This paper is concerned with the mental health literacy (MHL) specifically about the anxiety disorders (ADs). The term MHL is defined as “knowledge and beliefs specifically about mental disorders which aid their recognition, management or prevention” [1]. MHL encompasses several aspects; including recognition of disorders and attitudes facilitating this and help-seeking behaviour [2]. It has attracted a great deal of interest over the previous decade. The research has gone in two directions: first, to look at MHL for very specific disorders like schizophrenia [3], depression [4] or the conduct disorders [5] and second, to look the MHL of certain groups like young people [6-8].

A plethora of work has been conducted into the MHL of depression and schizophrenia [4]. Recognition of depression is found to be very high, with schizophrenia close behind [2,3]. The majority of studies report successful recognition as “correct” labelling of a disorder; defined as using the currently accepted psychiatric terminology. Using vignette methodology it was found that 75.6% “correctly” recognised depression [4]; similarly “correct” identification for depression has been reported at 97.2% and schizophrenia at 61% [9]. Through reporting on treatment beliefs, GP’s or family doctors were found to be considered the best help for depression [4]. Worryingly, over 43% of participants considered the cause of depression to be weakness of character, which could indicate why, despite a high recognition of the disorder, little is still known about its cause and which may affect help-seeking patterns. However, 73.5% thought the cause of depression to be genetic and 90% due to childhood abuse.

Research into the MHL concerning any or all of the ADs seems scarce. ADs include generalised anxiety disorder (GAD), panic disorder (with and without agoraphobia), obsessive compulsive disorder (OCD) and posttraumatic stress disorder (PTSD) (APA DSM-IV-R).The new proposals for the DSM-5 due to be published in 2013, have separated agoraphobia as a distinct disorder separate from panic disorder. The criteria have also proposed to remove OCD and PTSD from anxiety disorders into their own separate class of disorders [10]. However, in terms of this paper, OCD and PTSD will be included due to their current diagnosis as anxiety disorders.

ADs are amongst the most common mental disorderswith prevalence rates ranging from 13.6% and 28.8% in Western Countries [11]. An overview of results has been provided from 14 mental health surveys including one which reported anxiety disorders as the most prevalent class of disorders at 18.1% [11,12]. These studies also report that specific phobias and social phobia (or social anxiety disorder) are the most common of anxiety disorders. Similarly reports show that in US adolescents specific phobia, social phobia and separation anxiety are the most common disorders; and research reports an even higher prevalence rate of 39.1% for anxiety disorders overall [13].

A few MHL studies have focused on very specific anxiety disorders, such as one which investigated social phobia and PTSD recognition rates, not taking into account the other anxiety disorders [14]. Using phone interviews rates of recognition for PTSD were reported to be similar to that of schizophrenia (with around 1/3 providing the “correct” label) but only 9.2% of participants “correctly” labelled social phobia; a much lower rate. Participants more commonly used labels such as anxious, shy and low self-confidence for social phobia; from which the researchers concluded they were less likely to see it as a mental disorder.

Knowledge of PTSD has been compared to depression using vignette methodology [15]. This study found a significant difference between PTSD and depression in terms of a “correct” differential diagnosis (67.5% to 94.4%), “correct” prescription and treatment options. Similarly the prevalence of PTSD was found to be underestimated by GPs [16], and rather than referring to the preferential treatment of psychological therapy most medicated patients used SSRI’s (serotonin-specific reuptake inhibitors). Both papers suggest even in primary care with trained professionals there are some gaps in mental health literacy for PTSD, particularly in concern with treatment beliefs, which would imply the general public’s knowledge for PTSD treatment would be even worse since they do not have training.

Some looked at MHL of the anxiety disorders, looking at social phobia, GAD, Panic Disorder and OCD [17]. Using vignette methodology the researchers asked three questions. Results showed variability in “correct” labelling, with depression, OCD and social phobia all “correctly” assigned their label by greater than 86% of participants, whereas panic disorder was assigned by 47.7% and GAD by 41.5%. Other than panic disorder, the most popular assigned label was “medical problem” (27.6%), and for GAD 41.4% chose “general life stress” to label the problem. From this the authors concluded OCD and social phobia were recognised at similar rates to that of depression. This finding contradicts other research [14] which found poor recognition for social phobia, but this could be due to both very different methodologies and samples. A limitation of some of the research, which may have explained the high recognition for social phobia, was the use of a list of labels for the participant to choose from; this may have inflated recognition rates compared to an open-ended question being used [17]. Moreover, beliefs on whether the person should seek professional help varied across disorders and were dependent on symptom attribution; with a higher proportion suggesting treatment if they thought the problem was caused by mental illness compared to factors such as stress or personal weakness.

The current study attempts to study MHL about the ADs in a population of young people. One issue regarding MHL which has not yet been investigated is whether the gender of the character described in the vignette (to be known as “vignette gender”) affects MHL. With anxiety disorders in particular, there is a much higher prevalence amongst women than men [18]. It was found that more women met the criteria for an anxiety disorder in their lifetime, with one in three women compared to 22% of men. It was also found anxiety disorders are more disabling in women than men; measured by a greater illness burden (number of doctor visits and days missed off work). The research did not find any significant gender differences for social phobia, though some have found a significantly higher rate in women [19]. Similarly, when examining the after-effects of earthquakes female gender was found to be a significant risk factor for developing PTSD [20]. This finding was corroborated by researchers who found PTSD 10 months after an earthquake in 51.7% of females compared to 25.7% males in a population of students [21]. This higher prevalence in females suggests that people would recognise anxiety disorders better in females than male vignettes.

Past research has examined what factors predict levels of MHL including age, education level and topics studied, gender and contact with particular patient groups [7]. Numerous studies have found female gender is a significant predictor of MHL [6,22,23] though results remain equivocal with not all studies finding a sex difference in MHL.

Studies have examined the importance of a good level of MHL, particularly focusing on how this often leads to greater help-seeking. “Correct” labelling of mental disorders has been found to predict greater levels of appropriate help seeking behaviour [24,25]. Others found a diagnosis of an affective or anxiety disorder was given by 93% of GP’s when patients presented psychologising attributions, compared to only 23% when presenting somatising attributions; thus showing symptom attribution as a significant predictor of diagnosis [26].

Given the previous literature it was predicted that the recognition, treatment recommendations and perceived adjustment would vary significantly across the anxiety disorders described (H1). Next, that there would be a significantly higher proportion of “correct” labelling and sympathy for vignettes with female characters (H2). Finally it was predicted that females would demonstrate significantly higher mental health literacy than males (H3).

Of the 317 participants in the current study, 103 were first year Undergraduate Psychology students at the beginning of their course (and who had received no lectures on clinical psychology or the anxiety disorders) and 214 were non-students recruited from the general (London) population using opportunity sampling. 51.7% were female, with a mean age of 35.65 years, ranging from 18 - 80 years (SD 16.98 yrs). 63.4% of participants were British White, 23.7% Asian, 1.6% African-Caribbean and 11.4% other. In terms of marital status; 32.8% were married, 4.4% cohabiting, 47.3% single, 6.6% divorced, 3.2% widowed, 5% reported “other” and the remaining 0.6% did not report. 11.7% of participants had been treated for a (minor) psychiatric illness, but were fully recovered, and 65.6% reported having personal experience of knowing someone with a mental illness.

In the non-students group, 27.6% held a secondary school certificate as their highest qualification, 22.45 held a BA or BSc, 21.5% held “other higher qualification” (e.g. a-level/12^th grade or equivalent), 16.8% held an MA, MSc or PhD, 3.7% held only a school certificate (10^th grade), 3.7% were in full time education (e.g. an internship/army training/work-related courses), 3.3% held no academic qualifications and the remaining 0.9% chose not to report their qualifications. 87.4% of the non-student group had not studied Psychology formally at any point.

The questionnaire consisted of 8 vignettes describing the anxiety disorders; OCD, PTSD, GAD, specific phobia, social phobia, panic disorder, agoraphobia and separation anxiety disorder (available from the second author). The vignettes were selected from different sources; the vignettes for GAD, panic disorder, social phobia and OCD were taken from Coles and Coleman’s (2010) paper and the PTSD vignette was taken from Munro et al. (2004). The remaining vignettes were adapted from books: the agoraphobia vignette [27]; the separation anxiety disorder vignette [28]; and the vignette for specific phobia [29]. These were evaluated by two clinical psychologists, experts in the anxiety disorders, as face valid. Because most had been used in previous studies it is assumed that they have both content and construct validity. [7]

Each vignette was adapted to have both male or female persona, with all other characteristics and descriptions in the vignette remaining the same. There were thus two versions of the questionnaire; each participant got either one randomly with half the vignette characters as male or female. They were around 90 - 150 words long and written in English. An example for the female version of agoraphobia is given:

Celia, a 29-year-old female has fears of being in public places such as supermarkets, buses or trains in case she has an attack of anxiety and collapses. Since they began seven years ago, the symptoms have waxed and waned. At present she is housebound unless accompanied by her husband or friends. Even when accompanied she sometimes becomes anxious and may panic. She worries considerably before each outing.

After each vignette participants were asked 9 questions about the character to measure the aspects of mental health literacy. The first being an open ended question asking “What, if anything, would you say is X’s main problem?” Responses to this question were then coded into “yes” or “no” dependent on whether they provided the “correct” name of the disorder; the use of the currently accepted Psychiatric terminology. The responses were also coded into categories, in order to establish the most common labels given for each disorder. A second coder was used to ensure the reliability of coding.

The following 8 questions used 7-point Likert scales. For questions 2 - 7 the scale consisted of 1 = not at all to 7 = extremely. Questions 2 and 3 asked participants how distressing and difficult to treat they thought the problem would be, and question 4 how sympathetic they were towards the person described. Questions 5 to 8 asked about how well adjusted they felt the person was; including ratings of happiness (Q5), success at work (Q6) and personal relationships (Q7). The higher the rating (closer to 7) the better adjusted the participant thought the person was. A total adjustment score was also calculated by taking the average of the three questions. Question 8 asked on a scale of 1 = not at all to 7 = definitely, whether the person should seek help for their problem. Question 9 provided participants with a list of treatment options; none, friends, parents, other family members, GP, Psychologist/Psychiatrist, books and internet. Participants had to rate on a scale of 1 = not very likely to 7 = very likely for each option as a possible source of treatment.

Ethical approval for the study was sought and received from the departmental committee. Participants were recruited by paid post-graduate students mainly in public places (railway stations, coffee bars) whose task was collect 300 participants of mixed, sex, age and ethnic background. Around 30% refused the invitation primarily because of time constraints. Participants were given an instruction sheet and provided informed consent. Questionnaires were administered on paper, and took approximately 15 minutes to complete. After finishing participants were debriefed with a debriefing form and thanked for their participation. They were not remunerated for their participation.

Reliability analyses involved a secondrater (postgraduate) coding 10% of questionnaires using the same coding system as the author (correct/incorrect). Results demonstrated excellent reliability: Kappa 0.81.

Vignette Recognition of Anxiety Disorders

Figure 1 shows the anxiety disorders in descending order for “correct” labelling; use of the current Psychiatric terminology. It is evident that the ability to name a disorder varies, with high levels of recognition of OCD compared to very poor recognition of panic disorder, GAD, separation anxiety disorder and social phobia. Chi-square tests revealed that for all disorders, the proportion of participants “correctly” or “incorrectly” labelling the disorder was significantly different (p < 0.004). In all cases the majority gave “incorrect” labels, except OCD where the majority “correctly” labelled the disorder. Moreover, in order to assess the first hypothesis and test whether recognition (“correct” labelling) differed between the disorders, a Cochran’s Q test was used. This revealed a significant difference in the proportion of vi-