AIMS

The Biostatistics and Psychometrics group undertakes research into the structure and pattern of symptoms of particular disorders such as schizophrenia and depression.

It is also involved in the development of psychometric scales and screening tests of mental illnesses and research into the best methods to analyse data from clinical trials and interventions.

In addition to its own research program, the Biostatistics and Psychometrics group supports the work of other teams within the Institute by advising researchers on the design of experiments and trials and methods of statistical analysis.

HIGHLIGHTS

A National Depression Index for Australia
There is evidence that Australians, like residents of many Western countries, are becoming more depressed. Support for this belief comes from the difference between rates of depression in younger versus older people, and the comparison of earlier epidemiological studies with those conducted more recently. But is this apparent trend due to people being more willing to express their symptoms, more vigilant GPs, more readily available and better treatments, or due to a real increase in population levels of depression?

Unlike other serious illness where progress may be measured simply by recording reductions in the number of people dying of the illness, this is inappropriate for mental health disorders.

There is no register of cases of depression. Indeed, many people with the condition do not seek help. Also, many sufferers who have symptoms of depression but whose problems fall below the cut-off level for a formal diagnosis may be significantly disabled by their problems.

These factors mean that the only reliable statistics on the mental health of the entire population must come from surveys. However, few surveys have the resources to collect the information required to make formal, clinical diagnoses of mental health disorders.

Thanks to the work of organisations such as beyondblue and the Institute’s own Depression Awareness Research Program, the recognition and treatment of depression is increasing.

Ultimately, the question that must be asked is whether these programs have an effect at the population level. The National Depression Index (NDI) has been developed to track changes in the population and in groups over time, and to enable groups within the population to be compared and groups who may be at elevated risk of depression to be identified.

How the NDI was developed
Working with Professors Ian Hickie (beyondblue) and Tony Jorm (ANU), Andrew Mackinnon undertook the statistical development of the NDI. Confirmatory factor analysis was used to identify items most strongly related to depression from a scale developed to screen for non-specific psychiatric distress (the Kessler ‘K-10’ scale). This scale had been administered in the large National Survey of Health and Wellbeing carried out throughout Australia in the late ’90s.

This survey also included an interview that yielded diagnoses of depression according to DSM-IV criteria, so the chosen items could be ‘calibrated’ against clinical diagnoses. Factor scores were calculated for each respondent, locating them on a depression ‘dimension.’ A problem with these scores is that they have no intuitive meaning – it would be difficult to look at the mean factor score of a group and say whether it represented high or low levels of depression.

These values would be meaningless to health professionals and members of the general public. To overcome this limitation, a transformation was developed that yielded values that can be readily interpreted as risk of depression relative to the Australian population.

Using another large sample representative of the Australian population – the Australian Bureau of Statistics National Health Survey carried out in 2001 – the index was calibrated so that a value of 100 represents the ‘average’ risk: values greater than 100 represent higher risk of depression and values less than 100 indicate lower levels.

Although the steps may seem convoluted and complicated, they make maximum use of a very small amount of information and produce an index that, we hope, will be meaningful to health professionals and the general public.

The National Health Survey is repeated on a regular basis and will contain the necessary items to calculate the NDI. It will therefore be possible to track depression in the Australian population in the years to come. As the NDI is based on only six short questions, it can also be used in other more specialised surveys to evaluate the depressive status of particular sections of the population.

The figures to the right demonstrate some of the analyses possible with the NDI. Differences in NDI values are illustrated across income and employment status groups, by sex. In the first figure, a comparison of income groups reveals that those in the lowest incomebracket, especially males, are at highest risk of depression, and that as income increases, risk of depression decreases.

The data in the second figure suggest that females who are unemployed and seeking full-time work are at a highly elevated risk of depression compared with others.

PERSONNEL

Andrew Mackinnon attended the Australasian Society for Psychiatric Research Conference in Christchurch, New Zealand, in December 2003. Professor Helen Christensen, Deputy Director of the ANU Centre for Mental Health Research, visited the Institute.