The purpose of this project is to show the feasibility of linking two community service program data sets, and to provide information about the characteristics and service use of the users of both programs (at the national level) for policy makers involved in developing and monitoring needs-based services. The two government programs chosen for this purpose were Disability Services (DS) and Home and Community Care (HACC), as both have a national minimum data set describing the client population and services provided. It is also expected that a group of clients would access both services.
The purpose of this paper is to describe the linkage process used to identify the extent of joint DS and HACC program use, and the methods used to derive data items for analysis. A detailed discussion of the DS and HACC programs and analysis of the linked data are presented in the companion AIHW report People using both Disability Services and Home and Community Care 2010-11 (AIHW 2014), released with this report.
In both the DS and HACC data collections, clients are identified via the statistical linkage key SLK-581 which consists of five letters of name, full date of birth, and sex. While SLK-581 generally distinguishes well between individuals, it is possible for two individuals to have the same key. Hence, use of SLK-581 protects the privacy of individuals as they cannot be re-identified from the key. Data linkage between DS and HACC clients was undertaken using key-based linkage centred around SLK-581. This method maximises the value of the SLK-581 for linkage and has been used previously in a number of projects.
In 2010-11, there were 314,000 SLK-581 person identifiers in the DS data collection and 934,000 in that for HACC. Missing components for SLK-581 can mean that there is insufficient data for data linkage. Just over 8% of DS person identifiers and 6% of HACC person identifiers had insufficient or unreliable data for matching. As a consequence, 288,000 DS clients and 874,000 HACC clients were included in the key-based linkage process.
Overall, there were 55,000 matches between DS and HACC clients. These accounted for 17.5% of all DS clients and 19.0% of matchable DS clients (that is, those with sufficient information to be included in the linkage process). Among the much larger set of HACC clients, 5.9% of all HACC clients and 6.3% of matchable HACC clients matched to DS clients. As expected from the different target groups of the two programs, HACC clients who were also DS clients tended to be younger than other clients. On the other hand, the relatively small group of DS clients aged 65 and over were more likely than others to be accessing both services.
When records are matched by SLK-581, they are assumed to relate to the same service user. In the majority of cases, demographic information on these records will be the same. However, even within data sets, in some cases some information on two or more matching records may be inconsistent. Hence, a range of edits were carried out to ensure consistency of client demographic data within the two data sets before matching and analysis.
Just as data collected on client characteristics at different times for the same data collection can vary, so, too, can information collected on client characteristics for different data collections. Hence, differences in distributions derived for people accessing both DS and HACC services-using information from the DS NMDS and HACC MDS, respectively- may point to problems in collecting and recording some data items consistently. Indigenous status appears to be one such variable in the current analysis.