4363.0.55.001 - National Health Survey: Users' Guide - Electronic Publication, 2007-08
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 14/09/2009
Page tools:
 Print Page
 Print All
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
This document was added or updated on 17/09/2009.
Sample selection
Interviewers Questionnaires Response rates Input coding Coding of health items
Type of medication Alcohol consumption Output data file Weighting, Benchmarking and Estimation procedures
Benchmarks SCOPE The 2007-08 NHS collects information by personal interview from usual residents of private dwellings in urban and rural areas of Australia, covering about 97% of the people living in Australia. Persons in scope of the survey were those identified by an adult within each sampled private dwelling as a usual resident of that dwelling. Private dwellings are houses, flats, home units, caravans, garages, tents and other structures being used as a place of residence at the time of the survey. The survey excludes residents of:
Overseas visitors staying or intending to stay in Australia for 12 months or more were in scope, as were non-Australians (other than those above) who were working or studying in Australia, and their dependants. SAMPLE DESIGN AND SELECTION Sample design The 2007-08 NHS was conducted using a stratified multistage area sample of private dwellings. Decisions on the appropriate sample size, distribution and method of selection rested on consideration of the aims of the survey, the topics it contained, the level of disaggregation and accuracy at which the survey estimates were required, and the costs and operational constraints of conducting the survey. The sample was designed to provide:
To achieve these design objectives, the State and Territory sampling fractions were set as shown in the following table, which also depicts the corresponding expected number of fully responding households. The sample selection procedures described below result in every dwelling in the same State or Territory having a known probability of selection, equal to the State or Territory sample fraction.
Actual numbers of fully responding households are available in the Response Rates section of this chapter. Within selected dwellings, a random sub-sample of residents was selected as follows:
The exclusion of persons living in non-private dwellings (approximately 3% of the population) is unlikely to impact on the estimates included in this publication. The exclusion of the Australian population living in very remote areas (approximately 1%) has little impact on national estimates, and has only a minor impact on any aggregate estimates that are produced for individual States and Territories, except for the Northern Territory where the excluded population accounts for over 24% of persons. Sample selection The area-based selection ensures that all sections of the population living in private dwellings within the geographic scope of the survey were represented by the sample. Each State and Territory was divided into geographically contiguous areas called strata. Strata are formed by initially dividing Australia into regions, which are formed within State/Territory boundaries, and which basically correspond to the Statistical Division or Subdivision levels of the Australian Standard Geographical Classification (ASGC)(cat. no. 1216.0). Each stratum contains a number of Population Census Collection Districts (CDs) containing on average about 250 dwellings. In capital cities and other major urban or high population density areas, the dwelling sample was selected in three stages:
In Hobart, parts of Darwin and some strata of high population growth, the CD stage of selection is omitted leaving only two stages of selection. In strata with low population density each stratum was initially divided into units, usually corresponding to towns or Statistical Local Areas (SLAs), or combinations of both, and one or two units were selected from each stratum. Within selected units, the sample of dwellings was arrived at in the same manner as outlined for high population density areas. In total a sample of approximately 20,000 households was selected which, taking account of an expected rate of sample loss (e.g. vacant dwellings, dwellings under construction etc.) of 14% and non-response of 10%, was designed to achieve the desired sample of about 15,570 fully responding households. To take account of possible seasonal effects on health characteristics, the sample was enumerated over a 11 month period from August 2007 to June 2008. Collection Districts were allocated randomly over four sub-periods - August 2007 to September 2007, October 2007 to December 2007, January 2008 to March 2008, and April 2008 to June 2008. DATA COLLECTION Information was obtained in the 2007-08 NHS by trained ABS interviewers, through a personal Computer Assisted Interview (CAI) with adult members of selected households in scope of the survey. For the purposes of this survey, a household was defined as one or more persons, at least one of whom is aged 15 years and over, usually resident in the same private dwelling. Aspects of data collection are discussed below under the headings: interviews, interviewers and questionnaire. Interviews In the 2007-08 NHS, selected households were initially approached by mail informing them of their selection in the survey and advising them that an interviewer would call to arrange a suitable time to conduct the survey interview. (This was not possible for a small number of households where the ABS did not have an adequate postal address.) A brochure providing some background to the survey, information concerning the interview process, and a guarantee of confidentiality were included with the initial approach letter. General characteristics of the household were obtained from a responsible adult member of the household (ARA). This information included the number and basic demographic characteristics of usual residents of the dwelling, and the relationships between those people (e.g. spouse, son/daughter, not related). The ARA was also asked to nominate the person in the household who was best able to provide information about children in the household and household income. From the information provided by the ARA about the household composition, the survey instrument established those persons in scope of the survey, and, on a random basis, selected one adult and one child (where applicable) to be included in the survey. If the dwelling contained only usual residents aged 15-17 years, two people were randomly selected. A personal interview was conducted with the selected adult (where possible), and an adult was asked to respond on behalf of selected children under 15 years of age. In some instances, adult respondents were unable to answer for themselves because of old age, illness, intellectual disability or difficulty with the English language. In these cases, a person responsible for them was interviewed on their behalf, provided the interviewer was assured that this was acceptable to the subject person. Where there were language difficulties, other persons in the household may have acted as an interpreter if this was suggested by the respondent. If not, arrangements were made where possible for the interview to be conducted either by an ABS interviewer fluent in the respondent’s own language, or with an ABS interpreter. In contrast to previous NHSs, it was assumed that children aged 15 to 17 years would be interviewed in person, however, should a parent or guardian request it, an adult was interviewed on their behalf. This adult, who may or may not have been the selected adult respondent in the household, is referred to as the Child Proxy. In order to obtain a personal interview with appropriate respondents, interviewers made appointments to call back as necessary to the household. In some cases appointments for call backs were made by telephone, however all interviews were conducted face-to-face. Interviews may have been conducted in private or in the presence of other household members according to the wishes of the respondent. Interviews were only conducted on Sundays at specific respondent request. Although it is desirable to spread interviews across all days of the week, interviews were conducted on days that suited respondents. In cases where a respondent initially refused to participate in the survey, a follow-up letter was sent and a second visit was made to the respondent, usually by an office supervisor, to explain the aims and importance of the survey, and to answer any particular concerns the respondent may have had. No further contact was made with the respondent if they refused at the second approach to participate. Persons missed from the survey through non-contact or refusal were not replaced in the sample. Interviewers Interviewers for the 2007-08 NHS were recruited from a pool of trained interviewers with previous experience on ABS household surveys. Those selected to work on this survey underwent further classroom training and were required to satisfactorily complete home study exercises. All phases of the training emphasised understanding of the survey concepts, definitions and procedures in order to ensure that a standard approach was employed by all interviewers concerned. Each interviewer was supervised in the field in the early stages of the survey and periodically thereafter to ensure consistent standards of interviewing procedures were maintained. In addition, regular communication between field staff and survey managers was maintained throughout the survey via database systems set up for the survey. Interviewers were allocated a number of dwellings (a workload) at which to conduct interviews. The size of the workload was dependent upon the geographical area involved and whether or not the interviewer was required to live away from home to collect the data. Interviewers living close to their workload area in urban areas usually had larger workloads. Overall, workloads averaged 25-30 dwellings, to be enumerated over a two-week period. Questionnaire The Computer Assisted Interview (CAI) instrument that was used for the 2007-08 NHS was based on the 2004-05 NHS, modified as appropriate to incorporate new and changed survey content. Information collected included:
The questionnaire was designed to be administered using standard ABS procedures for conducting population interview surveys, with regard to the particular aims of the survey and the individual topics within it, and the methodological issues associated with those topics. Other factors considered in designing the questionnaire included the length of individual questions, the use of easily understood words and concepts, the number of subjects and overall length of the questionnaire, and the sensitivity of topics. Where appropriate, previous ABS questions on the topics covered were adopted. The CAI instrument allows the following:
The questionnaire was fully field tested to ensure:
The questionnaire employed a number of different approaches to recording information at the interview:
To ensure consistency of approach, interviewers were instructed to ask the interview questions exactly as written. In certain areas of the questionnaire however, interviewers were asked to use indirect and neutral prompts at their discretion, where the response given was, for example, inappropriate to the question asked or lacked sufficient detail necessary for classification and coding. This occurred particularly in relation to type of medical condition where interviewers were asked to prompt for a condition if a treatment or symptom was initially reported. The NHS 2007-08 questionnaire and related prompt cards are available from the ABS website under the 'Downloads' tabs of this Users' Guide and the National Health Survey: Data Reference Package, 2007-08 (cat. no. 4363.0.55.002). MEASURES TO MAXIMISE RESPONSE In any sample survey, responses should ideally be obtained from all selected units, however there will always be some non-response, when people refuse to cooperate, cannot be contacted, or are contacted but cannot be interviewed. It is important that response be maximised in order to reduce sampling variability and avoid biases. Sampling variability is increased when the sample size decreases, and biases can arise if the people who fail to respond to the survey have different characteristics from those who did respond. The ABS sought the willing cooperation of selected households. Measures taken to encourage respondent cooperation and maximise response included:
Through call-backs and follow-up at selected dwellings, every effort was made to contact the occupants of each selected dwelling and to conduct the survey in those dwellings. Interviewers made five call-backs before a dwelling was classified as a ‘non-contact’ (three call-backs in non-metropolitan areas). Call-backs occurred at different times during the day to increase the chance of contact. Once contact had been made at a dwelling, the interviewer completed all necessary questionnaires where possible. If any persons who were selected to be included in the survey were absent from the dwelling when the interviewer called, arrangements were made to return and interview them. Interviewers made return visits as necessary in order to complete questionnaires for selected persons in scope of the survey. In some cases, a selected respondent within a dwelling could not be contacted or interviewed, and these were classified as individual non-contacts. As noted above in the Data Collection: Interviews section of this chapter, respondents who refused to participate were usually followed-up by letter and a subsequent visit by an office supervisor. RESPONSE RATES A total of 19,979 private dwellings were selected in the sample for the 2007-08 NHS, reducing to an active sample of 17,426 dwellings after sample loss in the field stage. A fully responding household is one in which all parts of the interview were completed for all persons in on scope and coverage. An adequately responding household is the same as a fully responding household, with the addition of legitimate 'don't know' or refusal options. Adequately responding households for the 2007-08 NHS also included respondents who did not answer all or any of the income questions and those that refused to provide their height or weight, or be measured. Full response details are provided below:
Completed questionnaires were obtained for 20,788 persons in fully/adequately responding dwellings, as shown below:
INPUT CODING Input coding refers to the categorisation of country of birth, main language spoken at home, occupation, industry and industry sector, educational qualification and relationship within a household. This coding was performed by ABS Regional offices. Coding of alcohol consumption, type of long-term medical conditions and medication type was undertaken later in the ABS central office (see Coding of health items below). A brief outline of the input coding undertaken follows:
CODING OF HEALTH ITEMS In addition to the general coding of population characteristics outlined above, the following items were office coded:
This coding was undertaken by coding staff specifically recruited and trained for the task; all coding was centralised in the Canberra office of the ABS. Initially all records were run through an automatic coder for each of these items. The auto-coders sought exact matches between text recorded in the questionnaires, and text entries in the coders. Cases which could not be coded by the auto-coders were coded manually using the Computer Assisted Coding (CAC) systems. Rigorous quality control processes were applied throughout to ensure that the coding process met agreed standards. A brief outline of the coding is provided below. Further information about the CAC and auto-coder systems and how they were applied in the survey can be provided on request. Coding of medical conditions All reported long-term medical conditions were coded to a list of approximately 1000 conditions, which was built into both the auto-coder and the CAC system. Conceptually the coding process involved locating the reported condition in the coder, and recording the corresponding 3 digit ABS input code. In practice it was a more complex task and a query data base was established where coders could register any problems they came across, and where a solution could be posted. This provided coders with both a response to specific coding issues, and a resource for dealing with future problem cases. The code list used for the 2007-08 NHS was that initially compiled for use in the 2001 NHS by the Family Medicine Research Centre, University of Sydney, in association with the ABS. This was also used for the 2004-05 NHS. Conditions classified at the full level of detail are not generally available for output from the survey; however, they can be regrouped in various ways for output. The standard output classification, developed for the NHS, is based on the International Classification of Diseases: 10th Revision (ICD-10). A copy of the NHS output classification is provided in Appendix 2: Classification of Medical Conditions. Coding of type of medication The 2007-08 survey collected information on medications used by respondents in the two weeks prior to the survey for mental health and wellbeing, asthma, heart and other circulatory conditions, arthritis, osteoporosis and diabetes. The questionnaire provided space to record the names of up to three medications used in the reference period for asthma, arthritis, osteoporosis, mental health conditions and diabetes; up to five medications used for mental wellbeing; and up to 12 medications used for heart and circulatory conditions (up to three medications for up to four conditions). The coding process involved assigning a 4 digit generic type of medication code to each medication name recorded. An auto-coder and a CAC system were developed incorporating the names of medications readily available in Australia and commonly used for the nominated conditions. The lists of medication names were based on the lists prepared for the 2001 and 2004-05 NHS, updated as appropriate with reference to the World Health Organisation’s Anatomical Therapeutic Chemical (ATC) Classification and the Australian Medicines Handbook. Respondents were encouraged to refer to the medication packet, bottle, etc., when reporting, but may have reported from memory, and may have reported medications by their brand, trade or generic names. Some allowance was made in the coding process for the nature of the information reported; e.g. respondents not sure of the medication name, mispronounced medication name, interviewer misspelling of names, etc. The classification of generic type of medication used in the 2007-08 NHS is based on the World Health Organisation’s Anatomical Therapeutic Chemical (ATC) Classification (and associated coding indexes) and the Australian Medicines Handbook. Details of the classification used are provided in Appendix 3: Classification of type of medication. Brand name information is not available for output from this survey. Coding of alcohol consumption In the 2007-08 NHS, information about alcohol consumption was recorded against ten general categories of alcoholic drinks: beer, wine, champagne/sparkling wine, ready-to-drink spirits and liqueurs, spirits, liqueurs, fortified wine, cider, cocktails and other alcoholic beverages. Beer was further categorised to light, medium or full strength, and wine was categorised to white, red or low alcohol. Details of the type, brand and quantity (number and size of drinks) of each drink consumed on (up to) the last three days in the week prior to the day of interview were recorded, with up to 15 entries possible for each type of alcohol consumed. Quantities were recorded in terms of standard measures where possible; otherwise a description of the quantity consumed was recorded by interviewers. Interviewers recorded details about the brand or name of the drink to assist in coding. The autocoding and CAC systems were used to calculate in millilitres the amount of pure alcohol contained in the drinks reported. This system, which has been used since the 2001 NHS, took information about the type of alcoholic drinks consumed (including brand name for common drinks), and the size and number of drinks consumed, and applied a conversion factor to obtain the amount of pure alcohol consumed. Conversion factors tailored to specific drinks/drink types were included in the system, and default factors for each of the eight broad types of alcoholic drinks used in the survey were included for cases where more detailed information had not been recorded at interview. EDIT CHECKS During office processing of the data, checks were performed on records to ensure that specific values lay within valid ranges and that relationships between items were within limits deemed acceptable for the purposes of the survey. These checks were also designed to detect errors which may have occurred during processing and to identify cases which although not necessarily errors, were sufficiently unusual or close to agreed limits to warrant examination. Data available from the survey are essentially ‘as reported’ by respondents. In some cases it was possible to correct errors or inconsistencies in the data which were originally recorded through reference to other data in the record; in other cases this was not possible and some errors and inconsistencies remain on the data file. OUTPUT DATA FILE Information from the survey was stored on the computer output file in the form of data items. In some cases, items were formed directly from information recorded in individual survey questions, in others, data items have been derived from answers to several questions (e.g. the item ‘self-reported body mass’ is derived from reported height and weight). Some items have been derived from the reported information in conjunction with information obtained from other sources (e.g. in deriving the health risk associated with the reported level of alcohol consumption as defined by the National Health and Medical Research Council (2001)). In designing the output data file, the aim was to create a file which was similar to the 2004-05 data file, but simplified where possible. The result is a 7 level hierarchical data output file, down from 9 levels in 2004-05. The structure of the file is as follows:
A hierarchical data file is an efficient means of storing and retrieving information which describes one to many, or many to many, relationships; e.g. a person may report multiple conditions, and may report use of multiple medications for all/some of these. Data about households and families are contained as individual characteristics on person records. While estimates are also available at the household level, estimates at the family level are not available from this survey. A full listing of output data items available from the survey can be accessed on the ABS web site, under the Downloads tab of this Users' Guide. Once processing and validation of the data were complete, person and household weights were derived and inserted into each responding person’s record to enable the data provided by these persons to be expanded to obtain estimates relating to the whole population within scope of the survey (see below). To enable standard error values for the estimates to be produced, 60 replicate weights were included (see Technical Note). WEIGHTING, BENCHMARKING AND ESTIMATION PROCEDURES Weighting Weighting is the process of adjusting results from a sample survey to infer results for the total in-scope population. To do this, a 'weight' is allocated to each sample unit, i.e; a person or a household. The weight is a value which indicates how many population units are represented by the sample unit. For the 2007-08 National Health Survey, separate person and household weights were calculated, as only one adult and one child per household were enumerated. The steps used to derive person and household weights are described below. Initial Household Weight The first step of the weighting procedure was to assign an initial household weight to fully responding dwellings. The initial household weight was calculated as the inverse of the probability of the household's selection in the sample. For example, if the probability of a household being selected in the survey was 1 in 600, then the household would have an initial weight of 600 (that is, it represents 600 households). The initial household weight was then adjusted as described below. Adjustment for Period Clusters of households were randomly assigned to one of four sub-periods that divided up the 11 month reference period, to ensure that possible seasonal differences for health variables would be minimised. The four sub-periods were:
Note that the lengths of the sub-periods are 2 months, 3 months, 3 months, and 3 months respectively. To ensure that each calendar quarter contributed equally to yearly estimates, the initial household weights were adjusted for sub-period length. Initial Person Weights After obtaining adjusted initial household weights, initial weights were assigned to fully-responding persons based on the sub-sampling scheme deployed within households. Initial person weights were calculated by inflating the person's adjusted household weight by the probability of the person being selected. For persons 18 years and over, the household weight was multiplied by the number of adults aged 18 years and over in the household, and for persons 0-17 years old, the household weight was multiplied by the number of children aged 0 to 17 years in the household. Non-response adjustment In developing the survey weights, information available for responding and non-responding households was used by the ABS to conduct quantitative investigations into explicit non-response adjustments. No explicit non-response adjustment was made to the weighting for the 2007-08 NHS, however, as the effect of the investigated non-response adjustments to the estimates was negligible. Benchmarks Person and household weights adjusted to quarter are calibrated to independent estimates of the population of interest, referred to as 'benchmarks'. Weights calibrated against population benchmarks ensure that the survey estimates conform to independently estimated distributions of the population rather than to the distribution within the sample itself. Calibration to benchmarks helps to compensate for over- or under-enumeration of particular categories of persons and households, which may occur due to the random nature of sampling or non-response. Calibration to Household Level Benchmarks The household benchmarks used in the 2007-08 NHS weighting were preliminary household estimates for December 2007 based on the 2006 Census of Population and Housing, scoped to the NHS. The calibration levels used for benchmarking were State by part of State by household composition (numbers of persons 0-14 years old, numbers of persons 15 years and over). Calibration to Person Level Benchmarks The person benchmarks used in the 2007-08 NHS were preliminary population estimates for December 2007 based on the 2006 Census of Population and Housing, scoped to the NHS. The calibration levels used for benchmarking were State by part of State by (typically 5 year) age groups by sex. Document Selection These documents will be presented in a new window.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
