Epidemiology and decision-making in theileriosis control

B.D. Perry, P. Lessard, A.W. Mukhebi and R.A.I. Norval

International Laboratory for Research on Animal Diseases
P.O. Box 30709
Nairobi, Kenya

Several methodologies are currently available for the control of theileriosis (reviewed by Young et al., 1988). Immunization by the infection-and-treatment method (Radley et al., 1975) is the latest to be added to the list. The choice of a particular strategy by a farmer or a government should depend on the prevailing epidemiological and socioeconomic circumstances in a given area. Although infection and treatment offers the prospect of more widespread control of the disease, no single method is cost-effective or appropriate in all parts of a region in which the disease occurs. The severity of theileriosis, and hence the justification for controlling it, varies with differences in epidemiology, livestock management and farming systems. Young et al. (1988) have recently reviewed the methodologies available for tick-borne disease control and proposed an integrated approach to their application in Africa. This paper examines circumstances under which three basic theileriosis control methods might be used and proposes a conceptual framework with which a decision might be made on the best method to use in a given area.

CHOICE OF CONTROL METHOD

For the purpose of this discussion, disease control interventions and the characteristics of cattle populations have been simplified. Four mutually exclusive control options are considered, although in reality combinations of these would probably be used in many circumstances. These options are: immunization using the infection-and-treatment method, treatment of clinically affected animals with antitheilerial drugs (Dolan, 1981), the control of Rhipicephalus appendiculatus with acaricides, and no action at all. The consequences of these options (in this paper no action is considered an option) are measured in terms of disease incidence (the number of new cases in a given time period) and case-fatality rate (the proportion of new cases that die). It is assumed that immunization and vector control will influence disease incidence, whereas treatment will influence the case-fatality rate. Table 1 indicates the possible result of these simplified control actions in terms of disease incidence and case fatality rates in cattle populations of differing susceptibility. For the purposes of this discussion, it is also considered that immunization, treatment and vector control can be either effective or ineffective. The circumstances likely to govern the efficacy of each action are technical, logistic and socioeconomic.

Table 1. Possible effects of different theileriosis control actions

Action		Cattle population status
		Very susceptible	Partially susceptible	Low susceptibility
Immunize	Effective	Incidence low***	Incidence v. low***	Incidence v. low
	Ineffective	Incidence medium-high*	Incidence low-medium*	Incidence low
Treat	Effective	Case-fatality low***	Case-fatality low***	Case-fatality low
	Ineffective	Case-fatality medium-high*	Case-fatality medium-high*	Case-fatality low
Vector control	Effective	Incidence low***	Incidence v. low**	Incidence v. low
	Ineffective	Incidence medium-high*	Incidence low-medium*	Incidence low
Do nothing		Incidence high Case-fatality high	Incidence medium Case-fatality high	Incidence low Case-fatality low

* Indicates incremental change in status from "do nothing".

Immunization

A knowledge of the Theileria parva strains occurring in the area is necessary and there should be demonstrated protection against field challenge by the immunizing stock. There should be long-term availability of both the immunizing stock and the chosen antitheilerial drug. The vaccine should be affordable and there should be an effective and accessible veterinary service to administer and monitor the immunization. The lack of one or more of these requirements will result in ineffective immunization.

Treatment

Clinical cases of theileriosis must be promptly identified, because to be effective treatment must be applied early in the course of the disease (Dolan et al., 1984; Chema et al., 1986). Such prompt action will require a good diagnostic ability and high standards of management by the farmer as well as proximity to (or good communications with) effective veterinary services. Antitheilerial drugs should be available on a long-term basis and the treatment should be affordable. Again, the lack of one or more of these criteria will result in ineffective treatment.

Vector control

For controlling theileriosis by controlling R. appendiculatus with acaricides, the dipping or spraying should be carried out twice a week during periods of nymphal and adult tick activity. This generally requires a high level of management to be successful. Acaricide and water must always be available and the strength and efficacy of acaricide at delivery should be regularly monitored. The acaricide should also be affordable. Where these rather stringent requirements cannot be met, the vector, and hence disease, control will be ineffective.

ASSESSMENT OF POPULATION SUSCEPTIBILITY

A knowledge of cattle population susceptibility is clearly vital before choosing a disease control action. In general, this susceptibility will depend on the breed/type of animals present and on the challenge, particularly to young stock, by T. parva-infected ticks. A simple model of the effect of these parameters on disease incidence, case fatality and serum antibody prevalence rates is given in Figure 1. In circumstances of fairly constant challenge of indigenous Zebu cattle with infected ticks throughout the year, situations of endemic stability may exist where little or no clinical disease is observed (e.g., Moll et al., 1984; Young et al., 1986; Morzaria et al., in press). At the other extreme, exotic and grade livestock kept under even low R. appendiculatus challenge will be highly susceptible. The variety of situations between these extremes are determined by variations in genetic susceptibility of certain breeds or types and their crosses (Young, 1981; Dolan et al., 1982) and endemic instability in indigenous Zebu breeds where infected tick challenge is low or markedly seasonal (Moorhouse et al., 1986; D.L. Berkvens, personal communication). Situations are further complicated where livestock populations comprise both exotic and indigenous breeds and where different management systems (affecting tick challenge) are used within the same herd.

DECISION-MAKING

In the preliminary model presented, the selection of a control option and the susceptibility status of the target population provide an indication of expected outcomes. The assessment of prevailing circumstances, such as availability of veterinary services and standard of management, to evaluate expected efficacy of the intervention further defines the possible outcome. Table 1 shows incremental changes in incidences and case-fatalities that may result from the intervention when compared with doing nothing. This simple deterministic model demonstrates that control options should be tailored for given sets of circumstances and that the range of efficient disease control actions available is considerably narrower than at first appears.

Figure 1. Hypothesized relationship between disease incidence, antibody prevalence and case-fatality under increasing infected-tick challenge in two groups of cattle with different susceptibilities.

Where substantial data are available, this decision process can be enhanced by the use of decision analysis (reviewed in an animal health context by Erb, 1988). This technique evaluates the effect of alternative interventions on predetermined parameters and incorporates an assessment of risk. In this context, risk is taken as an evaluation of the probability of the successful (and unsuccessful) application of a technique. Decision analysis is often confined to the technical parameters describing and quantifying the effects of successful application of an intervention and uses data that come from previous research (such as the effect of immunization on disease incidence), but which, where definitive data are available, may be broadened to incorporate key logistic and socioeconomic parameters. However, many of the logistic and socioeconomic parameters do not lend themselves to quantification.

Figure 2 is a decision tree illustrating the alternative actions emanating from a decision "node". Each strategy then encounters a probability node. At this point it is assumed that the strategy has a given probability of a favourable outcome (P) or a mutually exclusive unfavourable outcome (1 - P). The assignment of probability values is based on experimental evidence or expert opinion. Where probability values are in doubt, the selection of probable best and worst (that is, a sensitivity analysis) can be made. The indicators of favourability of the intervention will be predetermined parameters. These parameters may include disease incidence rate, case-fatality rate, seroconversion rate in a given age stratum of the target population, or productivity (such as weight gain, milk yield and ploughing efficiency). Parameter values may then be translated into economic terms. The values of favourable and unfavourable outcomes (such as incidence rate and economic effect) are then multiplied by their respective assigned probability values. The resultant values for favourable and unfavourable outcomes are then summed to provide a net effect of the intervention studied. The same process is then applied to the other interventions under consideration and the net effect of each intervention compared. At this stage the most favourable interventions should be assessed for their socioeconomic consequences before final decisions are made. This simple framework for decision analysis can be modified to assess a variety of different outcomes, depending on the availability of data and the quality of decision required.

Figure 2. Simplified decision tree for estimating alternative East Coast fever control strategies.

REFERENCES

Chema, S., Waghela, S., James, A.D., Dolan, T.T., Young, A.S., Masiga, W.N., Irvin, A.D., Mulela, G.H.M. and Wekesa, L.S. (1986). Clinical trial of parvaquone for the treatment of East Coast fever in Kenya. Veterinary Record 118: 588-589.

Dolan, T.T. (1981). Progress in the chemotherapy of theileriosis. In: Irvin, A.D., Cunningham, M.P. and Young, A.S., eds. Advances in the Control of Theileriosis: Proceedings of an international Conference Held at the International Laboratory for Research on Animal Diseases, Nairobi, 9-13 February 1981. The Hague: Martinus Nijhoff, pp. 186-208.

Dolan, T.T., Njuguna, L.N. and Stagg, D.A. (1982). The response of Bos taurus and Bos indicus cattle types to inoculation of lymphoblastoid cell lines infected with Theileria parva schizonts. Tropenmedizin und Parasitologie 33: 57-62.

Dolan, T.T., Young, A.S., Leitch, B.S. and Stagg, D.A. (1984). Chemotherapy of East Coast fever: parvaquone treatment of clinical disease induced by isolates of Theileria parva. Veterinary Parasitology 14: 103-116.

Erb, H.N. (1988). The benefit-cost analysis of disease control programs. In: Lessard, P. and Perry, B.D., eds. Investigation of Disease Outbreaks and impaired Productivity. Veterinary Clinics of North America, Vol. 4. Philadelphia: W.B. Saunders, pp. 169-181.

Moll, G., Lohding, A. and Young, A.S. (1984). Epidemiology of theileriosis in the Trans-Mara Division, Kenya: husbandry and disease background and preliminary observations on theileriosis in calves. Preventive Veterinary Medicine 2: 801-831.

Moorhouse, P.D.S., Musisi, F.L., Mwase, E.T. and Snacken, M. (1986). The epidemiology of bovine theileriosis in Zambia: results of a longitudinal study in Southern Province. In: Proceedings of the 4th International Symposium on Veterinary Epidemiology and Economics. Singapore: Singapore Veterinary Association, pp. 389-391.

Morzaria, S.P., Musoke, A.J. and Latif, A.A. (1988). Recognition of Theileria parva antigens by field sera from Rusinga, Kenya [abstract]. The Kenya Veterinarian 12: 8.

Radley, D.E., Brown, C.G.D., Burridge, M.J., Cunningham, M.P., Kirimi, I.M., Purnell, R.E. and Young, A.S. (1975). East Coast fever. 1. Chemoprophylactic immunization of cattle against Theileria parva (Muguga) and five theilerial strains. Veterinary Parasitology 1: 35-41.

Young, A.S. (1981). The epidemiology of theileriosis in East Africa. In: Irvin, A.D., Cunningham, M.P. and Young, A.S., eds. Advances in the Control of Theileriosis: Proceedings of an International Conference Held at the International Laboratory for Research on Animal Diseases, Nairobi 9-13 February 1981. The Hague: Martinus Nijhoff, pp. 38-55.

Young, A.S., Leitch, B.S., Newson, R.M. and Cunningham, M.P. (1986). Maintenance of Theileria parva parva infection in an endemic area of Kenya. Parasitology 93: 9-16.

Young, A.S., Groocock, C.M. and Kariuki, D.P. (1988). Integrated control of ticks and tick-borne diseases of cattle in Africa. Parasitology 96: 403-432.