Abstract
Jorm et al. [1] defined ‘mental health literacy’ as referring to the ‘knowledge and beliefs about mental disorders which aid their recognition, management or prevention’. This group [1] developed a strategy for examining the extent to which two mental disorders (i.e. depression and schizophrenia) were correctly identified from separately presented vignettes, and for then assessing respondents' beliefs about a range of issues, in particular the helpfulness of a range of interventions for each condition, and the likely outcome for those disorders with and without treatment. A series of studies was then initiated in Australia. The first survey was undertaken in 1995 and involved a national sample (recruited by the Australian Bureau of Statistics), with more than 1000 respondents being questioned about each vignette [1]. Subsequently, they recruited large samples of general practitioners, psychiatrists and clinical psychologists, and examined the extent to which members of those professional groups rated the likely helpfulness of varying professional and non-professional help and of pharmacological and non-pharmacological interventions [2]. The two databases also allowed the beliefs of the general public and the professional groups to be compared [3]. Such comparisons identified a number of distinct differences between beliefs held by health practitioners and the general public, with such information having the potential to shape mental health education campaigns ‘to help close the gap between professional and public beliefs’.
Results of the Australian studies indicate that the methodological approach can provide important data about (i) knowledge of mental disorders; (ii) beliefs about determining causes and risk factors [4]; (iii) judgements about the utility of a range of therapeutic options; (iv) views about the prognosis for each disorder: whether treated or untreated; and (v) attitudes concerning long-term functioning and discrimination [5]. It appeared appropriate to test the broad methodological approach in another country, both to allow comparisons between responder groups in Singapore and between those groups and comparable Australian respondents. To that end, a series of studies have been commenced in Singapore. Data have been obtained from psychiatric staff, while subsequent studies will involve general practitioners and the general public.
Singapore is a small island (646 km2) 2° north of the equator and with a resident population of 3.1 million (comprising some 77% Chinese, 14% Malays, 7% Indians and 1% others), with a working language of English. Psychiatric services are predominantly public. The principal psychiatric facility, Woodbridge Hospital, has nearly 3000 inpatient beds, while the other psychiatric units (both general hospital and private facilities) have less than 200 beds. Woodbridge Hospital has some 7000 admissions per year, a large staff (including some 40 psychiatrists and over 500 nurses), who are also responsible for hospital, outpatient and community-based services. Thus, a survey of Woodbridge staff could be expected to provide representative information about professional mental health literacy in the region, both for local relevance and for comparison against the Australian surveys of mental health professionals.
In an independent publication [6], Parker et al. contrasted responses of Woodbridge staff subgroups, and largely examined the comparative prevalence of varying interventions in terms of their perceived ‘helpfulness’. Here we focus on data returned by the psychiatrists in comparison with all other professional staff, and then compare responses of those psychiatrists with Australian psychiatrists.
Method
A questionnaire was developed, essentially based on the Australian methodology, but with two key extensions. Respondents received vignettes identical to the two developed by Jorm et al. [2] to meet ICD-10 and DSM-IV criteria for major depression and schizophrenia, with each vignette detailing information about a ‘Mr A’. In addition, we developed the following vignette for mania: ‘Mr A is 27 and lives with his parents. He has been employed for most of the time since leaving school, but has recently left his job as a salesman. He has never taken any illicit drugs. His parents state that in the last three weeks he has been extremely active, requiring less sleep and not appearing tired, being over-talkative and disinhibited and — on occasions — quite irritable. He claimed to have invented a machine for curing cancer and wished to go to the U.S. to sell it. When stopped by his parents he became violent, and they called the police’.
The Australian study did not have a diagnostic option set; instead questions about diagnosis were open ended and later coded into categories. Here we allowed only one diagnostic option, and presented ‘stress’, ‘depression’, ‘schizophrenia/paranoid schizophrenia’, ‘mania’ and ‘anxiety’ and a number of additional options to respect some regional explanations of mental illness, as well as available resources, in Singapore which might well be chosen by the Singaporean general public. Thus, diagnostic categories of ‘mental weakness’, ‘physical weakness’ and ‘being possessed’ were added, while the range of professional and non-professional facilities and resources was extended considerably.
Four groups of professional staff were derived. As the hospital had a large number of nurses, we elected to survey both psychiatrically and generally trained nurses, with members of both groups receiving one vignette only for questionnaire completion. Occupational therapists, clinical psychologists and social workers formed a combined allied health group, and the psychiatrists the fourth group. Due to the comparatively fewer staff in the latter two groups, those members were requested to complete questionnaires in response to all three vignettes, each distributed with a fortnight interval to decrease memory effects. Questionnaires were to be completed anonymously and returned via ‘post boxes’, thus disallowing knowledge about individual responses.
The survey occurred in early 1999, with an 81% response rate to the 495 questionnaires distributed. Questionnaire components and analytic strategies will be detailed in the next section, with analyses contrasting responses by the psychiatrists (PSYs) with all other staff (i.e. OSs).
Results
Diagnostic accuracy
Table 1 documents the response rates by staff member discipline (determined by colour coding of forms prior to distribution).
Details on questionnaires returned (and distributed) for each vignette
Diagnostic ‘accuracy’ was assessed by comparing the rates of selecting the correct vignette diagnosis from the eight nominated options. The depression vignette was correctly identified by 95% of the PSYs and 80% of the OSs, the mania vignette by 100% and 73%, and the schizophrenia vignette by 96% and 91%, respectively. Incorrect diagnostic ‘spillage’ varied. For depression, 5% of the PSYs and 16% of the OSs chose the ‘stress’ diagnostic option while the ‘other’ option was nominated by the remaining 3% of the OSs. For mania, incorrect diagnostic options chosen by the OSs were schizophrenia (12%), depression (5%), stress (4%), anxiety (3%), being possessed (2%), mental weakness (1%) and ‘other’ (1%). For schizophrenia, none of the PSYs offered a specific alternate diagnosis, while for the OSs, 4% nominated depression, while 1% (for each) chose mania, stress, anxiety, mental weakness and ‘other’.
Intervention preferences
Respondents were required to nominate which option would allow Mr A to be best helped, with 10 options listed (i.e. consulting a family doctor, a psychiatrist, a psychologist, a church minister, a counsellor, a traditional healer, attending a polyclinic, talking things over with friends or family, first recognising that he has a problem, and taking medication). For depression, 84% of the PSYs as against 34% of the OSs judged that seeing a psychiatrist was the best option. Eleven per cent of the PSYs and 23% of the OSs elected for Mr A first having to recognise that he had a problem. The only other option chosen by the PSYs was medication (5%), while the OSs residual ‘best help’ options ranged widely (21% for talking things over with friends or family; 8% see a family doctor; 7% see a counsellor; 2% see a psychologist; 1% see a church minister; 1% take medications; and 1% other).
For mania, the only option chosen by the PSYs was to see a psychiatrist. By comparison, 60% of the OSs chose that option, as against 21% electing for first recognising the problem, 6% (each) for talking things over with friends or family, and going to a family doctor, with 3% for taking medication, 2% (each) for seeing a psychologist and going to a polyclinic, and 1% for seeing a counsellor.
For schizophrenia, 92% of the PSYs and 76% of the OSs nominated seeing a psychiatrist as the best option. Eleven per cent of the OSs and 4% of the PSYs elected for Mr A first recognising that he had a problem. Of the remaining PSY-chosen options, 4% recommended going to a polyclinic. Of the remaining OS-chosen options, 4% (each) favoured a family doctor or medication, 2% talking things over with friends or family, while 1% (each) chose the counsellor and psychologist options.
Helpfulness of interventions
Respondents were required to rate the extent to which a range of resources, medications, activities and therapies might be helpful or harmful. The list (detailed in Tables 2 and 3) contained all options included in the Australian surveys and additional ones to reflect Singapore services and culture. Here, coding options allowed each intervention to be scored as 3, if rated as ‘helpful’, 2 if rated as ‘neither’ or ‘depends’, and 1, if rated as ‘harmful’, while any ‘not known’ response option was scored as missing data. We first tabulated helpfulness percentage rates as judged by the whole sample for each option, and then tabulated and compared mean helpfulness scores of the two contrast groups by use of the Mann-Whitney U-test.
Ratings by psychiatrists (PSYs) and all other staff (OSs) of judged helpfulness (assessed as percentages) of differing activities and therapies for those with depression, mania and schizophrenia, and comparisons of mean scores generated by PSYs and OSs (scaled as 3 = helpful, 2 = neither or depends and 1 = harmful)
Ratings by psychiatrists (PSYs) and all other staff (OSs) of judged helpfulness (assessed as percentages) of differing resources and medications for those with depression, mania and schizophrenia, and comparisons of mean scores generated by PSYs and OSs (scaled as 3 = helpful, 2 = neither or depends, and 1 = harmful)
Aggregated helpfulness rates (considering all respondents' views) identified psychiatrists and general practitioners rating highly as resources for all three disorders, while traditional healers, pharmacists and religious leaders returned consistently low ratings. Disorder-specific differentiation was most evident in relation to depression, where in comparison to the two psychotic disorders, counselling and assistance from friends and relatives returned differentially higher helpfulness ratings. For treatment of depression, the highest ratings (i.e. more than 50%) were returned for antidepressant medication, relaxation therapy, psychotherapy, benzodiazepines and sleeping pills. A mood stabiliser approached that criterion, electroconvulsive therapy (ECT) and antipsychotic medication were not frequently endorsed as helpful, while herbal and purging medicines, hypnosis, special diets, acupuncture and alcohol were rarely endorsed. In terms of activities and therapies, reading about people with similar problems, socialising, cutting out alcohol, being admitted to a psychiatric ward, having a holiday and rest received high endorsements.
For the management of both mania and schizophrenia, a psychiatrist was most clearly endorsed as the most helpful resource person, with moderate endorsement rates for general practitioners, hospital doctors, psychologists, social workers, family and friends, and minimal endorsement for traditional healers, religious leaders and pharmacists. As medications, antipsychotics were almost invariably rated as helpful for schizophrenia, and with moderately high endorsement for benzodiazepines and sleeping pills; while for mania, mood stabilisers, antipsychotic medication and benzodiazepines were again endorsed moderately highly. Electroconvulsive therapy (and antidepressant medication) received low endorsement for both of these conditions. As established in relation to the depression vignette, there was virtually no support for traditional or alternative healing medicines. Psychiatric hospitalisation received high endorsement, and moderately high endorsement was evident for reading about people with similar problems, ceasing alcohol, rest, psychotherapy and relaxation therapy.
Turning to direct comparison of ratings returned by the psychiatrists and other staff, mean scores provide additional information about highly helpful options (mean scores of, or close to, 3.0), intermediate options (mean scores of, or close to, 2.0) and highly harmful options (mean scores of, or close to, 1.0). Formally significant differences were evident for each disorder. For depression, the PSYs were more likely (than the OSs) to rate all other medical staff (i.e. doctors, polyclinic doctors, specialist doctors, and trending for general practitioners) and psychologists as more helpful, and less likely to rate counsellors as helpful. They returned higher helpfulness rates for ECT and mood stabilisers, but were less likely to endorse relaxation and socialisation as helpful. For mania, the psychiatrists viewed consulting counsellors, family members and friends, and psychotherapy as less likely to be helpful, but were significantly more likely to rate psychiatric hospitalisation and ECT as helpful. They were significantly less likely to view becoming physically more active, getting out more and receiving hypnosis as helpful, but were more likely to so rate ceasing alcohol.
For schizophrenia, the PSYs were more likely to nominate attendance at a polyclinic as helpful, while the OSs rated psychologists, counsellors and close friends as more likely to be helpful. The PSYs viewed psychotherapy, antidepressants, mood stabilisers, relaxation therapy and hypnosis as less helpful than did the OSs. The OSs viewed becoming more physically active and taking a holiday more optimistically than did the PSYs, while the OSs viewed the use of alcohol as significantly more harmful.
Finally, we examined whether the groups differed in nominating differing numbers of options as helpful. The PSYs nominated a greater number of resources as helpful for the depression vignette than did other staff (i.e. 174/304 vs 834/1712, χ2 = 7.5, p < 0.01), but no difference was identified for the mania and schizophrenia vignettes. The only other differences were in relation to the schizophrenia vignette only, with the psychiatrists nominating a smaller set of medication options (i.e. 64/225 vs 348/945, χ2 = 5.6, p < 0.05) and of activities and therapies (i.e. 107/425 vs 604/1785, χ2=13.6, p < 0.001) as helpful.
Perceived impact of the disorder
Respondents were required to judge the likelihood of certain outcomes for Mr A after receiving help and in comparison ‘to other people in the community’. The PSYs and the OSs did not differ in rating (across each diagnostic vignette) the likelihood of Mr A having a good marriage, being a caring parent and being creative or artistic. On the remaining outcomes (i) no differences were identified for the depression vignette; (ii) on the mania vignette, the PSYs judged a greater chance of Mr A taking illegal drugs (mean scores of 1.7 vs 1.3, p<0.01) and drinking alcohol to excess (1.7 vs 1.3, p < 0.05); while (iii) for schizophrenia, the PSYs judged Mr A as being less likely to understand other people's feelings (1.7 vs 2.2, p < 0.05), and to be more likely (p < 0.001) to be violent (1.8 vs 1.2), drink too much alcohol (1.8 vs 1.2), take illegal drugs (1.9 vs 1.2), have poor friendships (2.4 vs 1.6), and attempt suicide (2.3 vs 1.4). Comparison across the vignettes revealed a clear and consistent pattern for the PSYs to rate a diffusely worse outcome for the schizophrenia vignette and the best outcome for the depression vignette, with outcome intermediate for the mania vignette. A similar trend, but not quite so clear cut, was evident in ratings returned by the OSs.
Beliefs about discrimination
Requested to judge whether Mr A would be discriminated against by others in the community if they knew about his problems, the respective rates nominated by the PSYs were all significantly higher than nominated by the OSs, being 68% vs 32% for depression, 91% vs 70% for mania, and 100% vs 73% for schizophrenia.
Prognosis
Staff were requested to estimate the most likely prognosis, both with and without Mr A receiving appropriate professional help, and with options listed in Table 4. As detailed in Table 4, in the absence of professional help, the most commonly elected option was for the person's condition to worsen (i.e. for depression, chosen by 63% of the PSYs and 76% of the OSs; for mania, by 65% of the PSYs and 82% of the OSs; and for schizophrenia, by 92% of the PSYs and 94% of the OSs). For depression and for mania, the PSYs were more likely than the OSs to choose the recovery with reoccurrence option, while few of the OSs and none of the PSYs chose the ‘no improvement’ option. Both groups returned distinctly more optimistic prognoses in response to rating outcome if professional help was provided, with no vignette returning any judgement of Mr A either worsening or failing to show any improvement. Here, 37% of the PSYs (as against 15% of the OSs) voted for a full recovery without further problems for the depression vignette, although two-thirds of both groups favoured the recovery with reoccurrence risk option. For mania and for schizophrenia, the majority in both groups favoured recovery with reoccurrence, with few electing for the partial recovery option.
Prognosis provided by psychiatrists (PSYs) and all other staff (OSs) if patients either did not receive, or did receive professional help
Comparison of psychiatrist groups
The broad study design also allowed comparison of responses returned by the Singapore psychiatrists with those of the 1128 Australian psychiatrists surveyed by Jorm et al. [2,3], at least in regard to the two vignettes used in both studies, and with one-half receiving each separate vignette. In terms of diagnostic accuracy, 95% of the Singaporean and 97% of the Australian psychiatrists correctly identified the depression vignette, while the respective rates for the schizophrenic vignette were 96% and 79%. While, for the second, an additional 14% of the Australian psychiatrists mentioned ‘psychosis’, the Australian respondents were not given a diagnostic option set.
In terms of the helpfulness (and harmfulness) of a range of professional and non-professional help, Table 5 data allow comparison only for those options included in both surveys. For depression, percentage helpfulness rates for resources were very similar, apart from the Singaporean psychiatrists being somewhat more likely to judge antianxiety agents/benzo-diazepines, sleeping pills and antipsychotic agents as helpful. Helpfulness rates for the activities and therapies were very similar: apart from the Singapore psychiatrists being less likely to rate ‘getting out more’ and being more likely to rate ‘self-help books’ as helpful. Table 5 reports equivalent comparison data for schizophrenia. Rates for judging professional resources as helpful were very similar, while the Singaporean psychiatrists were less likely to nominate those without some level of psychiatric training (including family and friends) as helpful. For medications, both groups were in agreement about anti-psychotic medication, while the Singapore psychiatrists were much more likely to judge sleeping pills, somewhat more likely to judge antianxiety agents, and somewhat less likely to judge antidepressants as helpful. For activities and therapies, there were again more similarities than differences, but the Singapore psychiatrists were somewhat less likely to judge relaxation therapy and psychotherapy as helpful, and more likely (as for depression) to rate self-help books as helpful. The comparable rates for ECT (i.e. 40% vs 37%) are worth noting.
Comparison of responses by Singaporean (Sing.) and Australian (Aus.) psychiatrists: percentage reporting study variable helpful
In terms of disorder-generated discrimination, psychiatrists in both regions rated schizophrenia as invariably resulting in discrimination, while two-thirds judged that depression would generate discrimination: identical estimates. In terms of outcomes, we first consider ‘negative’ ones. The rating by the Singaporean psychiatrists for depression was (from most to least common): attempting suicide, taking of illegal drugs, violence, poor relationships and drinking too much, whereas for the Australian psychiatrists, the order was attempting suicide, drinking too much, poor relationships, illegal drugs and violence. For schizophrenia, the order of the Singaporean psychiatrist ratings was: poor relationships, attempting suicide, illegal drugs, violence and drinking too much alcohol, while for the Australian psychiatrists, it was attempting suicide, poor relationships, illegal drugs, drinking too much and violence. In terms of ‘positive’ outcome parameters, for both groups, ‘understand other's feelings’ rated distinctly higher for those with depression than all other options, which rated comparably. For schizophrenia, both groups showed little spread across the various outcomes.
In terms of the likely prognosis, with and without professional help, few differences were evident. For untreated depression, 63% of the Singaporean psychiatrists as against 49% of the Australian psychiatrists voted for worsening or no improvement, with 11% in both groups voting for full recovery (with or without reoccurrence) and 26% of Singaporean psychiatrists versus 36% of Australian psychiatrists for partial recovery (with or without reoccurrence). For treated depression, 37% of the Singaporean and 36% of the Australian psychiatrists voted for full recovery without reoccurrence, 63% and 59% for full recovery with reoccurrence, and with minimal representation across other categories (i.e. 0% vs. 2%).
For untreated schizophrenia, 96% of the Singaporean psychiatrists and 91% of the Australian psychiatrists voted for worsening or for no improvement. For treated schizophrenia, the respective rates for full recovery (with or without reoccurrence) were 76% versus 53%, for partial recovery (with or without reoccurrence) were 24% versus 46%, and 0% versus 0.2% for no improvement or worsening, suggesting a single difference: that the Singaporean psychiatrists were more likely than the Australian psychiatrists to judge that treatment would lead to full episode recovery rather than to a partial recover.
Discussion
The actual number of Singaporean psychiatrists surveyed was low, risking false negative results in any formal statistical analyses (particularly in comparing Singaporean psychiatrists' responses with the Australian data set), and thus argued more for qualitative interpretation. While the psychiatrists were comparatively few in number, they have service responsibility for managing the bulk of psychiatric disorder in Singapore, as detailed in the introduction [6], and it is therefore likely that the survey has obtained representative views of staff working in the public mental health sector. The small number of psychiatrists and allied health staff necessitated us asking them to respond to all three vignettes, risking a number of response biases. Our Table 1 data indicate that a number did not respond to all vignettes, an issue with somewhat unclear impact.
The methodology has a number of advantages in identifying literacy issues, of both intrinsic interest and in terms of considering the need for, and any focus of, educational and training issues. Limitations are also evident. Thus, vignettes (by design) provide limited background information for assessing helpfulness, prognostic and outcome parameters, particularly when vignettes represent a cross-sectional clinical presentation, rather than allow responses to its evolution and to the impact of first-line and subsequent interventions. Options, such as resources and medications, are not hierarchically or sequentially organised. Thus, the percentage rate for judging ECT as effective (for whatever vignette) does not represent the extent to which ECT is viewed as an initial or prioritised intervention.
We first summarise differences between the Singaporean psychiatrists and their comparison mental health professional staff. Both groups were highly likely to identify the schizophrenia vignette accurately, while the psychiatrists were superior in identification of the depression and mania vignettes. In terms of the most appropriate option for providing the ‘best help’ to Mr A, the PSYs clearly chose seeing a psychiatrist (rates of 84–100%), while the OSs evidenced quite variable first-preference rates for that option (i.e. 76% for schizophrenia, 60% for mania, and 34% for depression). For depression, the OSs chose a broad, and seemingly reasonable, set of alternate options (such as talking things over with family or friends, or going to a family doctor). A commonly chosen option by the OSs was that ‘Mr A must first recognise that he has a problem’, which implies insight and perhaps motivation, and is perhaps a difficult requirement for a psychotic patient, whether with mania or schizophrenia.
In terms of rating professional and non-professional help, both groups judged psychiatrists as extremely likely to be helpful; the psychiatrists, however, favoured a mental health professional-weighted model rather than a strict and narrow ‘medical model’, while the non-psychiatrist raters were more likely to rate highly the contribution of other professionals as well as of family and friends. In rating medications, both groups were highly likely to rate certain psychotropic drugs as helpful, but offer little or no support for medications such as tonics, purging drugs or herbal medicines. However, quite distinct differences were evident between the psychiatrists and other staff in judging the utility and/or effectiveness of psychotropic drugs for the differing conditions.
Electroconvulsive therapy received relatively low rates of endorsement across the whole group, but was judged by the psychiatrists as being more helpful for both depression and mania. The lack of a clear disorder-specific role for ECT could well reflect rating the utility of ECT for ‘major depression’ rather than for melancholic or psychotic depression (and therefore argue for not necessarily rating its helpfulness as particularly high in the vignette presented), as well as the reality that ECT can be of some help in a percentage of patients with mania. The psychiatrists clearly judged hospitalisation as likely to be of greater help for the two psychotic conditions than did the non-psychiatrists, while the latter were more likely to rate a number of less specific strategies (e.g. becoming physically more active, getting out and about more, relaxation therapy, hypnosis) as well as psychotherapy for the two psychotic conditions, suggesting that the non-psychiatrists had a wider conceptual approach. Both groups, however, judged other non-specific strategies as highly helpful, such as (for all three disorders) cutting out alcohol, not using it to relax, reading about others with similar problems and attending courses on relaxation, indicating support for a relatively pluralistic approach.
In terms of outcome, the psychiatrists demonstrated a clearer differential in the comparative prognosis of the three disorders, with schizophrenia being viewed as having the worst (and depression the best) general outcome, and with a similar differential demonstrated in relation to data on likely stigmatisation. In terms of many of the recognised sequelae of schizophrenia (e.g. impact on relationships, chance of suicide), the psychiatrists appeared to make the more accurate estimates of illness impact.
In assessing prognosis, both groups (but more so for the non-psychiatrists) were distinctly pessimistic about outcome without professional help, seemingly discounting the significant percentage of patients who spontaneously recover from a depressive episode, and both were somewhat optimistic in their estimates of the impact of professional help, in that failure to improve or worsening of the conditions was not conceded. However, it must be remembered that subjects were required to choose the most likely outcome.
We have identified a number of views held by Singaporean psychiatrists, which, by themselves, might appear region-specific. Comparison, however, with ratings generated by Australian psychiatrists revealed more similarities than differences, perhaps reflecting the parallel educational influences on psychiatrists in both regions. Not only is this general conclusion of some intrinsic interest, but it points to the advantages of undertaking mental health literacy studies in differing regions, as it can allow any effects emerging from clinical experience as well as from social and cultural factors to be examined.
The clearest example of consensual rating, and perhaps the most important one, is in relation to estimating prognoses for major depression and for schizophrenia. We have noted the seemingly overly pessimistic views about untreated, and seemingly overly optimistic views about professionally treated, major depression. A significant percentage of depressed patients experience spontaneous remission, particularly those with non-psychotic and non-melancholic expressions (i.e. ‘major depression’). Conversely, clinical experience alone reveals that a significant percentage of patients with ‘major depression’ fail to respond to trials of numerous antidepressant strategies. However, when options chosen by the Singaporean and Australian psychiatrists are compared, we found very similar option profiles (e.g. 100% of the Singaporean and 95% of the Australian psychiatrists electing for full recovery of episode). Allowing only a single option (i.e. the most likely outcome) may well disguise differences that could emerge if psychiatrists were instead required to estimate what percentage of any number of Mr As with major depression would be likely to take up the allowed prognostic options. As shaped, inviting ‘most likely’ estimates may generate quite sharply pessimistic and optimistic group responses, and not gauge the variable impact of treatment across individual patients. This raises the important caveat that must be conceded in mental health literacy studies, particularly those comparing responses by professional staff and community samples, that differences may in part be methodological. Without careful consideration of the data generated by the Singaporean psychiatrists alone, this issue might have been minimised or missed and, of greater concern, we could have judged them as excessively pessimistic about the natural history of depression and overly optimistic about psychiatric treatment. Comparative analyses with data generated by the Australian psychiatrists were then of distinct importance in establishing quite similar profiles of responses across the two regions. Thus, comparison between groups is likely to be of more importance than any modal response of a particular group.
Footnotes
Acknowledgements
We thank Dr Ang Ah Ling, Medical Director, for supporting this study, together with other members of the Mental Health Literacy Working Party (Drs Mahendran, Lee and Yap; Madam Yeo), together with all Woodbridge Hospital staff who contributed to the survey.