Awareness. Interaction. Direction. (A.I.D.): evaluation of a novel mental health awareness training

ABSTRACT

Objective

The current study evaluates the effectiveness of Awareness. Interaction. Direction. (A.I.D.), a novel Mental Health Awareness Training (MHAT) that was developed to address training limitations of some existing MHATs.

Method

Between 2022 and 2023, data was collected from 1198 participants who were from a university, working for K-12 schools, or community members in the Mid-Western United States. Participants self-selected or were assigned by their supervisor to complete the A.I.D. training. Prior to and after completing the training, participants answered questions pertaining to mental health knowledge, their confidence in implementing the A.I.D. action plan, and personal stigma. Paired responses for the mental health knowledge questions before and after the A.I.D. training were compared using McNemar’s test. Linear mixed models were used to compare confidence and personal stigma outcomes before and after the A.I.D. training.

Results

After completion of the A.I.D. training, participants demonstrated a significant increase in mental health knowledge and confidence in implementing the A.I.D. action plan and a significant decrease in personal mental illness stigma. Significant differences by age and gender were identified.

Discussion

The current study demonstrated the novel A.I.D. training is an effective training. A.I.D. provides a MHAT option that is brief and that can be delivered in different formats (e.g. virtual or in-person) to best accommodate the individuals receiving the training. This presents organisations with another MHAT option.

KEYWORDS:

• Mental health
• gatekeeper training
• mental health awareness training
• mental health knowledge
• evaluation
• stigma

Introduction

Mental health concerns have been increasing in recent years in the United States (US; Weiner, 2022). In 2017–2018, 19% of all US adults experienced a diagnosable mental, behavioural, or emotional disorder based on Structured Clinical Interviews using the 4th edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV); that was an increase of 1.5 million people from the previous year (Reinert et al., 2021). Then, the Covid-19 pandemic exacerbated the general trend of declining mental health (Reinert et al., 2021). By 2021, 57.8 million US adults reported living with any mental illness, equating to 22.8% of all US adults (Substances Abuse and Mental Health Services Administration [SAMSHA], 2022). This trend has continued to worsen, with 33% of US adults reporting experiencing anxiety or depression in early 2023 – extremely elevated compared to 11% of people reporting anxiety or depression prior to the COVID-19 pandemic (Kaiser Family Foundation [KFF], 2023).

Among the 57.8 million US adults living with any mental illness, only 47.2% obtained services within the past year (SAMHSA, 2022). Reasons for not seeking or obtaining professional mental health services are vast (National Alliance on Mental Illness, 2017). A study completed by the Veteran Cohens Network, in partnership with the National Council for Mental Wellbeing, took a comprehensive look at the access to mental health care in the US and found 42% of respondents listed high cost or having inadequate health insurance as their number one barrier for not seeking treatment (National Council for Mental Wellbeing, 2018). In addition, approximately 150 million individuals currently reside in areas that have been federally designated as mental health professional shortage areas (Weiner, 2022). The aforementioned study found 38% of respondents needed to wait at least a week or longer to receive treatment and nearly half (46%) of respondents either had to or knew someone who must drive an hour or more to receive services (National Council for Mental Wellbeing, 2018).

However, there are also more malleable barriers to receiving professional mental health services, such as a lack of awareness and knowledge on how to seek help as well as perceived social stigma surrounding mental illness (National Council for Mental Wellbeing, 2018). Approximately 31% of Americans have expressed being worried regarding potential judgement from others when they have discussed receiving mental health help (National Council for Mental Wellbeing, 2018). In fact, 21% of Americans have lied about seeking mental health help to avoid judgement from others (National Council for Mental Wellbeing, 2018).

Gatekeeper/mental health awareness trainings

Gatekeeper trainings, or Mental Health Awareness Trainings (MHAT), are becoming more popular to address some of these latter reasons why individuals do not seek professional mental health services, such as lack of knowledge and stigma. MHATs help prepare society (especially those with no mental health background) at the individual and population level to teach them how to adequately recognise and then engage with someone who is experiencing mental distress (SAMHSA, 2023). MHATs are also sometimes called gatekeeper trainings, because they target so-called ‘gatekeepers’ in the community (e.g. caregivers, teachers) who regularly interact with various individuals and therefore can more easily recognise when someone is experiencing a crisis or mental distress (U.S. Department of Health and Human Services Office of the Surgeon General & National Action Alliance for Suicide Prevention, 2012). The rationale is that by targeting gatekeepers who interact with large and diverse populations, MHATs can have a greater total impact on the community.

There are many pre-existing MHATs. One of the most prominent is Mental Health First Aid (MHFA), which is actively delivered in twenty-four countries (MHFA International, 2023). MHFA was originally created in Australia in 2001 and has been offered in the US since 2008 (MHFA, 2023). MHFA is currently offered as a blended format (online asynchronous pre-work followed by an instructor-led session) or instructor-led session for groups of up to 30 participants. It is a relatively lengthy training, with Adult MHFA taking 7.5 h and Youth MHFA taking 6.5 h. However, it has demonstrable efficacy. Recent meta-analyses have found that while short-term outcomes for MHFA are promising (e.g. increase in mental health knowledge, quality of intended helping behaviours, and actual use of the ALGEE plan; decreases in mental health stigma), long-term changes have had mixed findings (El-Den et al., 2020; Forthal et al., 2022). While MHFA is a well-liked, comprehensive, and effective training, its core limitation is its lengthy format that makes it inaccessible for populations who have very limited time.

Kognito, a practice-based digital learning training designed to improve mental health knowledge and awareness, was another MHAT that offered a range of online interactive simulations (e.g. Kognito At-Risk Mental Health for Students) that took an hour or less to complete (Kognito, n.d.). As part of the training, participants interacted with a virtual individual to build the capacity to have real-life conversations about wellness (Kognito, n.d.). Kognito has been established as an effective gatekeeper training; however, a gap is left as Kognito sunset the brand as of August 2023 (Kognito, n.d.). Kognito was particularly useful because it was offered solely online and allowed the flexibility to complete the training at your own pace (Kognito, n.d.). Numerous studies have shown Kognito improves an individuals’ likelihood of intervention and preparedness to interact with someone who is experiencing mental distress (Coleman et al., 2019; Smith-Millman et al., 2022).

Question, Persuade, and Refer (QPR) is another prominent MHAT. QPR was developed by Paul Quinnett in 1995 as an emergency mental health intervention to assist individuals experiencing suicidal ideation (QPR Institute, n.d.). QPR can be completed online asynchronously or in-person with groups of up to 35 participants. Like MHFA and Kognito, many studies have shown that QPR is effective in increasing individuals’ knowledge of suicide and increasing their ability to intervene if an individual is showing signs of suicide (Mitchell et al., 2013; Litteken & Sale, 2018; Aldrich et al., 2018). However, a major limitation of QPR is that it focuses solely on suicide prevention training and does not provide an overview of mental health in general.

In order to address limitations presented by current MHATs (see Table 1 for summary), Kent State University’s (KSU) Center for Public Policy and Health (CPPH) conceptualised and developed Awareness. Interaction. Direction. (A.I.D.) in 2019 based on the need for an efficient, highly modifiable training that would be conducive to the programmatic needs of a variety of organisations. As previously mentioned, popular and widespread MHATs of the time had fixed and very rigid requirements (e.g. time requirements, set topics, specified class sizes, specific formats – in-person or virtual, synchronous or asynchronous). To make MHATs readily available, easy to implement, and modifiable, the CPPH developed A.I.D.

Table 1. Comparison of MHATs.

	Content	Length	Modality	Class size
MHFA	Global mental health and illness	6.5–7.5 h	Hybrid asynchronous online + synchronous in-person, in-person	Up to 30
Kognito	Mental distress	1 h	Asynchronous online; discontinued in 2023	Unlimited
QPR	Suicide prevention	1 h	Asynchronous online, in-person	Unlimited (asynchronous) or up to 35 (in-person)
A.I.D.	Global mental health and illness	1–1.5 h	Asynchronous online, synchronous online, in-person	Unlimited (asynchronous) or up to 50 per instructor (in-person)

A.I.D. is a modular MHAT program designed to be delivered in 60-to-90 min by a trained instructor(s). The training's content was guided by Social Cognitive Theory (Bandura, 2000) and the Broaden and Build Theory of Positive Emotions (Conway et al., 2013) and developed by faculty and staff in the CPPH. The training covers mental health and mental illness in general. Participants learn the A.I.D. action plan, which teaches participants the knowledge, skills, and resources to actively assist individuals in mental distress by (1) being Aware of and recognising signs and symptoms of mental distress, (2) effectively Interacting with a person in mental distress, and (3) Directing a person in mental distress to the appropriate mental health resources (i.e. self-help strategies, crisis text lines, or professional help). Participants have the opportunity to apply the A.I.D. action plan to a practice scenario. A variety of situational practice scenarios are available for the instructor to choose from in order to select one that is most relevant to the participants being trained. The training also discusses mental health terminology and mental illness stigma. If requested, instructors can add the self-care section, which includes the definition of self-care and a discussion of types of self-care. A.I.D. can accommodate group sizes of any number, though a class size of around 30 people is ideal. Larger trainings benefit from having multiple instructors (e.g. a class of 100 was led by two instructors), which allows for breakout groups for discussions. The training can be conducted in-person or virtually (synchronously), where an instructor provides a lecture-type format and involves participants by posing discussion questions and providing a practice scenario where participants can practice using the A.I.D. action plan. A.I.D. can be delivered asynchronously by having participants watch a recorded version of the training.

The aims of the present study are to evaluate participant learning and attitudes before and after completion of the A.I.D. training to determine:

1. Whether mental health knowledge increases;

2. If confidence in implementing the steps of the A.I.D. action plan improves;

3. If personal stigma surrounding mental illnesses decreases.

Materials and methods

A.I.D. training

Participants in the United States completed the in-person or online training course that lasted 60 to 90-minutes. The course is broken down into three sections and has specific learning objectives for each section. Section 1 is ‘Terminology’, and the learning objectives are (1) Define mental health, stress and mental distress, (2) Learn about the mental health continuum, (3) Define mental health problem and mental health disorder, (4) Learn responses to stress and mental health problems, and (5) Define stigma. Section 2 is ‘Stigma’, and the learning objectives are (1) Discuss why people with mental illness are stigmatised, (2) Discuss why stigma is a problem, and (3) Discuss ways to reduce or prevent stigma. Section 3 is ‘Learning A.I.D.’, and the learning objectives are (1) Learn the three steps of the A.I.D. action plan and (2) Practice the A.I.D. action plan.

Participants

Students, faculty, and staff at a large public university in the Mid-Western United States, as well as K-12 school personnel and community members in that university’s county, participated in the study. Participants were recruited to take the A.I.D. training by email, flyers, social media, or being assigned to complete the A.I.D. training by their supervisor. Surveys were administered either in-person by the instructor on the day of the class or online a few days before the training via Qualtrics. Completion of the surveys was optional and not a requirement to participate in the training. Therefore, not everyone who completed the training completed the surveys. Data was collected from 2022 to 2023 from a total of 1198 participants prior to and after completion of the A.I.D. training. Participants were not compensated for their time; however, university employees were eligible to receive training credits toward a university-wide annual training initiative. Prior to data collection, the study was submitted to the Kent State University Institutional Review Board (IRB Protocol 47, Mental Health Training Evaluation).

From the initial total sample (N = 1,198), 77 participants were dropped due to only completing the pre-test survey, leaving our analyses with a maximum sample of 1,121 participants. Additionally, the initial training sessions’ pre- and post-test surveys did not include the mental health confidence or mental health knowledge questions (N = 162), decreasing the sample for applicable analyses (Tables 3–5). Finally, there was some additional missing data due to respondents skipping individual questions. We identify some implications of this missingness further in the Discussion below.

Demographic data

A variety of demographic data was collected. Participants were asked, ‘What is your current age?’ and responded with an open-ended format. Racial-ethnic identity was measured by asking participants, ‘What is your race/ethnicity?’ and providing the following response options (1) non-Hispanic White, (2) non-Hispanic-Black, and (3) other. Gender was measured by asking, ‘What is your gender?’ and providing the following response options (1) woman, (2) man, (3) transgender, and (4) other. Due to small sample size, responses (3) and (4) were collapsed into ‘other’ (3) gender in analyses.

Mental health knowledge

Mental health knowledge was assessed by five questions developed by the research team (Appendix A). The questions were written to reflect information about mental health discussed during the training. Participants either got each multiple-choice question incorrect (0) or correct (1). Questions were scored by research assistants using the answer key. The low value for Cronbach’s Alpha (0.54 on pre-training survey and 0.50 on post-training survey) indicated the measure is not sufficiently consistent; thus, subsequent analyses were performed on individual questions and not as a scale.

Mental health confidence scale

The Mental Health Confidence Scale (Appendix B) consists of three questions developed by the research team. The questions assess how confident the participant feels on a scale of 1–5, with higher values representing greater confidence, in implementing the three steps of the A.I.D. action planrecognising and being aware of signs of mental distress, interacting with someone exhibiting mental distress, and directing someone with mental distress to the appropriate professional resources. The value for Cronbach’s Alpha was 0.84 and 0.91 on the pre-training and post-training surveys, respectively, indicating good reliability. Participants’ scores on the mental health confidence scale were only calculated and included for analysis if they answered all scale questions at both time points.

Personal stigma scale

The Personal Stigma Scale (PSS) is a nine-question scale designed to assess personal attitudes and stigma towards mental illness. The scale was adopted from the personal stigma items of the Depression Stigma Scale (DSS; Griffiths et al., 2004) and asks participants to indicate how much they agree with each statement (‘Strongly Agree’ = 1 to ‘Strongly Disagree’ = 5). Whereas the DSS asked specifically about depression, the PSS modified the questions to focus on mental health in general. For example, ‘People with depression could snap out of it if they wanted’ on the DSS was changed to ‘People with mental illness could snap out of it if they wanted’ for the PSS. Additional examples of statements participants were asked to rate included, ‘Mental illness is a sign of personal weakness’ and ‘People with mental illness are dangerous’. Mean scores range from 1 to 5, with higher scores representing greater personal stigma. The value for Cronbach’s Alpha was 0.85 and 0.89 on the pre-training and post-training surveys, respectively, indicating good reliability. Participants’ scores on the personal stigma scale were only calculated and included for analysis if they answered all scale questions at both time points.

Statistical analyses

Data was analyzed using Stata 14 statistical software (StataCorp., 2015). The descriptive statistics were done using mean, median, standard deviation (SD) and interquartile range (IQR) for continuous variables and percentages for categorical variables. McNemar’s test was used to evaluate differences in proportions between pre-training and post-training mental health knowledge questions. The Wilcoxon signed rank test was used to evaluate pre–post changes in the Mental Health Confidence Scale and the Personal Stigma Scale. Unadjusted and adjusted linear mixed models were used to compare Mental Health Confidence Scale and the Personal Stigma Scale outcomes before and after the A.I.D. training. For all statistical analyses, a P-value of < 0.05 was considered statistically significant.

In each analysis, we only retain complete cases.¹ While multiple imputation methodologies retain a higher sample size, they can introduce bias in samples like this one where the missing data is predominantly in the outcome variables and we cannot assume that the missing Y values are ignorable (see reviews in Hughes et al., 2019; Von Hippel, 2007). Thus, the sample size for each analysis below varies based on the number of missing respondents for each model’s variables. This approach maintained an adequate sample size for statistically significant power (e.g. the smallest sample for a complete case analysis is 743, Table 3), though we identify some limitations of our conclusions due to missing data in the Discussion.

Results

Demographics

Table 2 displays frequencies or means for demographics variables. Participants ranged from 17 to 71 years old. The average age of the study participants was 30.35 years (SD = 14.64 years). The vast majority of participants were non-Hispanic White (82.02%) and ciswomen (64.03%).

Table 2. Demographic characteristics of the participants (Total N = 1,121).

Variables	Mean (SD) or n (%)
Age (N = 1,002)	30.35 (14.64)
Race/Ethnicity (N = 1,068)
Non-Hispanic White	876 (82.02%)
Non-Hispanic Black	71 (6.65%)
Other	121 (11.33%)
Gender (N = 1,062)
Ciswoman	680 (64.03%)
Cisman	361 (33.99%)
Other	21 (1.98%)

Note: The sample size for each variable varies due to missing data.

Mental health knowledge

McNemar’s test was conducted on all five mental health knowledge questions since both time (pre-training and post-training) and response (incorrect and correct) are binary variables. The results presented in Table 3 suggest a significant increase in mental health knowledge post-training, as the proportions of correctly answered responses for each mental health knowledge item does change over the course of the intervention (all p-values < 0.001).

Table 3. Assessment of percentage of change in mental health knowledge.

Mental health knowledge	Pre-test		Post-test		% of Correct Change	McNemar’sp-value
	Correct	Incorrect	Correct	Incorrect
	%	%	%	%
Q1. Definition (N = 884)	46.38	53.62	82.92	17.08	36.54	<0.0001
Q2. Continuum (N = 913)	90.47	9.53	97.15	2.85	6.68	<0.0001
Q3. Stigma (N = 894)	73.38	26.62	92.73	7.27	19.35	<0.0001
Q4. Causes (N = 809)	88.88	11.12	92.21	7.79	3.33	0.0003
Q5. Myth (N = 743)	84.39	15.61	93.27	6.73	8.88	<0.0001

Note: Only complete cases are analyzed, leading to variable sample size per question due to missing data.

Mental health confidence scale

The Wilcoxon signed-rank test indicated there was a statistically significant improvement in confidence in implementing the steps of the A.I.D. action plan from pre-training [Median (IQR) = 10 (8–12)] to post-training [Median (IQR) = 12 (10–14)], p < 0.0001; Table 4. Results from the linear mixed models demonstrated that time (pre-training or post-training), age, and gender significantly affected mental health confidence (Table 5). Participants reported an average increase of 2.110 [95% CI: 1.980–2.240] points on the mental health confidence scale post-training compared to pre-training. With every 1-year increase in age, mental health confidence decreases by 0.037 [95% CI: −0.048 to −0.026]. Compared with ciswomen, participants who identified as cismen scored an average .547 [95% CI: −0.863 to −0.272] points lower on the mental health confidence scale. This analysis found no significant racial-ethnic differences in mental health confidence, though see the Discussion for further detail on this finding, nor did it find significant differences mental health confidence scores between ciswomen and other genders.

Table 4. Assessment of differences in mental health confidence and personal stigma among participants.

Scale	Pre-training		Post-training		z-Value	Wilcoxon signed rank test p-value
	Mean (SD)	Median (IQR)	Mean (SD)	Median (IQR)
Mental health confidence	9.84	10	11.95	12	23.10	<0.0001
(Range3–15; N = 891)	(2.66)	(8–12)	(2.29)	(10–14)
Personal stigma scale	16.89	16	15.28	14	16.80	<0.0001
(Range9–45; N = 946)	(5.40)	(13–20)	(5.90)	(11–19)

Note: Only complete cases are analyzed, leading to variable sample size due to missing data.

Table 5. Associations between time and demographic characteristics on the changes of mental health confidence in a linear mixed model.

	Model 1 (N = 891)			Model 2 (N = 882)
	Estimate	P-value	95% CI	Estimate	P-value	95% CI
Time
Pre-training	Ref	–	–	Ref	–	–
Post-training	2.111	<0.0001	1.983 to 2.244	2.110	<0.0001	1.980 to 2.240
Age				−0.037	<0.0001	−0.048 to −0.026
Race/ethnicity
Non-Hispanic White				Ref	–	–
Non-Hispanic Black				0.315	0.303	−0.285 to 0.915
Other				−0.254	0.236	−0.718 to 0.209
Gender
Ciswoman				Ref	–	–
Cisman				−0.547	<0.0001	−0.863 to −0.272
Other				−0.153	0.763	−1.147 to 0.842

Note: Only complete cases are analyzed, leading to variable sample size due to missing data.

Personal stigma scale

The Wilcoxon signed-rank test indicated there was a statistically significant reduction in personal stigma scores from pre-training [Median (IQR) = 16 (13–20)] to post-training [Median (IQR) = 14 (11–19), p < 0.0001; Table 4]. Results from the linear mixed models demonstrated that time (pre-training or post-training), age, and gender significantly affected personal stigma (Table 6). Participants scored an average 1.871 [95% CI: −2.106 to −1.636] points lower on the personal stigma scale post-training compared to pre-training. With every 1-year increase in age, personal stigma scores increased by 0.049 [95% CI: 0.025–0.072]. Next, compared with non-Hispanic Whites, other racial-ethnic groups scored an average 1.842 [95% C: I0.796–2.889] higher on the person stigma scale. Finally, compared with ciswomen, cismen reported scored an average 3.817 [95% CI: 3.106–4.529] points higher on the personal stigma scale. This analysis did not find significant differences in personal stigma scores between non-Hispanic Whites and non-Hispanic Blacks, though see the Discussion for further detail on this finding, nor did it find significant differences in personal stigma scores between ciswomen and other genders.

Table 6. Associations between time and demographic characteristics on the changes of personal stigma in a linear mixed model.

	Model 1 (N = 946)			Model 2 (N = 877)
	Estimate	P-value	95% CI	Estimate	P-value	95% CI
Time
Pre-training	Ref	–	–	Ref	–	–
Post-training	−1.763	<0.0001	−1.990 to −1.536	−1.871	<0.0001	−2.106 to −1.636
Age				0.049	<0.0001	0.025 to 0.072
Race/ethnicity
Non-Hispanic White				Ref	–	–
Non-Hispanic Black				−.043	0.950	−1.304 to 1.304
Other				1.842	<0.0001	0.796 to 2.889
Gender
Ciswoman				Ref	–	–
Cisman				3.817	<0.0001	3.106 to 4.529
Other				−2.179	0.058	−4.437 to 0.078

Note: Only complete cases are analyzed, leading to variable sample size due to missing data.

Discussion

Increases in mental health concerns, coupled with only about half of the people living with any mental illness in the US receiving professional mental health services (SAMHSA, 2022), has led to offering MHATs to attempt to reduce barriers to seeking help. A.I.D. is a novel MHAT created in 2019 to address limitations (e.g. length, format) imposed by other MHATs, such as MHFA, the discontinued Kognito, and QPR. The current study aimed to assess the effectiveness of A.I.D. by comparing participant learning and attitudes before completing and after completing the training.

After participants completed A.I.D., there was a significant increase in correct responses to each of the five mental health knowledge questions. The questions assessed various aspects of mental health relating to mental health terminology, definition of mental illness stigma, and myths regarding mental illness. The results are consistent with other MHATs that have demonstrated increases in knowledge after the training (Aldrich et al., 2018; Litteken & Sale, 2018; Mitchell et al., 2013), which could reduce barriers to seeking professional mental health services (National Council for Mental Wellbeing, 2018).

A second aim of the study was to assess confidence in implementing the A.I.D. action plan’s steps. Similar to Banh et al. (2019) demonstrating improvement in confidence in improving mental health help after completion of MHFA, participants indicated being significantly more confident in implementing the steps in the A.I.D. action plan after the completing the training, as measured by the mental health confidence scalerecognising and being aware of when someone is exhibiting signs of mental distress, engaging and interacting effectively with someone exhibiting signs of mental distress, and directing someone exhibiting signs of mental distress to appropriate resources. However, older participants and cismen reported not being as confident in implementing the A.I.D. action plan (with older adults scoring an average of 0.037 points lower per additional year of age and cismen scoring an average of 0.0547 points lower than ciswomen on our mental health confidence scale; Table 5). The content of the A.I.D. training was not created specifically for any one age group or gender but was instead designed to be inclusive of a diverse group of participants. However, this finding may indicate the need to review and revise training content to be more specific and inclusive of all age groups and genders or have specific scenarios and examples that can be utilised with specific populations. Moving forward with the development of A.I.D., making such demographically targeted modules will be a priority.

The final aim of the study was to assess the effects of the A.I.D. training on personal stigma surrounding mental illness. After completion of the training, participants reported a significant 1.763-point average reduction in personal stigma. However, there were some notable demographic differences in personal stigma (Table 6). First, older participants reported higher levels of personal stigma compared to younger participants, with an average increase in .049 points on the personal stigma scale per additional year of age (e.g. a participant twenty years older would score an average 0.98 points higher on the personal stigma scale) (Table 6). Mental illness has a long history of being stigmatised in our society (Rössler, 2016). Due to various mental illness stigma reduction efforts (e.g. national mental illness stigma reduction campaigns, celebrities sharing their personal stories of mental illness), we have witnessed a generally higher level of acceptance of people living with some mental illnesses, such as depression, though some illnesses remain highly stigmatised (e.g. schizophrenia; Pescosolido et al., 2021). However, older adults, who have lived longer and potentially have had more experience with negative stereotypes regarding mental illness, may continue to have elevated rates of personal stigma. Second, other racial-ethnic groups were found to score an average of 1.842 points higher on the personal stigma scale compared to non-Hispanic Whites. This dovetails with the general trend of racial-ethnic groups outside non-Hispanic Whites reporting greater mental illness stigma in the United States (Misra et al., 2021). While we found no significant differences in personal stigma between non-Hispanic White and non-Hispanic Black participants, this finding may have been influenced by missing data issues (discussed below). Third, cismen scored an average of 3.817 points higher on the personal stigma scale than ciswomen (Table 6). This finding echoes prior research, which has long identified cismen as more stigmatising towards mental illness than ciswomen (e.g. Mann & Himelein, 2004; Pescosolido et al., 2021). A.I.D. content may need to be reviewed and revised to be more beneficial for all participants. Again, a priority for the development of A.I.D. moving forward will be creating modules that target specific audiences (e.g. older generations, diverse racial-ethnic groups, and cismen).

There are some key limitations to this study. First, the majority of participants in the study were predominantly non-Hispanic White college students and/or teachers/faculty/staff at a school or university. The A.I.D. training was offered as part of a Substance Abuse and Mental Health Services Administration (SAMHSA) grant, which targets specific catchment areas, which limited the ability to offer the training in other, more diverse areas. Furthermore, a logistic regression analysis of missingness found that non-Hispanic Black participants were significantly less likely to complete the mental health confidence scale and personal stigma scale; similarly, older participants were significantly more likely to not complete the mental health confidence scale (analyses available upon request). This introduces bias to analyses that used those variables and particularly problematises our findings regarding non-Hispanic Black respondents due to that demographic’s already small N in our sample (Table 2). Collectively, this means that the A.I.D. training’s effectiveness needs further study in more diverse populations. Additionally, because data was only collected from individuals who underwent A.I.D. training, this is an uncontrolled trial design, which are likely to overestimate effect sizes due to the lack of a comparison group (Nair, 2019). Finally, the post-training data was collected immediately after the A.I.D. training without additional follow-up, which means that the lasting effects of mental health confidence and personal stigma were not measured.

Future directions for researching the A.I.D. training include addressing the previously mentioned limitations (e.g. including a more diverse sample, adding a comparison group). In addition, analyses of training data by type of participant (e.g. student, faculty/staff) could provide insight into the training’s effectiveness for specific sub-populations. Furthermore, analysis of method of delivery (i.e. asynchronous compared to synchronous) could further guide the most effective delivery method.

The current study demonstrated the novel A.I.D. training is an effective MHAT. Unlike some MHATs (Table 1), A.I.D. provides a MHAT option that is brief, can be delivered in different formats (e.g. virtual or in-person), and can accommodate large group sizes. This training presents organisations with another MHAT option to best accommodate the needs of organisations and the individuals receiving the training.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Substance Abuse and Mental Health Services Administration.

Notes

1 Supplemental linear mixed models of all cases using maximum likelihood estimation methods for missing data were also run for both the mental health confidence and personal stigma analyses, with nearly identical results to the complete case analyses presented here. These supplemental analyses are available upon request.

Appendices

Appendix A

Mental health knowledge questions

1. ‘A state of successful performance of mental function, resulting in productive activities, fulfilling relationships with people and the ability to change and cope with adversity’ best defines:

   • Mental health *

   • Resilience

   • Self-efficacy

   • Positive psyche

   • Unsure or don’t know

2. Which of the following statements is more accurate?

   • Mental health and wellness can range on a continuum from ‘healthy’ to ‘crisis’ *

   • Mental health refers only to severe issues such as major depression or schizophrenia

   • Self-help, such as meditation is not helpful to individuals suffering with a mental illness

   • Unsure or don’t know

3. Which of the following best describes ‘mental health stigma’?

   • Viewing an individual in a negative way because they have a mental health issue *

   • Believing that mental illness is a biological condition that can only be helped with medication

   • Not being able to afford therapy for a mental health problem

   • Unsure or don’t know

4. One of the main causes of mental illness is lack of self-discipline and will power.

   • True

   • False *

   • Don’t know

5. A person with a mental illness is more dangerous than a person without a mental illness.

   • True

   • False *

   • Don’t know

Note. * Indicates correct answer.

Appendix B

Mental health confidence scale

Read each of the five statements below and then circle your level of agreement from 1 (‘do not agree at all’) to 5 (‘strongly agree’).

Table

Currently, I am confident that I can … Recognise and be aware of when someone is exhibiting signs of mental distress.	1 Do Not Agree at All	2	3	4	5 Strongly Agree
Currently, I am confident that I can … Engage and interact effectively with someone exhibiting signs of mental distress.	1 Do Not Agree at All	2	3	4	5 Strongly Agree
Currently, I am confident that I can … Direct someone exhibiting signs of mental distress to appropriate resources (e.g. self – help information, crisis hotline number, counselor).	1 Do Not Agree at All	2	3	4	5 Strongly Agree