|Page tools: Print Page Print All|
5 The scope excluded:
6 The sample frame for the SHC was the Medicare Enrolment Database (MEDB). The sample was selected from this frame by the Department of Human Services (DHS) in accordance with a stratification and allocation specified by the ABS.
7 As people were in scope of the SHC if they saw a GP at least once in the 12 months prior to selection (23 November 2015), there may have been people who saw a GP at least once in the 12 months prior to enumeration (April 2016 to June 2016) who were not in scope as they did not visit a GP between November 2014 and November 2015. Similarly, there may have been people who visited a GP in the 12 months prior to selection but did not visit a GP in the 12 months prior to enumeration who were in scope.
8 The SHC sample was designed to support estimates at the Primary Health Network (PHN) area level. A stratified random sample was used where the strata were based on the following variables:
9 People on the frame were assigned a PHN based on the postcode of their postal address as recorded on the MEDB. A correspondence between postal areas and PHNs was used to do this. As not every postcode is included in the ABS postal area classification, there were around 100,000 people who could not be allocated a PHN. At the sample design stage, these people were allocated to an unknown PHN category.
10 There were 8.8 million people in scope on the MEDB. A total sample of around 124,000 people was selected by sorting within stratum by number of GP visits and then applying a skip using a random start.
11 Also incorporated into the overall sample design was a requirement to oversample people who saw a GP more than 12 times such that the resulting sample consisted of approximately half people who saw a GP 12 or more times and half who saw a GP 1-11 times.
12 There were around 124,000 people selected for this survey. Of these, there were 35,495 responses, giving a response rate of 28.6%. In this survey it is not possible to distinguish between non-response and sample loss. For example, a person may have been selected to participate, but will not have received any survey materials due to an out-of-date address on the MEDB.
13 The following table contains response rates by the State or Territory that the person was selected in. Persons selected in the unknown PHN category have unknown State for selection.
Table 1.1 RESPONSE RATES
14 In the survey output, respondents are placed into the geographic regions (e.g. State, PHN) and SEIFA decile that correspond to their reported home postcode.
15 Data was collected by mail. In order to facilitate maximum response, a four stage mail-out approach was used. The four stages consisted of:
16 In each phase of the mail out, a cover letter from the DHS was included, explaining that the DHS had not provided the ABS with any personal details of the selected person.
17 People with low English proficiency, or who had a disability which prevented them from completing the survey on their own, were able to complete the survey over the phone. People with low English proficiency were offered the option of an interpreter from the Translation and Interpreting Service (TIS National) who could facilitate a phone call with the ABS and translate as an ABS officer provided information or collected the participant's data over the phone.
WEIGHTING, BENCHMARKS AND ESTIMATION
18 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 enumerated person. The weight is a value which indicates the number of people in the population represented by the sample person.
19 The first step in calculating weights for each unit is to assign an initial weight, which is the inverse of the probability of being selected in the survey. For example, if the probability of a person being selected in the survey was 1 in 600, then the person would have an initial weight of 600 (that is, they represent 600 people).
20 The probability of being selected was based on the stratification variables: number of GP visits (1-11 or 12 or more), PHN, ‘low’, ‘medium’ or ‘high’ IRSAD, sex and age group (five-year groups from 45 to 79 years of age, then 80 years and over). Persons in the number of GP visits (12 or more) had a higher chance of selection than those with 1-11 GP visits.
21 In order to determine the stratum of selection, self-reported values were used for PHN, SEIFA, sex and age as the ABS does not have access to any personal information from the frame. While self-reported frequency of GP use was also available, analysis of data for persons who consented to have their MBS data released for the study indicated that this was not a reliable indicator of the frequency of GP use (derived from MBS claims) as was used for sample selection. For this reason, the, the initial weight was calculated in the following ways:
22 For weighting purposes, the PHN and SEIFA allocated to each record were based on the postcode of the respondent’s reported address, which in some cases differed from the postcode, PHN and SEIFA used in selection.
23 Weights calibrated against population benchmarks ensure that the survey estimates conform to the distribution of the MEDB population rather than the distribution within the sample itself. Calibration to population benchmarks helps to compensate for over or under-enumeration of particular categories of people/households which may occur due to either the random nature of sampling or non-response.
24 The survey was benchmarked to counts of the in-scope population at November 2015 from the MEDB for PHN (based on postal address postcode) by sex by 10 year age groups to age 75 and over.
25 Survey estimates of counts of people are obtained by summing the weights of people with the characteristic of interest.
26 The Census and Statistics Act, 1905 provides the authority for the ABS to collect statistical information, and requires that statistical output shall not be published or disseminated in a manner that is likely to enable the identification of a particular person or organisation. The requirement means that the ABS must take care and make assurances that any statistical information about individual respondents cannot be derived from published data.
27 Perturbation is used in this publication to minimise the risk of identifying individuals in aggregate statistics. Perturbation involves a small random adjustment of the statistics and is considered the most satisfactory technique for avoiding the release of identifiable statistics while maximising the range of information that can be released. These adjustments have a negligible impact on the underlying pattern of the statistics. After perturbation, a given published cell value will be consistent across all tables. However, adding up cell values to derive a total will not necessarily give the same result as published totals.
RELIABILITY OF ESTIMATES
28 All sample surveys are subject to error which can be broadly categorised as either:
29 Sampling error is the difference between the published estimate, derived from a sample of people, and the value that would have been produced if all people in scope of the survey had been included. For more information refer to the Technical Note.
30 In this publication, estimates with an RSE of 25% to 50% are preceded by an asterisk (e.g. *3.4) to indicate that the estimate has a high level of sampling error relative to the size of the estimate, and should be used with caution. Estimates with an RSE over 50% are indicated by a double asterisk (e.g. **0.6) and are generally considered too unreliable for most purposes.
31 Margins of Error are provided for proportions to assist users in assessing the reliability of these data. The proportion combined with the MoE defines a range which is expected to include the true population value with a given level of confidence. This is known as the confidence interval. This range should be considered by users to inform decisions based on the proportion. Proportions with an MoE of greater than 10 percentage points are preceded by a hash (e.g. #40.1) to indicate the range in which the true population value is expected is relatively wide.
32 Non-sampling error may occur in any collection, whether it is based on a sample or a full count of the population such as a census. Sources of non-sampling error include: non-response; errors in reporting by respondents or recording of answers by interviewers; and errors in coding and processing data. Every effort was made to reduce the non-sampling error by: careful design and testing of the questionnaire; follow-up of respondents; and extensive editing and quality control procedures at all stages of data processing.
33 Non-response bias occurs where non-respondents may have different characteristics from those who did respond. While the collection is subject to non-response bias, it is not possible to reliably quantify this. The magnitude of any bias depends on the rate of non-response and the extent of the differences in characteristics between those people who responded to the survey and those who did not. See Table 1.2 below for a comparison with other ABS collections.
34 This collection was undertaken in a different manner from other ABS household surveys as it was a voluntary, mail-based collection with follow-up by mail. The collection achieved a response rate of 29%.
Comparability with other ABS surveys
35 As this is the first time the SHC has been collected, there is no time series data available.
36 Similar concepts to those in the SHC are found in other ABS surveys, for example the Patient Experience in Australia (cat. no. 4839.0) and the National Health Survey (cat. no. 4364.0.55.001). However, comparison with other ABS surveys should be undertaken with caution. There are several reasons why data from the SHC may not be comparable with other ABS collections:
37 Given the response rate, an investigation was carried out in order to try to understand whether the sample is representative of the in-scope population. While there are some differences between the sample distribution from the Survey of Health Care and other surveys by age, sex, SEIFA and PHN, these have been taken into account by the weighting process. The table below gives a comparison of proportions for a range of other variables from SHC, Patient Experience 2015-16 (PEx) and the National Health Survey 2014-15 (NHS) for the same scope (i.e. over 45 years and at least one GP visit in the previous 12 months). For ‘level of education’ and whether the person ‘had private health insurance’, the proportions were similar across the different sources. For some health related variables, the SHC generally had a higher proportion of people with poorer health. No adjustment to the weighting has been made for this potential bias.
Table 1.2 COMPARISON BETWEEN SHC, PEX AND NHS, PERSONS AGED 45 YEARS AND OVER WHO HAD AT LEAST ONE GP VISIT IN THE PREVIOUS 12 MONTHS(a)
Cells in this table have been randomly adjusted to avoid the release of confidential data.
(a) There are differences in the scope of the SHC and the PEx and NHS collections which must be considered when interpreting comparisons in this table. In particular, the SHC includes persons who live in non-private dwellings such as hospitals and nursing homes which are excluded from PEx and NHS. In addition, the NHS excludes persons in areas classified as Very Remote, and both PEx and NHS exclude those in discrete Aboriginal and Torres Strait Islander communities.
NHS and PEx data in this table are restricted to persons aged 45 years and over and who reported having at least one GP visit in the 12 months. SHC data are restricted to persons aged 45 years and over and who reported having at least one GP visit in the 12 months prior to being selected.
More detailed information about the methodology for these two collections can be found in the NHS Explanatory Notes (cat. no. 4364.0.55.001) and PEx Explanatory Notes (cat. no. 4839.0).
(b) There were differences in the methods used to collect long term health conditions between the collections. Both PEx and SHC used a tick box question, PEx with eight and SHC with 14 options. The NHS uses a much more extensive set of questions and is not comparable.
(c) Language other than English spoken at home is not collected in PEx.
Interpretation of results
38 There were a number of issues unique to this collection which may affect the interpretation of the results. These include:
Editing of survey data
39 With the mode of delivery being a paper form survey, there was some input editing required. Below are some examples.
40 Geographic classifications were applied to the survey data based on the respondent’s reported home postcode, using correspondences between the geography of interest and ABS Postal Area geography.
41 Standard ABS Geographies were classified according to the Australian Statistical Geography Standard (ASGS): Volume 1 - Main Structure and Greater Capital City Statistical Areas, July 2011 (cat. no. 1270.0.55.001).
42 Remoteness areas are classified according to the Australian Statistical Geography Standard (ASGS): Volume 5 - Remoteness Structure, July 2011 (cat. no. 1270.0.055.005).
43 Primary Health Networks (PHNs) are a classification developed by the Department of Health, see Primary Health Networks in the Glossary. The correspondence between PHN and ABS Postal Area geography was used to relate a person’s postcode as listed on the MEDB to a PHN.
44 Where a postcode crossed a PHN boundary the entire postcode was allocated to the PHN with largest proportion of people living in it. There was a slight exception where a postcode crossed a state boundary; in this case individuals were manually coded to the state they reported as their address.
Socio-economic Indexes for Areas (SEIFA)
45 The survey uses the 2011 Socio-economic Indexes for Areas (SEIFA).
46 SEIFA is a suite of four summary measures that have been created from 2011 Census information. Each index summarises a different aspect of the socio-economic conditions of people living in an area. The indexes provide more general measures of socio-economic status than is given by measures such as income or unemployment alone.
47 For each index, every geographic area in Australia is given a SEIFA number which shows how disadvantaged that area is compared with other areas in Australia.
48 The Survey of Health Care publication uses the Index of Relative Socio-Economic Advantage and Disadvantage, derived from Census variables related to advantage and disadvantage such as high and low household income, lower educational attainment, unemployment, jobs in relatively skilled or unskilled occupations and dwellings without motor vehicles.
49 SEIFA uses a broad definition of relative socio-economic disadvantage in terms of people's access to material and social resources, and their ability to participate in society. While SEIFA represents an average of all people living in an area, it does not represent the individual situation of each person. Larger areas are more likely to have greater diversity of people and households.
50 For more detail, see the following:
PRODUCTS AND SERVICES
51 Data Cubes containing all tables in Excel spreadsheet format can be found on the ABS website (from the Downloads tab). The spreadsheets present tables of estimates and proportions, and their corresponding relative standard errors (RSEs) and margin of error (MoE).
Customised data requests
52 Special tabulations of the data are available on request. Subject to confidentiality and sampling variability constraints, tabulations can be produced from the survey incorporating data items, populations and geographic areas (including state and territory level data), tailored to individual requirements. These are provided in electronic form. A list of data items from the 2016 Survey of Health Care is available from the Downloads tab.
54 ABS surveys draw extensively on information provided by individuals, businesses, governments and other organisations. Their continued cooperation is very much appreciated and without it, the wide range of statistics published by the ABS would not be available. Information received by the ABS is treated in strict confidence as required by the Census and Statistics Act 1905.
55 The survey and study were funded by the AIHW, and conducted by the ABS. This publication was jointly prepared and released by the ABS and the AIHW.
56 The ABS and AIHW also acknowledge and thank the DHS for its assistance in the sample selection and postage process of the study.
57 The SHC used variations of questions sourced from other national and international non-ABS surveys. The ABS and AIHW would like to acknowledge the following organisations:
58 Specific questions in the SHC that were based on questions from these source organisations are listed in the table below:
Table 1.3 SOURCE ORGANISATIONS AND SOURCE QUESTIONS USED IN SURVEY OF HEALTH CARE SOURCE ORGANISATION
These documents will be presented in a new window.