Health service use in the year before and after intentional self-harm among ex-serving ADF members

The previous sections described overall service use, including the proportion of ex-serving members accessing each service, the types of services used, and average number of health services used. However, those analyses considered each service type in isolation and did not show how individuals used several services in combination or whether distinct patterns of engagement existed across the cohort.

This section reports the results from a modelling technique called latent class analysis (LCA). This was conducted to identify subgroups (or groups) of ex-serving members who were admitted for self-harm based on health service use in the year before and after the admission. This data-driven analytical approach groups individuals based on their patterns of engagement in the year before and after across indicators of health service use, derived from hospital (ED and admitted care), PBS/RPBS and MBS data. The six indicators used in analysis included GP attendances, specialist attendances, allied health attendances, ED presentation, hospital admission and PBS/RPBS services.

Understanding these patterns provides insights into how ex-serving members interacted with the health system before and after an intentional self-harm admission, helping to identify potential opportunities for intervention across different service settings.

The LCA identified three groups of health service users, based on the frequency and pattern of service use across all six indicators. These groups were identified separately for the year before and the year after the self-harm admission and were labelled as low, medium, and high users. The groups were derived by the data-driven grouping of individuals with similar patterns and intensity of service use rather than through allocation based on any specific thresholds for each group.

Both the self-harm cohort and comparison cohort had three groups. The groups of ex-serving members in the self-harm cohort were:

• Low users: ex-serving members who on average had lower use of health services. In the year before 16% were in this group and in the year after this proportion had decreased to 10%. As an example, people in this group had average GP service use of 3.4 in the year before and 2.1 in the year after self-harm.
• Medium users: ex-serving members who had moderate use of health services. In the year before 65% were in this group and in the year after this proportion had decreased to 61%. As an example, people in this group had average GP service use of 10.0 in the year before and 10.1 in the year after self-harm.
• High users: ex-serving members who on average had higher use of health services. In the year before 19% were in this group and in the year after this proportion had increased to 29%. As an example, people in this group had average GP service use of 21.9 in the year before and 17.6 in the year after self-harm.

The distribution of ex-serving members in the comparison cohort groups was stable in the year before and after. In comparison to the self-harm groups, there was a greater proportion of ex-serving members in the low user group and smaller proportion in the high user group. For more information see Supplementary table S6.1 (see data).

Relative to the self-harm cohort a much smaller proportion had accessed a mental health service in each group. For example, in the year before admission, 44% of those in the self-harm cohort had accessed a mental health service compared with only 4% in the comparison cohort. This pattern was the same for the medium user group (93% in the self-harm compared with 42%) and high user groups (98% compared with 75%).

The frequency of health service use by service type is shown in Table 3 for ex-serving members in the self-harm cohort in the year before admission. Across the analysed health services in the self-harm cohort, the combined volume of health service use in the medium user group is more than 7 times higher than the low user group and the high user group is 2.5 times higher than the medium user group.

A significant proportion of the health service use of each group was due to PBS services. For low users, this was 31% in the year before and 63% for the medium and high users. Low users had high GP use (57%) relative to their overall service use. These patterns were broadly similar for the comparison cohort.

In summary, the LCA revealed distinct subgroups of ex-serving members based on health service use volume before and after a self-harm admission. The shift toward higher service use after the admission, particularly in mental health services, underscores the importance of targeted, coordinated care strategies that address complex service needs across the continuum of care.

Source: AIHW Veterans Health Dataset (VHD), July 2010–June 2020

Which ex-serving members are using more health services?

Analysing the demographic and service-related characteristics of the three health service use subgroups provides information on who is more likely to use more health services. In the year before, persons in the self-harm cohort who were high health service users were more likely to be female, aged 55 years and over, Officer rank, or have one or more comorbidities. These were similar in the comparison cohort except for Officer rank (Figure 7).

Persons in the self-harm cohort who were low users were more likely to be aged 25 to 34 years, have served in the Navy, separated involuntarily for retention not in service interest or other involuntary reasons, separated less than a year ago or had no comorbidities. These were similar to the comparison cohort except for serving in the Navy (Figure 7).