Specificity of Disease Terminology
Interpretation and Utilisation of Results
Confirmation of Results
In disease control and eradication programmes, surveillance is the key to success. Effective surveillance is essential to insure the appropriateness and timeliness of interventions, whether they are movement control, treatment, vaccination or stamping out. Thus, the effectiveness of surveillance is a major determinant of the benefit-cost equation. A multifaceted surveillance programme is essential to ensuring the sensitivity, specificity and timeliness of the overall surveillance activity.
For diseases that are of popular concern, participatory disease search is a sensitive method that should be employed as part of active surveillance. Livestock owners, particularly pastoralists, are relatively adept clinical diagnosticians and can readily recognise major disease problems present in their area. Questionnaire and participatory techniques are excellent methods for assessing the epidemiological risk level of an area or for tracing owner reports back to (or forward to) active outbreaks for laboratory investigation.
In apparently disease free areas, participatory disease searches can be used to confirm the absence of disease. Information gained through participatory disease searches on husbandry systems, grazing patterns and trade practices can be used to establish epidemiological risk levels for different disease introduction scenarios. Concerning endemic areas, data collected on livestock movement, animal contact and service delivery can suggest mechanisms of endemicity. This information can form the basis for the design of more appropriate control strategies. Participatory investigation can also help to establish the risk levels of disease spread out of endemic areas.
The participatory disease search begins with collection of background data on the community from secondary sources. Just as with general disease surveys, available literature and key informants should be utilised to obtain an overview of the community. Decision making structures, herding groups and interview strategies need to be identified.
In participatory disease search, it is very important to avoid tarmac bias. It would be better to introduce an isolation bias (see Sources of bias on p. 6). Animal health services are usually more accessible near roads and district towns. If one is really searching for disease - hunting for cases - try to identify the lacunae in service delivery. Check secondary sources and ask key informants for the communities that lack health posts or almost never receive vaccination. In preliminary interviews incorporate mapping exercises and ask about remote communities and grazing areas. Make a list of isolated communities and try to fit them into your itinerary. Be prepared to camp and to walk.
The investigation team should carry along a basic epidemiological field kit as described in Annex 1.
As with other participatory appraisals, a checklist should be developed which acts as a reminder for the appraisal team as to what subjects must be covered. In the case of disease searching, one is looking for very specific information. It is extremely important that this narrow objective is not communicated to the respondents. The interview should be introduced as an animal health study and begin with a broad animal health question. One can focus on the species of interest from the outset, but not the disease under study. For example, if one is searching for CBPP, one could introduce the study as a general appraisal of animal health issues concerning cattle. A sample PDS checklist is presented in Box 2.
Box 2: Sample PRA Participatory Disease Search Checklist
Do not introduce the target disease as a subject of discussion. Discuss general disease issues until the respondents raise the target disease as a problem.
1. Introduce the PDS team as a group conducting an animal health appraisal.
If by the end of the interview the target disease has not been
mentioned, the PDS team can ask directly about the disease in question. However,
it is significant that the respondents did not identify the disease as a problem
and the notes should clearly indicate that the PDS team introduced the
This process of systematically working through current disease problems and major historical disease problems should require between 30 and 45 minutes. However, it is essential that you do not communicate your interest in a specific disease early in the interview. This general approach has the advantage that you will be able to understand the importance of the target disease in relation to other animal health problems. Also, the general information will give you background and allow a better appraisal of the quality of information provided by the respondents. Finally, it takes a few moments for participants (the appraisal team and respondents) to relax and establish confidence. When the key issue is finally addressed, it will be just one more topic in a broad discussion. The respondents will state their knowledge on the subject without attempts to impress, please or lure the appraisal team.
Once the disease of interest is mentioned, the appraisal team should ask the respondents to describe the disease. If the description is accurate and more or less consistent with the descriptions provided by other groups of respondents, the appraisal team can proceed to inquire about occurrence of the disease. The respondents should be probed on the following themes, if they have not already volunteered the information in their descriptions:
1. Have they had personal experience with the disease or did they learn about it from others?If the respondents have not had direct personal experience, their comments are much less significant as it largely reflects what they heard from others. In this case, their comments should be noted as hearsay or rumours. Such information may provide useful leads for further investigation, but should not be utilised as core data for epidemiological analysis.
2. If they have had personal experience with the disease, when, where and in whose cattle did they observe the disease?
3. What were the general circumstance at the time of the event (grazing conditions, water availability, security conditions, livestock contacts with other communities and wildlife, trade links, etc.)?
4. Has the disease occurred at any other time? Repeat questions 1 to 3, as time permits, for each previous occurrence of the disease of interest.
Only specific reports, which give dates, locations and specify contact herds by name, or community, should be used. Vague statements that were not supported by specific plausible accounts detailing animal contact and clinical descriptions should be discounted. Accounts should be assessed for both internal consistency and cross-checked with other independent reports.
Descriptions of direct observation of disease, especially in the respondents own cattle, are significant statements. This information forms the core data of the analysis. First-hand descriptions should be noted and tabulated. Later, this data can be analysed for trends.
Concerning assessing reports, the following weighting criteria may serve as an example for the categorisation of volunteered reports from individuals who can accurately describe the disease:
1. A first hand report from a herder of cases in his own herd.The individual reports that are internally consistent from categories I, II and III can be tabulated and mapped to build a consensus view of the historical and recent incidence of the disease in the region.
2. A first hand report of clinical cases observed by a veterinarian or veterinary assistant.
3. A report of cases directly observed by a cattle owner but in others cattle.
4. A second-hand report or hearsay from veterinarians, herders, public officials or elders who did not actually see the disease. This last information should be noted and may be used as leads to assist the team in selecting locations for further study, but should not be used as data in the construction of maps and time lines or other forms of analysis.
Key informants who are not veterinarians or cattle owners often provide very important information and leads. The character of the informant and his community reputation in regard to livestock knowledge are useful guidelines to assessing the quality of information. Further on, one can test the information provided by key informants in livestock owner interviews.
The disease search team must carefully define the terms used by the community in relation to the disease being studied. There may be a specific term for the disease in the local language that can be translated directly. If the disease has different clinical forms such as severe and mild or acute and chronic, the community may have different terms and be able to distinguish between these. Alternatively, the community may only be able to recognise one clinical form and other forms may pass as undifferentiated disease.
For some clinical syndromes, the community may use a non-specific term that may include several similar diseases.
The best approach to this issue is to prepare case definitions for each term that the community is using relative to the disease of interest. There may be some overlap of terms and one disease term may be contained within a broader term. This is an area where Venn diagramming may be of assistance. The disease search team can then compare the communities case definitions with accepted case definitions used in western veterinary medicine.
The information collected through interviews and exercises will contribute the building up of time lines, maps and scenarios of clinical disease occurrence. In addition, information on risk factors and the clinical and epidemiological behaviour of the target disease in the communitys livestock will be assembled.
The appraisal team must be able to organise the data to demonstrate identifiable trends for the study to be meaningful. Examples of trends could be stabily endemic disease with small numbers of young stock being affected each year or epidemic waves of disease passing through locations. If patterns cannot be identified in the results, the appraisal may not be meaningful.
The ability to diagnose disease is dependent on the nature of disease entity being searched for, the existence of pathognomic signs or lesions, and the similarity of differential diagnosis. For most diseases, it is not possible to arrive at a final diagnosis based on the clinical description of one or a few cases. There are epidemiological methods for categorising and linking cases that can strengthen the index of suspicion and in some cases make a definitive diagnosis possible. The following categories may be useful in the analysis of results:
Clinically compatible events: Cases or outbreaks that meet the case definition for the target disease. These may also be termed suspect events.The concept of epidemiological diagnosis is very powerful. In medical epidemiology, it is accepted that cases epidemiologically linked to confirmed cases are themselves confirmed (CDC, 1990). For some diseases (i.e. varicella, mumps, etc.), two or more epidemiologically-linked and clinically compatible cases in the absence of laboratory confirmation are accepted as confirmed cases (CDC, 1990; CDC 1998). Another useful category is probable events. These are clinically compatible case for which a high index of suspicion exists. In probable events, the burden proof shifts from proving that the event is in fact the disease in question, to proving that it is not.
Epidemiologically linked events: Clinically compatible cases or outbreaks that are temporally or spatially related. Either a chain of transmission can be shown or reasonably assumed.
Epidemiologically characteristic events: An aggregate of clinically compatible cases or outbreaks that behave in the population in a manner consistent with the disease in question. Knowledge of the behaviour of the clinical syndrome in populations (the epidemiology of the syndrome) is needed before a syndrome can be categorised as epidemiologically characteristic.
Unfortunately, international veterinary epidemiology as it is applied today has not reached the same degree of sophistication, as has medical epidemiology. Although we have international reporting requirements for List A and List B diseases, we do not yet have sophisticated case definitions and clinical diagnostic guidelines for these diseases. Case definitions and diagnostic criteria require a significant amount of thought, consultation among concerned professionals, and field-testing. A clinical outbreak definition is under discussion for stomatitis-enteritis and will be incorporated in the next section. For other diseases, concerned experts are encouraged to begin formulation exercises.
If participatory disease searches lead to the detection of active disease, it is recommended that the cases be appropriately sampled for laboratory confirmation. If the disease search has been thorough and representative cases have been confirmed in the laboratory, then an extensive chain of epidemiologically linked, confirmed events may have been established. Laboratory confirmation will also confirm the interpretation that the appraisal team has made of the communitys terminology.
In the absence of laboratory confirmation, a positive disease search will demonstrate a series of disease compatible events that exhibit trends. Depending on the nature of the disease and the nature and extent of EVK, the appraisal may be able to argue that the data represent a probable or even confirmed occurrence of disease. The appraisal will need to substantiate its arguments based on the known clinical and epidemiological behaviour of the disease and draw on epidemiological and clinical diagnostic techniques.
McCracken (1988) has suggested that the two uses of the results of a topical RRA are as a working hypothesis for action-oriented development programmes or as a research hypothesis for further investigation by other methods. This would suggest that the results of PDS are applicable as working hypothesis to guide the activities of disease control programmes or as hypotheses on the basic epidemiology and ecology of disease for testing in academic research.