As noted, the Timor-Leste Family Study used a retrospective, cross-sectional study design to compare the physical, mental and family health of families of ADF members who deployed to Timor-Leste with those attributes of a comparison group of families of ADF members who did not deploy to Timor Leste.
The Centre for Military and Veterans’ Health developed nominal rolls (listings of names, demographic details, and contact details) for all ADF members who deployed to Timor-Leste between 1999 and 2010 and all members who did not deploy to Timor-Leste in the same period. Individuals on the comparison nominal roll were frequency matched to those on the Timor-Leste roll by sex and Service.
The study team selected members who were listed as being in a relationship and aimed to randomly sample 4,000 members from each roll to create the Timor Leste sample and the comparison sample. The random sample, however, incorporated both proportional and oversampling of certain groups. Proportional sampling occurred in the case of participants in another study, the Centre for Military and Veterans’ Health Military Health Outcomes Program, which was conducted at the same time as the Timor-Leste Family Study.
The Military Health Outcomes Program
MilHOP is a Defence-funded program of studies examining the health and wellbeing of serving and ex-serving ADF members. The aim is to learn about the types of health problems and related symptoms that are relevant to ADF members in order that Defence can better respond to such problems in the future. MilHOP takes in the Health and Wellbeing Study, the Middle East Area of Operations (MEAO) Health Study, the MEAO Prospective Study, and the MEAO Health Study: Mortality and Cancer Incidence Studies.
In an effort to avoid overburdening ADF members with studies, the Centre for Military and Veterans’ Health decided, in consultation with the Departments of Veterans’ Affairs and Defence, to create links between the participants in the Timor-Leste Family Study and those in MilHOP. This involved ADF members who were part of the two studies being able to consent to the following:
linking of their MilHOP data with the Timor-Leste Family Study so that their participation in the Family Study involved completing only 10 questions
allowing CMVH to use the nominal roll contact details of their partner to invite their partner to participate in the Family Study. (Partner contact details provided to Defence may be used only with permission of the ADF member; this permission allowed CMVH to contact partners directly.)
Because the MilHOP ‘consenters’ were more likely to be currently serving permanent members, proportional sampling also occurred for serving and ex serving members so as to reflect the actual size of the serving and ex-serving Timor-Leste veteran populations (75 per cent serving; 25 per cent ex-serving).
Female ADF members and members from the Royal Australian Navy and the Royal Australian Air Force were oversampled because these groups are small compared with male ADF members and numbers in the Australian Army respectively. Compared with the rest of the military population, Army males were more frequently deployed to Timor-Leste. Oversampling from these groups allowed sufficient power to detect any differences in the analyses that were based on either sex or Service.
The target of 4,000 ADF members per sample was not reached because there were insufficient numbers in the groups that were oversampled—for example, female RAAF officers who had deployed to Timor-Leste.
The Timor-Leste Family Study’s Human Research Ethics Committees granted approval for the study to contact the partners of ADF members and invite them to participate if the ADF member agreed. CMVH obtained ADF member agreement through consent forms. For MilHOP respondents who consented to partner contact, the Family Study team mostly had partner contact details from the nominal rolls. For non-MilHOP respondents (and MilHOP respondents whose partner contact details were not in the nominal rolls), the team obtained partner contact details by asking the ADF member to provide their partner’s details on their Family Study consent form if they wished to. Figure 3.1 summarises the sampling process.
The target of 4,000 ADF members in each of the Timor-Leste and comparison samples was based on an assumption of the minimum participation required by partners and ADF members (25 per cent) and a sample size requirement of 1,000 ADF member–partner pairs. This sample size was calculated to have adequate statistical power to detect a range of differences in health outcomes between the two equal-sized groups using population baseline health outcome for small and large differences in outcomes between the Timor-Leste and comparison partners. For example, the 2004–05 National Health Survey (Australian Bureau of Statistics 2006) found that nine per cent of females aged 25–44 years reported fair or poor general health. Assuming that comparison partners’ reported health was similar to the ABS national findings and the study achieved a 25 per cent participation rate in both partner groups, the study would have 98 per cent power (strong) to detect an absolute difference of 6 per cent between Timor-Leste and comparison partners.