Research Design
A population-based, longitudinal prospective survey study was conducted with the general population in Hong Kong aged 18 and above. Quantitative telephone surveys were administered at three time points: T1 baseline, (December 2020–January 2021, before the implementation of the Vaccination Programme), T2 follow-up (June 2021–July 2021, about 4–5 months after the start of the Vaccination Programme), and T3 follow-up (21 December 2021 to 21 January 2022). For this study, only cross-sectional data collected from the T3 follow-up wave were used. During this period, a local outbreak of the Omicron variant virus was in its early stage with increasingly infected cases, and government announced to tighten social distancing measures since January 7, such as closing public entertainment venues and restricting dine-in arrangements to daytime before 6 p.m. with a maximum of two people per table.
Participants And Data Collection
The inclusion criteria were Hong Kong residents at the age of 18 or above and able to communicate in Cantonese or Putonghua. A structured questionnaire was administered via telephone interview by a team of trained research assistants. The telephone numbers were randomly selected using a multi-stage procedure as applied in previous population-based telephone surveys in Hong Kong. To select the randomized telephone numbers for the calls, a local directory covering prefixes of both the landline and mobile telephone numbers in Hong Kong was used. The prefixes randomly selected were used as ‘seeds,’ for generating another set of numbers, using the ‘last digit plus/minus one/two’ method, to form the second half of the telephone numbers for making the calls. After removing duplicated numbers, a total of 6000 telephone numbers randomly sequenced were used for further random selection. For telephone calls to a number with more than one eligible individual in the household, the last birthday method was adopted to identify the participant. Verbal informed consent was obtained from each of the participants. Research ethics approval was obtained from the Human Subject Ethics Sub-Committee of The Hong Kong Polytechnic University (Approval Number: HSEARS20200814002). In the first wave (T1) of data collection, a total of 1255 completed cases were obtained, followed by 1003 in the second wave (T2), and 803 in the third wave (T3). The sample obtained was weighted, based upon the gender and age distribution of the Hong Kong population reported by the Census and Statistics Department in 2021.
Measures
Depression is the dependent variable measured by the Chinese version of the 9-item Patient Health Questionnaire-9 (PHQ-9).[20] The psychometric properties of the scale in Chinese populations were previously confirmed.[20] For each of the item, the participants were asked to rate their answer along a four-point scale ranging from 0 (not at all) to 3 (nearly every day). A higher total score, ranging between 0 to 27, indicates a higher level of depression. A total score of 0–4 indicates no depressive symptoms, 5–9 mild depressive symptoms, 10–14 moderate depressive symptoms, 15–19 moderately-severe depressive symptoms, and 20–27 severe depressive symptoms.[21] A Cronbach’s alpha coefficient of .961 was reported in this study. A cut-off score of 5 and above was used to differentiate the normal ones (0–4) and ones with mild or higher level of depression (5 and above).[21]
Independent Variables
Hassles brought by the COVID-19 pandemic referred to the level of disturbance experienced and was measured by asking the participants to indicate the level of disturbance experienced using a five-point scale ranging from 1 (none) to 5 (all the time). The 10 items covered daily areas related to leisure activities, economy, employment or study, social activities and interactions with family members, use of services, going out, travelling, and use of computer technologies, etc. A higher total score along the scale range of 10 to 50, indicated a higher level of hassles experienced.
Awareness of pandemic was represented by a few variables measuring perceived severity toward the pandemic, level of concern toward the pandemic, perceived risk of being infected by COVID-19, current experience in contracting COVID-19, and previous experience of being infected in past pandemics.
Perceived severity of COVID-19 was measured by the items ‘COVID-19 is a severe disease’. Concern over COVID-19 was measured by the item ‘You are concerned about having COVID-19’. Current experience with COVID-19 was measured by three items: 1) ‘You have family member infected or suspected to be infected by COVID-19’, 2) ‘You have friends infected or suspected to be infected by COVID-19’, and 3) ‘Someone in the building you reside is infected or suspected to be infected by COVID-19’. Previous experience of being infected in past pandemics was measured by two items: 1) ‘You were infected by SARS, avian influenza, or swine flu’, and 2) ‘You had family members or friends infected by SARS, avian influenza, or swine flu’. All items were rated along a 5-point Likert scale, ranging from ‘strongly disagree’ to ‘strongly agree’. Perceived risk was measured by the question asking the participants to rate the chance that the participant himself/herself be infected by COVID-19 along a five-point scale ranging from ‘no chance at all’ (1) to ‘very great chance’ (5).
Two trust variables were included to represent trust in authority and trust in healthcare professionals in controlling the spread of COVID-19. The former was measured by two items asking the participants to rate their levels of trust in the politicians and governmental officials respectively. Trust in healthcare professionals was measured by one item for rating levels of trust in healthcare professionals. Each item was rated on a 11-point Likert scale, from 0 (lowest level of trust) to 10 (highest level of trust). Demographic variables examined included gender, age, highest level of education, and economic activity status (i.e., being active or inactive).
Data analysis
SPSS 26.0 was used for data analysis. In inferential data analysis, p-values smaller than 0.05 were deemed statistically significant. Descriptive statistics were used to examine the frequency distribution and mean of the variables concerned. To examine the predictors of depression, hierarchical logistic regression analysis was conducted. The dependent variable was depression, and the scores were grouped in to being depressed (using the cut-off point of 5). The first blocks of independent variables included: (i) demographic background (i.e., gender age, education attainment, and economic activity status); (ii) perceived severity of the COVID-19 pandemic, concern toward the COVID-19 pandemic, current experiences of COVID-19, previous experiences of other pandemics and epidemics, perceived risk of being infected, and hassles due to COVID-19, and iv) trust in politicians and government officials, and trust in medical professionals. Hierarchical linear regression was employed to test the moderating effect and conduct the simple slope test. To avoid multicollinearity, variables to be test were centralized.