General discussions and recommendations
Following the conclusion of the presentation sessions, workshop participants divided into two discussion groups. The first group discussed modelling of three biological processes involved in vector-borne and other parasitic diseases, namely immunogenetics, parasite transmission and livestock production and ecology. This group also considered issues relating to modelling science in general. The second group took a different approach and considered the modelling needs within the three major disease complexes of relevance to the workshop, tick-borne diseases, trypanosomiasis and helminthiasis. The two groups were then reunited to rationalize their deliberations and develop appropriate recommendations.
Discussion series I
There was a vigorous discussion on the approaches to and possibilities of modelling infection, immune responses and vaccine design. The meeting then focused on the modelling of Theileria parva infection within the host and a simple model for the cytotoxic T. lymphocyte (CTL) response was outlined. The possible effect of a vaccine based on a sporozoite antigen (p67) of T. parva, to which the response detected is thought to be antibody mediated and mainly infection blocking, was discussed. The case was put strongly that it was possible to quantify rates of infection and immunity processes and quantify the relationships between processes. Models can and have been developed for infection and immune responses and they can be valuable aids in exploring vaccine approaches. Even if models are not particularly accurate initially with regard to infection and immunity, they are powerful tools for exploring and explaining mechanisms, and raising questions to be answered by experimentation. It was recommended that ILRAD explore modelling approaches to evaluate immune responses to T. parva with a view to supporting vaccine development strategies.
The double expresser model for immune response-driven antigen variation in trypanosome infections, presented in the Host-Parasite Interaction session, was discussed in relation to rates of development and possible differences in responses between N'Dama and Boran cattle. The in vitro expression of antigenic variants could provide important information invalidating the theoretical basis of this model.
Differences in parasitaemia profiles between N'Dama and Boran cattle were described by ILRAD scientists and questions were raised as to whether these differences would be due to differences in parasite replication rates. This could be tested in the model.
Further discussion on trypanosomiasis and the responses of trypanotolerant cattle populations identified a range of variables which could and should be modelled, such as primary response affinity of antibodies and rates of development of antigenic variants. The task of obtaining the full range of antigenic variants of any particular stock (or clone) for the latter was considered daunting.
There was some discussion amongst immunologists as to the likely mechanisms underlying resistance to trypanosomiasis in different cattle populations.
Significant improvement in understanding of the role of a modelling approach and its potential for investigating and explaining biological problems at a molecular or cellular level was an important outcome of the workshop and a closer interaction between biologists and modellers was encouraged by all participants.
The discussion started by considering whether it would be possible to model the important components of parasite transmission in both tick-borne diseases and trypanosomiasis with a single model. The view was expressed that this could be achieved for many aspects of the transmission of vector-borne infections, simply by varying a standard set of parameter values in the model. The discussion then considered what types of problems could be addressed by such 'generic' models.
The view was expressed that these could effectively examine the comparative importance of different parameters in the dynamics of disease transmission (such as the relative importance of carrier status in determining tick infection rates under different conditions of endemic instability in T. parva infections), but that they would not address some of the more fundamental questions which constitute much of ILRAD's research. For these questions (such as the level of vaccination coverage required to protect different populations under specific conditions), more complex simulation models may be necessary.
Ways of further developing and validating the presented model for T. parva (developed by Medley and colleagues) were discussed. It was recommended that available data sets of theileriosis dynamics from eastern Zambia (Berkvens) and Zanzibar (Woodford) should be applied to the model, along with future data sets currently under collection in coastal and highland Kenya.
The suitability of the model T3HOST as another starting point for modelling tick population dynamics was discussed. Some of the group were or the opinion that it was the most realistic of the available tick dynamics models, but its lack of general availability constrained independent verification of that by modellers. Doubt was expressed in some quarters that it would serve as a panacea for modelling tick dynamics, let alone parasite transmission.
Possible collaborations were then discussed. There was an optimistic sentiment expressed by the modellers that existing comparative approaches in use for other vector-borne infections could easily be applied to tick-borne diseases, and various ways of achieving this were tabled. These included the setting up of specific collaborative arrangements for different processes with particular modelling groups and the option of widely distributing available data to all modellers, who would then each explore them with their own approaches.
LIVESTOCK PRODUCTION AND ECOLOGY
The group recognized various areas that require strengthening if modelling systems are to be used to support the development of sustainable agricultural production systems in developing countries. There is an immediate and priority need to address the lack of data in certain areas; to quantify the economic losses due to morbidity and mortality from target diseases; to identify appropriate epidemiological studies, particularly in regard to disease transmission levels in various ecological situations; and to identify the availability of data sources that may be useful to modellers.
In order to address these areas, it was recommended that FAO integrate more closely the activities within the Animal Production and Health Division and, in collaboration with ILRAD, ILCA and others, further computerize its data management systems making appropriate use of GIS, remote sensing and modelling.
It was also recommended that FAO/ILRAD/ILCA and other relevant international agencies, in collaboration with national institutions, support the adaptation of existing models on epidemiology and control of parasitic diseases and livestock management to African conditions.
The group considered that these organizations should promote and provide support for strengthening the capabilities of national veterinary services in data gathering and where possible standardized equipment and procedures should be adopted.
Finally, it was recommended that FAO examine ways to retrieve and collate historical data on livestock production systems making databases mote accessible for modelling.
The perceived modelling requirements of ILRAD and FAO were discussed. Both organizations clearly have a variety of needs for modelling, but these need to be prioritized. ILRAD participants identified some specific needs and examples of these are:
· immune responses against trypanosomes and Theileria;
· antigenic variation;
· the pathogenesis of anaemia;
· transmission dynamics of parasites;
· host-parasite interactions;
· disease impact assessment; and
· evaluation of alternative control options.
FAO identified other needs, particularly in the area of livestock production models and the delivery of interventions. The group considered that there is scope for ILRAD and FAO to collaborate further in the design of decision-support systems and control strategies based on models. It was recommended that FAO gives priority to the development and transfer of models which will directly enhance livestock production, and that FAO promotes awareness of the value of proven modelling systems.
The review of modelling procedures appropriate to meet these requirements were considered and it was the general feeling that analytical procedures appear to be appropriate for parasite transmission and immune response studies, whereas simulation/database/spreadsheet models are more suitable for the areas of livestock production, vector population dynamics and disease control. In an attempt to identify relevant approaches, it was stressed that simple approaches should be undertaken initially which could increase in complexity as the problem and solution demanded. It was considered very important to have modelling under consideration at an early stage in project development. It was recommended that ILRAD and FAO consult with other centres of excellence in developing and promoting models.
It was noted that model development is time consuming and is undertaken against a background of rapidly changing technologies. Solutions and methods will have a finite generation time. It was recommended that different techniques be used to address similar problems and that the use of generic models be explored where possible in order that effort invested in modelling one disease may be applied to others.
Data requirements are often only known once model construction has been initiated. The group considered that models should be initiated where possible on the basis of existing data. For model development, database collection should focus on spatial and temporal data that will enable models to be tested over a range of circumstances. It was recommended that FAO considers ways of making digital databases more widely and easily accessible.
The group noted that there are issues concerning ownership, accessibility and preservation of data which need to be addressed.
It was agreed that collaboration in model development, in terms of modellers and biological scientists, and in terms of ILRAD, FAO and other collaborators, should be initiated early in any project preparation phase. Most collaboration arises from individual contacts and in particular where there is a critical mass of resources. To assist this, some participants considered that the establishment and dissemination of registers of modellers and models, complete with status reports of model development, could be a valuable role for FAO. Finally, with respect to modelling approaches, it was recommended that FAO play a convening role in training people in the use of models, and contribute to the provision of resources for such training.
Discussion series 2
It was agreed that if good quality data are available it should be possible to model many aspects of the immune processes and the response to immunization. Models should, therefore, be further developed and attention focused on how to quantify processes in the immune response and the relationships between these processes.
Models have been developed, or are being developed, for relevant aspects of parasite transmission. In the face of changing tick and TBD control strategies and policies, this should be given priority.
It was the general opinion among the modellers that all models should be as simple as possible initially, growing in complexity as components are understood and others identified. This could eventually lead to the integration of disease transmission models, vector population dynamics models and livestock production models which when combined will form the basis for interactive decision support systems.
In discussing the deliberations on livestock production (above) it was considered necessary for governments to establish national priorities (in terms of diseases and other constraints) for future priority attention. The priorities should be based on benefits to be gained from available and potential control measures rather than value of theoretical losses. Specifically, in the case of TBDs, there is an urgent need to quantify both direct and indirect production losses (for both morbidity and mortality and the inability to use land or introduce improved cattle).
At all levels, both modellers and statisticians should interact with scientists in both the laboratory and the field. It was emphasized that for progress to be made, there is a need for greater understanding and respect of each other's disciplines.
There are certain specific research areas in ILRAD's trypanosomiasis research program that could be addressed through a modelling approach. The view was expressed that at present the development of diagnostics was not a priority area for modelling. Likewise it was considered that modelling of vaccine development should be left for another forum to discuss, given the limited progress in research on vaccine development.
The interest, therefore, focused on the possible modelling of chemotherapy and trypanotolerance which is of relevance to the following research areas.
· Trypanosome genetics, including relevant areas in chemotherapy, drug resistance, and marker identification.
· Parasitaemia and other aspects of trypanosome infection dynamics within the blood stream remains unclear and therefore controversial. Several of the modellers expressed an interest in working with these data.
· The role of trypanosome immunology in trypanosome population dynamics is still uncertain. It was suggested that research in the immunology field could benefit from modelling expertise outside ILRAD. Since there is a variety of alternative hypotheses on immune mechanisms, it might be best to consider collaboration with more than one group/individual.
In order to obtain a more complete picture of the epidemiology of trypanosomiasis, it was suggested that data from several representative field sites should be collected to study the relative role of different variables and parameters play in contributing to variations in infection transmission rates. This was considered by some to be a relatively easy exercise intellectually, but will clearly involve considerable thought and commitment at the field sites. The obtained data could be used subsequently for the construction of a model describing trypanosomiasis transmission at each field site.
The importance of more effective monitoring of livestock production systems in Africa was stressed and it was agreed that the ILCA/ILRAD Trypanotolerance Livestock Network could be utilized and possibly act as a prototype for similar networks for collecting and analysing data on trypanosomiasis impact to livestock.
Through the modelling of livestock production systems, it should be possible to develop decision-support systems (DSS). It is important to realize that DSSs can be targeted at individual farmers, communities of farmers, governments and donors. Recommendations from a DSS to farmers may not necessarily be in the interests of ecosystems or societies on a long-term basis. It was considered that the Socioeconomics Program of ILRAD could make a major contribution to the development of DSS appropriate for these different client groups in the control of livestock diseases in Africa. It is also vital that national research and implementation organizations are involved at an early stage as partners in this development, and not just as recipients of the products of DSS analysis.
The importance of managing resources for sustainable production was discussed and it was agreed that FAO clearly has an important role in developing and promoting activities which would result in sustainable solutions to resource management. In this FAO has to address and resolve the conflict between the producers (as advised by the DSS analysis) and consumers (not only the broad society within which the producers live and sell their produce, but the world community). In the short term FAO is interested in the collection, collation and analysis of existing data sets and in developing the methodologies for analysing those data to establish the constraints and environmental impact of tsetse, trypanosomiasis and tsetse control throughout Africa.
Initially the discussion focused on the currently available models for the management of helminth control programs and the possible adaptation of these to African conditions. Some of these models (analytic differential equation models) are used for anthelmintic management strategies and they are not site specific or species specific, nor do they have a livestock management component. Others are simulation models. The modellers expressed their optimism regarding their future adaptation if sufficient databases are available.
The rapidly developing problem of anthelmintic resistance, which has reached emergency proportions in some areas of the world, was debated. Modelling in this area has been initiated in several institutions and existing models might be useful in the African context to predict the consequences of indiscriminate use of anthelmintics. It was recommended that this be pursued by FAO.
Among the alternatives to the use of chemicals to control helminths is breeding for genetic resistance to helminth parasites. Work is under way in several regions comparing the resistance qualities of indigenous breeds with those of exotic breeds; initial experiences in Australia indicate that modelling will be a useful tool in this discipline.
It was agreed that more knowledge was needed regarding the impact of helminth parasites in goats and modelling could be better utilized in order to obtain more understanding on the population dynamics of goat parasites.
The understanding of the interaction between different parasitic diseases in any one host is limited. The need for establishing data in this field and the potential use of modelling was discussed.
The importance of being able to integrate the models on helminth epidemiology and control with models on livestock production was stressed.
It was recommended that FAO support the ongoing development of models for determining the minimum effective use of anthelmintics in strategic control programs and the prevention of the development of anthelmintic resistance.