Building National Data Sets - An ABS Perspective

Presented by Barbara Dunlop, First Assistant Statistician,
Social and Labour Division, Australian Bureau of Statistics.

Introduction

This paper presents an ABS perspective on the development of State-based population health surveys. The paper outlines the role of the ABS as Australia's national statistical agency and discusses recent ABS activities in the field of population health data. It then considers the potential benefits to national population health data which could derive from development of a State survey program and the role which ABS could play in realising that potential.

Role of the ABS

As Australia's national statistical agency, the ABS has a key role in meeting the statistical requirements of the Australian Government , the State and Territory governments, and the community at large. The functions of the ABS include:

• collection, compilation, analysis and dissemination of statistics;
• co-ordination of statistical activities;
• development and promotion of statistical standards; and
• assisting other official agencies in statistical matters.

ABS Population Statistics Program

Much of ABS population statistics work is concerned with describing the social well-being of the population. Underlying the notion of social well-being is a range of fundamental human needs and aspirations, each of which can be linked to an area of social concern. The ABS framework for social statistics is built around the following areas of social concern:

• population
• family and community
• education and training
• income
• culture-leisure
• health
• housing and neighbourhood
• work
• crime and justice
• employment
• transport and travel

The framework also includes another dimension which focuses on a variety of population groups of special interest to the community and to governments, because of their special need or disadvantage, etc. This dimension of the framework includes older people, children, youth, families with children, long-term unemployed, lone parents, people with disabilities, carers, recipients of various government benefits, low income earners, Aboriginal and Torres Strait Islanders, and people whose language background is other than English.

The ABS population statistics program aims to ensure that statistics are available at regular intervals for each area of social concern and for all relevant key population groups so as to:

• provide a backdrop on society and benchmarks against which social progress can be monitored, and
  
  
• support the policy development, program delivery and program evaluation of key Commonwealth and State government agencies,

The broad strategies used to provide the range and quality of population, social and labour statistics required to inform decision making and community discussion include the 5-yearly censuses of population and housing; an extensive program of household surveys and employer surveys; statistics from administrative sources; statistical analysis and reporting, and integration of statistics across collections.

Role of the ABS in health statistics

In the field of health statistics, the ABS recognises the statistical functions and responsibilities of other agencies, such as the Australian Institute of Health and Welfare, the Department of Health and Aged Care and State/Territory health authorities. ABS sees its role as complementing the activities of these other agencies by providing statistical leadership and co-ordination; and by contributing in those areas or in those ways it is best able, either directly through running collections, or by assisting other agencies in statistical activities, development of standards, etc.

ABS has a reduced, but continuing, role in data collections that obtain data as a by product of administrative processes. However, the activity through which it is best able to contribute is direct data collections. Such collections include, for example, industry and population based collections, national level and State or Territory specific collections in the ABS household survey program, as well as user funded collections.

As the national statistics agency, the ABS has certain comparative advantages in the conduct of household surveys. These include:

• effective legislation for collecting statistics and protecting the confidentiality of the data collected
• the ability to undertake complex large-scale statistical activities
• an established and highly trained interviewer workforce
• responsibility for the co-ordination of statistical activities across government agencies
• development and maintenance of a large range of statistical standards and classifications
• integrity of its published statistics which are not influenced by political expediency
• statistics not generally released unless they meet minimum quality levels
• positive relationship with data providers as a result of a strong focus on the provider load, sensitivity and privacy issues.

For a survey to be included in its national survey program, ABS considers that the survey should collect information which:

• is required at national and/or State/Territory level
• will have a major influence on significant government decisions
• has a high public interest component
• cannot readily be collected to the required standard by others
• cannot readily be extracted from other data sources (eg administrative)
• follows standardised question methodology, where possible
• is acceptable to respondents in terms of load and intrusiveness
• can be obtained from respondents at a reasonable standard of quality
• achieves acceptable levels of response.

Health information from ABS household surveys

Surveys on health topics conducted by ABS in the last 5 years have included:

• The 5 yearly National Health Survey (NHS), last conducted in respect of 1995;
  
  
• Periodic topic-specific surveys conducted in association with the ABS Monthly Population Survey (MPS), at national level and in individual States and Territories. Topics recently addressed nationally include children's immunisation and private health insurance;
  
  
• Health topics included in the quarterly ABS Population Survey Monitor (PSM); and
  
  
• User funded surveys, or components of surveys eg National Nutrition Survey, Survey of Mental Health and Wellbeing.

A full list is provided at Attachment 1.

Over this period, the program has delivered a large range of population health data. The data sets obtained have reflected the priorities identified in consultation with data users. Characteristics of the program have been:

• a commitment by ABS to high data quality, through development and testing of survey instruments, sound sample design and selection, a highly trained interviewer workforce, field procedures to ensure low non-response rates, thorough editing of data, and so on.
  
  
• delivery of data sets for States and Territories which are directly comparable with each other, and with national benchmarks. Depending on the level of detail required, national data sets are generally more reliable than those for States and Territories.
  
  
• delivery of data sets which can be analysed for a wide range of population groups. This has been facilitated through large sample sizes and the range of demographic and socio-economic data collected in association with the health variables. The use of standard modules covering these characteristics also enables analysis between surveys (including non-health surveys) at the population group level.
  
  
• accessibility of the data obtained to all with an interest in health information, even though the surveys are developed primarily to meet policy data needs of government.

The ABS health survey program, however, has been unable to meet the large range of data requirements identified for it. In trying to stretch the capacity of the program, the last NHS became one of the largest, most complex and most expensive population surveys conducted by the ABS, and resulted in delays in releasing survey results. This has served to illustrate the trade-off necessary between the timeliness of output and the size and complexity of the data set involved.

The program of ABS funded health surveys has been designed to provide data at the national level, and for individual States and Territories. Through sample supplementation in the NHS, some States and Territories have sought to improve the reliability of State estimates, and to meet requirements for broader level regional estimates. ABS understands that this approach has been viewed as a cost-effective means of meeting needs in some States, but not in others. The inability of the ABS to release unit record data at regional level on confidentiality grounds has been a further concern to some State users of ABS survey data.

Review of the household survey program

The ABS is currently undertaking a review of its household survey program to ensure it:

• provides the range of statistics required to monitor the well-being of Australians and important target population groups, and
  
  
• addresses the highest priority requirements of key Commonwealth and State government agencies.

This review is in response to deficiencies identified in the current program. Demands being placed on the program well exceed the capacity of the ABS survey vehicles and ABS resources to develop and conduct surveys - as a consequence there are some important gaps in the statistics that are available. Also, there is a lack of flexibility in the program resulting from either commitments to regular surveys in particular areas or from constraints imposed by the survey vehicles. Finally, the stop/start approach to some user funded surveys is causing significant problems for the scheduling of surveys and for the maintenance of a skilled survey workforce.

As part of the review the ABS is examining options for improving the capacity and flexibility of the current household survey program, which could lead to a significant revamping of the current collection vehicles and the topic content and mix of the program. In addition, ABS is exploring opportunities to enter into longer term partnerships with users for additional funds to extend the program to meet specific user needs. It should be noted that the ABS is not intending to reduce its total funding for the household survey program and the review is being conducted on this basis.

Within this general review of the household survey program, the ABS is reviewing its program of health surveys so that it better meets the data needs of clients in a timely and cost effective manner. The main focus of this latter review will be the NHS and options for redeveloping or replacing it.

The specific aims of the review of the health survey program are to develop a strategy for the collection and dissemination of household-based population health data by ABS which:

• provides the range of statistics required to monitor the health of Australians and important target groups;
  
  
• better meets the current priority data needs of Commonwealth and State/Territory governments and other data users, in terms of the data content, frequency and geographic focus;
  
  
• is flexible to changing data needs;
  
  
• contributes within a wider population health information strategy, and fits appropriately within the context of population surveys conducted by other agencies at national, state and substate level; and
  
  
• conforms to certain general parameters.

The general parameters within which the strategy is being developed are:

• that the information proposed for inclusion in surveys is appropriate for collection by the ABS (eg must be apolitical in terms of content; be expected to receive high levels of cooperation by selected households; not unduly invade personal privacy; and yield high quality data).
  
  
• that the surveys must be able to be conducted within the constraints of ABS survey vehicles and methodologies, and be consistent with the outcomes of the larger review of the ABS household survey program.

In the context of the review of the household survey program, and the specific review of the health survey program, over 100 agencies have been contacted. These included Commonwealth and State and Territory Health authorities, health professional and industry groups, health researchers, and agencies representing health consumer and community perspectives. The Review sought information about each agency's data requirements; specifically the relative priority of health topics and indicators, the frequency at which data are required and the geographic level at which data are required. The Review also sought information on population groups of prime interest and information on how data from previous ABS surveys have been used, to provide ABS with a better understanding of data requirements.

Information sought from ABS health surveys

At the time of preparing this paper not all replies to the review had been received and/or processed. However, from those replies which have been processed some impressions can be gained of the data requirements which agencies are looking to ABS surveys to address. These are briefly summarised below:

• highest priority demographic and socio-economic characteristics required are: sex, age, Indigenous status, country of birth, language, labour force status, income, education and disability status.
  
  
• highest priority population groups for which data are required are: women, men, Indigenous people, older people and those with low income.
  
  
• health topics of highest priority are: prevalence of health priority areas; recent illness; long-term conditions; hospital episodes (including casualty, outpatients and day clinics); doctor and dental consultations; smoking, alcohol consumption and exercise.
  
  
• for most topics, the number of responses indicating a requirement for annual data, and for 5 yearly data, are similar.
  
  
• most responses, across topic areas, have indicated a requirement for both national and State/Territory level data.
  
  
• although a number of topics not previously covered in ABS surveys have been identified in responses, none have yet emerged as a high priority for future surveys, apart from disability.

While it is recognised these are summary impressions only, they do suggest at the broadest level that:

• except in regard to Indigenous statistics, the previous program delivered data for the population groups of most interest. The ABS is currently developing a survey strategy for Indigenous statistics, and a discussion paper is expected to be available shortly. That strategy will be a key element in finalising a strategy for future health surveys.
  
  
• the previous survey program addressed most topics of highest priority.
  
  
• surveys in the ABS program are considered a key source to meet needs for data at both State and national levels. It is accepted that the current sample design does not deliver detailed results for the less populous States and the Territories.
  
  
• there is a requirement (currently unmet) for data more frequently than previously provided on a regular basis by the program.

It is noteworthy that a review of requests for unpublished data from the 1995 NHS shows that the level of interest in national data is about 3 times greater than State data. Perhaps this in part is a further reflection of the inability, noted above, of the previous surveys to adequately meet State needs (eg for regional data). Health topics most requested include long-term conditions, hospital episodes, use of medications, days off work, smoking, alcohol consumption and health insurance.

From an ABS perspective, it is against this background and the ABS survey strategy which will emerge from the review, that the program of State-specific health surveys will need to be considered. Draft proposals for a future ABS health survey strategy will be presented for discussion by the Australian Statistical Advisory Council in March 1999, with the strategy expected to be finalised mid year. Whatever the shape of the strategy which will emerge, and the particular surveys of health topics to be conducted under that strategy, the ABS remains committed to a strong continuing role in the collection and dissemination of population health statistics, which complements information available from other sources.

Common elements in State surveys

Although the review of the health survey program discussed above will assist in improving the contribution made by ABS surveys to population health information, it is clear that the capacity of the ABS program is limited, and will still fall short of data needs. It is equally clear that there are some types of data and some levels of data which the ABS program cannot address unless there is specific user funding and/or special supplementation of a survey, for example, data to address State specific health issues, or data to meet needs at the regional level. The development of State-based population surveys by other agencies will not only fill data gaps and meet priorities for individual State purposes, but also has potential to complement data available from ABS surveys and other sources, and to make an important contribution to population health data in Australia.

Realisation of that potential will depend on the degree of commonality between State collections: commonality in survey content, survey methodology, sample design, etc. This is not to suggest that all State surveys necessarily be the same; rather, that they share common elements, and where the same data are collected they are collected using a common set of questions and a consistent methodology.

One approach may be to have a "core" set of items, collected regularly (but not necessarily in each State survey), such that a common and directly comparable set of indicators is obtained for each State, and for areas within States as appropriate to State needs. If this "core" was further shared with ABS surveys, State results could be compared with national benchmarks. Alternatively, State indicators could be aggregated to form new national benchmarks.

The potential benefits of such an approach can be summarised as:

• providing inter-State and intra-State comparability of data.
  
  
• providing comparability with existing national benchmarks or creation of new national benchmarks.
  
  
• encouraging comparability over time through the "core" data set, enabling data to be used for monitoring purposes.
  
  
• addressing State specific data needs, with the added benefit that the data can be related to other benchmarks if required.
  
  
• sharing of development (and possibly some operational) experience, systems etc, and reducing overall costs to each State.

Fundamental to such an approach is a basic set of indicators which are agreed by States as a high priority for regular inclusion in their surveys.

The way ahead

ABS is committed to making an ongoing contribution to population health data. However, it is clear that ABS collections cannot (nor should they aim to) meet all needs for survey-based population health data. Rather, ABS surveys are part of larger data network, which relies on activities of other agencies.

The ABS can contribute in ways other than through direct data collection and dissemination activities, and sees substantial benefits that would be realised from effective co-ordination of State and national activities in this area. In undertaking its co-ordination role, the ABS has interests in all statistical collections undertaken by government agencies, including matters influencing outcomes for respondents, users of statistics, and use of public funds for statistical purposes. In line with these interests, the ABS can:

• assist in the planning and development of State surveys, in aspects of sample design, and collection methodologies, particularly in those aspects of State surveys common across States;
  
  
• play a role in the development and dissemination of a basic set of agreed indicators;
  
  
• advise on the use of standards to maximise the usefulness of data and minimise duplication and overlap, make ABS standards readily available, assist with the development of new standards, and promulgate developed standards.

The ABS is well positioned to coordinate these developments due to its statistical infrastructure, experience in the conduct of population surveys, and background in collecting health data.

Health information collected in ABS surveys

Authorised by: Executive Officer, NPHP
Contact: Darryl Kosch, nphp@dhs.vic.gov.au
Date of publication: 1999
Page last updated: 27 November, 2003

Copyright | Disclaimer | Privacy statement