkind, S., Gorfinkel, L.R., Hasin, D.S., 2022. Prospective effects of loneliness on missing data. Stat. Methods Med. Res. 22, 278–295. https://doi.org/10.1177/
frequency of alcohol and marijuana use. Addict. Behav. 124, 107115. https://doi. 0962280210395740.
org/10.1016/j.addbeh.2021.107115. Segrin, C., Domschke, T., 2011. Social support, loneliness, recuperative processes, and
Hodgson, S., Watts, I., Fraser, S., Roderick, P., Dambha-Miller, H., 2020. Loneliness, their direct and indirect effects on health. Health Commun. 26, 221–232. https://
social isolation, cardiovascular disease and mortality: a synthesis of the literature doi.org/10.1080/10410236.2010.546771.
and conceptual framework. J. R. Soc. Med. 113, 185–192. https://doi.org/10.1177/ Solmi, M., Veronese, N., Galvano, D., Favaro, A., Ostinelli, E.G., Noventa, V.,
0141076820918236. Favaretto, E., Tudor, F., Finessi, M., Shin, J.I., Smith, L., Koyanagi, A., Cester, A.,
Hofstede, 2024. Country comparison tool [WWW document]. Hofstede insights. URL. htt Bolzetta, F., Cotroneo, A., Maggi, S., Demurtas, J., De Leo, D., Trabucchi, M., 2020.
ps://www.hofstede-insights.com/country-comparison-tool. accessed 24. Factors associated with loneliness: an umbrella review of observational studies.
Holt-Lunstad, J., Smith, T.B., Baker, M., Harris, T., Stephenson, D., 2015. Loneliness and J. Affect. Disord. 271, 131–138. https://doi.org/10.1016/j.jad.2020.03.075.
social isolation as risk factors for mortality: a meta-analytic review. Perspect. Steptoe, A., Shankar, A., Demakakos, P., Wardle, J., 2013. Social isolation, loneliness,
Psychol. Sci. 10, 227–237. https://doi.org/10.1177/1745691614568352. and all-cause mortality in older men and women. Proc. Natl. Acad. Sci. U.S.A. 110,
Holt-Lunstad, J., Smith, T.B., Layton, J.B., 2010. Social relationships and mortality risk: a 5797–5801. https://doi.org/10.1073/pnas.1219686110.
meta-analytic review. PLoS Med. 7, e1000316. https://doi.org/10.1371/journal. Surkalim, D.L., Luo, M., Eres, R., Gebel, K., van Buskirk, J., Bauman, A., Ding, D., 2022.
pmed.1000316. The prevalence of loneliness across 113 countries: systematic review and meta-
Hughes, M.E., Waite, L.J., Hawkley, L.C., Cacioppo, J.T., 2004. A short scale for analysis. BMJ 376, e067068. https://doi.org/10.1136/bmj-2021-067068.
measuring loneliness in large surveys. Res. Aging. https://doi.org/10.1177/ Survey Monkey [WWW Document], 2024. SurveyMonkey Inc. URL. pt.surveymonkey.
0164027504268574. com (accessed 1.24).
IBGE, 2022. Censo 2022 [WWW Document]. IBGE. URL. https://censo2022.ibge.gov. Terracciano, A., Luchetti, M., Karakose, S., Stephan, Y., Sutin, A.R., 2023. Loneliness and
br/?localidade=BR. accessed 22. risk of Parkinson disease. JAMA Neurol. 80, 1138–1144. https://doi.org/10.1001/
Kroenke, K., Spitzer, R.L., Williams, J.B., 2001. The PHQ-9: validity of a brief depression jamaneurol.2023.3382.
severity measure. J. Gen. Intern. Med. 16, 606–613. https://doi.org/10.1046/ Valtorta, N.K., Kanaan, M., Gilbody, S., Ronzi, S., Hanratty, B., 2016. Loneliness and
j.1525-1497.2001.016009606.x. social isolation as risk factors for coronary heart disease and stroke: systematic
Leigh-Hunt, N., Bagguley, D., Bash, K., Turner, V., Turnbull, S., Valtorta, N., Caan, W., review and meta-analysis of longitudinal observational studies. Heart 102,
2017. An overview of systematic reviews on the public health consequences of social 1009–1016. https://doi.org/10.1136/heartjnl-2015-308790.
isolation and loneliness. Publ. Health 152, 157–171. https://doi.org/10.1016/j. Vancampfort, D., Lara, E., Smith, L., Rosenbaum, S., Firth, J., Stubbs, B., Hallgren, M.,
puhe.2017.07.035. Koyanagi, A., 2019. Physical activity and loneliness among adults aged 50 years or
Lim, M.H., Eres, R., Vasan, S., 2020. Understanding loneliness in the twenty-first century: older in six low- and middle-income countries. Int. J. Geriatr. Psychiatr. 34,
an update on correlates, risk factors, and potential solutions. Soc. Psychiatr. 1855–1864. https://doi.org/10.1002/gps.5202.
Psychiatr. Epidemiol. 55, 793–810. https://doi.org/10.1007/s00127-020-01889-7. von Elm, E., Altman, D.G., Egger, M., Pocock, S.J., Gøtzsche, P.C., Vandenbroucke, J.P.,
Lin, S.L., 2023. The “loneliness epidemic”, intersecting risk factors and relations to STROBE Initiative, 2007. The Strengthening the Reporting of Observational Studies
mental health help-seeking: A population-based study during COVID-19 lockdown in in Epidemiology (STROBE) statement: guidelines for reporting observational studies.
Canada. J. Affect. Disord. 320, 7–17. https://doi.org/10.1016/j.jad.2022.08.131. Ann. Intern. Med. 147, 573–577. https://doi.org/10.7326/0003-4819-147-8-
Lucke, J.F., 2005. The α and the ω of congeneric test theory: an extension of reliability 200710160-00010.
and internal consistency to heterogeneous tests. Appl. Psychol. Meas. 29, 65–81. Wenig, V., Heumann, E., Stock, C., Busse, H., Negash, S., Pischke, C.R., Heinrichs, K.,
Luhmann, M., Buecker, S., Rüsberg, M., 2023. Loneliness across time and space. Nat Rev 2023. Associations of loneliness with mental health and with social and physical
Psychol 2, 9–23. https://doi.org/10.1038/s44159-022-00124-1. activity among university students in Germany: results of the COVID-19 German
Lumley, T., 2020. In: Survey: Analysis of Complex Survey Samples. student well-being study (C19 GSWS). Front. Public Health 11, 1284460. https://
Manfro, P.H., Pereira, R.B., Rosa, M., Cogo-Moreira, H., Fisher, H.L., Kohrt, B.A., doi.org/10.3389/fpubh.2023.1284460.
Mondelli, V., Kieling, C., 2023. Adolescent depression beyond DSM definition: a WHO, 2022. Physical activity [WWW Document]. URL. https://www.who.int/news-roo
network analysis. Eur. Child Adolesc. Psychiatr. 32, 881–892. https://doi.org/ m/fact-sheets/detail/physical-activity, 10.19.23.
10.1007/s00787-021-01908-1. WHO ASSIST Working Group, 2002. The alcohol, smoking and substance involvement
Mann, F., Wang, J., Pearce, E., Ma, R., Schlief, M., Lloyd-Evans, B., Ikhtabi, S., screening test (ASSIST): development, reliability and feasibility. Addiction. https://
Johnson, S., 2022. Loneliness and the onset of new mental health problems in the doi.org/10.1046/j.1360-0443.2002.00185.x.

488