Technical Advisory Group Meeting
for the Psychosocial Effects of
Conflict-Related Trauma
August 1-2, 2000
Introduction
While the field of cross-cultural assessment and treatment of the psychosocial effects
of conflict-related trauma is gaining increasing attention in the scientific literature,
there is still enormous room for research and the development of ‘best
practices’ in the discipline. The Technical Advisory Group meeting of August
1-2, 2000 was hosted by World Vision, co-facilitated with Johns Hopkins University and
sponsored by CERTI (Linking Complex Emergencies Response and Transition Initiative, a
program sponsored by USAID and others). The meeting was called to provide technical
feedback on recent research initiatives conducted in Rwanda, and to discuss the best ways
forward. Participants in the meeting represented World Vision, Johns Hopkins
University, Harvard University, Columbia University, Tulane University, Randolph-Macon
College, Christian Children’s Fund, the Commission on Mental Health Services (based
in Washington DC) and the American Red Cross. The diversity in backgrounds among
participants ensured that the process benefited from a broad spectrum of approaches and
expertise.
The primary objective of the TAG meeting was to review the recent work by Paul Bolton
(Johns Hopkins University) and Lincoln Ndogoni (World Vision International) detailed in
the report entitled Cross-Cultural Assessment of Trauma-Related Mental Illness.[1] As part of the CERTI
project, this work consisted of the development and field testing of a method for
assessing aspects of mental health across cultures. In the Rwanda field trial the
methodology was used to assess the prevalence, severity and functional impact of
depression among a rural Rwandan population. The methodology involves multiple
research techniques adapted to be within the technical and financial capacities of most
NGOs on the field level. 'The principal components include: ethnographic research on local
perceptions of function and mental health, creation of a locally relevant functional
assessment instrument and adaptation of existing mental health instruments to local
conditions using the ethnographic results, validity testing of these adapted instruments,
and use of these instruments in a community-based survey. In Rwanda the depression
section of the Hopkins Symptom Checklist was adapted to assess depression symptomotology.
The morning session on the first day was devoted to the discussion of the report’s
methodology. Prior to the meeting seven of the TAG participants had been asked to
review the report and provide advance written feedback. Specifically they were asked
to comment on the strengths and weaknesses of the methodology. For each weakness
they were also asked to make a recommendation for improvement. The meeting began
with a review of these comments. This led to a wide-ranging discussion touching on
many technical and practical issues. The afternoon session shifted to a discussion
of implications of the research for intervention programming to address the psychosocial
health of trauma-affected populations. The issues brought up in these discussions
are summarized below.
A full list of participant names and contact information is provided as an attachment.
Perceived Strengths of
the Methodology
Study Content
- Contribution to the field of
applied behavioral research on the psychosocial effects of trauma – This
work was applauded by all participants as a contribution to a vitally important field of
which the research base is still limited. Participants acknowledged that this is an
extremely difficult problem area to research, particularly while a systematized approach
is lacking. There is little empirical data and numerous methodological
challenges. Even basic terminology is not yet agreed upon in this multi-disciplinary
field. To have completed this study at this level of rigor was impressive. All
agreed that this was an excellent contribution to knowledge of the problem and to applied
assessment strategies.
- Emphasis on social function
- The value of the emphasis on social function was widely noted for several reasons.
Social function is thought to be easier to measure than symptomotology. It was
recognized that improvement of social function was likely to be the desired outcome of NGO
community-level psychosocial programs, rather than alleviation of the symptoms of mental
illness, and thus this assessment tool assists in the evaluation of program impact.
Rather than assuming an association between depression and impairment of function, this
study contributed to the examination of the relationship between these dimensions of human
affect and behavior. This work assists in clarifying the relationship between
depression and function in transitional (post-conflict) settings, an issue about which
there has been little research to date. The relevance of social function as an
outcome variable is supported by evidence that social productivity (i.e. returning to
work) probably helps to prevent depression and has some effect in alleviating depression,
in addition to the profound collateral benefit of furthering the reconstruction and
rehabilitation of the community. Additionally, the demonstration of a clear association
between depression and social function has implications for national development, and may
result in future funding and intervention initiatives.
- Ethnographic component
– The ethnographic component of the methodology allows the assessment process to be
contextualized in the Rwandan setting; this is essential for mental health work, which is
fundamentally dependent on the local culture. By studying the local classification
system of health concepts and practices, the team is then able to incorporate mechanisms
to detect both depression (as defined by the DSM), and the most similar local syndrome (in
this case a syndrome described in terms of severe grief) and to examine the relationships
between them. This approach allows researchers to examine the validity of western
health concepts in other cultures, to adapt and validate western instruments measuring
these concepts, and to explore ways to tailor interventions to be locally appropriate.
- Focus on depression
– The team agreed that depression was the type of psychosocial response most
appropriate to be the focus of this initial assessment. This is because W.H.O.
reports have identified depression as a major contributor to social disability, both in
developing and developed countries, and its prevalence is likely to be greatly elevated in
the crisis setting. It is believed to account for a large portion of longer-term
trauma-related mental illness, and may also have a high prevalence in non-crisis
settings. It was suggested, however, that future assessments should be broadened to
include other common forms of trauma-related mental illness.
Process
- Process –
The NGO and academic linkage was well appreciated by participants, and viewed to be a
model worth expanding in application in the future.
- Accessibility of the
methodology – It was noted that the methodology is within the resource
capacity of many NGOs, given sufficient external technical support where needed. The
methodology is also mobile – it requires only limited infrastructure and can be
implemented relatively easily.
- Participation and capacity
building – The reliance primarily on national staff to implement the
survey and the plans to train national staff in analytical methods in the future were
recognized as practices consistent with a commitment to local participation and capacity
building.
Dissemination
- Incorporation of multi-disciplinary
involvement – Participants felt the involvement of the TAG members in planning and
review of the methodology contributed to the quality of the project in terms of building
on expert knowledge in all aspects of the work.
Concerns and
Suggestions
Study content
- Omission of ethnographic
information – It appeared that much of the cultural information gained
in the ethnographic process was omitted from the report, despite the fact that a wealth of
such information was gained in that process. With regard to local concepts of
illness and health, comparatively little information on the process of translation between
western concepts and local concepts was provided. It would be very useful for
readers to understand exactly what information was gained regarding local taxonomies and
concepts of illness and health as a context for understanding the report’s findings,
and to understand the process by which the assessment team interpreted this information
and drew analogies to western concepts and language. For example, depression is currently
seen in western culture as a primarily biological phenomenon but in other cultures the
perception and experience of that illness may be wholly different – e.g. spiritual,
emotional or simply a normal reaction to life-events. Additionally the report should
include some exploration of the pre-disaster population in terms of culture, health,
mental health, traditional health practices, coping strategies, etc., for the purpose of
helping to assess what is ‘normal’ for that population.
- Terminology –
The terminology currently in use in the psychosocial / mental health field is ambiguous
and warrants definition. The lack of a clearly defined vocabulary for the
psychosocial health field results in confusion and masks the fact that there are
psychosocial health implications for all types of interventions, not just those in the
“psychosocial sector”. It was noted that the distinctions between the
terms ‘psychosocial health’ and ‘mental health’ have not been
adequately defined, a point that becomes particularly problematic with community-level
interventions that are designed to promote ‘psychosocial health’ among the
community as a whole, without targeting individuals identified to have clinical mental
illness. The definition of the core terms in this field would be a valuable
contribution to discussions of assessments of needs as well as programming.
- Clinical orientation
– If assessing clinical disorders implies clinical interventions only, then this
approach could be viewed as underemphasizing community resources that may promote healing:
social support networks, social movements to interpret and come to terms with traumatic
experiences, etc. This should be addressed in the report as NGOs are well placed to
identify local structures and traditions that may promote healing and that may serve as
the basis for more specialized mental health services.
- Spirituality and coping
strategies – Spirituality is an important aspect of life in many African
cultures, and is likely to be at the core of how people cope with trauma in Rwanda.
It was, however, not discussed in the report. The omission raises the question of
what may have been missed in this analysis. The report seems to assume that healthy
coping is inter-personal and directly addresses one’s trauma; this assumption may be
questioned. It was suggested that for the severely trauma-affected, some studies
have reported that overtly addressing one’s trauma may be counter-productive, and the
most well functioning are those who can “bury” their experiences and move on.
- Possible biases in reporting
– Several concerns were raised about potential sources of bias in the methodology
that should be addressed in the report. As discussed above, the process by which
translation of ethnographic information took place is not described in detail in the
report, and it is unclear how the assessment team ensured linguistic equivalence among
concepts in the study. This raises the possibility of the influence of biases or
inadequacies in the translation process. There is a tremendous risk of changing or
losing meaning in this translation process (translation between words, between worldviews
and between concepts). Attendees of the meeting stated that the validity of the
translation could be assessed in part by examining the extent to which the given measures
behaved in relation to each other and whether this was as expected. For example,
depression would be expected to be more common in women than men and to be associated with
social function. Both were found to be true in the study. The reviewers wished
to see more of such evidence of internal consistency. This would make them more
willing to accept that translations between concepts are likely to be correct. The
issue of the ethnicity of interviewers and respondents was also raised: within the Rwandan
context it is reasonable to expect that the relationship between interviewer and
respondent ethnicity could result in interviewing, reporting or recording biases.
Another possible bias derives from the tendency of respondents to represent issues or
underreport behaviors according to what the respondent believes the interviewer would like
to hear (i.e. SOCIAL desirability bias), such as minimizing reported alcohol
use. It is unclear how much this phenomenon, if present, is due to respondents’
religious beliefs, or World Vision’s religious affiliations, or both.
Additionally, the question “What are the main problems that affect people in this
Commune as a result of the genocide in 1994?” presumes that the respondent sees the
1994 war as a “genocide”, which implies a specific political perspective.
The word “genocide” is politically loaded in Rwandan culture, and the question
may lead to reporting bias depending on the perspective of the respondent. It
is possible that because the Rwandans in the study area have previous exposure to
intensive NGO activities, they may bias their information because of a perception that
doing so will improve their chances of gaining access to additional aid. Also, the
interviewers were local staff who would likely have experienced some of the same trauma
experienced by the study population, and may therefore have mental health issues which
could bias their techniques in their work (discussed further below).
- Generalizability
- The assessment eventually focused on Kinzenze (a predominantly Tutsi area) rather than
Butamwa (a predominantly Hutu area). It is possible that the pre-existing
psychosocial health program in Kinzenze and their proximity to the capital affected (differentially
compared to Butamwa’s population) the population’s self-awareness, their
understanding of health (particularly mental health) and Western language surrounding
health concepts. If Tutsis are more likely to want to talk about the genocide and
its effects, then the combination of this difference in reporting openness combined with
the focus on Kinzenze may bias the results in favor of Tutsi respondents’
perspectives and levels of psychosocial distress. The population studied in this
work are unique in the extreme conditions they have experienced as well as the extensive
contact that they have had with external agencies in recent history; this uniqueness,
combined with the potential sources of bias listed above, raise questions about the
generalizability of findings from this study to the broader population of trauma-affected
people in Rwanda, Africa and other sites of complex emergencies and war.
- Post Traumatic Stress
Disorder – Participants felt that it was appropriate to prioritize
depression over PTSD because of depression’s greater prevalence in a wider diversity
of settings and its equal, if not greater, role in social impairment. Yet the report
may imply that it is not important to assess PTSD, despite the fact that screening solely
for depression has a low sensitivity for detecting PTSD sufferers. It was concluded
that the scope of any assessment should be limited to issues and conditions which the NGO
is prepared to address in programs, but an assessment tool should be developed for PTSD
for future work in conflict-affected areas.
- Interventions
– It was noted that prior to conducting the assessment, World Vision staff could have
discussed internally the range of interventions which the organization would be willing to
implement in response to different types of results. This would have allowed the
assessment team to identify, throughout the course of the assessment, projects that would
be appropriate and feasible responses to information they encountered.
- Staff mental health –
The question was raised as to the effects on the interviewers’ mental health on the
research process, given that the local staff derive from the same population and likely
suffered similar exposure to traumatic events (in any case, it is now recognized that
international staff require debriefing regardless of previous exposure). How does WV
(or any other organization) know if its local staff are mentally well enough to serve as
interviewers in this type of research? Psychosocial work in conflict-affected areas
poses particularly stresses for NGO employees (following over-identification and empathy,
for example) which should be actively addressed by staff health policies. It was
agreed that psychosocial health programs (and other programs in general) should have a
staff health component built into the planning and budgeting.
- Social Functioning
– The assessment of social functioning was viewed to be a strength of the
study. However, reviewers raised concerns over the specific approach taken.
The approach resulted in respondents answering hypothetical questions rather than
reporting behaviors. The functionality questions attempt to control for the effects
of age, sex, lack of money and lack of assistance. To do so, they asked the
respondent to compare him/herself to someone of the same age and sex “who has no
problems,” i.e. with health, or poverty primarily. Reviewers were concerned
that this then requested respondents to compare themselves to a hypothetical person, which
may lead to inaccurate responses. It was suggested that a more complete review of
approaches for assessing social functioning (including proxy reporting, etc.) be
conducted, particularly approaches that may control for those variables in the analysis
rather than in the reporting. As a separate issue, the report seems to make certain
assumptions about functional and dysfunctional coping strategies. Despite the
ethnographic research at the beginning of the work, local perceptions of functional and
dysfunctional coping strategies were not provided.
- Causal analysis
– The methodology allows for the detection of correlations rather than the
directionality of causal relationships. Longitudinal studies or
quasi-experimental designs are needed to begin to address issues of causality.
- Test-retest reliability
– A sub-sample of respondents were re-interviewed and their responses evaluated for
test-retest reliability. The reliability was found to be adequate only
(Pearson’s correlation of 0.671 for the Depression scale and 0.574 for the function
questions). Given that in most scenarios the tool would be used repeatedly over
time, the low test-retest reliability points to the need for further research into the
reasons for this (such as test reactivity and mood variation).
Process
- Risk of retraumatization
– All participants were concerned about the risk of emotional retraumatization for
respondents, particularly if counseling were not made available during and after
questioning, and particularly if effective referral systems were not in place. This
risk gives rise to ethical questions which have not necessarily been fully resolved,
regarding the responsibility of agencies implementing trauma-related surveys.
Policies will need to be developed by World Vision to address this problem. It
should be noted that participants were not asked about traumatic experiences.
Therefore, the risk for harm may well be less than in more traditional studies of the
psychosocial effects of trauma.
- Involvement of the local
population in interpretation of study findings – The Rwandan communities
studied in this assessment do not appear to have been involved in the interpretation of
the results. Their involvement is likely to be of heightened importance in a field
of this nature. The utilization of international technical assistance (and therefore
the reliance on western concepts, approaches and tools) always poses a risk of the
disempowerment and marginalization of local populations. While this research incorporates
a participatory model, the involvement of local communities should extend through the
interpretation/analysis and recommendation phases.
- Time delay - The usefulness
of the survey tool may decrease with time as the population’s mental health and
coping mechanisms change following a traumatic event(s). The time delay between the
disaster and the needs assessment, and then between the assessment and intervention,
should be minimized to ensure the information is still as relevant and the population is
still actively involved in the process.
Dissemination
- Scope of project
– It was recognized that the report is excessively multi-purpose in terms of its
audience, and therefore does not address the needs and interests of specific
users. The authors should consider focusing solely on one audience, or
preparing multiple reports for dissemination. A particular concern was that many of
the technical concepts were beyond the training of many people who would be interested in
the report.
- Educational componentand
sharing the methodology – Because this study is a field test of a
methodology which is to be used by NGO staff on the field level, it would be very useful
if the report provided educational information about how the work was done, detailing in
particular decision-making and interpretation processes. The component could include
what problems were encountered in the field and how they were dealt with; a discussion of
the applicability and utility of the work in programming; and how translation was done. At
this point, most of the individuals focusing on developing this field are Western and
highly educated. It is our responsibility to describe the work with as much
practical detail as possible so that NGO field offices can use the document to implement
similar pieces of work. The more transparent the report, the more intelligently and
critically the users can apply the work to their own country programs. Ultimately,
what will be needed is a manual that both describes the underlying principles guiding this
research, and then a detailed ‘how to’ section, reporting forms, detailed
descriptions of random sampling, training and methods of evaluating interviewers.
- Team capacities
– The report should demonstrate that the assessment team was sufficiently technically
equipped to adapt assessment tools based on local cultural and psychological information.
- Statistics
– The statistical analysis in this study should be moved to appendices in any version
of the document aimed to field personnel.
Recommendations
Study content
- Inclusion of more extensive
mapping of local cultural belief systems – The documentation of local
belief systems, particularly regarding the classification and treatment of health
conditions, would contribute significantly to the interpretation and applicability of the
research. It would elucidate the cultural health system from which local
trauma-related syndromes were derived; it would also lay a foundation for similar
ethnographic research in the future. The report should include sufficient detailed
qualitative and practical information to facilitate the application of the methodology by
field-level NGO staff. The ethnographic information acquired using this methodology
is a valuable contribution to the understanding of local health beliefs and practices, and
this information should be provided in sufficient detail that such methods can be
effectively employed by a broader diversity of people, particularly at the field level.
- Inclusion of other forms of
mental illness – The tool should be adapted to include PTSD,
panic/anxiety disorder and perhaps other forms of mental illness that are associated with
psychosocial trauma. Qualitative research should be conducted to determine how these
illnesses are understood, categorized and detected in local cultures prior to conducting
assessments to detect prevalence levels.
- Assessment of trauma-related
mental illness in children – Tools for the assessment of mental illness
in children should be developed as well. This requires a review of literature and
practices regarding research tools and skills appropriate for children of different age
groups, as well as research into the links between parental mental health and the mental
health of their children. Reviewers acknowledged that this is a big task which has
not so far been addressed by any group. Current assessments of children largely
consist of assessing exposure to traumatic experiences and assuming mental health effects.
- Issues for future
assessments – Future assessments should look more closely at causal
relationships between variables, the impact of adult mental health on care taking, local
perceptions of coping mechanisms, and the links between the HIV/AIDS pandemic, mental
illness and function. There should be important lessons learned from HIV work, given
the substantial funding for AIDS-related programs.
Process
- Community involvement in
interpretation and programming – The community should be involved in the
interpretation of the results, particularly given the socio-cultural nature of mental
health.
- Site selection
– To facilitate the assessment of impact, the methodology should be used in areas
where mental health-oriented interventions by other agencies (e.g. UNICEF and World Vision
in the case of Kinzenze) are ongoing.
Dissemination
- Dissemination for multiple
audiences – It may be appropriate to produce seperate documents
addressing different aspects of the work (i.e. a technical document detailing the
analysis, and a more educational document describing the research process).
Eventually these documents should form the basis of a single manual which can be used by
NGOs after appropriate training in its use.
- Cost estimates
– Cost estimates should be provided for single and repeated assessments for planners
to determine the feasibility of using the method for their own country programs.
- Statement of objectives –
The objectives of the research should be listed clearly at the beginning of each report,
to limit the scope of the paper and prevent it from becoming unwieldy. This
introduction would also provide an opportunity to define all relevant terms, such as
psychosocial health, mental health, etc.
Additional
Discussion Points
Participants stressed the importance of maintaining an active network among researchers
and practitioners working in fields related to psychosocial health. The network
should span academia and NGO/UN field programs and clinical to public health approaches.
Since other groups may be developing similar approaches, the possibility of linking up
with these groups should also be explored. The continual sharing of knowledge is
essential to the development of the discipline, with the objective of formulating
best-practice protocols for use at the field level. This network could take several
forms: collaboration in the development of research techniques and tools as well as in
fieldwork; periodic technical meetings to discuss specific aspects of the discipline; an
electronic library (available on CD-ROM and the Internet) which pulls together literature
and tools relevant to the field; and an electronic list serve to facilitate communication
between people involved in this field on all levels.
If mental health support to field staff and beneficiaries/respondents is to be
considered an essential part of psychosocial programming in the future, it would be useful
to develop a discussion of the types of support appropriate for different groups. It
is likely that individual support needs will differ based on education, residence (rural
vs. urban), ethnicity, or other factors. A staff health component should be
built into most proposals.
Psychosocial work (assessment as well as programming) is charged with a number of
ethical issues, regarding which population groups are focused on, what language is used
for the survey, and whether one focuses on ethnicity in the work. All of these
issues may have political implications that must be considered in planning.
Conducting similar research into the mental health of children will be particularly
challenging technically and ethically. Ethical issues surround the retraumatization
of children and the identification of strategies most effective at assessing their mental
health. It may make more sense to work with external measures of function among
children rather than symptoms. If we are restricted to measuring behavioral
variables, is this sufficient? Research is necessary on the link between parental mental
health and their children’s mental health, to assess the assumption that assisting
mentally ill parents automatically helps their children in the form of better care
taking.
Psychosocial programs may be based upon several basic models. The model of
intensive care for the acutely mentally ill, while possibly assisting in the recovery of
severely affected individuals, is not adequate for improving mental health on the
population level. For the African context it was felt to be more appropriate to use
community-based approaches which target the entire community for assistance
(sensitization, group therapy, etc.), in combination with support for an effective
referral system, preferably through partnership with national governments or other
partners. It is important to recognize that in disasters or crises, it is very
possible for the mentally ill to fall through the social net where others can no longer
take care of them. The mentally ill may be overlooked by agencies giving assistance
because it is assumed that they are not very relevant to rehabilitation or
development. Mental health hospitals are normally avoided because they are thought
to be a place to go to die. Often those hospitals are not well staffed, patients are
not well cared for, and the patients may never leave. Additionally these centers may
actually be targeted in active conflict. In a community-level approach to
programming, the objective is to improve the functioning of the whole society rather than
the rehabilitation of the severely affected, and often after crises the whole population
is affected to varying degrees. A related technical challenge is that we
actually do not know how to screen for and identify the mildly/moderately affected in the
community. If intensive clinical care is to be provided, there are advantages of
using the Centers of Excellence model: training opportunities, the development of
productive relations with authorities, the presence of a referral system, positive effects
on the morale of staff, opportunities to pilot new treatment modalities, and the ability
to develop outreach programs which may have spin-off benefits for many years. The
establishment of Centers of Excellence would require that WV be committed to the country
program and confident of funding for the long term.
The capacity of traditional healers to diagnose and treat specific conditions (locally
defined) has by and large received only very limited attention. NGOs interested in
improving psychosocial health should investigate what traditional healers do particularly
well, and identify where western approaches may be more effective. It should be
noted that often people do not feel comfortable talking about traditional health
practices, or they aren’t allowed to discuss those practices publicly. There
are natural areas of complementarity between western health practitioners and local health
practitioners that should be identified in future research. NGOs should synthesize facets
of traditional healers, counselors / helpful community members, spiritual leaders
and western medicine to maximize program effectiveness.
Local involvement in research and programs are important for sustainability. The locus
of control has influence on the effectiveness of any program. In the field of
psychosocial health, local interpretation, planning and feedback are particularly
important.
NGOs must maintain flexibility in their work, understanding that assessment tools and
interventions addressing mental health may differ between countries.
The assessment tool used in this study is not suited to the acute phase of
crises. It is necessary to develop a rapid emergency mental health assessment tool
for initial planning, to be followed by the more in depth assessment as the situation
stabilizes. Protocols for conducting mental health needs assessments or
interventions in acute crises must be developed. This is a field for future research
and collaborative efforts.
Conclusions
While the meeting yielded many suggestions regarding refinement of the tools and
techniques used in Rwanda, the work was applauded for its overall technical strengths and
valuable contributions to the knowledge base in this field. Participants found the
collaborative work between NGOs and Universities to be a promising model for future work,
and felt that the TAG had been a very useful forum for academic-field interchange about
assessment and programming related to the psychosocial effects of conflict-related
trauma. All of the participants expressed enthusiasm to participate in collaborative
initiatives in the future.
Discussions are currently underway regarding ways in which this technical network can
be coordinated, supported and funded. Possible initiatives include TAG meetings,
electronic information-sharing groups, electronic libraries and collaboration in upcoming
projects. Additionally, three sites have been identified for application of the
methodology over the coming year, and plans are being made to develop an assessment tool
for children’s mental health. Numerous opportunities for collaboration will
arise over the coming year in the development of both assessment and programming
protocols.
