After an RRA (rapid rural appraisal) which produced a general picture of the health sector, it was decided to continue the appraisal, using the method known as RAAKS, or rapid appraisal of agricultural knowledge systems (Engel 1995). RAAKS is designed especially to study the social organization of knowledge, in particular the loss of knowledge. The methodology encompasses three stages: problem definition, analysis of constraints and opportunities, and action planning. Each stage will be described in detail below.

Defining the problem

The starting point of the RAAKS exercise was the objective formulated by the council of leaders: to strengthen indigenous health care. But this problem definition was valid for only one of the actors, so that all the actors involved first had to redefine the problem. This was done in three steps: identifying relevant actors, elaborating their 'mission statements', and determining those external factors which impacted on the system.

To identify relevant actors, all those taking part in the exercise contributed other actors, and then began the mapping of actors and their linkages.

Next, the persons representing an actor were grouped together in order to elaborate their objectives and their contribution to a health knowledge system. Thus the payes formed one group, the midwives another, the local leaders another, etc. The mission statements which were then formulated (see box 1) served as a reference which was frequently referred to during the rest of the appraisal.

The external limitations had been defined during previous interviews. The most important constraint proved to be the temporary migration of wage labour. In the first place, this disrupts traditional reciprocal relations, as the villagers do less work in the agricultural fields of the paye, while the paye continues healing for free, since this is regarded as a social, moral and ethical obligation. In addition, the absence of villagers--including the payes--also hampers the transfer of knowledge, which entails a lengthy initiation process lasting up to 20 years.

Analysis of constraints and opportunities

On the basis of the overall problem statement, the group was divided into teams (1) for fieldwork. Each team consisted of one professional to assist with the methodological part, one local facilitator, and the relevant actors in that community, all acting as investigators. The teams started with the integration analysis: elaborating a matrix of all relevant actors by all relevant actors. After filling in the line corresponding to the actor, the participants were asked to name other relevant actors, as well as the most important relation with all of the other actors. This wealth of information was then used in the actual mapping. All relevant actors were visited and asked to group actors with good contacts near each other, and to separate those who had little contact. The teams were not required to reach consensus in order to draw a map: like all visual tools, RAAKS stimulates the drawing of various images which can be compared and debated (see figure 1).

Actor analysis

Coordination analysis

This window helps to identify leadership. The information on who controls whom from the integration analysis was supplemented by interviews and prime-mover septagrams. These drawings, made by each actor, indicate the importance they attributed to certain factors or actors when it came to achieving their aims or improving the system. The inner circle signifies 'no importance at all', the outer circle represents 'total determination'. Thus, if the point on the radial axis is near the outside circle, the factor or actor is important, and capable of 'moving' the system towards a better--or poorer--state. Factors and actors were suggested by the interviewee and by the facilitators.

The paye suggested that the importance be quantified by asking: 'How many patients are affected by a factor or an actor?'

Communication analysis

This window, which is concerned with cultural and cognitive differences which impede effective communication, is designed to determine whether people speak the same 'language'. The exercise focuses on the meaning and nature of 'illness'. The first step was to translate the word into Guarani. The second was to discuss the various meanings attributed to it by the different actors, and to compare these with the meaning of 'illness' in Spanish, which is the language spoken by the professionals.

The analysis brought to light the different perceptions and cognitive structures put forward by the various actors. For the payes--and the Guaranis in general--illness is caused by bad spirits, the "imbaekua", directed by one person to another. The spirits are embodied in an insect that takes possession of the human body. Only the paye with his supernatural power can extract the insect during a ceremony or session. The paye makes his diagnosis by smoking a cigarette of coca leaves, blowing the smoke over the body of the patient and 'reading' the pattern. This method of diagnosis obviously differs from that of the village health workers, who are trained in Western perceptions of illness. These underlying differences are an obstacle to communication and integration.

There is, however, some integration of the systems. The Guaranis working in the modern health sector also recognize the role of bad spirits. They listed various causes of disease, and acknowledged that there are certain ailments they cannot heal; they often refer patients with certain diseases to a paye. Conversely, the paye will not try to heal tuberculosis, but will send the patient to the hospital. And even the local priest (Evangelical Church) acknowledges that illness can be caused by an insect, and believes in the supernatural powers of the paye and in the gods of nature. But he insisted that he placed his hope above all in the God of the Sky, who also has healing powers which can be invoked through prayer.

'Fields of action'

Before embarking on the action planning stage, the team discussed and synthesized the information gathered thus far. Departing from the social web of actors, we identified the systems or 'fields of action' referred to by the paye. Each system or field was formed on the basis of its social organization, leadership, mission, the power at play, and the knowledge processes involved. The concept proved to be a useful tool in fitting the various pieces of information together.

Let us take as an example the field of the paye. The action or mission statement is curing and preventing diseases, in humans as well as plants.

Leadership is in the hands of the paye, who controls the work of his helpers, and the oporo pisoa who helps patients to recover after the paye has extracted the 'insect' by giving them a massage and other treatments.

The power at play is supernatural, secret and concentrated.

There are a multitude of other actors, each specialized in certain diseases or spirits.

Several actors have special knowledge of medicinal plant and animal products, ranging from medicinal properties to the location and ecology of the species.

Knowledge acquisition is slow, and takes the form of initiation rather than training and extension.

The paye selects his successor, but very few people have the characteristics required--patience, silence, dedication.

In addition, there is the example of the hospital. The field of action of the hospital is healing.

It is organized hierarchically, with the hospital doctor as leader.

The social organization includes health workers. Many of them do not have the necessary links with the existing indigenous health system. The Red Cross project, which previously operated in the area, had installed their own social structure and infrastructure, thus ignoring the existing system and institutionalizing the lack of integration of the two systems.

The type of power or knowledge at play is technical and financial.

Knowledge is transferred in a top-down manner through training and extension. The health workers had to be trained from practically zero.

Action planning

The strength of the tool is that people start modelling without explicit reference to the real world. The model itself must be logical and systemic, so that it can be compared with other actors' models and with the existing situation. Next, action plans can be designed in order to improve the existing situation. To take the example of figure 4, the midwife had included 'supervision by a second person' in her model, while in reality this was not being done. After having discussed this discrepancy, it was decided to incorporate such supervision in the list of proposals.

In order to present and discuss more plans, the team that had analyzed the knowledge system in the community developed a matrix which stated for each planned activity who would be responsible, who would perform the activity, as well as how and, when and what the available resources were. The matrix summarized the proposals of different actors and the insights of the team, and served as an input for the plenary planning session.

The plenary session started with the presentation of the matrices of the four communities. Each matrix was discussed in the wider group. Next, the proposals were integrated and a detailed action plan was elaborated.

The action plan made it clear that improvements to the knowledge system would depend on the coordinated action of all actors involved: many planned actions involved sharing tasks and responsibilities among various actors. The most important recommendations were as follows:

• Include in the terms of reference for the doctor a sincere respect for indigenous health care, and make one of his/her tasks the coordination with the payes.
• Improve both administrative skills and transparency through the formation of a board charged with promoting and supervising indigenous health care, as well as generating funds for indigenous health care and for training.
• Facilitate meetings between the various actors to exchange information, experiences and expertise. During these meetings, traditional medicine health workers can be elected, who will work alongside the village health workers.
• Start a legal procedure to obtain recognition of the work of the payes.
• Ensure the effective economic and moral support of the whole community for village health workers, traditional medicine health workers, and payes.

Conclusions: the health care system

Conclusions: the methodology

The RAAKS methodology proved highly successful in bringing various actors together to discuss their problems, visions and views. The methodology allows for a creative use of tools and combines different approaches; it also makes possible the analysis of power issues. The local facilitators and the relevant actors were very positive about the methodology and its logic. For the professionals from outside the region, none of whom were familiar with health care, it was a quick and structured way to 'dive into the deep'. All this suggests that RAAKS is useful not only in strengthening specific on-going processes, but also as a rapid appraisal methodology--for identifying in-depth research topics.

The methodology could be improved or expanded with respect to the analysis of intercultural communication. The "logic-based stream of analysis" might be combined with and fed by a "stream of cultural analyses" (Checkland & Scholes 1990:29). More work needs to be done, especially with respect to the latter recommendation. A possible starting point is the actor-oriented approach, with its concepts 'the actor's life world', 'interface' and 'discontinuities' (Long 1989).

The challenge now is to help the local facilitators in using this set of tools to conduct an appraisal of their wider institutional environment. The council of leaders will play a crucial role in the implementation of the plans, as well as in the negotiation, acquisition and coordination of projects. Thus there is considerable benefit to be gained from a RAAKS-type appraisal performed not by 'outsiders' studying Guarani communities, but by 'insiders' who have made a study of the methods of the 'outsiders'(2).

References

Checkland, P. and J. Scholes (1990) Soft systems methodology in action. Chichester: John Wiley & Sons.

Engel, P.G.H. (1995) Facilitating innovation. Wageningen: Agricultural University Press.

Long, N. (ed.) (1989) Encounters at the interface: a perspective on social discontinuities in rural development. Wageningen: Agricultural University Press.

Acknowledgement

The author would like to thank the FAO comunication project and all Guarani participants.