4363.0.55.001 - National Health Survey: Users' Guide, 2001  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 27/05/2003   
   Page tools: Print Print Page Print all pages in this productPrint All  
Contents >> Chapter 2 - Survey Design and Operation

CONTENTS

Scope of the survey
Sample design and selection
Sample design
Sample selection
Sample deselection
Data collection
Interviews
Interviewers
Questionnaires
Measures to maximise response
Response rates
Data processing
Input processing
Input coding -
Country of birth coding
Coding of language spoken at home
Coding of occupation
Coding of industry and industry sector
Other input coding
Data entry
Output processing
Output coding
Coding of medical conditions
Coding of type of medication
Coding of alcohol consumption
Data entry
Computer processing
Estimation procedures
Benchmarks
Weighting specification
Age standardisation

Note: Except where indicated, the general information provided in this chapter is applicable to both the 2001 NHS(G) and 2001 NHS(I).


SCOPE OF THE SURVEY - 2001 NHS(G)

The 2001 NHS(G) covered urban and rural areas across all States and Territories of Australia; sparsely settled areas were excluded from the sample. Persons in scope of the survey were those identified by an adult within each sampled private dwelling as a usual resident of that dwelling A private dwelling was defined as a house, flat, home unit, caravan, garage, tent and any other structure being used as a private place of residence at the time of the survey. Non-private dwellings including hotels and motels, hostels and boarding houses were excluded. Also excluded were hospitals, nursing and convalescent homes, prisons, reformatories and single quarters of military establishments.

In general, a household comprised a group of persons living together in a dwelling who considered themselves to be separate from other people in the dwelling and who made regular provision to take meals together. All households within each sampled private dwellings were in scope of the survey, but not all members of those households were in scope. Persons living in Australia, but not usually considered part of the Australian resident population, and excluded from the scope of this survey were :
  • diplomatic personnel of overseas governments and non-Australian members of their households;
  • non-Australian service personnel stationed in Australia and their dependants; and
  • short term overseas visitors whose usual place of residence was outside Australia. Overseas visitors staying or intending to stay in Australia for 12 months or more were in scope.

Non-Australians (other than those above) working in Australia, or in Australia as students or settlers, and their dependants, were included in the survey scope.

Although the terms dwelling and household refer to conceptually different entities, for the purposes of the following discussion on weighting and estimation the terms are used interchangeably, appropriate to the sense of the particular point made.


SCOPE - 2001 NHS(I)

As with the 2001 NHS(G), the 2001 NHS(I) covered usual residents of private dwellings only. The 2001 NHS(I) covered sparsely settled and non-sparsely settled areas of Australia. Non-Indigenous persons were not eligible for selection in the supplementary Indigenous sample although, if they were the parent or guardian of an Indigenous child, they may have acted as spokesperson for the child (referred to as the child proxy).

In the 2001 NHS(I), only Indigenous households were considered in scope of the survey. An Indigenous household was defined as a household where at least one adult (18 years or over) of Aboriginal and/or Torres Strait Islander origin was usually resident. Households selected in non-sparsely settled areas were screened to identify Indigenous households for inclusion in the NHS(I) sample. Users should note that, in non-sparsely settled areas, Indigenous children living in households where there was no Indigenous adult usually resident did not have a chance of selection in the 2001 NHS(I), therefore resulting in under- representation of this group. Up to one quarter of all Indigenous children living in non-sparsely settled areas reside in such households. For more detail on this issue see the section 'Sample Design and Selection - 2001 NHS(I)' below.

SAMPLE DESIGN AND SELECTION


Sample design - 2001 NHS(G)

The 2001 NHS(G) 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 2001 NHS(G) sample was designed to provide:
  • relatively detailed estimates for each State, ACT and Australia;
  • relatively detailed estimates for capital city/rest of State areas within each State;
  • broad level estimates for regions within the more populous States;
  • estimates for those characteristics which are relatively common and sub-populations which are relatively large and spread fairly evenly geographically.

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/Territory sample fraction.

TABLE 2.1: State/Territory sample

NSW
Vic
Qld
SA
WA
Tas
NT
ACT

Approximate sampling fraction
1/520
1/440
1/400
1/245
1/290
1/135
1/335
1/95
Expected fully responding households
4,679
4,146
3,507
2,507
2,549
1,409
206
1,273




Within selected dwellings a random sub-sample of residents was enumerated as follows:
  • one adult (18 years of age and older);
  • all children aged 0 to 6 years;
  • one child aged 7 to 17 years.
This approach was expected to yield complete interviews from around 31,200 people; 20,000 adults and 11,200 children.


Sample Design - 2001 NHS(I)

The 2001 NHS(I) sample was designed to provide both additional Indigenous selections in non-sparsely settled areas and to include Indigenous people living in sparsely settled areas of Australia. The sample design was based on a broad dissection of Australia into non-sparsely settled areas and sparsely settled areas (defined as Statistical Local Areas (SLAs) with a dwelling density of less than 0.057 dwellings per square kilometre). The samples for non-sparse NHS(I) and sparse NHS(I) were designed separately, with each involving a multistage sampling process.
The 2001 NHS(I) was conducted using a stratified multistage area sample of private dwellings as outlined above for the 2001 NHS(G). The 2001 NHS(I) sample was designed to supplement the expected Indigenous sample selected in NHS(G) to provide estimates for those characteristics which are relatively common in the Australian Indigenous population.
    Within selected dwellings in non-sparsely settled areas a random sub-sample of residents was enumerated as follows:
    • one adult (18 years of age and older);
    • up to 2 children aged 0 to 17 years.
    Within selected dwellings in sparsely settled areas a random sub-sample of residents was enumerated as follows:
    • one adult (18 years of age and older);
    • up to 1 child aged 0 to 17 years.
    This approach yielded the following number of complete interviews from the total Indigenous sample:
    • Non-sparsely settled areas - 2,727
    • Sparsely settled areas - 954
    • Total Indigenous sample - 3,681(a)
      (a) These sample counts are based on the 'total Indigenous sample' which includes the 483 Indigenous respondents living in non-sparsely settled areas collected as part of the 2001 NHS(G) sample.


      Sample selection - 2001 NHS(G)

      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 Sub-Division levels of the Australian Standard Geographical Classification. These regions are then divided into Local Government Areas in State Capital City Statistical Divisions (metropolitan regions), and into major urban centres as well as minor urban and rural parts in non-metropolitan regions. 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:
      • a systematic sample of CDs was selected from each stratum with probability proportional to the number of dwellings in each CD;
      • each selected CD was divided into groups of dwellings or blocks of similar size, and one block was selected from each CD, with probability proportional to the number of dwellings in the block; and
      • within each selected block a list of all private dwellings was prepared and a systematic random sample of dwellings was selected. Dwellings selected were not contiguous, with 6-9 dwellings between each one selected.

      In Hobart, 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 Local Government Areas (LGAs) or combinations of both, and one or two units were selected from each stratum with probability of selection proportional to the number of dwellings in each unit. Within selected units, the sample of dwellings was arrived at in the same manner as outlined for high population density areas. The effect of this approach is that sample was not necessarily selected from each LGA, rather those selected represented neighbouring LGAs of similar geographical characteristics.

      In total a sample of approximately 26,960 households was selected which, taking account of an expected rate of sample loss (e.g. vacant dwellings, dwellings under construction etc.) of 13% and non-response of 15%, was designed to achieve the desired sample of about 20,000 fully responding households.

      To take account of possible seasonal effects on health characteristics, the sample was initially allocated equally to each quarter of the calendar year 2001. Selected CDs were randomly allocated in such a way as to ensure an acceptable compromise between an even spread of sample throughout the year and adequate workloads for interviewers. No workloads were allocated during the six week period from 28 July to 10 September 2001, as this was the enumeration period for the Census of Population and Housing and associated Post Enumeration Survey.


      Sample deselection - 2001 NHS(G)

      Owing to higher than expected survey enumeration costs a decision was taken in August 2001 to deselect sample from the 4th quarter of enumeration. This deselection was confined to sample selected to be enumerated in the months of October, November and December, and represented about 80% of the sample from those months. Enumeration of this reduced sample was undertaken in October and November. The deselection was distributed across States according to the distribution of the original sample; the sample in NT and ACT was not reduced. The final expected sample take after this deselection is shown in the table below.

      TABLE 2.2: Adjusted sample sizes

      NSW
      Vic
      Qld
      SA
      WA
      Tas
      NT
      ACT

      Expected fully responding dwellings:
      Capital city
      Balance of State
      Total
      2,510
      1,449
      3,959
      2,695
      1,038
      3,733
      1,535
      1,588
      3,123
      1,542
      548
      2,089
      1,694
      484
      2,179
      514
      639
      1,153
      152
      104
      *256
      1,378
      -
      *1,378

      * - these figures are higher than those in Table 2.1 above due to realised lower sample loss and higher response rates than those originally anticipated.

      As a result of this deselection, an adjustment to the weighting method was introduced; see Estimation Procedures, in this Chapter. Also, as a result of this sample reduction, the survey estimates relate to the period February to November rather than the full 12 months of 2001.

      Sample deselection was not applicable to the 2001 NHS(I).


      Sample Selection - 2001 NHS(I)

      In the non-sparse NHS(I), dwellings were selected using a stratified multistage area sample. Similar to most of the NHS(G) sample selection, Census collection districts (CDs) formed the first stage selection units. CDs were selected with a probability proportional to the number of Indigenous households in the CD as identified during the 1996 Census of Population and Housing. Therefore, CDs with a higher proportion of Indigenous households had a greater chance of selection. A random selection of dwellings within selected CDs were then screened to assess their usual residents' Indigenous status. Where a dwelling contained one or more Indigenous usual residents aged 18 years or more, one Indigenous adult (18 years of age or more) and up to two Indigenous children (0 to 17 years of age) were randomly selected to participate in the survey. This sample design has been accounted for in the weighting process.
      In the 2001 NHS(G), and as was the case for the 1995 NHS, all children of Aboriginal and/or Torres Strait Islander origin living in households in non-sparsely settled areas had a random chance of selection in the survey sample. However, in the 2001 NHS(I), in non-sparsely settled areas selected households were screened to identify only those households where at least one adult (18 years or over) of Aboriginal and/or Torres Strait Islander origin was usually resident. Therefore, Indigenous children living in households in non-sparsely settled areas where there was no Indigenous adult usually resident (up to one quarter of all Indigenous children in non-remote areas reside in such households) did not have a chance of selection in the NHS(I).

      The estimation strategy involved weighting the total Indigenous sample obtained from both NHS(G) and NHS(I) to Indigenous population benchmarks at the age, sex and area of usual residence level. The weighting strategy did to some extent redress the undercoverage of Indigenous children in the NHS(I), but not completely. Therefore there is still a potential bias in estimates where the health characteristics of Indigenous children living in households where there are no Indigenous adults are substantially different to those of Indigenous children in households where Indigenous adults are resident. Analysis of this potential bias due to the under-representation of Indigenous children was not found to be significant in the context of the sampling error associated with the survey. In the 2004-05 Indigenous Health Survey, field procedures will be changed to provide for adequate representation of Indigenous children in households with no resident Indigenous adult.

      In the sparse NHS(I), the sample was obtained from a random selection of discrete Indigenous communities and outstations across Australia using information collected in the 1999 Community Housing and Infrastructure Needs Survey (CHINS). Within selected communities and outstations, a random selection of dwellings was made. Within selected dwellings, one Indigenous adult (18 years of age or more) and up to one Indigenous child (0 to 17 years of age) were randomly selected to participate in the survey.

      For the non-sparse NHS(I), approximately 91% of households identified with in-scope Indigenous residents responded to the survey. This response rate does not take into account the 6.5% of households in the screened component of the sample that were unable to be contacted to establish the Indigenous status of the occupants. For the sparse NHS(I), approximately 87% of in-scope households responded to the survey.

      DATA COLLECTION

      Information was obtained in the 2001 NHS by trained ABS interviewers, in the main through personal interviews with an adult member of selected households in scope of the survey. Aspects of data collection are discussed below under the headings: interviews, interviewers and questionnaires.


      Interviews - 2001 NHS(G)

      In the 2001 NHS(G) 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. A brochure, providing some background to the survey, information concerning the interview process and a guarantee of confidentiality was included with the initial approach letter (Primary Approach Letter). For a small number of households where the ABS did not have an adequate postal address, this was not possible.

      At the initial visit by the interviewer, a household form was completed from information provided by a responsible adult member of the household (ARA). This form sought details of the number and basic demographic characteristics of usual residents of the dwelling and established those persons in scope of the survey.

      One adult respondent was selected from each dwelling. This was a usual resident aged 18 years or more whose birthday was closest after the date of interview. If two or more adults had the same birthday the adult respondent was that person whose first name came first in alphabetical order. A personal interview was conducted with the selected adult where possible. If the dwelling contained no usual residents aged 18 years or more the dwelling was not enumerated.

      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. Information was collected by proxy for just over 1% of adult respondents. 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 for the interview to be conducted either by an ABS interviewer fluent in the respondent’s own language or with an ABS interpreter.

      An adult was interviewed on behalf of children selected within the dwelling. All children aged 0 to 6 years were enumerated, and one child aged 7 to 17 years, selected on the basis of birthday closest to date of interview. The ARA was asked who the best person in the dwelling was to report information about children in the dwelling. This person is referred to as the Child Proxy. There was only one child proxy per dwelling. If the child proxy had difficulty in answering questions on behalf of a child, children may have been asked questions, but only with a parent or guardian's consent.

      The relationship of the child proxy and the children for which they were reporting was as below:

      TABLE 2.3: Relationship of proxy to child

      Relationship
      % of children selected

      Mother / stepmother
      78.8
      Father /stepfather
      18.6
      Other relative
      2.0
      Other
      0.6


      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 desirable to spread interviews across the other days of the week, interviews were conducted on days to suit respondents. The result was a bunching of interviews on Mondays, Tuesdays and Wednesdays, as shown below.


      TABLE 2.4: Percentage of adults interviewed

      Monday
      Tuesday
      Wednesday
      Thursday
      Friday
      Saturday
      Sunday
      Total

      23.3
      23.7
      20.4
      15.5
      9.6
      7.1
      0.5
      100.0


      At the completion of the interview, female respondents aged 18 and over years were invited to complete an additional questionnaire relating to additional women’s health issues. Participating women completed the questionnaire themselves and returned the form to the interviewer at that time in a sealed envelope provided. These questionnaires were not left with respondents for later collection or mail back. To assist in overcoming language difficulties, this questionnaire was made available in seven commonly spoken languages other than English; Arabic, Chinese (Traditional), German, Greek, Italian, Spanish and Vietnamese. In other cases where the respondent was unable to complete the form themselves (e.g. literacy problems) the interviewer administered the form, ensuring first the respondent understood the nature of the questions asked, and had agreed to participate.

      In total, conduct of the survey averaged 45 minutes per household, which included completion of the household form, all personal (and/or proxy) interviews and completion of the supplementary women’s health questionnaire.

      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.


      Interviews - 2001 NHS(I)

      In the non-sparse 2001 NHS(I) Primary Approach Letters (PALs) were not used for selected households. However, for sparse NHS(I) ABS Regional Offices contacted communities and clinics by telephone prior to enumeration commencing and a Community Approach Letter was used as confirmation.
      Persons aged 18 years or more were interviewed personally, with the exception of persons who were too sick or otherwise unable to respond personally. Persons aged 15 to 17 years were interviewed with the consent of a parent or guardian; otherwise a parent or guardian was interviewed on their behalf. For persons aged under 15 years, information was obtained from a person responsible for the child (referred to as the child proxy).

      There were, however, a number of differences in the data collection methods in sparsely and non-sparsely settled areas for the 2001 NHS(I). In non-sparsely settled areas, adult females were invited to complete a small additional questionnaire covering specific supplementary women's health topics. This additional questionnaire was voluntary and self-enumerated. Of the Indigenous women invited to complete the supplementary questionnaire in non-sparsely settled areas, 91% responded.

      In sparsely settled areas, standard household survey approaches were modified to take account of language and cultural issues. In addition, interviewers worked in teams of two, one male and one female, to collect the survey information. Male interviewers collected personal information from male respondents, and female interviewers collected personal information from female respondents. The interviewers were accompanied, wherever possible, by local Indigenous facilitators, preferably one male and one female, who assisted in the conduct and completion of the interviews. The Indigenous facilitators explained the purpose of the survey to respondents, introduced the interviewers, assisted in identifying the usual residents of a household and in locating residents who were not at home, and assisted respondent understanding of the questions where necessary.

      In addition, the survey content collected in sparsely settled areas was limited to those topics for which data of acceptable quality could be collected. Some questions were reworded to assist respondents in understanding the concepts. In sparsely settled areas, a subset of the supplementary women's health topics was not self-enumerated but was collected through face-to face personal interview with adult female respondents who were informed of the potential sensitivity and voluntary nature of these additional questions.


      Interviewers - 2001 NHS(G)

      Interviewers for the 2001 NHS(G) were primarily 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 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.


      Interviewers - 2001 NHS(I)

      Interviewers for the 2001 NHS(I) were primarily recruited from a pool of trained ABS interviewers who had previous experience with Indigenous surveys, particularly in Indigenous communities within sparsely settled areas. Those selected to work on the NHS(I) undertook NHS specific classroom training and were required to satisfactorily complete home study exercises. In addition, all interviewers selected to work on the 2001 NHS(I) underwent Indigenous Cultural Awareness training. All phases of the training emphasised an understanding of the survey concepts, definitions and procedures in order to ensure that a standard approach was employed.

      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.


      Questionnaires - 2001 NHS(G)

      Four questionnaires were developed for and used in the 2001 NHS(G), as outlined below.
      • Household Form - this form was used to collect basic demographic data about usual residents of the household (e.g. sex, age, date of birth, birthplace, Indigenous status, marital status) and details of the relationship between individuals in each household. This information was obtained from any responsible adult within the household. The form was also used to identify the selected adult respondent for the dwelling, and the child proxy where applicable. The form was also used by interviewers to record their calls made to the dwelling and the subsequent response status of the household in the survey (e.g. fully responding, refusal, vacant dwelling, etc.).
      • Personal Interview Adult Questionnaire - a single questionnaire was used to collect information from adults about their demographic and socio-economic characteristics and health characteristics such as health-related actions they had taken, long-term illness conditions experienced, selected lifestyle behaviours, etc.
      • Personal Interview Child Questionnaire - a single questionnaire was used to collect information about each child; this included their demographic and (for older children) their socio-economic characteristics and various health characteristics. This questionnaire was also used to collect some demographic and socio-economic information about the child proxy. Where the child proxy was also the selected adult in the dwelling, some of these details were copied from the selected adult's record; where the child proxy was not the selected adult the information was separately obtained from the child proxy and recorded on the child's form.
      • Women’s Health Supplementary Form - this form was provided at the completion of their interview to female respondents aged 18 years and over. It contained questions relating to specific women's health issues and was completed by the respondent in writing and returned to the interviewer in a sealed envelope. This approach was adopted in recognition of the potential sensitivity of the topics covered. Instructions to respondents on how to complete the form were contained on the form itself. In general, this form was to be completed without interviewer intervention, but interviewers may have assisted respondents if queried regarding the applicability of specific questions to themselves, etc.

      The Adult and Child questionnaires were designed to be administered using standard ABS procedures for conducting population interview surveys, having regard to the particular aims of the survey and of the individual topics within it, and to 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, sensitivity of topics, etc. Where appropriate, excerpts from previous ABS surveys on the topics covered were adopted. All questionnaires were fully field tested to ensure:
      • they were adequately addressing the data requirements from the survey i.e. they obtained the data required in the most effective and efficient way;
      • there was minimum respondent concern about the sensitivity or privacy aspects of the information sought, effective respondent/interviewer interaction, acceptable levels of respondent load, etc; and
      • the operational aspects of the survey were satisfactory e.g. arrangement of topics, sequencing of questions, adequacy and relevance of coding frames, etc.

      The questionnaires employed a number of different approaches to recording information at the interview:
      • questions to which responses were classified by interviewers to one or more of a set of predetermined response categories. This approach was used for recording answers to the more straightforward questions, where logically a limited range of responses was expected or where the focus of interest was on a particular type or group of response (which were listed in the questionnaire, with the remainder being grouped together under ‘other’);
      • questions to which responses were recorded by interviewers as reported, for subsequent classification and coding by office staff during processing. This style of question was used for the potentially more complex topics such as type of illness condition, type of medication used, type and quantity of alcohol consumed, etc;
      • questions asked in the form of a running prompt i.e. predetermined response categories were read out to the respondent one at a time until the respondent indicated agreement to one or more of the categories (as appropriate to the topic) or until all the predetermined categories were exhausted; and
      • questions asked in association with prompt cards. Printed lists of possible answers to the question were handed to the respondent who was asked to select the most relevant responses. By listing a set of possible responses (either in the form of a prompt card or a running prompt question) the prompt served to clarify the question or to present various alternatives, to refresh the respondent’s memory and at the same time assist the respondent to select an appropriate response.

      To ensure consistency of approach, interviewers were instructed to ask the interview questions as written in the questionnaire. 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 illness, where interviewers were asked to prompt for a condition if a treatment or symptom was initially reported.

      Copies of all the NHS questionnaires and related prompt cards are available from the ABS website.


      Questionnaires - 2001 NHS(I)

      Seven questionnaires were used in the collection of data for the 2001 NHS(I). These included:
      Three questionnaires which were developed for sparsely-settled areas NHS(I)
      (i) Household form
      (ii) Adult form (which included some questions on the Women's Health topic)
      (iii)Child's form; and

      Four questionnaires (those used in the 2001 NHS(G)) for non-sparsely settled areas NHS(I)
      (iv) Household form
      (v) Adult form
      (vi) Child's form
      (vii) Women's supplementary health form

      Copies of these forms are available on the ABS web site under the Health and Indigenous theme pages.

      The survey content for the non-sparse NHS(I) was the same as the content included in the 2001 NHS(G) - however, the following information was not collected for the NHS(I):
      • Country of Birth - Question 6 on the NHS(G) Adult and Child's forms
      • Year of Arrival - Question 7 on the NHS(G) Adult and Child's forms
      • Mental Health - Questions 250 to 269 on the NHS(G) Adult form.

      The total survey content for the sparsely settled areas NHS(I) was a subset (approximately 50%) of that collected in non-sparsely settled areas NHS(I) and the 2001 NHS(G), and was limited to those topics for which data of acceptable quality could be collected. In addition, some of the questions were reworded to assist respondents in understanding the concepts. As a result, separate questionnaires were developed for the sparsely settled areas NHS(I). The survey content for the sparsely settled areas sample included the following health topics:
      • health related actions such as visits to hospitals, doctors, dentists and other health professionals
      • health risk factors such as smoking, alcohol consumption and adult immunisation
      • health status indicators such as asthma, injuries, diabetes, cancer, cardiovascular conditions, hearing and sight problems
      • women's health such as cervical cancer screening, breastfeeding and contraception.

      In addition to the topics not collected in non-sparsely settled areas, topics such as private health insurance, asthma symptoms, exercise, nutrition and child's immunisation were not collected in sparsely settled areas.

      A detailed list of the data items available from the 2001 NHS(I) and sample copies of the household forms and questionnaires are available on the ABS web site under the Health theme page. Appendix 13 in this Users' Guide provides a broad outline of the 2001 NHS(I) topics and data items and indicates whether they were collected in sparsely-settled areas.

      MEASURES TO MAXIMISE RESPONSE - 2001 NHS(G)

      Response to a survey can be considered in two parts:
      • obtaining contact with the respondent and information from the respondent; and
      • ensuring that the information obtained is as accurate and relevant to survey objectives as possible.

      This section deals with the first of these shown above. Measures taken to ensure accuracy and relevance of the data (e.g. survey procedures, questionnaire design, interviewers, etc.) have previously been outlined in this chapter.

      In any sample survey responses should ideally be obtained from all selected units; in practice 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:
      • advice to selected households in the 2001 NHS(G), initially by letter, explaining that their dwelling had been selected for the survey, the purposes of the survey, its official nature and the confidentiality of the information collected. The letters gave advance notice that an ABS interviewer would call, and provided an ABS contact number for more information if required. An information brochure, specially designed for selected households was provided with the initial approach letter. This procedure could not be followed for a small number of households where the ABS did not have an adequate postal address, and was not adopted for households included in the supplementary sample of Indigenous people.
      • stressing the importance of participation in the survey by selected households. Each selected dwelling (and its residents) represented a number of others in that local area in that State and in Australia. Each household selected, represented a number of others like them in size, composition, location, occupation, lifestyle and health. The cooperation of those selected was important to ensure all households/persons were properly represented in the survey and properly reflected in survey results;
      • stressing the importance of the survey to the planning and provision of health services and facilities to meet Australia's health needs; and
      • stressing the confidentiality of all information collected. The confidentiality of data is guaranteed by the Census and Statistics Act 1905; under provisions of this Act the ABS is prevented from releasing any identifiable information about individuals or households to any person, organisation or government authority.

      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 at least three call-backs in rural areas and at least five in urban areas before a dwelling was classified as a ‘non-contact’. Call-backs occurred at different times during the day to increase the chance of contact. Once contacts 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, the selected adult or designated child proxy within a dwelling could not be contacted or interviewed, and these were classified as individual non-contacts.

      Respondents who refused to participate were usually followed-up later by letter and a subsequent visit by an office supervisor. Completed questionnaires were obtained where possible.

      MEASURES TO MAXIMISE RESPONSE - 2001 NHS(I)

      In addition to the procedures followed in the 2001 NHS(G) outlined above, special measures were taken to maximise the response rates from households selected in the 2001 NHS(I).

      The collection methodology used in the sparsely settled areas NHS(I) was adapted to be culturally appropriate for Indigenous persons living in these areas. However, the ABS was conscious not to change the collection methodology to the extent that data collected in sparsely settled areas would not be comparable with that collected in non-sparsely settled areas NHS(I) or the 2001 NHS(G).

      The special measures used in sparsely settled areas NHS(I) for interviews, including the use of 'teams' of one male and one female interviewer, the use of facilitators and the use of personal face to face interviews for the supplementary women's questions, are outlined above under 'Interviews - 2001 NHS(I)',

      The final collection methodologies and data content for the 2001 NHS(I) were determined through field testing, analysis of testing outcomes, and consultations with Indigenous respondents and people in the Indigenous health field.


      Field testing

      Separate testing programs were undertaken in non-sparsely settled areas and sparsely settled areas of the 2001 NHS(I) as outlined below.


      Non-sparse NHS(I)

      A pilot test was conducted in October 2000 to refine and adapt, where necessary, standard ABS interviewer procedures for the enumeration of Indigenous respondents in non-sparsely settled areas. The purpose of the test was to:
      • assess the screening procedures used to identify Indigenous households within selected Collection Districts
      • conduct test interviews with members of Indigenous households and assess respondent reaction to the survey
      • evaluate the appropriateness of the separate, voluntary, self-enumerated Women's Health Form (WHF) for Indigenous women.

      A focus group attended by Aboriginal women, including health workers, from regional NSW was also held to test the suitability of the WHF. The women were observed, with permission, as they filled in a WHF and then asked about certain issues such as the sensitivity of collecting and disseminating this women's health information, any literacy problems, any problems understanding concepts and questions, and overall reaction to the form. It was found that the younger women in the focus group i.e. those aged approximately 18 to 30, experienced few difficulties completing the form. They found it easy to complete, understood all the question terms and definitions and did not find any of the questions to be too sensitive. Some of the women had difficulty with recall questions, especially Q29 (age first started taking the pill). For cases where they couldn't remember the exact period they provided as estimate. Older women (those aged 50 and over) were more likely to have difficulty completing the WHF because of:
      • literacy problems
      • unfamiliarity with the completion of such forms
      • a general mistrust of government motives for the collection of this information.

      It was thought that the proposed collection methodology (i.e. ABS interviewers approaching households and interviewing selected respondents) would result in a poor response, both in terms of the number of likely refusals and the quality of information provided. To overcome this, it was suggested that Aboriginal Health Workers should accompany interviewers or complete the interviews themselves.


      Sparse NHS(I)

      An extensive testing program, comprising pre-testing, a pilot test and a dress rehearsal, was undertaken in 10 remote Indigenous communities across Australia from May 2000 to February 2001. The purpose of the tests were to:
      • assess the suitability and appropriateness of survey concepts, data items, question wording and terminology
      • gauge respondent, and community, reaction to the survey
      • conduct test interviews and assess the suitability of the personal interview collection methodology
      • assess the feasibility of alternative collection methodologies, e.g. collecting selected health data from clinic records
      • assess operational procedures, such as using both a male and a female interviewer, the use of local Indigenous people to assist interviewers, the random selection of respondents, etc.
      • consult with staff working in remote area Indigenous health clinics in regard to a suitable collection methodology and appropriate content.


      Analysis of testing outcomes

      Outcomes from the field tests, and the resulting test data, were analysed separately for non-sparsely settled areas and sparsely settled areas of the NHS(I).


      Non-sparse NHS(I)

      Overall, the survey was well received by respondents who participated in the pilot test. Respondents had little difficulty understanding the concepts and terms used in the questionnaires. Interviewers noted that the WHF was generally well received by female respondents, with only a few respondents requesting assistance and clarification (due to literacy levels) from interviewers when completing the form. Others asked family members to help them. The content of the WHF was not perceived as being overly sensitive or inappropriate.

      Following the field test, the ABS recommended the inclusion of the WHF in the non-sparsely settled component based on:
      • the high priority given by users to the collection of Indigenous women's health information
      • the issue of sensitivity with the collection of the topics being adequately addressed through the use of the voluntary, self-completion form and the provision of an envelope for the completed form
      • the overall success of the enumeration of the WHF during testing
      • the fact that the ABS has published similar information on women's health in the past without adverse reaction.

      Special consideration was given to the testing and analysis of mental health questions in the non-sparsely settled pilot test. The 2001 NHS(G) used the Kessler Psychological Distress Scale 10 (K-10) module of questions to collect information on mental health. These questions were tested in the non-sparsely settled areas pilot test to determine their appropriateness for the Indigenous population. Analysis of the results of the pilot test use of K-10 was inconclusive in regard to Indigenous respondents' reactions to the appropriateness and sensitivity of the questions. However, user consultations indicated that the Kessler 10 questions being used in the 2001 NHS(G) may not be meaningful or culturally appropriate for the Indigenous population without substantial modification. As a result, the K-10 was not used in the 2001 NHS(I).


      Sparse NHS(I)

      The experience and knowledge gained through the testing program in sparsely settled areas was invaluable to the development of the survey, and was used to revise survey content, questionnaires and operational procedures. The survey was well received by the communities and respondents who participated in each field test. The use of local people to assist the interviewers was also very successful in facilitating respondent acceptance of the interviewers and participation in the survey.

      RESPONSE RATES - 2001 NHS(G)

      After sample deselection as outlined above, a total of 21,891 private dwellings were selected in the sample for the 2001 NHS. This reduced to an active sample of 19,408 dwellings after sample loss in the field stage.

      The final response status for these dwellings is shown below:


      TABLE 2.5: Final response rates

      % of active sample - dwellings

      Full refusal
      2
      Fully non-contact
      3
      Fully non-response, for other reasons
      3
      Part refusal / non-contact
      1
      Fully responding
      92
      Total
      100 (19,408)



      Completed questionnaires were obtained for 26,863 persons in these dwellings, as shown below.


      TABLE 2.6: Completed questionnaires

      Age group (yrs)
      All ages


      0 - 6
      7 - 17
      18 and over
      Capital city
      Rest of State
      Total

      Number of records
      NSW
      1,020
      1,025
      3,953
      3,771
      2,227
      5,998
      Vic
      909
      915
      3,712
      3,912
      1,624
      5,536
      Qld
      814
      777
      3,142
      2,267
      2,466
      4,733
      SA
      449
      475
      2,092
      2,234
      782
      3,016
      WA
      514
      577
      2,204
      2,549
      746
      3,295
      Tas
      296
      268
      1,166
      729
      1,001
      1,730
      NT
      92
      72
      262
      271
      155
      426
      ACT
      373
      369
      1,387
      2,129
      -
      2,129
      Aust.
      4,467
      4,478
      17,918
      17,862
      9,001
      26,863



      RESPONSE RATES - 2001 NHS(I)

      For non-sparse and sparse NHS(I) the final response status is shown in the Table 2.7 and Table 2.8 below:

      TABLE 2.7: Non-sparse NHS(I) - Response rates for households

      State
      Fully responding(a)
      Non-response(b)
      Sample loss(c)
      Total households

      NSW
      340
      56
      13
      409
      Vic
      51
      7
      1
      59
      Qld
      397
      2
      0
      399
      SA
      67
      14
      7
      88
      WA
      164
      19
      2
      185
      Tas
      27
      1
      1
      29
      NT
      44
      6
      0
      50
      ACT
      2
      0
      0
      2
      Total
      1,092
      105
      24
      1,221


      Total % response (including sample loss) = 89%
      Total % response (excluding sample loss) = 91%
      (a) Fully Responding - all schedules fully complete for all selected Usual Residents.
      (b) Non-Response - e.g. full refusal, part refusal, full non-contact, part non-contact, etc.
      (c) Sample loss - e.g. out on scope, vacant dwelling, etc.

      TABLE 2.8: Sparse NHS(I) - response rates for households

      State
      Fully responding(a)
      Non-response(b)
      Sample loss(c)
      Total Households


      Qld
      128
      7
      10
      145
      SA
      11
      4
      17
      32
      WA
      132
      22
      92
      246
      NT
      247
      45
      171
      463
      Total sample
      518
      78
      290
      886


      Total % response (including sample loss) = 58%
      Total % response (excluding sample loss) = 87%
      (a) Fully Responding - all schedules fully complete for all selected Usual Residents.
      (b) Non-Response - e.g. full refusal, part refusal, full non-contact, part non-contact, etc.
      (c) Sample loss - e.g. out on scope, vacant dwelling, etc.

      DATA PROCESSING

      A combination of clerical and computer-based systems was used to process data obtained in the 2001 NHS. These are outlined below.


      Input processing

      Clerical checks were initially applied by interviewers to ensure the completeness and consistency of the questionnaires before being returned to the ABS. Depending on the nature of any errors, omissions or inconsistencies identified by interviewers they may have contacted the respondent for clarification, or more commonly instead made a note of any such problems and provided additional comment about individual questionnaires as appropriate when returning the questionnaires to an ABS office for processing.

      All questionnaires were again checked on receipt in ABS State offices to ensure interviewer workloads were fully accounted for and that all questionnaires and other documents for each household and respondent were completed and present, as appropriate.

      It was at this stage that households requiring follow-up, because of refusal, non-contact or language difficulties were identified for further action. Questionnaires were then assessed for adequacy. This essentially was a determination of whether or not sufficient data had been obtained to make that questionnaire useable. Questionnaires assessed as adequate were input coded, OMR scanned, computer edited, and included on the NHS database; these steps are outlined below. Child questionnaires were assessed as inadequate and discarded if an accompanying adequate adult questionnaire for that household was not present.


      Input coding

      Input coding refers to the coding of country of birth, language spoken at home, occupation, industry and industry sector and coding of the relationships within the household and related characteristics (e.g.social marital status). This coding was performed by office assistants in ABS State offices, prior to data entry. Coding of alcohol consumption, type of long-term medical conditions and medication type (output coding) was undertaken later in processing; see Output Coding below. A brief outline of the input coding undertaken follows:
      • Country of birth coding - The survey questionnaire listed 10 most frequently reported countries: interviewers were instructed to mark the appropriate box, or if the reported country was not among those listed to write in the name of the country for subsequent office coding. Written responses were classified to the Standard Australian Classification for Countries. A copy of the standard output categories at the minor group level of this classification is contained in Appendix 7. Information on country of birth was not collected for the 2001 NHS(I).
      • Coding of language - The survey questionnaire listed 10 frequently reported languages spoken at home. Interviewers were instructed to mark the appropriate box, or if the reported language was not among those listed, to write in the name of the language for subsequent office coding. Language was coded to the Australian Standard Classification of Languages. A copy of the standard groupings for output from this classification is contained in Appendix 8. Note that the sparse 2001 NHS(I) Adult and Child's forms listed only 4 categories - 'Only English', 'Aboriginal', 'Torres Strait Islander' and 'Other'.
      • Coding of occupation - In the 2001 NHS, occupation relates to the main job held by employed respondents at the time of their interview. Occupation was office coded, based on a description of the kind of work performed, as reported by respondents and recorded by interviewers. Occupation was coded to the four-digit (unit group) level of the Australian Standard Classification of Occupations (1222.0).
      • Coding of industry and industry sector - In the 2001 NHS, industry and industry sector relates to the main job held by employed respondents at the time of their interview. These were office coded based on the name of employer and the respondent's description of the business or service carried out at the respondent's workplace. Industry was coded to the four-digit level of the Australian and New Zealand Standard Industrial Classification (1292.0).
      • Other input coding - Limited information from the Household Form about the composition of the household was transcribed to individual respondent questionnaires by interviewers. More detailed information about the composition of households, the relationships between members and some characteristics of members was entered onto each adult and child questionnaire by Office assistants during initial processing and before data entry. Items derived from information on the Household Form, or copied directly from it included relationship of the selected adult, child or child proxy to other household members in non-family or mixed family households, social and registered marital status, age composition and size of household, student status and parent status of adult household members. Some administrative information was also recorded via this process.

      Further details of all input coding are contained in the Office Assistant's Instructions developed for this survey; these are intended as internal documents only and hence have not been provided as part of this Users' Guide. However, further information including excerpts from these Instructions may be made available on request as appropriate.


      Data entry - 2001 NHS(G)

      Following coding, all data from the adult and child questionnaires was progressively entered onto a computer file via an OMR (Optical Mark Reader) system. This system read in the data, created a computer record for each respondent, and incorporated an extensive range of computer edits to check that logical sequences had been followed in the questionnaires, that all necessary items were present and that relationships between items as originally recorded by the interviewer or respondent were valid. Errors encountered in this process were resolved at input stage, and amendments applied to the new file. All questionnaires were then returned to Canberra for further coding.

      The Women's Health Supplementary Forms were not OMR forms; these were sent to the Canberra office of the ABS for key data entry via an INSPECT (INputS ProcEssing & Collection Template) system built for this purpose. This system incorporated input edits. Data from these questionnaires were combined for each person with data from their main questionnaire for further processing.


      Data entry - 2001 NHS(I)

      Input data for the non-sparse NHS(I) sample was captured in an identical manner to the 2001 NHS(G) sample, using the same processing system.

      The sparse NHS(I) forms were processed separately. As with the WHFs, the sparse forms were not OMR forms and were sent to Central Office for key data entry via the INSPECT system mentioned above. A record was created for each respondent. Extensive edit checks were built into this system. These edits checked that logical sequences had been followed, all necessary items were present, that specific values lay within valid ranges and that relationships between items as originally recorded by the interviewer were valid. All sparse forms were also manually checked by ABS staff and where errors were encountered they were resolved and amended. All 954 sparse forms were manually checked and amended where necessary.


      Output processing - 2001 NHS(G)

      Once correct (clean) the data file created from data entry was progressively passed on for output processing. This stage involved clerical coding/data entry of coded data for some fields, more detailed and complex editing of data, derivation of output data items, and the inclusion of weights for the production of estimates. These activities are briefly outlined below.


      Output coding

      Clerical 'output' coding was undertaken on the following health items:
      • medical conditions
      • type of medication
      • alcohol consumption.

      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. A Computer Assisted Coding system (CAC) was developed for each of the three items above and a detailed quality control process was introduced to ensure that the coding process met agreed standards.


      Coding of medical conditions - All reported long-term medical conditions were coded to a list of approximately 1000 conditions, which was built into a CAC system. Conceptually the coding process simply involved locating the reported condition in the CAC system, and recording the 3 digit code allocated. 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 was compiled for use in the NHS by the Family Medicine Research Centre, University of Sydney, in association with the ABS. Conditions classified at the full level of detail will not generally be available for output from the survey; however, they can be regrouped in various ways for output. Three standard output classifications, developed for the NHSs, are available:
      • a classification based on ICD10,
      • a classification based on the International Classification of Primary Care (ICPC), and
      • a classification similar to the ICD9-based classification of conditions used in the 1995 NHS. This is intended to provide users with the ability to compare 2001 and 1995 results.

      A copy of each of these output classifications is provided as Appendixes 3, 4 and 5. Further information about the CAC system and how it was applied in the survey can be provided on request.


      Coding of type of medication - Whereas the 1995 NHS aimed to collect information about all medications used in a two week period, the 2001 survey collected only information on medications used for asthma, cancer, heart and other circulatory conditions and diabetes, and type of medications used for mental well-being. Information recorded about medications used for mental well-being was recorded in the questionnaire in broad type groups and was not further office coded.

      The questionnaire provided space to record the names of up to three medications each used for asthma, cancer and diabetes, and up to 12 medications used for heart and circulatory in the reference period. The coding process involved assigning a 4 digit generic type of medication code to each medication name recorded.

      A CAC system was developed incorporating the names of medications readily available in Australia and commonly used for the nominated conditions. The system drew on information from the World Health Organisation’s Anatomical Therapeutic Chemical (ATC) Classification, the Australian Medicines Handbook, information provided by officers of the Department of Health and Ageing and information provided by health professionals from related fields. 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. Further information about the CAC system and how it was applied in the survey can be provided on request.

      A classification of generic type of medication based on the World Health Organisation’s Anatomical Therapeutic Chemical (ATC) Classification (and associated coding indexes) and the Australian Medicines Handbook was developed for use in this survey. Although all medications reported as used for the designated conditions were recorded, the classification of generic type focussed on those types most commonly used for those conditions; other medications were classified to more general categories within the classification. Details of the classification used are provided in Appendix 6. Brand name information is not available for output from this survey.


      Coding of alcohol consumption - In the 2001 NHS(G), information about alcohol consumption was recorded against seven general categories of alcoholic drinks: low alcohol beer, medium strength beer, full-strength beer, wine, spirits, fortified wine and other alcoholic beverages. Details of the quantity of each of these drinks consumed on (up to) the last three days in the week prior to the day of interview were recorded. Quantities were recorded in terms of standard measures where possible; otherwise a description of the quantity consumed was recorded by interviewers. Interviewers were encouraged to record further details about the brand or name of drink where possible to assist in coding.

      A CAC system was developed to calculate in millilitres the amount of pure alcohol contained in the drinks which respondents reported they had consumed. This system used information about the type of alcoholic drinks consumed (including brand name for common drinks), and the size and number of drinks consumed; a conversion factor was applied to this information 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 seven broad types of alcoholic drinks used in the survey were included for cases where more detailed information had not been recorded at interview.

      The coding of alcohol consumption involved a number of separate steps:
      • checking the information recorded and reclassifying (where appropriate) consumption recorded in the ‘other alcoholic beverages’ category;
      • converting reported measures to standard measures as required; and
      • for each type of drink consumed for each of the three previous days, converting the quantities of drink consumed to the quantity of alcohol contained in those drinks.

      Look-up tables were provided within the CAC system to assist in the conversion process for beers, spirits, wines and fortified wines. These look-up tables listed the amount of pure alcohol contained in a given number of standard containers/measures of various sizes. Conversion factors were applied in cases not provided for in the look-up tables.

      Further information about the CAC system and how it was applied in the survey can be provided on request.


      Output processing - 2001 NHS(I)

      The non-sparse NHS(I) forms went through identical output processing as outlined above for the NHS(G) forms.

      Sparse NHS(I) data was processed separately from the non-sparse NHS(I) data. A sparse NHS(I) file was created which incorporated the special requirements of the sparse NHS(I) data, as detailed below, and aligned as closely as possible to the non-sparse NHS(I) file.

      In some cases, there were different question formats, forms or different codes used for input processing between the 2001 NHS(G)/non-sparse NHS(I) and the sparse NHS(I). In these cases the sparse NHS(I) output file was revised to reflect the 2001 NHS(G)/non-sparse NHS(I) structure or coding system.

      In a number of cases, due to differences in the way data were collected, a number of 2001 NHS data items had to be redesigned for sparse NHS(I) output. In these cases a sparse NHS(I)-specific data item was produced, which reflected the concept of the non-sparse NHS(I) data item as much as possible but does not exactly match the data item (in most cases, categories may be missing or combined together).


      Data entry

      Following the clerical checks and coding outlined above, information was entered by coding staff on to the main data file. Only information from the main interview questionnaire was entered at this stage.

      Prior to the merging of the newly entered data on to the master data file, manual and electronic edit checks were applied to ensure that the data entry carried out was syntactically correct and that all fields requiring a code had one allocated.


      Edit checks

      Further computer edits were applied to each record at original data entry and in subsequent computer processing. 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 this survey. The edits were also designed to detect errors which may have occurred during processing (e.g. during data entry, clerical coding) and to identify cases which although not necessarily errors, were sufficiently unusual or close to agreed limits as to warrant examination.

      Periodically throughout computer processing, the data were output to spreadsheets, frequency counts and tables containing cross-classifications of selected data items for checking purposes. These were aimed at identifying any problems in the input data, which had not previously been identified, errors in derivations and other inconsistencies between related items. In the final stages of processing additional output editing was undertaken to ensure that estimates conformed to known or expected patterns and were broadly consistent with data from previous NHSs or from other (including external) data sources, allowing for methodological and other factors which might impact comparability.

      Any errors detected in the data or derivations were checked and subsequently corrected where possible. While all reasonable care was taken to ensure the data are correct some errors may survive to the final data file; further information about data quality issues are contained in Chapter 7: Data Quality and Interpretation of Results.

      Data available from the survey are essentially ‘as reported’ by respondents. The processing procedures and edit checks outlined above were designed primarily to minimise errors occurring during processing. In some cases it was possible to correct errors or inconsistencies in the data which was originally recorded through reference to other data in the record; in other cases this was not possible and some apparent errors and inconsistencies remain on the data file.


      Edit checks - 2001 NHS(I)

      Lengthy, in-depth statistical validation was conducted on the output file. This involved internal consistency checks, examining rates and percentages, and comparing 2001 data with 1995 NHS(I) data for consistency and expected trends, examining the quality of data obtained in the 2001 NHS(I), particularly in sparsely settled areas, and checking the 2001 NHS(I) rates against comparable sources of Indigenous health data for consistency. Any apparent inconsistencies detected in the data or derivations were checked and subsequently corrected where appropriate. Particular attention was paid to the quality of data collected in sparsely settled areas. For a detailed explanation of additional quality assurance performed on this data refer to "Specific data quality issues for the sparse NHS(I)" in Chapter 7.


      Computer processing

      Information from the questionnaires 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 ‘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 National Health and Medical Research Council).

      Data from this survey are not available at household, family or income unit level. However some items were compiled at these levels (e.g. household composition and size), and are contained as individual characteristics on person records.

      A full listing of output data items is available on the ABS web site. Information concerning the SAS computer code underlying some items can be made available on request.

      Once processing and validation of the data were complete, factors or weights were 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.

      ESTIMATION PROCEDURES - 2001 NHS(G)

      Note: The principles discussed below relate to both the NHS(G) and NHS(I) but unless specifically identified all figures and examples presented relate to the NHS(G) sample only.

      Several sets of estimates are available from the 2001 NHS :
      • total Australian population estimates, which are based on the total sample of 26 863 Indigenous and non-Indigenous records from the 2001 NHS
      • non-Indigenous estimates which are based on the 26 380 non-Indigenous records from the 2001 NHS(G)
      • Indigenous estimates which are based on the total of Indigenous records from the 2001 NHS(G) plus all Indigenous records from the 2001 NHS(I) (a total of 3 681 records).

      This is illustrated below:

      TABLE 2.9: Estimation procedures

      From 2001 NHS(G) sample

      From 2001 NHS(I) sample

      Population
      benchmarks
      Output estimates
      Non-Indigenous
      records
      Indigenous
      records
      Indigenous
      non-sparse records
      Indigenous
      Sparse records

      2001 NHS(G)
      data file
      26,380
      483
      Indigenous & non-IndigenousTotal Australian population and
      Non-Indigenous population
      2001 NHS(I)
      Indigenous data file
      483
      2,244
      954
      IndigenousIndigenous population



      The following sections describe the estimation procedure for the 2001 NHS; separate procedures were followed for the production of estimates for the Indigenous population. Although separate Indigenous and non-Indigenous benchmarks are used in weighting the 2001 NHS(G), Indigenous estimates compiled from 2001 NHS(G) records only will not be released. This is because the sample of Indigenous people in the 2001 NHS(G) is very small, and a significant proportion of the Indigenous population did not have a chance of selection in the survey (those living in sparsely settled areas).

      Estimates from the 2001 NHS(G) are derived using a procedure which combines information collected in the course of the survey, in responses to the survey, and concerning the propensity of selected sample units to respond, with independently available information concerning the underlying populations. As a result, the 2001 NHS(G) estimates of population conform to population counts at broad age, sex, part of State (Capital city/rest of State), Indigenous status and State/Territory levels.


      Benchmarks

      Unlike previous NHSs which were benchmarked to population estimates averaged over the period of the survey, the 2001 NHS was benchmarked to the estimated population (adjusted for the scope of the survey) as at 30 June 2001. The total Indigenous sample from the 2001 NHS was benchmarked to the estimated Indigenous population at 30 June 2001. These estimates were based on the 2001 Census of Population and Housing.

      Person level benchmarks were used, excluding persons not resident in private dwellings. Conceptually as persons in sparsely settled areas did not have a chance of selection in the NHS(G) they should be removed from these population benchmarks. However this is difficult to do accurately and the benchmarks used for NHS(G) include persons resident in sparsely settled areas, except in NT. The effect on survey estimates from this is considered negligible as the relative proportion of the states' population resident in sparse areas is very small. In NT benchmarks excluding sparsely settled areas have been used for NHS(G) as persons resident in sparsely settled areas comprise a significant proportion of the Territory's population. The weighting methodology also required the use of household benchmarks for non-response adjustments - see below. 2001 Census-based household estimates were not available for use in time for this survey, and 1996-based estimates were used instead.


      Weighting specifications

      To obtain person-based estimates, expansion factors or 'weights' were inserted into responding persons' records to enable the data provided by these persons to be expanded to provide estimates relating to the whole population within the scope of the survey. The strategy for deriving person weights involved a number of steps as 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 i.e. in the NHS(G) the inverse of the relevant state sample fraction as noted in the sample design section above

      The initial household weight was then adjusted for the NHS(G) as described below.


      Adjustment for period of enumeration

      Note: This section does not apply to the NHS(I).

      As outlined previously in this Chapter, due to a higher than anticipated response rate and associated higher costs some sample was deselected in the 2001 NHS(G). As the deselection took place in the later stages of enumeration the sample was not evenly distributed between quarters as was originally intended.

      Traditionally the NHS is conducted over a 12 month period to account for seasonal variation. The aim of adjusting the initial weights for period of enumeration is to have each quarter represented equally, regardless of the relative size of the sample in that quarter. This affected each state differently. The following is a table showing the number of responding dwellings in each state, for each quarter.

      TABLE 2.10: Responding dwellings, by State/Territory and quarter

      Quarter
      NSW
      Vic
      Qld
      SA
      WA
      Tas
      NT
      ACT

      Number of records
      1
      1,252
      1,134
      945
      646
      702
      369
      72
      388
      2
      1,234
      1,201
      1,006
      660
      711
      355
      56
      351
      3
      1,214
      1,142
      1,014
      649
      660
      383
      64
      341
      4
      254
      235
      177
      137
      131
      59
      70
      357

      % of records in each State
      1
      31.7
      30.6
      30.1
      30.9
      31.9
      31.7
      27.5
      24.4
      2
      31.2
      32.4
      32.0
      31.6
      32.3
      30.5
      21.4
      25.3
      3
      30.7
      30.8
      32.3
      31.0
      30.0
      32.9
      24.4
      24.6
      4
      6.4
      6.3
      5.6
      6.6
      5.9
      5.1
      26.7
      25.7



      For all the states but ACT and NT (where the sample was quarantined from deselection) the sample in the fourth quarter is much smaller than for previous quarters. To prevent an exaggerated period of enumeration adjustment from using four adjustment periods, the fourth quarter sample was amalgamated with that of the third quarter. The period of enumeration adjustment was then set such that the sample from the first two quarters each contributed 30% of the total weight, and the amalgamated sample from the third and fourth quarter contributed 40% of the total weight. The adjustment factor for quarters 1 and 2 was calculated as 0.3*(total sample for the state)/(quarter sample the for state), and for the amalgamated third and fourth quarters the factor was calculated as 0.4*(total sample for the state)/(quarters 3 and 4 sample for the state). For ACT and NT the adjustment was set such that each quarter contributed an equal weight. The adjustment factor for each quarter was calculated as 0.25*(total sample for territory)/(quarter sample for territory).

      The adjustment for period of enumeration is therefore:

      adjusted weight = (initial weight due to state skip)*(adjustment factor relevant to state, quarter)


      Adjustment for Household Composition

      Note: This section does not apply to the NHS(I).

      The aim of this adjustment was to counter non-response biases in target variables which are affected by household composition. The adjustment involved calibrating the adjusted initial weights to a set of benchmarks incorporating a household composition dimension. The calibration redresses under or over representation of households based on their composition resulting from both the random nature of sampling and/or systematic differential non-response patterns.
      The household benchmarks used were estimates as at 30 June 2001 projected from the 1996 Census. The benchmarks used for 2001 NHS excluded sparsely settled areas of NT . Two levels of household benchmarks were used in the calibration:
      • state by household composition; and
      • state by part of state (capital city/rest of state)

      Household composition classes were defined by combinations of the number of adults (one, two, and three or more) and children (none, one, and two or more) in the household. The calibration process involved adjusting the input household weights as little as possible such that they aggregated to the marginal household benchmark totals as specified above. The weights were not calibrated to agree with benchmarks at the state by part of state by household composition level. Benchmarks were not specified at this level as the small number of responding households in some of the cells would have led to unstable weights.


      Initial Person Weights

      The next step in the procedure was assigning initial weights to fully responding persons. This involved taking into account the survey the respondent was selected in i.e. NHS(I) or NHS(G), and the subsampling scheme deployed within households. In NHS(G) for example, if one adult, one child aged 7-17 years, and all children aged 6 years and younger were enumerated in selected households. Initial person weights were calculated by inflating the person's adjusted household weight by the probability of the person being selected. For persons aged 6 years or younger the person weight was equal to the household weight. For persons aged 7-17 years the household weight was multiplied by the number of residents in the household aged 7-17 years. Similarly for persons aged 18 years and over.


      Calibration to Person Level Benchmarks

      The final step in the weighting procedure was calibrating the initial person weights to a set of person level benchmarks. The calibration to benchmarks ensures that the sample survey estimates agree exactly with independent measures of the population at specific levels of disaggregation. In addition, the calibration reduces the impact of differential non-response bias at the specific levels of disaggregation, and also reduces sampling error.

      The person benchmarks used in the NHS were estimates for June 2001 projected back from the 2001 Census of Population and Housing results.

      The benchmarks used for NHS(G) excluded sparsely settled areas of NT. Two levels of person benchmarks were used in the calibration for the NHS(G):
      • state by part of state by age group by sex; and
      • state by Indigenous status.

      Two levels of Indigenous person benchmarks were used in the calibration:
        • state by sparse/non-sparse; and
        • age group by sex.

        The age group categories in the NHS(G) were set such that there was a reasonable sample size in each weighting class at the state by part of state by age by sex level. The age categories adopted are shown below:


        NTOther states


        0-40-4
        5-145-9
        15-3410-14
        35-5415-19
        55 and over20-24
        25-29
        30-34
        35-39
        40-44
        45-49
        50-54
        55-59
        60-64
        65-69
        70-74
        75 and over


        The age group categories in the NHS(I) were set such that there was a reasonable sample size in each weighting class at the age by sex level. The age categories adopted are shown below:

        0-4
        5-9
        10-14
        15-19
        20-24
        25-29
        30-34
        35-39
        40-44
        45-54
        55 and over

        The calibration process involved adjusting the input person weights as little as possible such that they aggregated to the marginal person level benchmark totals as specified above. The use of the Indigenous benchmarks was to enable improved estimates for the non-Indigenous population, which was a key component of the weighting strategy for the 2001 NHS(G).


        Age-standardisation - 2001 NHS(I)

        Age standardisation techniques have been used in some of the tables in the publication National Health Survey: Aboriginal and Torres Strait Islander Results, Australia (cat. no. 4715.0) to remove the effect of the differing age structures in the Indigenous and non-Indigenous populations for 2001, and over time. The age structure of the Indigenous population is considerably younger than that of the non-Indigenous population on average, and age is strongly related to many health measures. Therefore, estimates of prevalence which do not take account of age may be misleading. The age standardised estimate of prevalence is that which would have prevailed should the Indigenous and non-Indigenous populations have the standard age composition. In publication tables, the standard population used for age standardising was the total Australian population at 30 June 2001 based on the 2001 Census of Population and Housing, adjusted for the scope of the survey.

        The direct age standardisation method was used. The formula is as follows:
        Formula

        where Cdirect = the age standardised estimate of prevalence for the population of interest,
        a = the age categories that have been used in the age standardisation,
        ca = the estimate of prevalence for the population being standardised in age category a, and
        Psa = the proportion of the standard population in age category a.

        The age categories used in the standardisation for this publication were 5 year age groups to age 44, then 45-54, and 55+ years of age.


        Chapter 1 - Introduction

        Chapter 2 - Survey Design and Operation

        Chapter 3 - Health Status Indicators

        Chapter 4 - Health Related Actions

        Chapter 5 - Health Risk Factors

        Chapter 6 - Population Characteristics

        Chapter 7 - Data Quality and Interpretation of results

        Chapter 8 - Data Output and Dissemination
        Appendix 1 - Glossary of Terms Used

        Appendix 2 - Sample Counts and Weighted Estimates

        Appendix 3 - Classification of Long-term Medical Conditions: Based on ICD-10

        Appendix 4 - Classification of Long-term Medical Conditions: Based on ICD-9

        Appendix 5 - Classification of Long-term Medical Conditions: ICPC Based

        Appendix 6 - Classification of Type of Medication

        Appendix 7 - Classification of Country of Birth

        Appendix 8 - Classification of Language Spoken at Home
        Appendix 9 - Classification of Occupation

        Appendix 10 - Classification of Industry of Employment

        Appendix 11 - Classification of Types of Alcoholic Drinks

        Appendix 12 - Standard Errors

        Appendix 13 - Content of the 2001 National Health Survey (Indigenous)

        Appendix 14 - List of Abbreviations




        Previous PageNext Page