Abstract
Background:
Individuals with depression may not seek treatment for their symptoms due to several types of barriers to treatment. In support of the growing research on mental health care access and the role of social media, this study aimed to increase knowledge of these barriers among social media users.
Methods:
Participants were recruited from several social media platforms, including Instagram, Facebook, Twitter, Reddit, Tumblr, and online depression forums. Eligible participants had endorsed having posted about feeling sad or depressed on social media, or followed social media groups that post about depression-related topics. Participants completed an online survey about their depression symptoms, interest in treatment, and potential barriers to accessing treatment.
Results:
Of the participants reaching criteria for depression, those with major depression were more likely to seek out treatment, to report an unmet need for treatment, and have a higher risk of suicide. For participants with major depression, barriers to treatment were more likely to be attitudinal, while participants with mild depression experienced more structural barriers.
Conclusions:
This study demonstrates several barriers to treatment that occur for individuals struggling with depression, and that online platforms are effective mediums to recruit individuals with depression symptoms who seek mental health support.
Keywords: depression, social media, barriers, mental health
Introduction
Depression is a mental illness that impacts millions of people in the United States each year. In a survey of over 46,000 Americans, researchers found that over one in twelve individuals surveyed screened positive for depression but less than two thirds of that subgroup received any sort of treatment for depression (Olfson et al., 2016). In addition, there are significant different utilizations of treatment among different socioeconomic groups, with the lowest being among men, the uninsured, and racial/ethnic minorities. The risk of untreated depression is immense. Studies have illustrated the presence of psychiatric disorders within seventy to ninety percent of individuals dying of suicide, with specifically depression in fifty to seventy five percent of those instances (Hawton et al., 2013; Cavanaugh et al., 2003).
There are many reasons as to which an individual may not seek treatment for depression. Studies have found that attitudinal barriers are the most significant barriers in most cases of untreated depression. Other common barriers to treatment are structural barriers such as finances or difficulties with availability (Andrade et al., 2014). The stigma of mental illness, whether the perceived stigma that individuals may feel from others, or the self-stigma of one’s own internalized feelings about mental health, result in social/cultural factors that can contribute to attitudinal barriers (Andrade et al., 2014; Corrigan et al. 2002). Thus, it is incredibly important to identify new strategies that can overcome some of these barriers and engage individuals in treatment.
Social media platforms, ranging from Facebook to Twitter, are modern mechanisms for persons struggling with mental illness to discuss related suffering and receive informal help (Yan & Tan, 2014). The emotional content discussed on social media potentially differs from content featured in face-to-face conversations due to the presence of fewer real-world barriers, such as stigma (Lachmar et al., 2017; Burns et al., 2009). Therefore, social media may serve as a conduit for researchers and practitioners to identify persons with symptoms of depression and to use digital interventions to connect persons to treatment resources (Kelleher et al., 2018).
In support of the growing research on mental health care access and the role of social media, this study aimed to increase knowledge on barriers to treatment for depression among social media users. The study specifically examined characteristics of and barriers to treatment among social media users who were currently not in treatment and across severity of depression symptoms. Results have implications for engaging social media users reporting depressive symptoms in technology-based treatment interventions.
Methods
Study sample and recruitment
From October 2016 through August 2018, we intermittently recruited individuals who were networking about depression-related content on social media to take a cross sectional survey about their mood and depressive symptoms as well as their social media use. Participants were recruited from several social media platforms, including Instagram, Facebook, Twitter, Reddit, Tumblr, and online depression forums. Several recruitment methods were used, including private messaging individuals who were networking on depression-focused groups (Reddit) or who were posting about depression (Twitter, Tumblr), posting about our study on depression-focused groups (Reddit, other online forums), and using advertisements targeted to individuals networking about depression-related topics (Instagram, Twitter, Reddit, Facebook).
Recruitment materials included a link to our study website, where individuals could then click on a link taking them to our eligibility survey. Eligible participants were ≥15 years old, U.S. residents, and endorsed having posted about feeling sad or depressed on social media, or followed/subscribed to social media accounts/groups that post about depression-related topics. Those who were eligible provided online consent and continued to the full survey. Due to minimal risks associated with taking this online survey, parental consent was waived for 15 to 17 year olds. Qualtrics (Qualtrics, 2019) was used to develop and distributed the online survey, which could be completed using either a computer or mobile device. Participants were incentivized with a 1 in 15 chance to win a $100 Amazon.com gift card. This study procedure was reviewed and approved by the Institutional Review Board at the principal investigator’s home institution.
A total of 643 participants were eligible and consented to participate in our survey. While we used tactics to help prevent machine responses (Captcha) and repeat survey takers (Qualtrics’ feature “Prevent Ballot Box Stuffing” which prevents participants from taking the survey from the same web browser), we used additional data cleaning steps to ensure high quality responses for analysis. We first removed individuals who did not progress at least 50% of the way through the survey (n=210) because they would not have completed important survey items (e.g., depression symptoms). We also removed 21 individuals who took <8 minutes to take this survey (lowest 5th percentile of completion times), as short completion times are a good method to indicate meaningless survey data (Leiner, 2013). Finally, we also removed 23 participants who had illogical response patterns or duplicate responses. This resulted in a total of 389 participants for analysis (median survey completion time 19.7 minutes, inter-quartile range [IQR] 15.4–27.1 minutes).
We further focused the current study on participants with major depression and who had depression symptoms in the past two weeks (n= 165).
Measurement
Depression symptoms were assessed using the Patient Health Questionnaire (PHQ-9; Kroenke et al., 2001) scale, a nine-item instrument used to screen for depression severity based on the DSM-IV depression criteria. The PHQ-9 score ranges from 0 to 27, with a threshold score of ≥ 10 indicating probable major depression (sensitivity and specificity both 88%) and ≥ 5 indicating probable mild depression (Kroenke et al., 2001). The PHQ-9 has demonstrated excellent internal reliability (for example, α=.89 in a primary care setting; α=.86 in an obstetrics–gynecology clinic), test retest reliability, validity, and feasibility (Kroenke, et al., 2001; Wulsin, Somoza & Heck, 2002). For these reasons, the PHQ-9 is often used in both research and clinical practice settings (Kroenke & Spitzer, 2002).
Participants were asked whether they are “currently seeking mental health treatment”. Those who gave a denial answer were subsequently asked “In the last 12 months, was there ever a time when you thought you should seek treatment for depression, but you did not go?” to assess perceived unmet need for treatment. Those giving an affirmative response were queried about the reasons for not seeking the treatment. The items used to assess the barriers to treatment were modified from the treatment utilization section of National Institute on Alcohol Abuse and Alcoholism (NESARC) questionnaire. The NESARC is a national survey conducted by the National Institute on Alcohol Abuse and Alcoholism (NIAAA; Grant et al., 2014) and has been widely cited (Cohen, Feinn, Arias, & Kranzler, 2007; Kerridge et al., 2017; Vaeth, Wang-Schweig, & Caetano, 2017). Table 2 details the items. Guided by the categorization of barriers to substance use treatment by Kaufmann et al. (2014), barriers were grouped into three categories: (1) structural, (2) financial, and (3) attitudinal (concerning stigma, treatment, and depression).
Table 2.
Characteristics of participants who perceived unmet need for depression treatment, by severity of depression (N = 165 unless otherwise noted)
| Total Sample n (%) |
Major depression (n=87) n (%) |
Mild depression symptoms (n=78) n (%) |
p | |
|---|---|---|---|---|
| Barriers to treatment | ||||
| Attitudinal-Treatment | 106 (64) | 66(76) | 40 (51) | .001 |
| Attitudinal-Disorder | 97 (59) | 51 (59) | 46 (59) | .963 |
| Attitudinal-Stigma | 89 (54) | 58 (67) | 31 (40) | .001 |
| Structural | 82 (50) | 36 (41) | 46 (59) | .024 |
| Financial (n=126; adults only) | 58 (46) | 29(48) | 29 (45) | .742 |
| Number of barriers to treatment | .234 | |||
| No barriers | 8 (5) | 2 (2) | 6 (8) | |
| Few barriers (1–2 barriers) | 66 (40) | 34 (39) | 32 (41) | |
| Moderate barriers (3–5 barriers) | 91 (55) | 51 (59) | 40 (51) | |
| Received treatment for mental health in lifetime? | .047 | |||
| Treatment at some point in lifetime | 108 (65) | 63 (72) | 45 (58) | |
| No prior treatment | 57 (35) | 24 (30) | 33 (42) | |
| Risk of suicide | <.001 | |||
| High | 126 (76) | 76 (87) | 50 (64) | |
| Low | 39 (24) | 26 (20) | 18 (44) | |
| Demographics | ||||
| Gender (n=157) | .879 | |||
| Male | 103 (66) | 54 (65) | 49 (66) | |
| Female/Otherc | 54 (34) | 29 (35) | 25 (43) | |
| Age (n=156) | .014 | |||
| 15–17 | 37 (24) | 26 (32) | 11 (15) | |
| 18+ | 119 (76) | 56 (68) | 63 (85) | |
| Race/Ethnicity (n=156) | .030 | |||
| Non-Hispanic White | 92 (59) | 55 (67) | 37 (50) | |
| Other | 64 (41) | 27 (33) | 37(50) | |
| Student status (n=157) | .849 | |||
| Yes | 73 (47) | 38 (46) | 35 (47) | |
| No | 84 (54) | 45 (54) | 39 (53) | |
| Employ status (n=157) | .098 | |||
| Yes | 91 (58) | 43 (52) | 48 (65) | |
| No | 66 (42) | 40 (48) | 26 (35) | |
| Annual household income (n=155) | .004 | |||
| <$25,000 | 75 (48) | 50 (60) | 25 (35) | |
| $25,000 −$74,999 | 55 (35) | 25 (30) | 30 (42) | |
| ≥$75,000 | 25 (16) | 8 (10) | 17 (24) | |
| Recruiting social media platform | 0.01 | |||
| 31 (19) | 18 (21) | 13 (17) | ||
| Forums | 1 (1) | 1 (1) | 0 | |
| 48 (29) | 27(30) | 21 (29) | ||
| 61 (37) | 22 (25) | 39 (50) | ||
| Teenline | 2 (1) | 2 (1) | 0 | |
| Tumblr | 11 (7) | 8 (9) | 3 (4) | |
| 11 (7) | 9 (10) | 2 (3) |
Demographics, risk of suicide, and treatment history were included as covariates, many of which have been associated with experiencing barriers to depression treatment (Alegria et al., 2008; Givens et al., 2007; Tija et al., 2005). Demographic variables included gender, age, race/ethnicity, school status, employ status and annual household income.
Statistical Analysis
Data were analyzed using SAS Version 9.4 for Windows (SAS, 2013). Analyses proceeded in several steps. First, we compared three mutually exclusive groups on demographic characteristics and risk of suicide: those who did not seek depression treatment and did not perceive a need, those who did not seek treatment and perceived a need for such treatment (perceived unmet need), and those who sought treatment. Those who sought treatment and those with perceived unmet need were compared to those who did not seek treatment or perceived unmet need as the reference category using separate Pearson chi-square test. Next, we compared two severity of depression groups (major depression group and mild depression group) among individuals who reported having perceived unmet need for depression treatment on demographic characteristics as well as the number of barriers to treatment, treatment history, and the risk of suicide using Pearson chi-square test. Finally, we used logistic regression models to compare the barriers to treatment between the two depression groups. Both unadjusted and adjusted analyses were conducted. We first adjusted demographic covariates and then further adjusted suicide risk and treatment history.
Results
Participant characteristics
Of the 348 individuals with depression in our sample, half (51%) met the PHQ-9 criteria for major depression, and the other half (49%) met the criteria for mild depression. The majority of participants were male (70%), non-Hispanic White (58%), and the median age was 22 years (minimum 15, maximum 70). Characteristics of the sample by the severity of depression and treatment seeking status are shown in Table 1. Among individuals with major depression, those who sought treatment for depression were more likely than those who did not perceive such need or did not seek treatment to identify as non-White (49% vs. 33%, p<0.05). Additionally, among individuals with mild depression, those who perceived unmet need for depression treatment are more likely than those who did not perceive such need or did not seek treatment to be employed (65% vs. 40%, p<0.05) and those who sought treatment for depression were more likely than those who did not perceive such need or did not seek treatment to have high risk of suicide (77% vs. 64%, p<0.05).
Table 1.
Characteristics among individuals with major depression and mild depression symptoms who sought depression treatment or perceived an unmet for such treatment (N=348 unless otherwise noted)a
| Major depression (n=179) | Mild depression symptoms (n=169) | |||||
|---|---|---|---|---|---|---|
| Did not seek treatment or perceive unmet need (n=14) n (%) |
Perceived unmet need (n=87) n (%) |
Sought treatment (n=78) n (%) |
Did not seek treatment or perceive unmet need (n=26) n (%) |
Perceived unmet need (n=78) n (%) |
Sought treatment (n=65) n (%) |
|
| Gender (n=335) | ||||||
| Male | 11 (79) | 54 (65) | 56 (74) | 14 (56) | 49 (66) | 50 (79) |
| Female | 3 (21) | 29 (35) | 20 (26) | 11 (44) | 25 (34) | 13 (21)* |
| Age (n=333) | ||||||
| 15–17 | 6 (43) | 26 (32) | 26 (35) | 8 (32) | 11 (15) | 10 (16) |
| 18+ | 8 (57) | 56 (68) | 49 (65) | 17 (68) | 63 (85) | 53 (84) |
| Race/Ethnicity (n=330) | ||||||
| White | 12 (86) | 55 (67) | 38 (51) | 13 (52) | 37 (50) | 38 (62) |
| Other | 2 (14) | 27 (33) | 36 (49)* | 12 (48) | 37 (50) | 23 (38) |
| Student status (n=335) | ||||||
| Yes | 8 (57) | 38 (46) | 43 (57) | 13 (52) | 35 (47) | 29 (46) |
| No | 6 (43) | 45 (54) | 33 (43) | 12 (48) | 39 (53) | 34 (54) |
| Employ status (n=335) | ||||||
| Yes | 5 (36) | 43 (52) | 25 (33) | 10 (40) | 48 (65)* | 39 (62) |
| No | 9 (64) | 40 (48) | 51 (67) | 15 (60) | 26 (35) | 24 (38) |
| Annual household income (n=330) | ||||||
| <$25,000 | 8 (57) | 50 (60) | 29 (39) | 4 (16) | 25 (35) | 26 (42) |
| $25,000 −$74,999 | 5 (36) | 25 (30) | 28 (38) | 13 (52) | 30 (42) | 22 (35) |
| ≥$75,000 | 1 (7) | 8 (10) | 17 23) | 8 (32) | 17 (24) | 14 (23) |
| Risk of suicide | ||||||
| High | 10 (71) | 76 (87) | 69 (88) | 14 (54) | 50 (64) | 50 (77)* |
| Low | 4 (29) | 11 (13) | 9 (12) | 12 (46) | 28 (36) | 15 (23) |
| Recruiting social media platform | ||||||
| 1 (7) | 18 (21) | 26 (33) | 3 (12) | 13 (17) | 24 (37) | |
| Forums | 0 | 1 (1) | 0 | 0 | 0 | 0 |
| 4 (29) | 27 (31) | 29 (37) | 8 (31) | 21(29) | 19 (30) | |
| 7 (50) | 22 (25) | 14 (18) | 13 (50) | 39 (50) | 20 (31) | |
| Teenline | 0 | 2 (2) | 1 (1) | 1 (4) | 0 | 0 |
| Tumblr | 1 (7) | 8 (9) | 6 (8) | 0 | 3 (4) | 0 |
| 1 (7) | 9 (10) | 2 (3) | 1 (4) | 2 (3) | 2 (3) | |
Stars indicate statistical significance levels from unadjusted logistic regression models.
The “perceived unmet need” and “sought treatment” groups are both separately compared to the “did not seek treatment or perceive unmet need” group;
p<0.05
Treatment seeking and perceived unmet need
Figure 1 presents the patterns of treatment seeking and perceived unmet need for treatment in the total sample and by the severity of depression. Within the sample of participants with depression, 41% reported seeking treatment and 47% reported perceiving an unmet need for such treatment. Compared to individuals with mild depression, those with major depression were both more likely to seek treatment (84% vs. 71%, p=0.029) and to report unmet need for such treatment (86% vs.75%, p=0.044).
Figure 1.
Treatment use by participants with perceived unmet need for depression treatment among the total sample and individuals with major depression and mild depression.
Barriers to depression treatment among participants who perceived an unmet need for treatment
Table 2 shows characteristics of the 165 participants, all of whom were not currently linked to treatment and perceived an unmet need for treatment stratified by the severity of depression. Almost all participants (95%) reported at least one type of barrier to treatment and more than half people (55%) reported moderate barriers (i.e., 3–5 barriers). Individuals with major depression were more likely than those with mild depression symptom to receive treatment for depression at some point in lifetime (72 % vs.58, p=0.047), to have a higher risk of suicide (87% vs. 64%, p<0.001), and to be adolescents (age 15–17; 32% vs. 15%, p=0.014). Those with mild depression, however, are more likely to identify as non-White than those with major depression (50% vs. 33%, p=0.03).
Barriers to depression treatment by the two groups of depression are presented in Table 3. The most common types of barriers reported were attitudinal toward treatment (64%). Compared to participants with mild depression, participants with major depression were more likely to experience attitudinal barriers toward treatment (76% vs. 51%, p=0.001), and stigma (67% vs. 40%, p=0.001). Individuals with mild depression, however, are more likely to have the structural barriers than those with major depression (59% vs. 41%, p=0.024). The prevalence of the attitudinal barrier toward depression and financial barriers did not differ significantly between those two groups.
Table 3.
Barriers to depression treatment among individuals with major depression and mild depression symptoms who perceived an unmet need for depression treatment
| Barriers | Total n (%) |
Major depression (n=87) n (%) |
Mild depression (n=78) n (%) |
P |
|---|---|---|---|---|
| Attitudinal – Treatment | 106 (64) | 66(76) | 40 (51) | .001 |
| Didn’t think anyone could help | 42 (25) | 26 (30) | 16 (21) | .168 |
| Tried getting help before and it didn’t work (adults only) | 48 (38) | 34 (56) | 14 (22) | <.001 |
| Was afraid they would put me into the hospital or of the treatment they would give me | 48 (29) | 33 (38) | 15 (19) | .008 |
| Hated answering personal questions | 34 (21) | 23 (26) | 11 (14) | .051 |
| Didn’t want to go | 37 (22) | 27 (31) | 10 (13) | .005 |
| Attitudinal – Disorder | 51 (59) | 51 (59) | 46 (59) | .963 |
| Thought the problem would get better by itself / Didn’t think the problem was serious enough | 72 (44) | 38 (44) | 34 (44) | .991 |
| Thought it was something I should be strong enough to handle alone | 84 (51) | 46 (53) | 38 49) | .594 |
| Attitudinal – Stigma | 89 (54) | 58 (67) | 31 (40) | .001 |
| Was too embarrassed to discuss it with anyone | 63 (38) | 44 (51) | 19 (24) | <.001 |
| Was afraid of what my boss, friends, family, or others would think | 53 (32) | 35 (40) | 18 (23) | .019 |
| A member of my family objected | 18 (11) | 13 (15) | 5 (6) | .079 |
| Was afraid I would lose my job (adults only) | 8 (6) | 6 (10) | 2 (3) | .120 |
| Structural | 82 (50) | 36 (41) | 46 (59) | .024 |
| Didn’t know any place to go for help | 35 (21) | 18 (21) | 17 (22) | .862 |
| Didn’t have any way to get there or couldn’t arrange for childcare | 29 (18) | 15 (17) | 14 (18) | .905 |
| Didn’t have time (adults only) | 39 (31) | 14 (23) | 25 (38) | .060 |
| The hours were inconvenient (adults only) | 11 (9) | 4 (7) | 7 (11) | .403 |
| Had to wait too long to get into a program (adults only) | 6 (5) | 3 (5) | 3 (5) | .937 |
| Financial | ||||
| Can’t afford it / not covered by health insurance (adults only) | 58 (46) | 29 (48) | 29 (45) | .742 |
Results of logistic regression models for major depression predicting each type of barrier to depression treatment are shown in Table 4. After adjusting for demographic covariates (Model 2), major depression was no longer significantly associated with the attitudinal barriers toward treatment. After further adjusting for suicide risk and prior treatment history (Model 3), only the association between major depression and attitudinal barriers toward stigma remained statistically significant (OR: 3.42, 95% CI: 1.34–8.70).
Table 4.
Odds ratios and 95% confidence intervals for major depression in unadjusted and adjusted logistic regression models predicting types of barriers to depression treatment
| Attitude-treatment OR (95%CI) |
Attitude-disorder OR (95%CI) |
Attitude-stigma OR (95%CI) |
Structural OR (95%CI) |
Financial OR (95%CI) |
|
|---|---|---|---|---|---|
| Model 1 | |||||
| Major depression | 2.99 (1.54–5.79) | 0.99 (0.53–1.83) | 3.03 (1.61–5.73) | 0.49 (0.26–0.91) | 1.13 (0.56–2.27) |
| Model 2 | |||||
| Major depression | 1.98 (0.95–4.14) | 0.81(0.40–1.63) | 2.87(1.28–5.98) | 0.31 (0.15–0.64) | 0.79 (0.35–1.79) |
| Age | 0.98 (0.95–1.02) | 0.99 (0.96–.102) | 0.94 (0.90–0.99) | 0.97 (0.93–1.01) | 0.99 (0.96–1.04) |
| Gender | 0.47 (0.22–1.03) | 0.94 (0.47–1.89) | 1.60 (0.76–3.34) | 0.98 (0.48–1.99) | 0.90 (0.40–1.99) |
| Race/ethnicity | 0.88 (0.41–1.87) | 0.74 (0.37–1.94) | 0.81 (0.38–1.69) | 0.62 (0.30–1.26) | 1.71 (0.74–3.98) |
| Student status | 1.78 (0.73–4.34) | 1.90 (0.85–4.27) | 1.27 (0.54–2.99) | 0.91 (0.40–2.04) | 0.96 (0.37–2.48) |
| Income | |||||
| Middle income | 0.57 (0.26–1.27) | 1.22 (0.58–2.57) | 0.96 (0.44–2.12) | 0.87 (0.41–1.86) | 0.57 (0.25–1.33) |
| High income | 0.27 (0.10–0.76) | 0.69 (0.26–1.82) | 0.79 (0.28–2.27) | 0.96 (0.36–2.57) | 0.11 (0.03–0.45) |
| Model 3 | |||||
| Major depression | 1.66 (0.68–3.99) | 0.99 (0.43–2.28) | 3.42 (1.34–8.70) | 0.62 (0.26–1.47) | 0.94 (0.36–2.45) |
| Age | 0.98 (0.95–1.02) | 1.01 (0.97–1.04) | 0.96 (0.92–1.01) | 0.98 (0.94–1.02) | 1.01 (0.96–1.04) |
| Gender | 0.37 (0.16–0.85) | 0.90 (0.44 −1.86) | 1.97 (0.88–4.40) | 1.16 (0.54–2.48) | 0.82 (0.36–1.89) |
| Race/ethnicity | 0.81 (0.37–1.77) | 0.71 (0.35–1.44) | 0.79 (0.36–1.72) | 0.61 (0.29–1.29) | 1.62 (0.69–3.81) |
| Student status | 1.98 (0.77–4.64) | 1.91 (0.85–4.32) | 1.31 (0.53–3.19) | 0.87 (0.38–2.01) | 1.01 (0.38–2.62) |
| Income | |||||
| Middle income | 0.55 (0.25–1.24) | 1.21 (0.57–2.55) | 0.95 (0.42–2.15) | 0.87 (0.40–1.92) | 0.56 (0.24–1.31) |
| High income | 0.26 (0.09–0/77) | 0.66 (0.25–1.77) | 0.63 (0.20–1.98) | 0.83 (0.30–2.29) | 0.11 (0.03–0.44) |
| Risk of suicide | 2.57 (1.03–6.44) | 1.42 (0.61–3.32) | 0.72 (0.28–1.85) | 0.60 (0.25–1.44) | 1.51 (0.56–4.07) |
| Treatment history | 0.94 (0.41–2.14) | 0.62 (0.28–1.32) | 0.26 (0.11–0.63) | 0.34 (0.15–0.73) | 0.73 (0.28–1.91) |
Reference group: Gender: male; Race/ethnicity: white; Income: low income;
Discussion
This study assessed characteristics and barriers to treatment among social media users with varying depressive symptom severities and perceived unmet needs for mental health treatment. Findings demonstrated that participants with symptoms of major depression had significantly more histories of mental health treatment and reported a greater risk of suicide than participants with symptoms of minor depression. Participants with symptoms of major depression were also significantly younger, White, and earned a lower income than participants with symptoms of minor depression.
Additionally, as compared to participants with symptoms of minor depression, we found that participants with symptoms of major depression significantly reported more concerns with mental health treatment, stigma, and structural barriers to receiving treatment, consistent with previously published studies (Olfson et al., 2016). Adjusting for demographics, suicide risk and treatment history, stigma surrounding depression and treatment was a significant barrier to services for participants with symptoms of depression and perceived unmet needs. This aligns with previously published studies illustrating that stigma and other attitudinal barriers play a tremendous role in preventing individuals from seeking treatment (Andrade et al., 2013).
The findings of this study have implications for how social media can be used to help minimize barriers to treatment for depression through mental health literacy. Specifically, engagement strategies may take into account how barriers to treatment vary by severity of depression symptoms. For example, participants with symptoms of major depression reported greater fears of hospitalization and treatment than participants with minor depression symptoms. Our findings also demonstrate that the social media users with an unmet need for depression treatment are characteristically White male adults who earn a lower household income. This is in line with research illustrating that young men without health insurance are among the lowest demographics for engagement of treatment of depression (Olfson et al., 2016). As previous literature in this area focuses on access to and engagement in treatment among minority populations (Brown et al., 2010), it may be important to engage this subpopulation of White males with depression into treatment and research as well to fill the gaps in the literature through the use of social media.
Our findings have implications for various mental health disciplines, both researchers and clinicians at the frontlines of care. As ill-managed depression can have fatal consequences (Hawton et al., 2013; Cavanaugh et al., 2003), it is especially important for all mental health professionals who work at all levels of care to be aware of the barriers to treatment for specific populations and to identify how social media and online forums may be mediums to engage more vulnerable individuals. For example, participants with mild depression were more likely to report structural barriers than those with major depression, and, therefore, may benefit from mental health support delivered on-line or via a phone application. Additionally, the results suggest the need for partnerships among mental health, computer science, and communications disciplines to collaboratively design applications on social media platforms and online forums that can appropriately encourage distressed users to seek help.
Limitations to the study are noted. Data was cross-sectional, and the study’s participants were self-selecting. As the sample was restricted to individuals 15 years and older and users of specific social media platforms, the findings may not be generalizable to younger teenage social media users or users who access and engage with less mainstream social media platforms and online-forums. As the recruitment from social media platforms and online forums did significantly differ by severity of depression symptoms, future research is needed to substantiate how mental health symptoms vary across frequent users of specific online mediums. Future study may also track changes in depression scores and engagement with mental health treatment over time and replicate the study with younger online users and/or users who access niche platforms and forums.
Considering the reported greater risk of suicide, social media users with major depression symptoms should be a priority for development and implementation for treatment engagement strategies (Bentley et al. 2014). The identification of individuals at risk for suicide is especially important in the light that only about half of individuals who attempted suicide received any sort of mental health treatment (Han et al., 2014). Participation in this study suggests that social media is a promising platform to recruit individuals with depression symptoms who want mental health support; therefore, social media also has the potential as a setting to engage persons with depression in help-seeking behavior (Jashinsky et al., 2014). Future studies may test the feasibility of technology and/or social-media based engagement strategies for treatment of depression among hard-to-reach demographics.
Acknowledgments
This work was supported by the National Institutes of Health [grant numbers R03 MH109024, K02 DA043657]. The first author is supported by the National Institute of Mental Health of the National Institutes of Health under Award Number T32MH019960.
Footnotes
There are no conflicts of interest to report.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author (PCR). The data are not publicly available as they contain information that could compromise the privacy of research participants.
References
- Alegría M, Chatterji P, Wells K, Cao Z, Chen CN, Takeuchi D & Meng XL, (2008). Disparity in depression treatment among racial and ethnic minority populations in the United States. Psychiatric Services, 59(11), 1264–1272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Andrade LH, Alonso J, Mneimneh Z, Wells JE, Al-Hamzawi A, Borges G, Bromet E, … Florescu S, 2013. Barriers to mental health treatment: results from the WHO World Mental Health surveys. Psychological Medicine, 44(6),1303–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bentley SM, Pagalilauan GL & Simpson SA (2014). Major depression. Medical Clinics of . North America, 98(5), 981–1005. [DOI] [PubMed] [Google Scholar]
- Brown C, Conner KO, Copeland VC, Grote N, Beach S, Battista D & Reynolds CF III (2010). Depression stigma, race, and treatment seeking behavior and attitudes. Journal of Community Psychology, 38(3), 350–368. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Burns JM, Durkin LA & Nicholas J (2009). Mental health of young people in the United States: What role can the internet play in reducing stigma and promoting help seeking? Journal of Adolescent Health, 45, 95–97. [DOI] [PubMed] [Google Scholar]
- Cavanagh J, Carson A, Sharpe M & Lawrie S (2003). Psychological autopsy studies of suicide: A systematic review. Psychological Medicine, 33(3), 395–405. [DOI] [PubMed] [Google Scholar]
- Cohen E, Feinn R, Arias A, & Kranzler HR (2007). Alcohol treatment utilization: findings from the National Epidemiologic Survey on Alcohol and Related Conditions. Drug and Alcohol Dependence, 86(2–3), 214–221. [DOI] [PubMed] [Google Scholar]
- Corrigan PW & Watson AC (2002). Understanding the impact of stigma on people with mental illness. World Psychiatry, 1(1):16–20. [PMC free article] [PubMed] [Google Scholar]
- Han B, Compton WM, Gfroerer J, McKeon R (2014). Mental health treatment patterns among adults with recent suicide attempts in the United States. American Journal of Public Health, 104, 2359–2368. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hawton K, Casañas I Comabella C, Haw C & Saunders K (2013). Risk factors for suicide in individuals with depression: a systematic review. Journal of Affective Disorders, 147(1–3), 17–28. [DOI] [PubMed] [Google Scholar]
- Givens JL, Houston TK, Van Voorhees BW, Ford DE & Cooper LA, (2007). Ethnicity and preferences for depression treatment. General Hospital Psychiatry, 29(3), 182–191. [DOI] [PubMed] [Google Scholar]
- Grant BF, Amsbary M, Chu A, Sigman R, Kali J, Sugawana Y, … & Chou PS (2014). Source and accuracy statement: national epidemiologic survey on alcohol and related conditions-III (NESARC-III). Rockville, MD: National Institute on Alcohol Abuse and Alcoholism. [Google Scholar]
- Jashinsky J, Burton SH, Hanson CL, West J, Giraud-Carrier C, Barnes MD & Argyle T (2014). Tracking suicide risk factors through Twitter in the US. Crisis, 35, 51–59. [DOI] [PubMed] [Google Scholar]
- Kaufmann C, Chen L-Y, Crum R & Mojtabai R (2014). Treatment seeking and barriers to treatment for alcohol use in persons with alcohol use disorders and comorbid mood or anxiety disorders. Social Psychiatry & Psychiatric Epidemiology, 49(9), 1489–1499. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kerridge BT, Mauro PM, Chou SP, Saha TD, Pickering RP, Fan AZ, … & Hasin DS (2017). Predictors of treatment utilization and barriers to treatment utilization among individuals with lifetime cannabis use disorder in the United States. Drug and Alcohol Dependence, 181, 223–228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kroenke K, & Spitzer RL (2002). The PHQ-9: a new depression diagnostic and severity measure. Psychiatric Annals, 32(9), 509–515. [Google Scholar]
- Tjia J, Givens JL & Shea JA (2005). Factors associated with undertreatment of medical student depression. Journal of American College Health, 53(5), 219–224. [DOI] [PubMed] [Google Scholar]
- Kroenke K, Spitzer RL & Williams JB (2001). The PHQ-9: validity of a brief depression severity measure. Journal of General Internal Medicine, 16, 606–613. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Leiner DJ (2013). Too fast, too straight, too weird: Post hoc identification of meaningless data in internet surveys. SSRN Electronic Journal, 10.2139/ssrn.2361661. [DOI] [Google Scholar]
- Olfson M, Blanco C & Marcus SC (2016). Treatment of adult depression in the United States. JAMA Internal Medicine, 176(10), 1482–1497. [DOI] [PubMed] [Google Scholar]
- Qualtrics. (2019). Qualtrics. Provo, UT: Qualtrics; Retrieved from https://www.qualtrics.com [Google Scholar]
- SAS. (2013). SAS computer software, Version 0.4. Cary, NC: SAS Institute, Inc. [Google Scholar]
- Tjia J, Givens JL & Shea JA (2005). Factors associated with undertreatment of medical student depression. Journal of American College Health. 53(5), 219–224. [DOI] [PubMed] [Google Scholar]
- Wulsin L, Somoza E& Heck J (2002). The feasibility of using the Spanish PHQ-9 to screen for depression in primary care in Honduras. The Primary Care Companion to the Journal of Clinical Psychiatry,4, 191–195. [DOI] [PMC free article] [PubMed] [Google Scholar]