The association between living alone and depressive symptoms and the role of pet ownership among Japanese workers | BMC Public Health

Study design

As part of a large cohort study (the Japan Epidemiology Collaboration on Occupational Health [J-ECOH] Study [11,12,13], a self-administered questionnaire survey was conducted at five participating companies of the J-ECOH via a paper-based questionnaire (three companies) or a web-based questionnaire (two companies) between April 2018 and February 2021. In the questionnaire, we asked about health-related lifestyles (including pet ownership and living arrangements) and depressive symptoms. We also obtained the data of annual health checkups that were conducted by the employers during the same fiscal year as the questionnaire survey.

We explained the objective of the survey in the first page (for the paper-based questionnaire) or window (for the online questionnaire) of the questionnaire. Consent was confirmed if a potential participant signed and returned the paper-based questionnaire or clicked the consent box of and completed the online questionnaire.

Study participants

Among the 17,078 workers whose health checkup data was available, 12,847 participated in the questionnaire survey (response rate: 75.2%). We excluded those with missing information about pet ownership, depressive symptoms, or covariates (n = 84), leaving 12,763 participants in the main analysis.

Outcome

Depressive symptoms were assessed using the 11-item Center for Epidemiological Studies – Depression Scale (CES-D-11). Participants were asked to respond to questions on 11 symptoms associated with depressive disorder that participants had experienced in the previous week with four response options including “rarely or none of the time (< 1 day)”; “some or a little of the time (1–2 days)”; “occasionally or a moderate amount of the time (3–4 days)”; and “most or all of the time (≥ 5 days)”. After reverse scoring two positive questions (“I enjoy every day” and “I can spend my time being satisfied about life”), we assigned a score ranging from 0 to 3 to each response and calculated the total score (0–33). The Cronbach’s alpha was 0.80 in this study. Then in accordance with previous studies [14,15,16,17], a score of ≥ 9 was defined as having depressive symptoms.

Exposure

We asked participants whether they lived with one or more of the following (i.e., allowing multiple responses): spouse (wife/husband); child(ren) of preschool age; child(ren) of primary school age or older; parent(s); sibling(s); grandparent(s); pet(s); and other(s). Those who answered in the negative to all the options except for that of a pet were defined as those living alone while those who answered in the affirmative to the question regarding pet ownership were defined as those with a pet. Based on these two variables (i.e., living alone and pet ownership), the participants were classified according to the following four categories: (1) living alone with pet(s); (2) living alone without a pet; (3) living with others and pet(s); and (4) living with others but without a pet.

Covariates

Information on sociodemographic variables was obtained via the questionnaires, including: age (in years); sex (male; female); marital status (married; not married); employment status (upper management; middle management; other); and educational background (9–12 years; 13–16 years; ≥ 17 years).

We also collected questionnaire information on lifestyle parameters that are known to be associated with depression/depressive symptoms (i.e., smoking and alcohol consumption) [18, 19]. Specifically, we asked the participants their smoking status and categorized them into the following groups, i.e., never, former, and current smoker. Alcohol consumption was estimated based on the questionnaire information on consumption frequency (i.e., never drink/quit; 1–3 days/month; 1–2 days/week; 3–4 days/week: 5–6 days/week; daily) and amount consumed per occasion (i.e., 0.5 go; 1 go; 1.5 go; 2 go; 2.5 go; 3 go; 3.5 go; 4 go or more; 1 go is the traditional Japanese unit and is equivalent to approximately 20 g of ethanol). We assigned the values to each response option for consumption frequency ranging from never drink/quit = 0; 1–3 days per month = 0.5; 1–2 days per week = 1.5; 3–4 days per week = 3.5; 5–6 days per week = 5.5; to daily = 7, which were then multiplied with the amount to compute daily alcohol consumption. The participants were categorized according to the following groups: did not drink, drank < 1 go/day; drank 1–1.9 go/day; drank ≥ 2 go/day.

Statistical analysis

To examine the association of the combinations of living arrangements and pet ownership in relation to depressive symptoms, we ran a Poisson regression analysis with a robust variance estimator to estimate prevalence ratios (PRs) and corresponding 95% confidence intervals (CIs) while adjusting for sociodemographic and lifestyle-related variables and study site (five sites) [20, 21]. We confirmed that there was no evidence of multicollinearity (i.e., the variance inflation factor was 1.88).

In accordance with a previous study in Japan that examined the association between living arrangements and the risk of suicide, and found an elevated risk of suicide even among those who did not live alone [10], we categorized participants according to 16 categories defined by the combinations of living arrangements, that is, living with spouse, children, and birth family (i.e., biological parents or siblings or grandparents) and pet ownership, and conducted a sensitivity analysis. Those who responded that they lived with other(s) but without specific details were excluded from this analysis (n = 217), resulting in an analytic sample of 12,569 participants.

Statistical analysis was conducted using the Stata® Version 16.0 statistical software package (StataCorp., College Station, Texas, USA).