Malaria epidemics hit Ethiopia every few years, killing many non-immune people. It is difficult to forecast these epidemics, especially in highland areas which normally have little or no malaria. Research carried out at Erasmus University in the Netherlands assessed different methods for predicting malaria outbreaks based on the disease's pattern of past incidence.

Epidemic early warning systems are badly needed in areas with normally low levels of malaria transmission. They can help health services prepare to deal with the sick and can direct prevention measures. Changes in weather conditions may lead to major epidemics in these areas. So efforts are underway to develop early warning systems that use weather records, climate forecasts and other factors. However, if they focus only on climate and ignore malaria and human population dynamics, they are unlikely to be very accurate.

In this study, researchers explored whether it is possible to predict malaria incidence just from past patterns of disease. They compared five methods using monthly records from health facilities in 20 areas in central and north-western Ethiopia. They found that ‘seasonal adjustment’ is the most accurate method. This method compares data from the last three months with seasonal averages to detect any changes from normal. It performs as well or better than more complicated statistical techniques.

They also found that:

• Malaria transmission is highly variable in most highland areas from season to season and year to year.
• It is easier to predict malaria incidence over a shorter rather than a longer time period.
• Seasonal adjustment is most accurate when using disease pattern data from the past three years.
• It is more difficult to predict incidence during the dry season (especially December to February) than the rainy season (June to August).

The researchers point out that problems with health service data make it difficult to estimate the true number of cases. For example, health facilities may operate above their normal capacity during epidemics. In addition, reported data do not include cases in remote rural areas receiving house-to-house treatment from travelling health workers.

Seasonal adjustment accounts for both normal seasonal changes and recent trends. However, it is not yet accurate enough for use in an epidemic early warning system, especially during the dry season. The researchers provide a simple description of the method which may be used or adapted by malaria control programmes in the absence of better ways of predicting epidemics. But they emphasise the need for a more accurate forecasting system that combines the past pattern of disease with the use of other predictors, such as temperature and rainfall.

Source(s):
‘Forecasting malaria incidence from historical morbidity patterns in epidemic-prone areas of Ethiopia: a simple seasonal adjustment method performs best’, Tropical Medicine and International Health 7 (10): 851-857, by T. Abeku, et al., 2002
HINARI subscribers can access the full-text article here. Full document.

Funded by: Department of Public Health Erasmus MC, The Netherlands; Netherlands Institute of Health Sciences; Trust Fund of Erasmus University, The Netherlands; and World Health Organisation

id21 Research Highlight: 26 March 2003

Further Information:
Tarekegn Abeku
Disease Control and Vector Biology Unit
Department of Infectious and Tropical Diseases
London School of Hygiene and Tropical Medicine
Keppel Street
London WC1E 7HT
UK

Tel: +44 (0)20 7612 7861
Fax: +44 (0)20 7580 9075
Contact the contributor: tarekegn.abeku@lshtm.ac.uk

Erasmus University, The Netherlands

Other related links:
'Controlling malaria in times of emergency: East Timor’s experience'

'Control panel - tools to prevent malaria epidemics in highland Africa'

See id21's collection of links relevant to infectious diseases.