About Young Minds Matter
A description of the survey aims, content and methods.
A description of the survey aims, content and methods.
Young Minds Matter: The Second Australian Child and Adolescent Survey of Mental Health and Wellbeing is part of the National Survey of Mental Health and Wellbeing initiative. This is an Australian Government Department of Health funded initiative comprising national surveys of adults in the general population in 1997 and 2007, those with psychotic illnesses in 1997-98 and 2010, and the first national child and adolescent survey.
The 1998 Australian Child and Adolescent Survey of Mental Health and Wellbeing was the first national survey anywhere in the world to focus on the mental health of children and adolescents. The survey found that mental health problems were relatively common, with approximately 14% of children and adolescents having mental health problems based on scores obtained from the Child Behavior Checklist (CBCL). It also found that only one in four children and adolescents with mental health problems had attended professional services in the six months prior to the survey.
Since 1998 there have been substantial changes in the way mental health care is provided in Australia. There have been significant changes in people’s understanding of mental health. A range of new programmes and services has been developed in the education and health sectors that provide information, support, and specialist help for young people and their families. Programmes such as KidsMatter and MindMatters promote mental health through the school environment. While each jurisdiction has taken its own approach, schools are generally providing higher levels of support, including access to specialist counselling. Through the Australian Government’s Better Access programme, more general practitioners have been trained in supporting the mental health needs of children and young people and their families, and they now have greater access to Medicare-subsidised psychological therapy. Launched in 2006, headspace has rapidly expanded and provides a range of support services for young people aged 12-25 years with emotional and behavioural problems and substance use problems. Additionally, several counselling services are now available by telephone, and a range of internet-based services has also been created that provide information and counselling support online.
A second national adult survey of mental health and wellbeing was conducted in 2007. The survey found approximately one in five adults had experienced a mental health problem in the preceding year, and for a substantial proportion of these adults, their problems first started in childhood or adolescence. While access to mental health services has improved considerably, it is still recognised that for many people there can be substantial delays in seeking help, at the time when early interventions may have the greatest impact.
The Second Australian Child and Adolescent Survey of Mental Health and Wellbeing not only provides data that complements other national surveys, but also presents a contemporary picture of mental disorders and services used by those young people who have mental health problems.
The Second Australian Child and Adolescent Survey of Mental Health and Wellbeing (Young Minds Matter) was designed to provide current information about the mental health and wellbeing of children and adolescents in Australia and the extent to which they use health and education services to obtain help with problems.
The main aims of the survey were to determine:
In order to fulfil these aims a broad range of information was collected from parents and carers, and adolescents themselves. This included diagnostic modules, questions specifically tailored for the Australian health care environment and schools, and sections dealing with current practices, including the use of the internet. Background information about the environment in which young people live and develop, including the functioning of families, the mental health of their parents and carers, and their experiences at school and online, was also collected.
The survey was based on a face-to-face interview with the parent or primary carer for each selected child. When the selected child was aged 11 years or older, they were also invited to complete a confidential questionnaire on a tablet computer.
The parent and carer questionnaire took on average 60 minutes to complete. All participants were asked questions in all modules. The parent and carer interview was made up of the following modules:
Young people aged 11 years and over were asked to complete a self-report questionnaire on a tablet computer in private. The questionnaire took an average of 35 minutes to complete. As some of the questions were considered inappropriate for younger children, some modules were only asked of young people older than a specified age. The youth questionnaire was made up of the following modules:
The SDQ was used in Young Minds Matter with permission of Professor Robert Goodman, youthinmind. www.sdqinfo.org
Mental disorders were assessed using specific diagnostic modules from the Diagnostic Interview Schedule for Children Version-IV (DISC-IV) and a specifically developed Impact on Functioning module. The DISC-IV is a standardised diagnostic instrument, which was developed under the auspices of the United States National Institute of Mental Health and is used worldwide to assess mental disorders in large samples. Disorder status is determined according to the criteria of the Diagnostic and Statistical Manual of Mental Disorders Version-IV (DSM-IV), which is the diagnostic classification system used clinically world-wide. The DISC-IV modules for seven disorders were included in the survey:
The Australian Government Department of Health funded Young Minds Matter and commissioned The University of Western Australia to undertake the survey through the Telethon Kids Institute in partnership with Roy Morgan Research. Data were collected from parents and carers and young people between June 2013 and April 2014 by trained lay interviewers from Roy Morgan Research using computer-assisted personal interviewing.
The survey was developed with guidance from a Survey Reference Group, comprising over 20 experts in mental health for children and young people (spanning the health, education and community sectors), as well as consumer and community representatives.
The survey aimed to collect information from 5,500 randomly sampled families who had children and adolescents aged 4-17 years from across the country, with the exception of very remote areas. An additional random sample of 800 families with adolescents aged 16-17 years was also undertaken to enable specific estimates to be produced for this age group. In total 76,606 households were approached and those eligible households with children in the age range were visited up to a maximum of six times.
Area based sampling was used to select both samples and where there was more than one child in a household, a child was selected at random by a computer. Parents and carers were interviewed in their homes about one randomly selected child or adolescent in the family. The interview was conducted with the parent or carer who identified as knowing most about the child or adolescent. If the selected child or adolescent was aged 11 years or older, the young person was asked to complete a confidential questionnaire on a tablet computer.
In total 6,310 parents and carers or 55% of eligible households responded and 2,967 or 89% of young people aged 11-17 years also completed a questionnaire.
Analyses showed the sample to be representative of Australian children and adolescents aged 4-17 years on the basis of the age and sex of the child, family structure, parental education, income and employment, housing tenure and country of birth of both children and adolescents and their parents and carers.
The survey sample was selected in two components. The first component was a randomly selected sample of 5,500 families with children aged 4-17 years from across the country.
The sample size was chosen to deliver reliable national estimates of the prevalence, burden and service use by sex and for the two age groups of 4-11 and 12-17 years.
In total 46,248 households were approached to achieve a calculated final sample of 5,500 based on an in-scope rate of 21% and a target response rate of 60%.
In recognition of the increased risks for a range of emotional and behavioural problems in the older teenage years, the main survey sample was supplemented by a second component of 16 and 17 year-olds. This comprised an additional random sample of 800 families with children aged 16-17 years. The sample size was determined to be sufficiently large for separate analyses of the data and to produce specific estimates for this age group.
The oversample aimed to allow for the results for just 16 and 17 year-olds to be analysed together with that for 16 and 17 year olds from the 2007 adult survey. The combination of data from both surveys will provide two perspectives on mental health problems in this age group—one based on questions specifically designed for teenagers and one based on questions designed for adults.
An additional 30,358 households were approached to achieve the required 800 interviews.
Area-based sampling was used to select both the main sample and oversample. This method involves selecting a sample of small geographic areas (in this case the Statistical Area 1 or SA1 area, designed by the Australian Bureau of Statistics). The SA1 areas were selected using information from the 2011 Census of Population and Housing provided by the Australian Bureau of Statistics. SA1 areas in very remote areas were excluded and those where there were at least ten families with children living at the time of the 2011 Census were eligible for selection.
The SA1s used for the main sample were also used for the oversample. A target of four interviews of randomly selected households in each of the 225 SA1s was predicted to deliver the required sample of 800 families. However, selection of an additional 41 SA1s was necessary to achieve the desired 800 interviews. In total 266 SA1s were selected.
Interviewers went to a randomly selected sample of houses in each SA1 to identify households that had one or more children aged from four to 17 years. In households where there was more than one child aged 4-17 years living there, the names of all children were entered into the interviewer’s hand-held computer, which then selected a child at random as the study participant for the survey.
The one percent most remote SA1s in Australia were excluded from the survey for practical reasons.
This survey cannot produce estimates of mental disorders and service use for Aboriginal and Torres Strait Islander peoples. Random sampling alone with the number of participants for this survey was not considered sufficient for generation of these data within acceptable confidence intervals. A separate Indigenous sample was not included as there are important cultural issues in appropriately measuring mental health and wellbeing in Aboriginal and Torres Strait Islander children that could not be addressed within the framework of the population survey. A separate study would need to be undertaken to assess the mental health of Aboriginal and Torres Strait Islander children and young people in a culturally appropriate manner.
The survey was based on a household sample and, as such, children who were homeless or in institutional care were not included. Although this is likely to be a small proportion of children and adolescents, these sub-populations are known to have a high-risk of mental health problems. However, the survey aimed to collect information on children who were temporarily away from home, such as at boarding school or in hospital. Parents and carers were asked to list all of the children who are usually part of the household, but were temporarily living away, along with those living there. These children had an equal chance of selection as the survey child as other children in the family due to the random selection of the survey child by computer. However, young people aged 11 years and older who were not living in the household at the time could not be asked to complete the self report questionnaire and therefore there is no additional information from these young people themselves.
Roy Morgan Research was contracted to conduct the survey fieldwork. A team of over 100 trained, lay interviewers were employed to complete fieldwork from 31 May 2013 through to 10 April 2014.
Interviewers were trained over three days prior to survey fieldwork. Training covered matters related to the conduct of the survey interviews, as well as some basic mental health literacy and information on the use of the diagnostic tool, the DISC-IV. A comprehensive training manual and standardised curriculum for the three days ensured uniformity of interviewer training across all Australian sites.
Interviewers were required to approach selected households in their areas up to six times or until they were able to gain contact with a resident adult. These contact attempts were required to be made over a mixture of weekday/weekend and day/evening calls. Once contact was made with an adult living in the household a short screener process was used to confirm that the household was in scope.
There were a number of strategies used to provide information about the survey, encourage participation once contact had been made and, in turn, maximise the response rate. These were as follows:
Participation in the survey was voluntary and written consent was required from all participants.
Initially, verbal consent was obtained from parents or carers to participate in the survey. If there was more than one child in the family aged between 4 and 17 years, the survey child was chosen randomly by computer. If the child was 11 years or older, parents and carers were also asked for their consent to approach the young person to ask whether he or she would complete a youth self-report questionnaire.
Paper consent forms were completed by both parents or carers and young people at the completion of the household interviews.
Participants were informed that they had the right to withdraw their consent for study participation or the use of their information at any time during the interview and up to 31 March 2014, at which time final validation was being completed and data would be de-identified for inclusion in the main unit record file. A decision was made to extend the period for fieldwork part of the way through the initial collection phase and, in turn, the cut-off for withdrawal of consent for use of information was extended to 30 April 2014 for participants in this later phase of fieldwork.
Specific protocols were developed to ensure that if any issues arose for participants in response to the survey, there were options available to assist them in receiving the information and support they required.
Training of interviewers included specific training in how to deal with difficult and distressing situations if these arose while the interviewer was in the household. Interviewers were instructed to give the participant space and offer to come back at another time to complete the interview. They were directed not to offer opinions or advice, but direct participants to services listed in the brochures and, if required, provide the 1800 number for the on-call study psychologist.
If the participant required support after the interviewer had left the household, a number of options were available. Brochures for both parents or carers and young people, which listed relevant helplines or other services accessible by phone or on the internet, were left with the household. The survey website provided further information about the survey and links to relevant mental health services. A 1800 telephone number staffed by Roy Morgan Research, which participants could call to discuss any aspect of the survey or issue arising from the interview, was available.
In addition, an 1800 number was provided to all participants which could be used to contact an on-call study psychologist. This service was available 24 hours a day during the survey fieldwork period.
The 1800 on-call study psychologist received a total of 5 calls over the duration of survey fieldwork. There were no adverse events reported as a result of the survey.
State and territory child and adolescent specialised mental health services were notified when the survey was going to be in the field in case calls were received from participants.
Interviewers also were supported, with access to the on-call study psychologist, specific debriefing procedures and support directly through Roy Morgan Research.
In total 6,310 parents and carers or 55% of eligible households participated in the survey. In addition 2,967 or 89% of young people aged 11-17 years for whom their parents or carers had given permission completed a questionnaire.
Based on data from the 2011 Census, about one in four Australian households contain one or more children aged 4-17 years. In calculating the response rate for the survey, it is necessary to account for the fact that not all households approached in the survey contained children within the sample age range. Interviewers approached each selected household at different times. However, there remained some households where it was not possible to make contact with anyone living in the household despite these attempts. Census data have been used to estimate the proportion of these households that would be likely to have contained children aged 4-17 years.
The survey was voluntary and some households refused to participate. When a householder refused to participate in the actual survey, the interviewer tried to obtain some basic demographic information, in particular whether there were any children living there. About two thirds of householders would not provide even this basic information. Again Census data have been used to estimate the proportion of these households that would have contained children aged 4-17 years.
The estimated number of non-contacts and refusals where eligibility could not be determined were added to the number of participants and the number of refusals who confirmed they had one or more children in order to estimate the overall response rate. Using this method, the overall response rate for the survey was estimated to be 55%. It is acknowledged that the inclusion of households where it was not possible to make contact makes this a strict, conservative estimate of the overall response rate. If this group of households is excluded from the estimate, the overall response rate increases to 60%.
Several approaches were used to assess how representative the sample was of children and adolescents in the Australian population.
Firstly, the demographic characteristics of children, adolescents and families who participated in the survey were compared with those of the total population of Australian families with children aged 4-17 years. The survey included a number of questions that matched questions in the 2011 Census. This enabled a comparison to be made between the survey sample and the Census figures based on the age and sex of the child, family structure, parental education, income and employment, housing tenure and country of birth of both children and adolescents and their parents and carers.
Secondly, response rates were calculated for each individual SA1 area. Statistical models were then used to test for differences in response rates by characteristics of the area, including the Socio-Economic Indexes for each area, remoteness, state and part of state, and a range of census-derived measures, including proportion of overseas born, proportion of people speaking languages other than English, proportion with limited proficiency in English, proportion of sole parent families, proportion of families with low household income, highest level of schooling completed, and proportion of households living in rented accommodation.
Finally, the survey included the Strengths and Difficulties Questionnaire (SDQ), which is used internationally to measure mental health problems in children and adolescents. The SDQ is designed so that approximately 10% of children and adolescents fall into the ‘abnormal’ range on the total difficulties score. Analysis of the data revealed that 10.1% of young people in the second survey did so. In addition, the SDQ results on the various scales were compared with those from another national survey, the 2004 British Child and Adolescent Mental Health Survey. The distribution of the parent and carer reported SDQ scores was found to very closely match those from the British survey.
Only two of all the factors examined were found to be associated with participation in the survey:
Survey data were weighted, based on information from the 2011 Census, to account for these patterns of participation in the survey.
The survey sample was representative of the population for all other demographic characteristics considered.
Data were collected for the survey using computer assisted interviewing. This ensures that questions were asked in the correct sequence, only valid answers were recorded, and that questions could not be inadvertently skipped and no answers recorded. Logic checks were programmed into the questionnaire to ensure that logically inconsistent answers could not be recorded. After data collection, a data editing phase was undertaken to check for any unusual values indicating possible data entry errors.
Participants had the option of answering ‘don’t know’ in some questions and they could refuse to answer any question in the survey. These responses were treated differently depending upon the section of the survey:
There was a negligible amount of missing data for most data items in the parent and carer questionnaire. The main exception was in relation to household income, which had the largest proportion of missing values, with approximately 4% of families either responding ‘don’t know’ or refusing to provide their household income.
There was a negligible amount of missing data for most items in the youth self-report questionnaire, particularly as participants were mostly not given an option to respond with ‘don’t know’ or to refuse to answer a question. The main exceptions were in relation to the questions on self harm and suicidal behaviours, for which participants were given the option of responding ‘prefer not to say’, and for height and weight. Young people who did not provide either their height or weight were excluded from analyses relating to Body Mass Index (BMI) and also low weight problem eating behaviour and binge eating or purging problem eating behaviour both of which include BMI status as part of the definition. There was no evidence that the individual eating behaviours that form part of the definition of these problem eating behaviours occurred any more or less frequently in young people who reported their height and weight compared with those who did not.
There were no open questions requiring data entry and no responses required clinical judgement.
Where cell sizes were less than five, these cells have been suppressed when presenting data in these charts and tables because of the low reliability of these estimates, and to protect confidentiality.
There are a number of substantive methodological and content differences between the first and second child and adolescent national surveys of mental health and wellbeing.
The first survey used modules from the DISC-IV to assess four mental disorders - major depressive disorder, dysthymic disorder, attention-deficit/hyperactivity disorder and conduct disorder. These were completed by parents and carers.
There were similarities, but also significant differences in how mental disorders were assessed in the second survey. Firstly, the DISC-IV was the primary diagnostic tool for both surveys. Major depressive disorder, attention-deficit/hyperactivity disorder and conduct disorder were also assessed in the second survey. However, the second survey also assessed four anxiety disorders, namely social phobia, separation anxiety, generalised anxiety and obsessive-compulsive disorder. Parents and carers completed DISC-IV modules for all seven disorders. Therefore direct comparisons of the 12 month prevalence of mental disorders overall cannot be made.
In the second survey young people completed the same DISC-IV major depressive disorder module as their parents and carers, providing an alternative perspective on whether they had the disorder.
The DISC-IV eating disorders module was included in the first survey. However, too few cases were identified for results to be published. For this reason the module was not included in the second survey, but rather problem eating behaviours were assessed using questions drawn from the Avon Longitudinal Study of Parents and Children.
Questions were also added to determine the impact of symptoms of particular mental disorders from which severity could be determined. This allows for comparison of the prevalence data by severity as is possible with the other National Surveys of Mental Health and Wellbeing, but was not possible with the first survey.
In the first survey mental health problems were assessed using three tools:
The Strengths and Difficulties Questionnaire (SDQ) was included in this second survey and the Child Behavior Checklist (CBCL) and Youth Self-Report, which was used in the first survey, were not. There were several reasons for this change. The SDQ is substantially shorter than the CBCL, while having equivalent measurement properties. This provided leeway for questions on other topics to be added to the survey. In addition, the SDQ has been included in a range of other collections in Australia, including as a routine measure in the National Outcomes and Casemix Collection and the Longitudinal Study of Australian Children. It has also been designated as the leading national Children’s Headline Indicator for social and emotional wellbeing in children and young people by the Australian Institute of Health and Welfare.
Young people aged 11-17 years also completed the Strengths and Difficulties Questionnaire (SDQ) Youth Self Report in the second survey rather than the Youth Self-Report based on the CBCL, which was used in the first survey.
Information on depression symptoms, which was collected in the first survey using the CES-D, was covered in more detail by the DISC-IV major depressive disorder module, which also provided a diagnosis.
Questions on service use were completely redesigned for the second survey. The approach taken was based on the approach used in the 2007 National Survey of Mental Health and Wellbeing of the adult population. This approach is based on identifying if services have been used in the past 12 months, if parents and carers or young people feel they needed help or support, whether they received the level of support they needed, and what barriers may have prevented them from using services they felt they needed.
The first survey asked about services used in the six months prior to the survey, whereas the reference period in the second survey was 12 months to match the period over which mental disorders were assessed using the DISC-IV and also to align with service data from the other National Survey of Mental Health and Wellbeing initiative surveys, including the adult survey. Direct comparison of service use over these two different timeframes is not possible and data presented on services used by young people with mental disorders is also not comparable due to the differences in disorders.
The types of services covered in the second survey were much broader than those asked about in the first, reflecting significant changes in the types of services available to young people. Questions were added about new health services, online services and information, and additional information was collected about services used in schools.
There have been minor changes to the scoring algorithms used in the DISC since 1998. The 1998 survey results for major depressive disorder, ADHD and conduct disorder have been recalculated using the latest algorithms to be comparable with the results from this second survey.