Remember me
Over 50 million Americans experience mental illness in a given year [,], but only one-third of those diagnosed with a mental health condition receive treatment from a specialist mental health care provider []. Individuals with common mental disorders face many barriers to adequate care, including limited numbers of specialist mental health care providers, geographic barriers due to therapists living mostly in urban areas, and an inability to pay for or use insurance [,]. In recent years, there has been an increasing interest in the potential of low-cost, digitally delivered psychological treatments (eg, internet-based guided self-help [GSH] and smartphone apps) to provide treatment remotely and for low cost, such that they can be disseminated at a larger scale than “traditional” treatments, such as one-on-one face-to-face therapy []. These digital mental health interventions (DMHIs) may have the potential to expand access to effective psychological treatment [] with a potentially revolutionary impact on public health, spurring a surge of attention across clinical research [,], the private sector [,], and government initiatives [-]. Robust evidence has now supported the efficacy of DMHIs in various formats, including internet-delivered cognitive behavioral therapy [], both GSH and unguided self-help formats [-], and smartphone apps [-]. However, less research has focused on evaluating the potential reach of DMHIs [,].
Evidence suggests that public use of DMHIs may be lower than is required for broad public health impact. For example, DMHIs may be difficult for the public to access [], underused by mental health care providers [], or simpler and less appealing to the public than treatment developers might predict []. Therefore, it is vital that DMHI researchers afford sufficient attention to the actual implementation of DMHIs. Indeed, Ramos et al [] cautioned that DMHI research risks mimicking the pitfalls of the “Decades of the Brain,” where large volumes of funding and time were spent on seemingly exciting innovations that made virtually no public health impact. Excellent work is underway to address this issue by identifying successful implementation strategies to maximize the reach of DMHIs, drawing from theoretical models and previous successes in implementation and dissemination science [,-]. However, it may also be important to consider factors in individuals’ decisions to uptake and adopt DMHIs, drawing from literature on treatment-seeking behavior and mental health service use [-].
Attitudinal Barriers to Treatment AccessThe scale of impact of DMHIs is dependent on the extent to which individuals with unmet treatment needs are broadly interested in and would use DMHIs. Although DMHIs are designed to circumvent structural barriers to treatment access (eg, cost, geographic availability of mental health care providers, need for transportation, and time commitment), their reach may still be limited by attitudinal barriers to treatment uptake. Attitudinal barriers are beliefs held by individuals that may affect their treatment-seeking behavior, such as the lack of perceived need for treatment, stigma, beliefs about the efficacy of psychotherapy, and the desire to handle a problem on one’s own. Importantly, attitudinal barriers are at least as common as structural barriers [-]. For example, across 24 countries in the World Health Organization (WHO) World Mental Health surveys, only 38% of the individuals with a 12-month mental disorder diagnosis reported a perceived need for treatment, making low perceived need the most commonly reported barrier to treatment use []. Therefore, the most common barriers to accessing traditional treatment may not be addressed by the innovations of DMHIs, which could greatly limit their potential public health impact.
Because most common attitudinal barriers (eg, stigma) are likely to impact any form of mental health treatment seeking, it may be likely that barriers commonly reported for traditional psychotherapy will generalize to DMHIs. However, this has rarely been studied directly. A small, promising body of literature currently provides some information on attitudinal barriers that are specific and often unique to DMHIs. However, much of this research often focuses on the acceptability of using DMHIs [,], especially pertaining to issues with technology specifically rather than revealing the impact that attitudinal barriers to general mental health treatment seeking might have on DMHI uptake. In a 2021 systematic review, Borghouts et al [] identified common barriers to user engagement with DMHIs across a variety of study types and DMHI formats (including telehealth treatment), in which information about barriers was often indirectly reported as secondary components of studies, such as clinical trials, or extracted from user reviews. Many of the reviewed studies investigated adherence and retention of participants receiving a DMHI rather than potential uptake among the general population. Across the studies, concerns specific to the use of technology were among the most commonly reported barriers for both participants and mental health care providers, such as privacy, digital literacy, and the ease of use. This literature advances our understanding of users’ experience of using DMHIs, which is essential for the design and dissemination of DMHIs that are acceptable and engaging from both user and provider perspectives. However, among 208 relevant articles, the authors identified only 5 studies that had assessed the relationship between participants’ broader mental health–related beliefs and engagement in DMHIs. For example, in a trial of a smartphone app for relationship stress among adolescents, perceived treatment needs and belief in treatment effectiveness were each associated with greater likelihood of use [].
These findings provided some support for the idea that attitudinal barriers to mental health treatment seeking may generalize to DMHIs. However, none of these studies were conducted with the primary purpose of investigating this relationship. They did not gather comprehensive data surveying commonly experienced access barriers, as is typically done in studies of perceived barriers to mental health care [-]. Therefore, this work did not explore the relative impacts of reduced structural barriers versus remaining attitudinal barriers when considering the potential reach of DMHIs. More research is needed to understand the extent to which attitudinal and structural barriers to help seeking may limit the reach of DMHIs to individuals with unmet treatment needs.
Low-Income Groups and Racial and Ethnic MinoritiesIn addition to hopes that DMHIs may expand treatment access in the general population, there is also an often-stated assumption that DMHIs have the potential to reduce racial, ethnic, and socioeconomic disparities in access to mental health care [-]. These groups are often underserved due to multifaceted, often systematic sets of access barriers. The relationship between income and actual treatment use may be more complex than is often assumed. For example, middle-income families may have lower access than low-income families due to the latter’s ability to use government-funded mental health services []. Nonetheless, the ability to pay for treatment or use insurance is one of the most commonly reported perceived barriers to mental health treatment seeking [,,], especially in the United States []. In addition to the ability to pay [], lower-income individuals face associated structural barriers, such as geographic restrictions, for example, specialty mental health care providers are twice as likely to be available in the highest-income communities relative to the lowest-income communities []. Socioeconomic disadvantages disproportionately impact racial and ethnic minorities, who are overrepresented in lower-income communities. In 2012, the rate of access to mental health care for non-Hispanic White American individuals was 20%, whereas the rates for Black, Hispanic, and Asian American individuals were 10%, 9%, and 5%, respectively []. It is very reasonable to assume that making DMHIs available can reduce these disparities by reducing the structural barriers that often contribute to them. However, only a limited body of research has collected data to substantiate this assumption, particularly in regard to racial and ethnic minorities [], suggesting that the reach of DMHIs risks replicating the same inequities seen in traditional treatments [,,].
It is essential to consider the role of attitudinal barriers in the ability of DMHIs to ameliorate mental health care disparities, given their great influence on treatment use in the general population. First, the impact of DMHIs on mental health disparities may be limited by attitudinal barriers because attitudinal barriers can limit the extent to which services are used even if they are made accessible. Second, some literature suggests that certain attitudinal barriers may disproportionately affect underserved groups, raising the concern that targeting disparities via the dissemination of DMHIs may disregard potential contributions of attitudinal barriers as an important mechanism of these disparities. For example, some studies suggest that individuals with lower incomes and lower educational attainment have a lower perceived need for treatment than individuals with higher incomes []. Similar findings regarding racial and ethnic minorities are mixed, partially because attitudes vary both across different minority groups and intersectionally within them. For example, Asian American and Black American individuals report lower perceived need for treatment than non-Hispanic White American individuals, but Hispanic American individuals report approximately the same level of need as non-Hispanic White American individuals []. This picture is further complicated by intersectional differences. For example, despite those group-level differences, Hispanic men report greater need for treatment than non-Hispanic White men, and Black women report approximately the same level of need as non-Hispanic White men []. Cultural beliefs about mental health treatment seeking also vary by immigration status (immigrant vs US born) and country of origin, often, despite similar racial identities [,].
Existing literature exploring racial differences in perceived barriers to treatment use is severely limited []. Some data exist regarding racial differences in interest in DMHIs. For example, Hispanic adults may be more interested in DMHIs than non-Hispanic White adults [,-]; however, researchers have suggested this may be attributable to greater rates of smartphone use rather than greater willingness to use treatment overall. Unfortunately, there is limited data to further clarify these patterns because individuals from both lower-income and racial and ethnic minoritized groups are often left out of DMHI trials, such that their preferences and attitudes may be overlooked in the design delivery and content of DMHIs [,,,]. Neglecting to understand actual interest in DMHIs among underserved and marginalized groups may limit the potential of DMHIs to attract members of these groups and serve their mental health needs [,,]. Because attitudinal barriers are such a significant deterrent to treatment seeking and treatment access, paying special attention to how attitudinal barriers may translate to DMHIs in these groups will be key to ensuring DMHIs reach them as intended.
This StudyExisting literature has focused on developing innovative approaches to address structural barriers (ie, DMHIs), leaving a substantial gap in the effort to address attitudinal barriers. Therefore, it is unclear to what extent attitudinal barriers that limit access to traditional psychotherapy might generalize to also limit access to DMHIs, reducing the potential reach of DMHIs to populations with unmet treatment needs. In this study, we aimed to conduct an exploratory investigation of the relationship between attitudinal barriers to traditional psychotherapy access and potential use of DMHIs. We placed an emphasis on currently underserved populations such as racial and ethnic minorities and low-income groups in order to focus on those whom DMHIs are most intended to target. We captured both participants’ interest in GSH and their self-reported likelihood of using it. This allows us to parse the appeal of the intervention from its actual potential to engage participants. Research applying the theory of planned behavior [] supports the use of behavioral intentions as a reliable predictor of mental health help seeking [,], and behavioral estimates may be an even stronger predictor than intentions [].
Study procedures were approved by the Indiana University Human Subjects and Institutional Review Board (#14172). Participants were provided with an informed consent document with information about our laboratory, the study’s purposes and procedures, payment, risks and benefits, confidentiality and data security, and their right to withdraw consent at any time. The document was written at an approximately seventh-grade reading level to ensure accessibility across differing literacy or English fluency. Participants were required to attest that they were aged ≥18 years before proceeding. Participants were paid at a rate of approximately US $10 per hour (US $3.02 total for a median time expenditure of 17 minutes and 59 seconds; IQR 13 minutes 53 seconds to 24 minutes and 18 seconds). First, we chose Prolific over other platforms such as Amazon Mechanical Turk (MTurk) due to its stated commitment to ethical treatment of its online workers, including its “ethical reward” payment policy [-]. Second, we designed our study to address established ethical considerations in DMHI research. For example, we used liberal inclusion and exclusion criteria in order to capture a broad population of individuals with treatment needs (eg, no suicidality exclusion) [,]. Finally, online DMHI research must carefully consider clinical risk monitoring []. We provided participants with crisis resources for suicidality as well as information about how to find noncrisis treatment and direct access to a free online self-help booklet (refer to the subsequent sections) [].
Web-Based Data CollectionData Collection PlatformParticipants were US adults recruited via the web-based crowdsourcing data collection platform Prolific (accessed from July 25to July 26, 2022) []. We chose Prolific because it is designed to improve upon common problems in internet-based human subjects research with crowdsourcing platforms, such as “bots” [], the lack of engagement, repeat responders, and nongenuine responses, which can create serious issues in the quality of health research []. In this regard, Prolific’s quality control is superior to other platforms such as MTurk. Studies show that responders from Prolific are mostly rated as “high quality,” significantly outperforming MTurk and even undergraduate students [,].
Data Quality ChecksWe used a series of methods, informed by previous literature, to reduce the risk that “fraudsters” and low-quality responders would pose a threat to the integrity of the data [,]. First, participants were required to pass a reCAPTCHA task (version 1; Google LLC) []. Second, participants completed a series of reading comprehension questions designed to screen out nonhuman or inattentive responders (refer to the Survey Procedure section). Participants who answered >1 question wrong after 2 attempts were eliminated in the data cleaning process. After data collection, we also removed responders who completed the survey too quickly (ie, 2 SDs below the mean completion time).
Survey ProcedureOverviewParticipants provided the data for this study in 2 separate surveys, which is the procedure required in order to implement a nondemographic eligibility requirement on Prolific. The first survey was primarily for screening purposes but also included several demographic characteristics, a measure of psychological distress used as the screening criterion (the Kessler-6 Psychological Distress Scale [K6]; refer to the Measures and Variables section []), and 2 other brief self-report measures not used in the present analyses. Participants who met the K6 eligibility criterion (scoring >5) were invited to complete the second survey, which included the following: (1) further demographic characteristics, (2) a description of a particular GSH intervention described in the subsequent sections, (3) questions about their past-year mental health treatment use, and (4) several other self-report measures not analyzed in this study.
GSH Intervention Description and QuestionsParticipants were provided information regarding a transdiagnostic GSH, Doing What Matters in Times of Stress (DWM), as it is delivered in clinical trials run by our laboratory. In the DWM intervention, trial participants are provided with a booklet developed by the WHO [] that teaches principles and skills from acceptance and commitment therapy (ACT) []. They are provided the option to either access the booklet online (via a publicly posted PDF on the WHO website) or receive a paper copy in the mail. Guidance is provided by “coaches” (trained research assistants and graduate students in our laboratory), who meet with trial participants weekly via Zoom (Zoom Communications Inc) for 15- to 30-minute sessions across 3 to 6 weeks [,]. All survey participants were shown an advertisement used in our laboratory’s clinical trials and an additional description of basic information about DWM.
Participants were then asked to predict their likelihood of using the DWM intervention in the following series of questions:
First, participants were asked if they believe that they would click on the advertisement and complete a 15-minute survey if they saw it on social media (“Do you think you would click on the link in the ad...”; “Most likely yes” or “Most likely no”).Those who answered “Most likely yes” to the aforementioned question were asked if they believed that they would then answer a call from a researcher and stay on the phone for approximately 30 minutes (“Do you think you would answer the call...”; “Most likely yes” or “Most likely no”).Those who answered “Most likely yes” to the phone call question were provided information about the trial. Those who answered “Most likely no” were provided the same information before rating their interest in the intervention later in the survey. All participants were required to pass a set of comprehension questions.After learning the information typically shared in the trial’s welcome call, participants were asked if they believed they would be likely to enroll in the intervention portion of the trial (“Do you think you would sign up for the treatment?”; “Most likely yes” or “Most likely no”).Those who answered “Most likely yes” were asked how many GSH sessions they believed they would attend (“How long do you think you would most likely stay in the treatment [how many weeks of calls with the coach do you think you would do]?”; “I would probably not attend any of the sessions” or “I would probably do the first few calls with the coach [1 to 2 weeks] but not finish all the ones I scheduled” or “I would probably do all the calls with the coach that I signed up for [3 to 6 weeks]”).Measures and VariablesPsychological DistressPsychological distress was measured via the K6 [], a 6-item measure assessing frequency of emotional distress over the past 30 days (eg, “During the past 30 days, about how often did you feel...Nervous,” “...Hopeless,” “...Worthless”), where higher scores indicate greater distress. The K6 is a reliable and valid measure of distress [,] that has been reported to have excellent internal consistency in previous work (Cronbach α=0.89) []. It appeared internally consistent in this study (ω=0.79). The K6 can be used to screen for both serious mental illness at a score of ≥13 [] and milder forms of emotional distress with lower thresholds []. While investigators differ in describing the lower cutoffs as mild, moderate, or mild moderate, evidence generally supports the interpretation of a score >5 as indicating at least mild mental health needs [,,].
Demographic CharacteristicsParticipants answered questions about age, race, ethnicity, gender, sex, sexual orientation, income, and education level (refer to the Results section).
Self-Reported Likelihood of GSH UseAfter reading about the GSH intervention, participants answered a series of questions about their hypothetical use of the intervention. The self-reported likelihood of GSH use or “likely GSH use” outcome reflects endorsement that a participant believes that they would be likely to complete at least 1 GSH session if offered the intervention (ie, either “I would probably do the first few calls with the coach [1 to 2 weeks] but not finish all the ones I scheduled” or “I would probably do all the calls I signed up for”). Participants who did not reach this question due to denying that they believed they were likely to click on the advertisement, answer the phone call, or agree to enroll in the trial were coded as deniers of likely GSH use.
Interest in GSHAfter answering the GSH use questions, participants were asked to rate their overall interest in the intervention on a 4-point scale (“Overall, does this treatment sound like something you would be interested in?”; “Not at all interested,” “Somewhat interested,” “Moderately interested,” or “Very interested”). The ordinal value on this 4-point scale is the GSH interest outcome.
Past-Year Psychotherapy UseFollowing the survey section regarding the GSH intervention, participants proceeded to a survey section focusing on their actual past-year mental health treatment experiences. Participants indicated whether they had attended psychotherapy in the past year (“I went to therapy: seeing a mental health professional such as a psychologist, counselor, therapist, or social worker”) from a checklist of various forms of help (medications, informal social support, self-help, and other).
Perceived Need for PsychotherapyParticipants who denied any past-year psychotherapy use were asked whether there was any point in the past year at which they “thought [they] might benefit” from psychotherapy (“Yes, therapy: seeing a mental health professional like a counselor, psychologist, or social worker”), in the same checklist format as mentioned earlier. Selecting this answer choice was considered an endorsement of the perceived need for psychotherapy barrier; leaving it blank was considered to indicate the lack of perceived need barrier.
Barriers to Psychotherapy UseIndividuals who endorsed perceived need for psychotherapy were then presented with questions regarding their reasons for not receiving psychotherapy despite believing that they might need it. They were asked to select all contributing barriers for not accessing psychotherapy (“Please check all reasons that were part of why you did not go to therapy”). The answer choices included 11 common barriers to mental health treatment adapted from the National Comorbidity Survey [,]. A list of the 11 barriers is given in the Results section. The full text of each answer choice is provided in . Text entry responses under “other” were manually coded to either match an existing answer choice (eg, “don’t have the money” into the “money/insurance” answer choice) or remain in the “other” category if no obvious match was present.
Primary Barrier and Primary Barrier TypeAfter selecting all barriers that contributed to their lack of past-year psychotherapy use, participants were presented with the same list of barriers and asked to indicate their primary barrier for not accessing psychotherapy (“Which was the biggest reason you didn’t go to therapy?”). We generated a primary barrier type variable by grouping individual primary barrier choices into 3 categories according to previous literature [-] and the authors’ judgment: attitudinal (eg, “Didn’t think it would work”), structural (eg, “Issues with money or health insurance”), and other (eg, “The problem went away by itself”). Denial of perceived need for psychotherapy was considered an attitudinal primary barrier. When participants selected “other” as their primary barrier, the text that they entered was coded into one of the existing categories for primary barrier type. They were recoded to be (1) attitudinal (eg, “too much history. how could I even begin to get a new person up to [speed]”), (2) structural (eg, “no privacy at home guaranteed for phone appointment”), or (3) other if did not clearly fit either category (eg, “anxiety”).
AnalysesAll analyses were performed in R software (version 4.3.2; R Foundation for Statistical Computing) []. First, we generated descriptive statistics for sociodemographic characteristics, psychological distress severity, and each of the primary outcomes (GSH interest and likely GSH use). We analyzed the relationship of individual characteristics (sociodemographic characteristics and psychological distress) with GSH interest via multivariate polychoric regression in the MASS package (version 7.3-60) [] and the relationship of these predictors with likely GSH use via multivariate logistic regression. For all analyses involving income as a predictor, 1 participant was omitted due to a missing value for income. We additionally conducted univariate analyses for race and income, in order to isolate these key sociodemographic variables emphasized in the DMHI literature via 2 univariate polychoric regressions for GSH interest and 2 univariate logistic regressions for likely GSH use. These univariate analyses were informative in addition to the multivariate analyses of sociodemographic characteristics because we were substantively interested in the potential for DMHIs to reach low-income groups and marginalized racial-ethnic groups regardless of whether other demographic characteristics (eg, education) account for the difference.
Next, we generated descriptives for past-year psychotherapy use and analyzed its relationship with individual characteristics via logistic regression. We analyzed past-year psychotherapy use as a predictor of each GSH interest and likely GSH use via polychoric regression and logistic regression, respectively, controlling for psychological distress in both analyses. For all polychoric regressions, P values are not reported because the analysis does not directly generate P values, and simulated P values may not be reliable; our interpretations of statistical significance were made from CIs. Next, for participants who denied past-year psychotherapy use, we generated frequency statistics regarding endorsement of (1) perceived need for psychotherapy, (2) all contributing barriers to psychotherapy use, (3) primary barrier to psychotherapy use, and (4) primary barrier type. We only compared the frequency of lacking perceived need versus other primary barriers, rather than versus the frequency of all contributing barriers, due to the structure of our survey. Because participants who denied perceived need were not given the opportunity to select additional barriers, the level of detail collected about this subgroup’s access barriers was lower than for the subgroup that selected multiple contributing barriers.
Finally, we grouped the primary barriers into 3 categories: attitudinal, structural, and “other” (refer to the Primary Access Barriers and Barrier Type section). We analyzed sociodemographic characteristics and psychological distress as predictors of primary barrier type via multinomial regression in the nnet package (version 7.3-19) [] to accommodate the 3-category outcome. Next, we analyzed primary barrier type as a predictor of each GSH interest and likely GSH use via polychoric regression and logistic regression, respectively, controlling for psychological distress in both analyses. Finally, we analyzed the relationship between each contributing barrier and each outcome via a series of univariate linear regressions.
Due to the high number of statistical tests, we applied Benjamini-Hochberg adjustment to P values across all analyses.
Transparency and OpennessWe report on how we determined our sample, all data exclusions, all manipulations, and all measures in the study. All data, analysis code, and research materials are available on the Open Science Framework website online []. This study’s design and its analysis were not preregistered. No other papers currently use these data.
Most of the 971 participants identified as non-Hispanic White (n=665, 68.5%) and heterosexual (n=688, 70.9%), with an approximately even gender split (women: n=538, 55.4%) and a median age of 32 (IQR 25-41) years. The median K6 score was 11 (IQR 8-15) of 0 to 24 possible points. Full sample characteristics are reported in .
Table 1. Sociodemographic characteristics for a sample of adults with psychological distress (N=971).VariableValuesAge (years), median (IQR)32 (25-41)Gender, n (%)aK6: Kessler-6 Psychological Distress Scale.
Most of the 971 participants (n=782, 80.5%) reported that they were at least “somewhat interested” in GSH. Nearly half (n=458, 47.2%) of the participants were at least “moderately interested,” and 17.1% (n=166) were “very interested.” However, a slightly greater proportion (n=189, 19.5%) of participants was “not at all interested.” We found that 38.6% (n=375) of the participants reported that they believed they were likely to complete at least one GSH session if it were offered to them.
Individual Characteristics as Predictors of GSH Interest and Self-Reported Likelihood of GSH UseNone of the sociodemographic characteristics, including psychological distress, were statistically significant predictors of self-reported interest in GSH. For self-reported likelihood of GSH use, only age met statistical significance at P<.05 after Benjamini-Hochberg adjustment (odds ratio [OR] 1.02, 95% CI 1-1.03; P=.045). gives full results for both outcomes.
Table 2. Results of 2 multivariate regression models: (1) polychoric regression predicting interest in guided self-help and (2) logistic regression predicting self-reported likelihood of using guided self-help, each from sociodemographic characteristics in adults with psychological distress (n=970a).VariableInterest in GSHbSelf-reported likelihood of GSH usecaEach model is based on 970 participants with complete demographic data because 1 participant had a missing value for income.
bGSH: guided self-help.
cThe reference level is denying the likely use of GSH, such that odds ratio >1 reflects an association with endorsing the likely use of GSH.
dOR: odds ratio.
eFor categorical variables (all except age and psychological distress), the P value shown is an omnibus P value calculated across levels of the variable.
fBenjamini-Hochberg adjustment was performed across all analyses.
gReference level.
hK6: Kessler-6 Psychological Distress Scale.
Past-Year Psychotherapy UseDescriptives and Demographics for Past-Year Psychotherapy UseWe found that about one-third (331/971, 34.1%) of the participants reported past-year psychotherapy use. Among sociodemographic characteristics, only educational attainment (P<.001) and sexual orientation (P=.04) had statistically significant relationships with past-year psychotherapy use (refer to for full results). Higher psychological use distress severity was significantly associated with greater odds of endorsing past-year psychotherapy use (OR 1.07, 95% CI 1.03-1.1; P<.001).
Table 3. Logistic regression predicting past-year psychotherapy usea from sociodemographic characteristics in adults with psychological distress (n=970b).VariableORc,d (95% CI)Adjusted P valued,eAge0.99 (0.98-1.01).45Gender.27aThe reference level is denying past-year psychotherapy use, such that OR >1 reflects an association with endorsing past-year psychotherapy use.
bThis model is based on 970 participants with complete demographic data because 1 participant had a missing value for income.
cOR: odds ratio.
dFor categorical variables (all except age and psychological distress), the P value shown is an omnibus P value calculated across levels of the variable.
eBenjamini-Hochberg adjustment was performed across all analyses.
fReference level.
gK6: Kessler-6 Psychological Distress Scale.
Interest in GSH by Past-Year Psychotherapy UseAmong those (n=640) who denied past-year psychotherapy use, 77.7% (n=497) were at least “somewhat interested” in GSH, 39.8% (n=255) were at least “moderately interested,” and 11.3% (n=72) were “very interested.” Nearly one-fifth (n=143, 22.3%) of the participants were “not at all interested.” By contrast, among those (n=331) who endorsed past-year psychotherapy use, 86.1% (n=285) were at least “somewhat interested” in GSH, 57.7% (n=191) were at least “moderately interested,” and 28.4% (n=94) were “very interested.” Only 13.9% (n=46) of these participants were “not at all interested.” The difference in GSH interest between groups was statistically significant, that is, those who had used psychotherapy in the past year were significantly more interested in GSH than those who had not (OR 2.38, 95% CI 1.86-3.06; P<.001). This effect was not accounted for by psychological distress severity, which was not a statistically significant predictor of GSH interest in this model (OR 1.02, 95% CI 0.99-1.04; P=.43). Refer to for complete descriptive statistics and for a visualization.
Self-Reported Likelihood of Using GSH by Past-Year Psychotherapy UseAmong those denying past-year psychotherapy use, approximately one-third (205/640, 32%) reported that they would be likely to complete at least one GSH session, whereas over half (170/331, 51.4%) of those endorsing past-year psychotherapy use reported that they would be likely to do so. This difference was statistically significant (OR 2.25, 95% CI 1.71-2.96; P<.001). This effect was not accounted for by psychological distress severity, which was not a statistically significant predictor of self-reported likelihood of GSH use in this model (OR 1, 95% CI 0.97-1.03; P=.92). Refer to for complete descriptive statistics and for a visualization.
Barriers to Psychotherapy AccessDescriptives for Barriers to Psychotherapy AccessApproximately one-third (206/640, 32.2%) of the participants denied that they “might benefit” from psychotherapy (ie, no perceived need). When the subgroup that endorsed perceived need (434/640, 67.8%) was given the opportunity to select multiple contributing reasons for not using psychotherapy, the most commonly endorsed barrier was “issues with money or insurance” (323/434, 74.4% of those endorsing perceived need). “Didn’t know where to go or who to see” (230/434, 53%) and “too busy/not enough time” (190/434, 43.8%) were the second- and third-most commonly endorsed barriers in this subgroup (refer to for frequencies of all barriers and for a visualization).
Table 4. Frequency of endorsement of all contributing barriers to past-year psychotherapy use in a checklist format among participants who endorsed perceived need for psychotherapy and denied past-year psychotherapy use (n=434), presented by barrier type.All contributing barriers to past-year psychotherapy use by barrier typeParticipants, n (%)aStructuralaParticipants were given the opportunity to select multiple barriers, such that individual participants may be counted several times in this table.
bEach barrier description is abbreviated from the original answer choice presented to participants. See for full text of each answer choice.
cParticipants were queried about perceived need for treatment before being presented a checklist of other potential barriers to treatment access. Those who denied perceived need were not given the opportunity to select multiple barriers and therefore are excluded from this table.
Interest in GSH by Individual Access BarriersThe lack of perceived need for psychotherapy was significantly associated with lower interest in GSH (for endorsing relative to denying perceived need: OR 2.11, 95% CI 1.55-2.88; P<.001). This effect was not accounted for by the effect of psychological distress severity, which was not a statistically significant predictor of GSH interest in this model (OR 0.99, 95% CI 0.96-1.03; P=.83). None of the other individual barriers had a statistically significant univariate relationship with interest in GSH. Refer to for results of all univariate models, for full descriptive statistics, and for a visualization.
Table 5. Results of univariate logistic regressions for outcomes (1) interest in guided self-help (GSH) and (2) self-reported likelihood of guided self-help, by endorsement of each self-reported barrier to past-year psychotherapy use in a sample of adults with psychological distress (n=640).BarrieraGSH interestSelf-reported likelihood of GSH useb
Comments (0)