In recent years, Western countries have introduced stringent measures to restrict the inflow of asylum seekers.1 From 1992 to 2005, Australia implemented a policy of mandatory detention of all asylum seekers who arrived by boat or without valid travel documents. Mental health professionals have raised concerns that prolonged detention can re-traumatise people fleeing persecution and abuse,1 issues that have been reiterated by successive Commissions of Inquiry,2-4 and by clinicians involved with detainees’ welfare.5-7 Until recently, proponents of the detention policy argued variously that detainees were not mentally ill, that any existing mental illness was related to pre-existing problems, and/or that the psychological care provided in detention centres was adequate.3,8 Calls by bodies such as the Australian Medical Association for independent research into detainees’ mental health have gone unheeded.9
The few studies conducted of detainees’ mental health have revealed elevated levels of psychiatric disturbance,10 with length of confinement appearing to be associated with a progressive deterioration in mental state.6,11,12 A study of detained asylum seekers in the United States suggested that release from detention with secure residency improved mental health status, at least in the short term.13 The security and certainty associated with obtaining a permanent visa may be pivotal in gradually reducing refugee-related symptoms of anxiety, depression and post-traumatic stress.14
About 90% of detained asylum seekers in Australia have been recognised as refugees after rigorous assessments based on the United Nations Convention on Refugees. However, since 1999, on release from detention, these refugees have been given temporary protection visas (TPVs), not permanent protection visas (PPVs) with the full entitlements of Australian citizens (Box 1).
Since 1993, Australia has hosted about 20 000 settler-arrivals from Iran and Afghanistan.15 These countries have been affected by political instability and conflict for over 20 years, with widespread human rights abuses occurring. Large numbers of citizens have fled. Iran and Afghanistan share strong historical and cultural links, both speaking Persian dialects.
The sample was recruited consecutively from the Early Intervention Program of the Service for the Treatment and Rehabilitation of Torture and Trauma Survivors (STARTTS) in Sydney, New South Wales. Resettlement agencies in NSW are required to refer recent refugees (both TPV and PPV holders) to the program, irrespective of their mental status or level of exposure to past trauma. The program provides a comprehensive assessment of the resettlement, general and mental health needs of refugees, offering a range of services including referral to other agencies.
During the study period (2002–03), about 48% of all off-shore humanitarian entrants (PPV holders) resettled in New South Wales attended the STARTTS Early Intervention Program. Estimates from STARTTS and community organisations indicate that most TPV holders (in excess of 70%) attended the Early Intervention Program.
One of us (S M), a Persian-speaking psychologist familiar with the relevant dialects, approached consecutive adults referred to the program (or the first adult to attend if a whole family was referred). Four people were excluded, as they intended to move interstate, leaving 116 consecutive participants. Study recruitment was terminated during 2003 after it become apparent that most of the Persian speakers eligible for TPVs had been released from detention.
The Human Research Ethics Committee of the then South Western Sydney Area Health Service approved the study. Participants were provided with an information sheet in the appropriate dialect detailing the aims of the study, and all signed consent forms.
We applied two widely used symptom measures: the Harvard Trauma Questionnaire,16 which assesses exposure to trauma and symptoms of post-traumatic stress disorder (PTSD) in refugee populations; and the Hopkins Symptom Checklist-25,17 a measure of depression and anxiety. Continuous scores were used for both measures to avoid applying case threshold scores (these have not been established for these ethnic groups). We also used the General Health Questionnaire (GHQ-30)18 as an overall measure of distress, and the Medical Outcomes Study Short Form (SF-12) mental and physical component summaries.19
To assess recent adverse life experiences, two previously used purpose-designed measures were applied: the Postmigration Living Difficulties Checklist,20,21 whose items have consistently discriminated between asylum seekers and refugees with secure and insecure status;20,21 and the Detention Experiences Checklist, which examines 64 adverse experiences of confinement and detention, with previous research supporting its criterion validity.11,22
All measures were translated using established translation and blind back-translation procedures.23 The original and back-translated versions were assessed for meaning equivalence, and minor discrepancies were reconciled by two Persian-speaking psychologists.
Univariate comparisons are presented with statistical contrasts using two-sample t-tests and Pearson χ2 tests. Fisher’s exact test was used to investigate differences for individual items of the Harvard Trauma Questionnaire and the Postmigration Living Difficulties Checklist, with a Bonferroni adjustment for the overall type I error rate (Harvard Trauma Questionnaire, 0.05/16 = 0.003; Postmigration Living Difficulties Checklist, 0.05/23 = 0.0022).24
At Step 2, the impact of postmigration variables was examined using two sets of models. In the first, the effect of TPV status was calculated, adjusting for significant variables from Step 1. As TPV holders were the only group with a history of immigration detention in Australia, this model effectively examined the combined impact of detention and TPV status, adjusting for significant demographic and premigration predictors from Step 1. In the second set of linear regression analyses, TPV status was replaced by three indices wholly or predominantly characteristic of that group: the total number of negative detention experiences, the total number of current living difficulties, and whether or not respondents were separated from their spouse as a consequence of being in Australia. These stresses showed almost perfect multicollinearity with TPV status, indicating their relevance to that group, but also precluding the entry of TPV status into the equation. A stepwise forward selection procedure was used to identify whether any of the three indices were independently associated with psychiatric status and disability after adjusting for the effects of age, sex and premigration trauma.
Standardised β coefficients (mean = 0; SD = 1) are given. These provide an estimate of the magnitude of the unique association of each independent predictor (holding constant the effects of the other independent variables in the model) on the dependent or outcome variable. The percentage of variance explained (R2) is presented as a measure of the overall strength of the multivariate association between the predictor variables and the dependent variable. Regression diagnostics were examined to ensure that the standardised residuals were bivariate normal and that the condition index for independent variables was < 15, indicating acceptable multicollinearity between the independent variables.25
Box 2 presents the demographic data for the study sample. The refugees with PPVs and TPVs had lived in the Australian community for a short time (a mean of about 3 months), but TPV holders had previously been held in immigration detention for an average of 12 months. Males predominated among TPV holders, typical of the TPV population in Australia. A substantially larger number of male TPV holders were geographically separated from their spouse.
Levels of premigration trauma based on the Harvard Trauma Questionnaire were not significantly different between TPV and PPV holders, although there was a trend for the former to score higher on several items. Over half of the sample reported periods of lack of food or water, ill-health without access to medical care, forced separation from families, and family members or friends being murdered. More than 20% of the sample reported experiencing serious injury, forced isolation, imprisonment or torture.
Levels of endorsement (stressful/very stressful) for the 23 items in the Postmigration Living Difficulties Checklist (since arrival in Australia [PPV holders] or release from detention [TPV holders]) are given in Box 3. The proportion of TPV holders experiencing difficulties exceeded that of PPV holders on all items assessed (P < 0.001). Separation from the family, worry about the family’s safety, being unable to return home in an emergency, and anxiety about repatriation were endorsed by over 90% of TPV holders.
All TPV holders who had been in immigration detention centres in Australia identified detention experiences causing serious/very serious stress. More than 95% of ex-detainees reported serious/very serious stress regarding fears of being sent home, being told by officers that they should return to their country of origin, and language difficulties while in detention. Other items endorsed as causing serious/very serious stress by more than 90% of the sample included separation from families, being interviewed by immigration officers, not receiving adequate medical treatment, exposure to acts of violence and brutality, seeing people make suicide attempts, and several items related to poor conditions in detention. Items endorsed at a lower level were nevertheless noteworthy, including being assaulted by officers (81%), being handcuffed during transport (71%), being woken during the night for head counts (85%), being forced to use unhygienic toilets (81%), and solitary confinement (60%).
The results of univariate analyses comparing TPV and PPV holders on indices of psychiatric status and functional impairment are shown in Box 4. TPV holders scored higher on all psychiatric symptom measures and had a lower score on the SF-12 mental component summary score, indicating greater functional impairment in this domain. There was some evidence of a higher level of physical functioning among the TPV sample (indicated by higher scores).
The trauma count derived from the Harvard Trauma Questionnaire was significantly associated with psychiatric and functional impairment across all of the six clinical domains assessed (ie, PTSD, depression, anxiety, GHQ-30, SF-12 mental and physical functioning). For PTSD, in the next step, TPV status emerged as the greatest single contributor, yielding a large standardised β value of 0.84. The overall predictor model for PTSD accounted for a substantial amount of the variance (73.9%), with TPV status making the greatest contribution (≈ 68%). In the other equations, a combination of age, sex, trauma and TPV status contributed substantially to indices of mental health and functional impairment, but not to the same extent as for PTSD (Box 5). TPV status and prior trauma exposure consistently contributed to scores on the Hopkins Symptom Checklist-25 depression and anxiety scales, the GHQ-30 score and the SF-12 mental component summary scale, controlling for the effects of other variables included in the model. The model for physical disability showed a distinctive pattern, with older age, female sex and trauma exposure, but not TPV status, making contributions.
Our study provides consistent evidence that the migration trajectory experienced by TPV holders, particularly adversity in detention and ongoing living difficulties, is accompanied by persisting and wide-ranging mental health problems and associated disability.
Our data and those from a parallel investigation22 represent the first systematic documentation of the mental health of refugees in Australia who have experienced detention followed by temporary protection. Assessing the impact of these experiences is important, as they are central to the strategies adopted by Australia, and, increasingly, other Western countries, for deterring the inflow of asylum seekers.
Investigations in detention centres in Australia6,10-12 and abroad13,26 have indicated that confinement under harsh conditions is associated with deterioration in the mental state of detainees. Studies among asylum seekers living in the community,20,21,27,28 have also shown the compounding effects of past trauma, postmigration living difficulties and insecurity about the future.
The limitations of our study include potential transcultural measurement error and sampling biases. Measures were specifically chosen because of their past psychometric performance when applied across cultures and we applied standard translation–back translation methods. A native Persian-speaker (S M) with considerable experience in transcultural research undertook all interviews. She clarified the meaning of items with participants wherever necessary. Although the Early Intervention Program provided access to about half of the off-shore humanitarian arrivals (PPV holders) and most of the TPV holders in NSW during the study period, the coverage of PPV holders varied substantially by subcategory, with non-sponsored refugees (the largest group) being well represented (83%), but sponsored PPV holders less well represented (15%). If this sampling bias had any effect, it would be to reduce differences between TPV holders and the PPV group, as the under-represented group is well supported on arrival by existing family and social networks, a fact that is likely to mitigate mental distress.29
TPV holders with insecure residency may be more likely to be plaintive in their attitudes and hence exaggerate their problems. However, the TPV group had been recognised as refugees and had only recently been released from detention, so there was no immediate personal benefit to them to obtain further documentation or publicity relating to their cases. They were fully aware that their residency status would only be considered some years later, and that refugee determinations would be based solely on whether they were deemed at risk of future persecution in their home countries, and not on the subject of inquiry in our study. Furthermore, there were no differences between PPV and TPV holders in reporting past trauma exposure and the overall levels of abuses reported are not excessive compared with other refugee groups studied in Australia20,21 or worldwide.30-33
1 Australia’s permanent and temporary protection visas
Permanent protection visas (PPVs): Australia’s off-shore humanitarian resettlement program provides PPVs to a quota of refugees screened overseas and those requiring humanitarian resettlement. They are provided with a comprehensive resettlement program, are entitled to apply for Australian citizenship after 2 years of residency, and have unrestricted access to all services and entitlements of Australian citizens.
Temporary protection visas (TPVs): From 1999, all asylum seekers arriving without valid travel documents, who were subsequently recognised as refugees, were issued with TPVs, which place the burden of proof on the holders to re-establish a need for continued protection under the Refugee Convention every 3 or 5 years, with the real possibility of involuntary repatriation. TPV provisions restrict access to employment assistance, English-language training, and higher education. TPV holders are not entitled to apply for family reunion and are not authorised to leave Australia for any reason.
2 Demographic characteristics of asylum seekers granted permanent protection visas (PPVs) and temporary protection visas (TPVs)
No. (%) of those married whose spouse is in the home country |
|||||||||||||||
Residency in Australia (mean no. of months living in the community) |
|||||||||||||||
Mean duration of detention (months) |
3 Living difficulties causing serious/very serious stress since release from detention (holders of temporary protection visas [TPVs]) or arrival in Australia (holders of permanent protection visas [PPVs])
All contrasts between TPV and PPV holders significant at P < 0.001 using Fisher’s exact test.
4 Comparison of scores of temporary (TPV) and permanent (PPV) protection visa holders on psychiatric outcome measures: univariate analyses
5 Standardised β coefficients (and P values) for multivariate regression analyses predicting symptom outcomes
Medical Outcomes Study Short Form (SF-12) |
|||||||||||||||
Model 2: Step 1 + TPV stressors |
|||||||||||||||
Received 30 June 2006, accepted 4 August 2006
Abstract
Objectives: To determine the impact of the Australian provisions for temporary rather than permanent protection for asylum seekers found to be genuine refugees.
Design and setting: A comparison of the mental health of Persian-speaking refugees with temporary (n = 49) versus permanent (n = 67) protection visas attending an early intervention program in Sydney, New South Wales, 2002–03.
Measures: Standard measures were used to assess past trauma, detention experiences, postmigration stresses, symptoms of post-traumatic stress disorder (PTSD), anxiety, depression and functional impairment.
Results: The two groups had experienced similar levels of past trauma and persecution. Nevertheless, holders of temporary protection visas (TPVs) returned higher scores on three psychiatric symptom measures (P < 0.001). Multivariate analyses showed that TPV status was the strongest predictor of anxiety, depression and particularly PTSD. Further analyses suggested that, for TPV holders, experience of past stresses in detention in Australia and ongoing living difficulties after release contributed to adverse psychiatric outcomes.
Conclusions: The sequence of postmigration stresses experienced by TPV holders appears to impact adversely on their mental health.