This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Mental Health, is properly cited. The complete bibliographic information, a link to the original publication on http://mental.jmir.org/, as well as this copyright and license information must be included.
A depression-awareness campaign delivered through the Internet has been recommended as a public health approach that would enhance mental health literacy and encourage help-seeking attitudes. However, the outcomes of such a campaign remain understudied.
The main aim of this study was to evaluate the effectiveness of an online depression awareness campaign, which was informed by the theory of planned behavior, to encourage help-seeking attitudes for depression and to enhance mental health literacy in Hong Kong. The second aim was to examine click-through behaviors by varying the affective facial expressions of people in the Facebook advertisements.
Potential participants were recruited through Facebook advertisements, using either a happy or sad face illustration. Volunteer participants registered for the study by clicking on the advertisement and were invited to leave their personal email addresses to receive educational content about depression. The participants were randomly assigned into two groups (campaign or control), and over a four consecutive week period, received either the campaign material or official information developed by the Hospital Authority in Hong Kong. Pretests and posttests were conducted before and after the campaign to measure the differences in help-seeking attitudes and mental health literacy among the campaign and control groups.
Of the 199 participants that registered and completed the pretest, 116 (55 campaign and 62 control) completed the campaign and the posttest. At the posttest, we found no significant changes in help-seeking attitudes between the campaign and control groups, but the campaign group participants demonstrated a statistically significant improvement in mental health literacy (
The present study provides evidence that an online campaign can enhance people’s mental health literacy. It also demonstrates the practicality and effectiveness of an online depression awareness campaign using a Facebook-based recruitment strategy and distribution of educational materials through emails. It is important for future studies to take advantage of the popularity of online social media and conduct evaluative research on mental health promotion campaigns.
Mental illness is a substantial contributor to the global burden of disease [
Along the same line, some research findings suggest that an increase in mental health literacy can affect one’s help-seeking intention [
Health promotion campaigns delivered through the Internet have become increasingly more common recently [
Nevertheless, a systematic review revealed that there were very few effectiveness studies of online campaigns intended to encourage young people’s help-seeking for mental health, and indeed most demonstrated no statistically significant impacts [
This study aims to evaluate the effectiveness of an online campaign targeted at enhancing help-seeking attitudes for depression in Hong Kong. Specifically, the study seeks to examine whether such a campaign can promote participants’ help-seeking attitudes for depression and mental health literacy. This study is the final stage of a multiple-phase, mixed-method research project aimed at developing a universal media campaign to encourage a help-seeking attitude for depression in Hong Kong. Hong Kong is a business-centric and fast-changing city with dynamic economic development and wide income gaps. These modern city characteristics have marked implications for the population’s mental health status [
Theoretical models are needed to guide campaign development and strategies [
In addition, the quality of the campaign material was examined. One central question was how to design visual material for online display that can draw the largest extent of viewers’ attention to the material? Previous studies have demonstrated that people with depression have an attentional bias against negative and sad facial expressions [
The study protocols were approved by the Human Research Ethics Committee for Non-Clinical Faculties, The University of Hong Kong. Trial registration for this kind of health attitude research is not required. We followed the Checklist for Reporting Results of Internet E-Surveys (CHERRIES) for reporting the development and findings of online surveys [
Participants who identified themselves as Hong Kong residents using traditional Chinese language and aged 18-59 years old in their Facebook profiles were recruited through paid advertisements on Facebook. Facebook has more than 845 million users around the world [
With reference to the recruitment procedure of the previous studies, the cost-per-click advertising model was used [
Two Facebook advertisements were posted during the period from January 29, 2013 to March 3, 2013, with more than 26 million prevalence of impressions representing the incidence of appearance on Facebook (ie, people may see multiple impressions of the same advertisement). A total of HK $17,887.87 (roughly US $2300) was spent on both advertisements that gained a total of 5405 clicks. The main advertisement copy was posted with the headline “Say no to depression” in Chinese. On the side, there was a short paragraph asking:
What would you do if you discovered that your friend might have depression? Click here and we will provide you with four episodes of electronic material related to depression for free.
With the same copy, two versions of the advertisements were made with the following attached pictures (1) a depressed woman covering her face (sad face), and (2) three cheerful young ladies (happy face) (
The Facebook advertisement (
The study’s internal validity was one of our concerns as participant recruitment and campaign execution were carried out completely on the Internet. In order to take into account possible duplications and to minimize the number of fake accounts or registrations, which are widespread concerns for online surveys and campaigns, participants were required to manually input the email registration after clicking into the Facebook advertisements. This email registration process and the following steps of sending questionnaires and campaign material through email helped minimize the chances of false identities and accounts. The IP addresses were also tracked to minimize duplication [
Using Cronbach’s alpha and confirmatory factor analysis, the reliability and validity in the measurements of the items were tested in a separate study that was reported elsewhere [
The duration of the campaign was six weeks. Participant registration, which was also the end of the posting period for the Facebook advertisement, was due on March 3, 2013. The web link of the pretest was then sent out for all registered participants’ email accounts. The cut-off date for completion of the pretest was March 18, 2013, with prompt reminders sent beforehand. Those who failed to fill in the pretest were considered as dropouts from the campaign.
Every Monday for the following four weeks, a short paragraph of campaign material on depression was sent out in email format to participants. Participants had been randomly assigned to two groups that received either originally developed material for this campaign or the official mental health material prepared by the Hong Kong Hospital Authority, which was used as a control comparison. Participants were not aware of which group they belonged to. The full set of online questionnaires is provided (
The happy and sad face images and their targeted Facebook users.
Example of the Facebook advertisement.
The campaign material was originally designed by the first author, who has a bachelor’s degree in journalism and has received mental health first-aid training, and was reviewed by the second author, who is a clinical psychologist and mental health researcher. All campaign material was written in Chinese. Four episodes of campaign messages were created with reference to the factors in the help-seeking theoretical model with each week featuring a background of depression, attitudes toward help-seeking, subjective norms, and perceived behavioral control [
An external web link was printed at the end of the email body for optional further reading, which provided additional information related to the week’s message. The additional information was adopted from existing health information sourced from the World Health Organization, Hong Kong Hospital Authority, and academic studies related to depression [
On the Monday of the sixth week, a link to the posttest was sent out in email format to all participants. In the final email, participants were advised to fill in the posttest only after they had finished reading the previous four weeks’ material. Participants were given a week to fill in the posttest. The posttest was due at the end of week 6. Both pretest and posttest were voluntary surveys without incentive. Participants who did not complete the questionnaires or did not fill in the required material were considered to be dropout participants.
The chi-square test was used to test the differences between campaign and control groups for the variables described in
Of the 383 Facebook users who had consented and registered with their personal email address to join the campaign, 197 (51/100, 51.0%) completed the pretest and were included in the campaign. They were randomly assigned into one of the two groups with 98 in the campaign group and 99 in the control group. After receiving four weeks of electronic campaign material, of the 116 participants that completed the posttest, 54 (55/100, 55.0%) were in the campaign and 62 (63/100, 63.0%) in the control group. The participants’ flow of this study is shown in
Flow chart of the study.
No significant difference in demographic characteristics was found between the two groups (
Participants were mainly young adults, 18-29 years old, who represent the most frequent Internet and social media users. Over 50% of them had attained the secondary school education level. Also, about 30% of the participants reported that they had received a diagnosis of depression before. More than 70% of the participants reported suffering from mild to extremely severe depressive symptoms.
Demographic characteristics of participants of the campaign and control groups.
Demographic characteristics | Campaign (N=55) | Control (N=62) | Difference (%) |
|
|
n (%) | n (%) | ||||
|
|
|
|
|
0.13 |
|
Male | 23 (43) | 17 (27) | 16 |
|
|
Female | 31 (57) | 45 (73) | -16 |
|
|
|
|
|
|
0.84 |
|
18-29 | 24 (44) | 22 (35) | 9 |
|
|
30-39 | 6 (11) | 8 (13) | -2 |
|
|
40-49 | 10 (19) | 11 (18) | 1 |
|
|
50-59 | 12 (22) | 19 (31) | -9 |
|
|
60 or above | 2 (4) | 2 (3) | 1 |
|
|
|
|
|
|
0.35 |
|
Primary school | 0 (0) | 4 (6) |
-6 |
|
|
Secondary school | 30 (56) | 32 (52) | 4 |
|
|
Undergraduate | 19 (35) | 21 (34) | 1 |
|
|
Postgraduate | 5 (9) | 5 (8) | 1 |
|
|
|
|
|
0.26 |
|
|
Yes | 13 (24) | 22 (35) | -11 |
|
|
No | 41 (76) | 40 (65) | 11 |
|
|
|
|
|
0.70 |
|
|
Yes | 25 (46) | 32 (52) | -6 |
|
|
No | 29 (54) | 30 (48) | 6 |
|
|
|
|
|
0.88 | |
|
Yes | 33 (61) | 36 (58) |
3 |
|
|
No | 21 (39) | 26 (42) | -3 |
|
|
|
|
|
0.36 | |
|
Normal | 13 (24) | 15 (24) | 0 |
|
|
Mild | 3 (6) | 6 (10) | -4 |
|
|
Moderate | 12 (22) | 16 (26) | -4 |
|
|
Severe | 10 (19) | 4 (6) | 13 |
|
|
Extremely severe | 16 (30) | 21 (34) | -4 |
|
While comparing the click-through rate to the optional further readings between groups, 69 participants from the campaign group (70%, 70/100) clicked through at least once during the four weeks, whereas only 40 participants from the control group (40%, 40/100) clicked through, suggesting that the participants in the campaign group were more willing to read further information than those of the control group (χ2
1=16.7,
The primary and secondary outcome measures in the help-seeking model are presented in
Posttest scores of the primary and secondary outcome measures of the help-seeking models in the campaign and control groups.
Outcome measure | Campaign (N=54) | Control (N=62) | |
|
|
Mean (SD) | Mean (SD) |
|
|
|
|
|
Attitude | a2.05 (0.46) | 2.07 (0.55) |
|
Subjective norm | 2.84 (1.46) | 2.56 (1.61) |
|
Perceived behavioral control | 2.49 (1.31) | 2.29 (1.33) |
|
Intention | 3.01 (1.48) | 2.73 (1.58) |
|
|
|
|
|
Perceived barrier | 4.35 (1.48) | 4.26 (1.69) |
|
Attitude toward treatment | 2.58 (1.11) | 2.47 (1.21) |
|
Action and view toward family and friends with depression | 3.40 (1.10) | 2.73 (1.58) |
a1 is the highest likelihood to seek help, and 7 is the lowest
We also tested which picture, happy- or sad-looking, received more clicks to join the campaign. The advertisement with the happy face gained 2936 clicks with 12,591,361 impressions, representing a 0.023% (0.023/100) click-through rate into the advertisement and HK $3.05 (US $0.39) per click. The sad face advertisement gained 2469 clicks out of 13,412,509 impressions, amounting to a 0.018% (0.018/100) click-through rate and HK $3.62 (US $0.46) per click. The total spending of the happy and sad face advertisements was HK $8,942.89 (US $1,146.50) and HK $8,944.98 (US $1,146.80), respectively. The click-through rate of the happy face picture was statistically significantly higher than the one of the sad face (χ2
1 = 75.1,
This study demonstrated the possibility of delivering an online mental health campaign that promotes help-seeking for depression. Although no significant change was detected in the primary outcomes of the campaign, which were designed to closely following the TPB, the increase of mental health literacy in the intervention group provides supportive evidence in the educational value of such campaigns. The result is consistent with a similar campaign evaluation study that found that social media campaigns can improve mental health literacy but have limited impact on attitudinal outcomes [
It is important to note that our campaign was particularity attractive to individuals with mild to severe depressive symptoms, irrespective of who were or were not receiving treatment. We found that 30% (30/100) of the participants had a diagnosis of depression, and more than 70% of participants reported that they were mildly or severely depressed. This might reflect that participants who were suffering from depressive episodes were particularly attracted to the depression-related advertisement on Facebook and campaign material. Since nearly half of the participants have sought professional help for their depressive episodes, their help-seeking intention might have been shaped by their own personal illnesses and treatment experiences. This may partly explain the non-significant changes in the primary outcomes of the help-seeking intention but rather significant improvement in media literacy [
Another interesting finding concerns the different responses by those receiving the two pictures in the Facebook advertisement. The Facebook advertisement with a happy face picture gained more clicks than did the one with a sad face. According to our analysis, no significant difference in demographics was found between those who had clicked through the two advertisements. However, we found more people who had a diagnosis of depression among those who clicked into the happy face advertisement than those who clicked into the sad face, which runs contrary to previous laboratory findings of depressive patient’s differential attention to emotional pattern [
Facebook is increasingly being used for participant recruitment for research, especially targeting specific groups like adolescents, groups with specific health needs, patients, and caregivers [
The following are limitations worth noting. The recruited subjects may be biased toward certain characteristics (ie, relatively more female respondents and depressed patients). The dropout rate of the campaign can be considered high, which is not uncommon in online subject recruitment. The results might not represent those who dropped out from the email registration and the pretest. Due to the complications of reading campaign material and receiving pretest and posttest data through email, the dropout rate was higher than expected, but duplicated and fake responses were minimized since each registration and all email content was accomplished by a unique link and email message, which can ensure good data quality.
The loss of samples resulted in lower statistical power and might partly account for some of the non-significant statistical tests. In an online setting, there is no simple way to ensure that the subject has read and fully comprehended the material. But this factor should not alter the final results because both intervention and control groups should have the same impact, if any. Although it is acknowledged that help-seeking intentions may not directly lead to actual help-seeking behavior, effective preventive measures are essential in response to the prevalent reluctance regarding help-seeking for depression [
With the small sample size of our study, the conclusions may not be optimally representative to the public’s view of help-seeking and depression. However, using a theory-driven and culturally-specific campaign design strategy is recommended because it enables systematic understanding of public knowledge and allows room for the development of audience-oriented material. Although the change in help-seeking intentions was not significant in the current study, future studies can build upon our findings to replicate the study, revise the model, or carry out tests with a larger and more representative sample size.
Under the backdrop that clinical depression is projected to become a major global disease burden in the future, encouraging help-seeking for depression should be an essential aspect for health service development. Online media can help reach large numbers of participants and get help-seeking messages across to a population. In this study, an online media campaign, with the support of a theory-driven approach and customized campaign material, was used as an innovative and practical platform for a mental health campaign to encourage help-seeking for depression and to increase the individual’s media literacy. Further research is needed to confirm the results.
Full set of online questionnaires.
Campaign Material (Originally in Chinese).
depression anxiety stress scale
Internet Protocol
theory of planned behavior
This study and AH’s Postgraduate Scholarship were supported by the General Research Fund, Research Grants Council, Hong Kong (HKU 744410H). The writing of this article was partly funded by the approval of sabbatical leave to PWCW granted by the Faculty Human Resource Committee (FHRC) of the Faculty of Social Sciences, The University of Hong Kong. We also acknowledge Mr. Chung-Hong Chan’s assistance in conducting statistical tests.
None declared.