Search strategy for identification of studies

Three authors (R.E., S.N., and A.A.) independently performed a literature search in two electronic databases: PubMed and PsycINFO. Two authors (R.E. and H.H) explored the Medline database for studies on MHL that had been published in any journal through July 2019 (without restriction to year of publication). A Boolean/phrase search was performed on each database, with search terms on the main concepts of interest: health literacy (concept 1), mental health (concept 2), and GCC countries (concept 3). (Details of the search strategy are included in the S1 Appendix).

Inclusion and exclusion criteria

Inclusion criteria were: (a) cross-sectional studies with no time restrictions, (b) written in English, (c) age 18 and above, (d) conducted in any of the GCC countries, and (e) at least one outcome measure of the main components of MHL: knowledge of mental illnesses and their treatment, stigmatizing attitudes towards mental illnesses, and seeking help for self and offering help.

Study selection

Four authors R.E., S.N., A.A., and H.H. independently screened titles and abstracts, and excluded studies that were not relevant to the topic. They reviewed the full-texts of articles. First, database searches were exported into a master folder. All titles and abstracts were screened by R.E. and H.H. and then screened by S.N. or A.A. to assess eligibility for full-text printing and screening of references. Further, these authors independently screened all excluded titles and abstracts. If there was a disagreement, it was discussed with M.B. or N.E. to reach a final decision.

Data extraction

Independent data extraction of studies was performed by all four authors (R.E., S.N. A.A. and H.H.), to compare data and reach consensus. The following were extracted from each one: country, title, authors, time of study, design, population group, sample size, outcome measures, and the Newcastle-Ottawa Scale (NOS) score of the study. M.B. reviewed and adjusted the tables.

Critical appraisal method

The quality of the studies and related bias were assessed by using NOS, adapted for cross-sectional studies [11]. This tool evaluated three quality parameters (selection, comparability, and outcome), divided across eight specific items. Each item on the scale was scored from one point, except for comparability, which can be adapted to the specific topic of interest to score up to two points. Thus, the maximum for each study is eight, with studies having less than four points identified as representing low quality. In order to minimize a subjective interpretation of bias in scoring the NOS, two independent authors should typically have scored each paper. All studies were assessed for quality in three domains: study selection, comparability, and outcome, with two authors (R.E. and H.H.) independently scoring the domains. When independent evaluations of the ranks differed between authors, they discussed the issues with a third author (A.A or S.N) to reach consensus.

Additionally, for the results synthesis in this systematic review, we have critically reviewed and scrutinized the results and discussion of each study for outcomes, limitations and bias that authors may have had highlighted in their studies. Following that we planned to combine and examine various related ideas in literature, to show how included results, outcomes, and limitations fit together, and present them in a unified form. Finally, we used the items from CASP tool to draw limitations that were faced in this review, and further elaborate on them in the discussion.

Outcomes

Studies must include at least one outcome measure, which was categorized as mental health attitudes (i.e., stigma, prejudice), knowledge of mental health (i.e., disorder and symptom recognition), or behavior regarding mental health (i.e., intended or actual help-seeking).

Characteristics of the studies and participants

Table 1 summarizes characteristics of reviewed studies: country, authors, study design, population, sample size, setting of study, measurement tools, and outcome. Overall, a total of 27 studies were included in the current systematic review, all of which were published after 2002. Regarding the setting, almost half the studies (13) were conducted in Saudi Arabia, six were conducted in the UAE, and four were conducted in Qatar (4). The remaining studies were conducted in Oman (2) and Kuwait (2). No studies could be found in Bahrain.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of the studies included in the systematic review.

https://doi.org/10.1371/journal.pone.0245156.t001

The included studies involved 16,391 participants who were primarily adults from the community. In six studies, participants were healthcare professionals [12, 19, 24, 26, 27, 37], and five studies were conducted among college students [16, 23, 25, 28, 31, 36]. In general, females constituted more than half of all study participants. Moreover, the majority of studies stated the definition of at least one component of MHL and explored the possible sociodemographic factors believed to influence the level of mental health knowledge, attitude, stigma, and MHL overall; specifically, gender [22, 31, 35], marital status, level of education [21], ethnic groups [33].

Common demographic predictors of a lower level in recognizing mental illnesses across studies were younger age, unemployment, illiteracy, and female gender. To the contrary, one paper from Saudi Arabia showed that gender and type of education (medical and non-medical students) were not significantly associated with the level of MHL. However, such findings cannot be generalized, as they included students from only one university [23].

Level of mental health literacy

Three terms: knowledge, attitude, and practice were utilized in KSA and Qatar studies. The outcomes were similar, and a low level of knowledge, attitude, and practice were seen in KSA [20, 21], UAE [27] and Qatar [33, 34]. However, health professionals and students showed an intermediate level of knowledge in KSA [22, 23] and UAE [26]. Few publications reported only the attitude towards mental health, despite that results were similar across the studies, which agreed on the high level of stigma and shame [26, 31, 38]. However, none were conducted to explore self-efficacy and MHL in the GCC region.

Almost all included studies in this review revealed an average level of MHL, with some studies showing lower levels than 50% of participants who could not recognize some common mental disorders. Also, the majority of papers revealed a high proportion of participants had negative attitudes toward mental disorders and people with mental illnesses. Some scholars focused on the identification of mental health disorders in general, their risk factors, common symptoms, and treatments or intervention options [22, 23, 31–38]. Others selected specific mental health illnesses, such as depression [21, 26, 27] bipolar illness [20], post-traumatic stress disorder (PTSD), or psychosis [26, 27].

Researchers in the UAE focused primarily on MHL among healthcare professionals and students. Results of the study on pediatric department medical staff revealed an average level of MHL and limited recognition of common mental disorders. Additionally, there was a significant association between any form of psychosocial distress and choosing the correct depression diagnosis (P = 0.01) [26]. In a second study of nurses, less than half of respondents accurately identified selected mental disorders in specific cases and their appropriate evidence-based interventions; data showed no significant association with problem recognition and beliefs about interventions [27], while nurses in a Kuwaiti study conveyed a significantly negative attitude toward mental illness [37, 38]. Healthcare participants from Saudi Arabia demonstrated a high level of knowledge on anxiety and depression, especially among family practitioners and specialists [19].

Quality of studies

Table 2 shows the scoring of studies using the NOS tool. First, selection of adult participants in reviewed studies was from different populations, but not usually justified. Most studies reported specific inclusion and exclusion criteria. Probability sampling was conducted in few studies, which renders results representative of the selected population, such as random sampling [38], cluster probability [26, 27], and multistage stratified sampling design [12, 19, 32–34]. The remaining studies utilized nonprobability sampling from the community [13–15, 22–25, 28–31].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Scoring of reviewed studies using Newcastle-Ottawa Scale.

https://doi.org/10.1371/journal.pone.0245156.t002

Moreover, a high level of non-response was noted in a few studies, such as 77% [26], 53% [27], and 69% [37], which could probably have been avoided if they analyzed the characteristics of non-respondents. Some studies where purely descriptive, with no comparison among different sub-groups, which compromised their data via the confounding effect [26]. Results were not subjected to multivariate analysis, such as structural equation modeling; they did not test causal theories of how individual difference factors or experiences may have influenced their results.

Studies were heterogeneous in terms of outcome measures. The vast majority of papers utilized valid and reliable measurement tools; these tools were piloted and tested within the populations, thereby making these studies less vulnerable to measurement bias. Some studies adapted validated tools such as community attitudes toward the mentally ill, the CAMI Scale [37, 38], while others used measurement instruments that fulfilled the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSMV) [26, 27].

Furthermore, Quantifying the differences in means of levels of mental health literacy that were reported across included studies was not possible due to differences in measuring the outcomes and the scales used in addition to the selection of certain mental diseases to assess the level of mental health literacy related to them across different population groups. For example, adult females, health care providers or general adults.

Comparisons with previous evidence

Many studies suggest that a higher level of MHL is expected in the general population in developed countries, compared to their counterparts in developing regions. For instance, Furnham and Hamid in a systematic review (2014), examined a decade of research on MHL and found that citizens in developed countries had greater MHL: for example, the majority of reviewed studies showed that people had a greater recognition of depression than schizophrenia [8]. Authors justified their results in that the research from developing regions often used medical jargon and obscure language for the lay individual; in addition, restricted use of a written format for assessments rather than other forms of communications, such as audios and visuals; could complicate the situation [8]. Our study focused on the public, in which average MHL was found; these individuals were unable to recognize symptoms of common mental illnesses (e.g., depression and schizophrenia). A number of influencing factors could account for an opposing conclusion, such as literacy levels, cultural boundaries, and religious beliefs.

Gender differences varied across the literature; however, we found no significant difference between men and women in the GCC region in levels of MHL overall. To the contrary, Western studies show that women display better MHL than men [40, 41]. Other studies indicate that men are less able to correctly identify symptoms of mental illness in case vignettes, but are more likely to suggest self-help treatments [42–44]. Authors rationalize this by claiming that women favor psychological explanations for causes of mental illness, and are more open to psychological interventions, which is not the case in the Arab region, where mostly both men and women tend to link emotional and mental issues with religious and social believes.

Knowledge about mental health illness is lacking among the public as well as health care providers, in the available literature nowadays, despite the general perception that health care providers are more equipped to deal with patients suffering from mental illnesses. In this review we found low levels of MHL among physicians, nurses, and other related health care workers. These findings are related to some studies that revealed limited knowledge and awareness of common mental disorders among health care providers, that coincides with false beliefs mostly acquired from their community and cultural environments [45, 46]. This might potentially create a real barrier in achieving accessible and effective mental health services to the public and especially to vulnerable groups within the GCC region who may struggle to reach out for their primary health provider voluntary.

Internationally, mental health interventions were used in health promotion models for individuals, schools, and communities [47]. A number of successful awareness campaigns that targeted the general public was reported in the UK, Norway, and the USA, in which media were used to reach larger audiences [48]. Additionally, an Australian trial set out to assess the effectiveness of an intervention called ‘First Aid,’ which showed remarkable results among adults in the workplace [49]. The use of such methods for improving MHL in developing countries requires further study. In this review, interventional studies were not identified in GCC countries, which can be explained by multilevel obstacles hindering strategy implementation, such as acceptability of interventions at the cultural level [50]. MHL levels in the GCC must be accelerated in the future.

This review was limited to peer-reviewed articles published in English, which could result in bias. Countries in the GCC uses Arabic as the first language spoken, however, English is the second Language spoken and the formal language used in medical field including both practice as well as academic and scientific professional research. There, is only a few numbers database that covers articles written originally in Arabic language, but they mostly cover only limited numbers of articles from wide range of topic areas such as art and engineering, but not exclusively on health nor mental health in specific. Hence choosing English was the best choice that would yield more studies in searching. While extensive search was conducted, it is possible that relevant articles were not identified, as authors were using specific scientific databases. The three databases selected were among the most commonly used in systematic reviews. Relating to similar literature on the topic from the region, these same databases were commonly and most frequently used. Never the less, including more data bases could have yield more articles, but it wasn’t feasible nor convenient to add more search engine in the allocated time of the review. Authors also did not include information from other sources, such as unpublished reports from educational institutions or relevant literature. Our authors did not contact other authors to clarify vague information in reviewed studies. Thus, the evaluations may poorly assess the study quality, when details are not included in the reports.