The mental health of Australians 2

11. Methodological issues and comparison of findings

Page last updated: May 2009

11.1 Estimating the true prevalence of mental disorders
11.2 Comparison with 1997 National Survey of Mental Health and Wellbeing
11.3 Comparison with other mental health surveys
11.4 Conclusions

11.1 Estimating the true prevalence of mental disorders

The 2007 National Survey of Mental Health and Wellbeing underestimates the true extent of mental disorders in the Australian population. The exact level of underestimation is unable to be determined, however it is considered to be small. The main reasons for this are as follows.

Firstly, the survey did not include modules to determine the prevalence of schizophrenia and other psychotic disorders, somatoform disorders, eating disorders, impulse-control disorders and personality disorders. These disorders are only likely to contribute a few extra percent to the prevalence of mental disorders in the total Australian population due to their lower prevalence and their likely overlap with other disorders covered in the survey.

Secondly, like other similar surveys of the general population, the 2007 survey only interviewed people living in households and not those in institutions, nursing homes, prisons, and other specialist settings. While these non-household groups cover populations known to have a higher likelihood of mental disorders, these people make up a relatively small proportion of the total population aged between 16 and 85 years. Therefore non-inclusion of these groups does not greatly affect the overall prevalence.

Thirdly, the interview asked about the symptoms of mental disorders at any time in the respondent's lifetime. It is possible that milder symptoms, or those that occurred a long time ago, may have been forgotten.

Lastly, the response rate of the 2007 survey was considerably lower (60%) than that for the 1997 survey (78%). It is possible that the people who did not participate may have had a higher likelihood of meeting diagnostic criteria for mental disorders. If this is the case then the lower response rate in the 2007 survey may have led to a greater underestimation in prevalence compared to the 1997 survey. An intensive non-response survey was carried out by the Australian Bureau of Statistics to examine the impact of non-response on the accuracy of prevalence estimates. The results revealed that there is unlikely to be any major impact at the aggregate level and that the results of the survey are considered representative of the Australian population in terms of standard demographic factors. Top of page

11.2 Comparison with 1997 National Survey of Mental Health and Wellbeing

The 2007 survey is the second national mental health survey carried out in Australia. A similar survey was carried out in 1997 and it is useful to reflect on the similarities and differences between the findings of the two surveys.

In 2007 one in five Australians experienced a mental disorder in the previous 12 months. The same prevalence was found in 1997. In 2007, as in 1997, anxiety disorders were the most prevalent mental disorder.

The 2007 survey also reinforced the 1997 findings that mental disorders are associated with significant levels of disability and distress. Around one in three people in 2007 who met diagnostic criteria for a mental disorder in the previous 12 months had seen a health professional for their mental health, a figure that is very similar to that which was found in 1997.

While it is possible to make comparisons between the 1997 and the 2007 surveys, such comparisons should be made in the context of the similarities and differences between the methodologies used in the two surveys.

With regard to the similarities, both surveys assessed mental disorders according to the criteria set out in ICD-10 (as presented in this report) and also in the DSM-IV. Both surveys focussed on the same set of common mental disorders and the order in which these disorders were covered in the interview was roughly the same in both surveys. Both surveys interviewed a representative sample of the Australian adult population living in households. In terms of service use, both surveys asked about contact with the same major categories of health professionals who are most likely to provide help for mental health problems (that is, general practitioners, psychiatrists, psychologists, other mental health professionals and other health professionals). Both surveys asked the same set of questions regarding the perceived need for mental health care.

While every effort was made to maintain comparability between the 1997 and the 2007 surveys, there are also a number of significant differences. Firstly, there are methodological differences between the instruments used in the two surveys. The 1997 survey used version 2.1 of the CIDI as the base diagnostic instrument, whereas the 2007 survey used the World Mental Health Survey Initiative version of the CIDI, version 3.0. Substantial modifications were made to the CIDI to create version 3.0. These include changes to the number and content of questions used to tap the diagnostic criteria, changes to the structure of the interview specifically with regard to the placement of diagnostic screener questions in a separate early module, and changes to the sequencing of questions within diagnostic modules. Even small changes to the wording of a questionnaire can result in large differences in the extent and type of information elicited from respondents. Therefore, caution should be exercised when making comparisons between surveys that use different diagnostic interviews.

Another major difference between the interviews used in the 1997 and 2007 surveys relates to the timeframe used to assess the diagnostic criteria for mental disorders. In the 1997 survey the timeframe was the 12 months prior to the survey. In the 2007 survey the timeframe was the respondent's entire lifetime. An estimate of 12-month prevalence from the 2007 survey was derived from a combination of the lifetime prevalence of mental disorders and the presence of symptoms in the last 12 months. This estimate is not based on a comprehensive assessment of all diagnostic criteria within the 12 months prior to the survey. It is difficult to determine the magnitude of bias (if any) associated with a 12-month prevalence estimate derived in this way.

The enumeration period differed between the two surveys with the 1997 survey taking place between May and August and the 2007 survey taking place between August and December. Seasonal differences in the prevalence or impact of mental disorders between these times of the year are considered unlikely.

It should also be noted that the 1997 survey interviewed people aged 18 years and over, while the 2007 survey interviewed people aged from 16 to 85 years.Top of page

11.3 Comparison with other mental health surveys

In recent years, mental health surveys have been conducted in at least 28 countries around the world, including the United States, France, Ukraine, Israel, India, China and New Zealand. Collectively, these surveys form the World Mental Health Survey Initiative. All these surveys make use of version 3.0 of the CIDI, thus enhancing the ability to perform cross-national comparisons of the prevalence and impact of mental disorders around the world.

Australia has one of the highest rates of mental disorders compared with these other countries. However, the findings are remarkably similar to those found in the nationally representative survey of mental disorders carried out in New Zealand in late 2003 and early 2004 as part of the World Mental Health Survey Initiative. Notwithstanding the fact that the New Zealand survey contained a somewhat different set of mental disorders and the data was reported with respect to the DSM-IV classification system, the prevalence of 12-month mental disorders in the New Zealand survey was remarkably similar to that found in Australia (20.7% compared to 20.0%). Anxiety disorders were also the most common class of mental disorder and the sex differences in prevalence follow the same patterns. The New Zealand survey also confirmed the high levels of disability associated with mental disorders.

11.4 Conclusions

The 2007 National Survey of Mental Health and Wellbeing provides unique data, particularly with regard to the prevalence and impact of mental disorders, and service use for mental health problems. Further analyses will provide invaluable information on the complex relationship between symptomatology, diagnosis, comorbidity, the experience of mental disorder, perceived needs for care and use of services, which can be used to guide service planning and mental health activities into the future.