Sample

The study was conducted in the Western Cape region of South Africa. Participants were drawn from four areas in the Western Cape where the TRC had taken closed (private) statements and held public hearings. Because 90% of all people who gave statements to the TRC were Black Africans (Truth and Reconciliation Commission of South Africa, 1998), the present study focused on African communities in these four areas.

Participants (n=134) were selected according to the TRC's definition of “a victim of a gross human rights violation” (Truth and Reconciliation Commission of South Africa, 1998). This included people who were themselves violated and those whose family members were violated. It also included several different categories of violations: killing of a family member; torture of self or family member; severe ill-treatment of self or family member; abduction of self or family member; disappearance of family member (without return); and associated violations (police detention, raid on property, damage to property, looting) to self or family member.

The sample was divided into three groups:

1. (a) survivors who had testified at a public hearing of the TRC (Public);

2. (b) survivors who had given a closed statement to a TRC statement-taker (Statement);

3. (c) survivors who did not give a statement or public testimony (None).

Limited resources and lack of access to a list of TRC deponents in the Western Cape precluded the use of random selection in this study. Participants were therefore recruited through a combination of media advertising and networking with community agencies and key community figures. Once the participants had volunteered, the snowball method was used to reach other participants. These sampling methods have been employed previously in empirical studies of the psychological effects of human rights abuses (Reference Başoğblu, Paker and PakerBaşoğlu et al, 1994; Reference Thompson and McGorryThompson & McGorry, 1995), but the use of a non-random sample implies some caveats on findings.

The length of time between giving a closed or public statement to the TRC and participation in this study was between 2 and 3 years.

Instruments

Each participant completed the following:

1. (a) A structured interview (developed for the study, details available from D.K. upon request) assessing exposure to the categories of human rights violations (HRVs) delineated by the TRC.

2. (b) The Mini-International Neuropsychiatric Interview (MINI; Reference Sheehan, Lecrubier and SheehanSheehan et al, 1998), which is a widely used clinician-administered structured diagnostic interview for DSM-IV (American Psychiatric Association, 1994) and ICD-10 (World Health Organization, 1992) psychiatric disorders. A history of psychiatric or psychotherapeutic treatment was also recorded.

3. (c) The Enright Forgiveness Inventory (EFI; Reference Al-Mabuk, Enright and CardisAl-Mabuk et al, 1995), which is a self-rating scale of the affective, cognitive and behavioural components of forgiveness that has been used in numerous empirical studies internationally (Reference Hebl and EnrightHebl & Enright, 1993; Reference Al-Mabuk, Enright and CardisAl-Mabuk et al, 1995; Reference Freedman and EnrightFreedman & Enright, 1996; Reference Coyle and EnrightCoyle & Enright, 1997).

4. (d) The Composite International Diagnostic Interview (CIDI) post-traumatic stress disorder (PTSD) event module (Reference KesslerKessler, 1994), which has been used to measure trauma exposure in different cultural populations (Reference WittchenWittchen, 1994). This instrument was used to determine traumatic experiences, other than HRVs, that the participants had been exposed to and that could have an impact on their psychiatric status.

Participants were also asked to rate their degree of religious commitment as low, moderate or high.

Each instrument was translated into Xhosa, the indigenous language of most African people in the Western Cape, and back-translated into English, in order to ensure linguistic and semantic equivalence. The research protocol was approved by the ethics committee of the institution to which the authors were affiliated.

Procedure

Written informed consent was obtained from each participant. Interviews were administered by Xhosa-speaking researchers (a psychiatric nurse and a psychologist who received training in the use of the clinical instruments) and were conducted in a community centre located centrally in each area. Participants were informed of their right to refuse to answer any question or to terminate the interview at any point. Participants who were identified as having a psychiatric disorder or who exhibited marked distress during the interview or in the days following the interview were offered a variety of referral options, including both medication and psychotherapy.

Sample characteristics

Exposure to HRVs

The average number of violations (to both themselves and to family members) to which participants were exposed was 8.4 (s.d.=5.4), ranging across the sample from a low of one violation to a high of 24. With regard to the types of violations experienced, 90% of the total sample had experienced a violation to themselves, 82% reported violations to a family member and 72% had experienced both. The number of violations to the participant him/herself ranged across the sample from 1 to 17, with an average of 6.1 (s.d.=4.6). The average number of violations to a family member ranged from none to 24, with a mean of 2.4 (s.d.=2.4). The frequency of each type of violation is shown in Fig. 1.

Fig. 1 Frequency of human rights violations (total sample, n=134). ^*Includes detentions, raids, looting and damage to property.

There were significant gender differences in exposure. Men had a significantly higher mean number of total violations than women (P=0.02). Although men had experienced a significantly greater mean number of violations to themselves than had women (P<0.0005), women had a significantly higher mean number of violations to a family member than men (P<0.0005).

Psychiatric diagnosis and TRC exposure

A χ² test was used to examine whether there were any differences between the three exposure groups in rates of PTSD. The result was not significant (χ² ₂=3.62, P=0.16). However, inspection of the contingency table showed that the proportion of PTSD was 23.8% in the Public group, 47.5% in the Statement group and 41.9% in the None group. Because the Public group seemed to have a substantially lower rate of PTSD than the other two groups, which were very similar to each other, we combined the Statement and the None groups and compared them with the Public group. The difference was not significant at the 0.05 level (χ² ₁=3.31, P=0.07).

No association was found between TRC exposure and depression (χ² ₁=1.63, P=0.44), or TRC exposure and anxiety disorders other than PTSD (χ² ₁=0.54, P=0.28).

Exposure to the TRC and forgiveness attitudes

One-way analyses of variation (ANOVAs) revealed no significant difference in mean levels of forgiveness between the three exposure groups (F _2,131=0.39, P=0.68). However, the Public group showed a distinctly different pattern of forgiveness from the other groups. Although participants in both the Statement and None groups tended to have a spread of forgiveness from low, through moderate, to high, participants in the public group tended to be either very forgiving or very unforgiving. The distribution of forgiveness scores had a bimodal appearance and the standard deviation of the Public group (105.6) was substantially larger than for the Statement group (81.01) or the None group (83.3). Those Public participants who clustered on the low end of the forgiveness scale were mostly female, whereas those that clustered on the high end were mostly male. Further investigation revealed that forgiveness score is associated significantly with gender (F _1,131=6.79, P=0.01), with females being less forgiving than males. A linear regression then simultaneously examined the contribution of gender, religion and type of victimisation (self only, family member only or both) and found that gender was the only variable associated significantly with forgiveness score.

Forgiveness and psychiatric diagnosis

One-way ANOVAs showed that depression, PTSD and other anxiety disorders were all significantly higher among participants with low forgiveness scores when compared with those with high forgiveness scores (P=0.01, P=0.03 and P=0.04, respectively).

Participation in the TRC and psychiatric status

There was no significant difference in the rates of depression, PTSD or other anxiety disorders for participants who gave public testimony, closed testimony or no testimony. Although some individuals may have experienced testifying as either distressing or relieving, it would appear that, for this sample as a whole, the process of giving either public or closed testimony to the truth commission did not have a significant effect on psychiatric health (it had neither a notable therapeutic effect nor a notable counter-therapeutic effect).

The apparent lack of any significant impact on symptoms suggests that the process of testifying at the TRC may be qualitatively different from that of testimony therapy in the clinical setting. Thus, it may be overly ambitious for truth commissions to have a ‘therapeutic’ goal, except at the broader national level. It may be argued also that the perceived absence of justice (i.e. punishment of perpetrators and compensation of survivors) in the TRC process, about which many survivors have protested (CSVR & the Khulumani Support Group, 1998; Reference HamberHamber, 1998), may have been a barrier to recovery. If justice is done, and seen to be done, psychological healing may be facilitated. Finally, the frequent and ongoing exposure to other traumas among this population may also explain the apparent failure of the TRC process to reduce the presence of psychiatric disorder. Any short-term relief associated with giving testimony is unlikely to be sustained in the face of chronically high levels of community trauma.

Participation in the TRC and forgiveness

Our findings indicate that a lack of forgiveness is related to poor psychiatric adjustment, although the causal nature of this relationship cannot be established here. It appears that being unforgiving, although an understandable moral response to being violated, also carries an increased risk of psychiatric morbidity. However, TRC participation was not associated with any difference in overall levels of forgiveness, indicating that additional interventions to promote forgiveness, such as some form of survivor-perpetrator mediation, may be required.

We noted a tendency for survivors who gave public testimony to be either very forgiving or very unforgiving. Public hearings may have been characterised by a process of self-selection (or selection by the TRC) of survivors who were either very forgiving or very unforgiving. Alternatively, it may be that the process of giving public testimony facilitates a high level of forgiveness when it is effective but a low level of forgiveness when it is ineffective. Because it was found that those public testifiers who were highly forgiving tended to be men, whereas those who were unforgiving tended to be women, the effectiveness of the public process may depend on the deponent's gender. Although women in the total sample were found to be significantly less forgiving than men, the split between very low forgiveness among women and very high forgiveness among men was only apparent in the Public group, and could not be explained by factors such as religion or the type of violation experienced. In general, the association between gender and forgiveness has received little investigation (Reference Worthington, Sandage, Berry, McCullough, Pargament and ThoresenWorthington et al, 2000), and future research in this area may provide a better understanding of the apparent gender effect found among public testifiers in our sample.

Current psychiatric needs of survivors

The high level of psychiatric disturbance coupled with the low treatment rate in our sample suggests that survivors in South Africa are currently underdiagnosed and undertreated by the mental health system. Anecdotal clinical lore in South Africa indicates that this may be due in part to a tendency among Black African patients to present with physical complaints (e.g. sleep disturbance or bodily pain), whereas a range of psychiatric symptoms remain unreported without more careful screening and assessment. The healing capacity of the truth commission process, both in South Africa and elsewhere, may be much enhanced by the provision of adequate, relevant mental health services. This should include routine screenings at primary care level for a history of human rights abuses among patients in contexts where such experiences are prevalent, and the training of primary care physicians and nurses regarding the psychiatric effects of human rights abuses as well as available treatment options. Given the high costs of medical and psychiatric care in a context where resources are scarce, as well as the arguably limited cultural applicability of Western treatment modalities, indigenous healers and existing community resources also have an important role to play in the recovery of survivors of human rights abuses in South Africa.

Limitations

Owing to the non-random sample and retrospective design employed in the current study, significant findings should be interpreted with some caution. In addition, several of the instruments used have been developed for use with Western populations and have not been validated on South African samples. They may therefore miss important cultural nuances in the expression of psychiatric illness and of forgiveness (Reference Keane, Kaloupek, Weathers, Marsella, Friedman and GerrityKeane et al, 1996). Finally, it has been argued that classic PTSD criteria may not adequately capture the full range of post-traumatic reactions. The notions of ‘complex PTSD’ (Reference HermanHerman, 1992) or ‘disorders of extreme stress’ (Reference Pelcovitz, van der Kolk and RothPelcovitz et al, 1997) may be more appropriate conceptualisations of responses to extreme chronic stress such as that to which South African survivors have been exposed. The use of random samples, prospective research designs, culturally sensitive instruments and a broader conceptualisation of post-traumatic reactions will enhance future research on the impact of truth commissions for survivors of human rights abuses.

Despite these limitations, our findings provide important preliminary data which have several implications for the structuring of future truth commissions and for intervention with survivors of human rights abuses. First, truth commissions may not be sufficient to reduce psychiatric symptoms and promote forgiveness. Truth commissions should form part of, rather than be a substitute for, comprehensive and ongoing therapeutic interventions for individual survivors. Second, a lack of forgiveness may be an important predictor of psychiatric risk among survivors of human rights abuses. Finally, the psychiatric needs of survivors of human rights violation in South Africa have not been addressed adequately by the existing mental health system. Access to mental health interventions that are culturally appropriate and that address the specific needs of survivors of human rights abuses is a vital adjunct to an effective truth and reconciliation process.

CLINICAL IMPLICATIONS

• ▪ Truth commissions may not be sufficient for the recovery of survivors; additional therapeutic interventions that are culturally appropriate and specifically address the needs of survivors of human rights abuses should supplement the truth commission process.

• ▪ A lack of forgiveness may be an important predictor of psychiatric risk among survivors of human rights abuses.

• ▪ In contexts with a history of state-perpetrated violence, routine screenings for human rights abuses and post-traumatic responses are recommended.

LIMITATIONS

• ▪ Non-random sample.

• ▪ Retrospective design.

• ▪ Use of instruments standardised on Western populations.