| Forecasting, warning and detection of malaria emergencies: a case study. |
| Hay S.I. et al. |
| The Lancet 2003 361:1705-1706 |
| Full Text |
 |
 |
 |
|
| 4 June 2003 |
|
| Commentary by Jonathan Cox jonathan.cox@lshtm.ac.uk and Tarekegn Abeku tarekegn.abeku@lshtm.ac.uk |
|
| Increasing
preparedness for malaria epidemics is a central tenet of
the Roll Back Malaria initiative. This translates to
specific goals of detecting 60% of malaria epidemics
within two weeks of onset and responding to 60% of
epidemics within two weeks of detection. Malaria early
warning systems (MEWS) are widely advocated as a
mechanism to achieve this preparedness in sub-Saharan
Africa. Hay et al.'s paper in The Lancet,
which summarizes a more detailed UNICEF report,
evaluates various indicators for epidemic forecasting,
warning and detection using the case of the 2002 malaria
emergency in Kenya. Results indicated that only rainfall
monitoring [through remotely sensed rainfall estimates
(RFE)] offered any sort of value in terms of early
warning (although the utility of other weather variables
was not evaluated). Questions concerning the predictive
skill and applicability of MEWS across a range of
malaria endemicities remain, but few would argue with
Hay et al.'s notion that simple analyses of
rainfall are informative in areas of unstable malaria
transmission.
|
| We may soon have the
opportunity to prospectively test the reliability of
rainfall monitoring data as an early warning indicator.
Recent RFE data indicate that many parts of East Africa,
including epidemic prone areas of Uganda, Kenya and
southern Ethiopia have experienced abnormally high
rainfall since March of this year
(http://edcw2ks21.cr.usgs.gov/adds/). Although Hay et
al. are sceptical about the value of seasonal
climate forecasts, it is worth pointing out that current
forecasts for the area predict a 40% chance of higher
than normal precipitation and a 55% chance of above
average temperatures for May-July (the chance of
near-normal conditions being 35 and 30% respectively)
(http://iri.columbia.edu/). With warm and wet conditions
likely to coincide with the traditional peak malaria
transmission season, the occurrence of malaria epidemics
in the coming months would not be unexpected. We would
echo Hay et al.'s call for increased epidemic
preparedness - and particularly the need to ensure
adequate supplies of antimalarial drugs. In addition to
RFE data, it would be prudent to also track ground-based
meteorological data for epidemic prone areas. Hay et al.'s results suggest that monitoring meteorological variables can provide planners with an objective, if relatively rough, means of epidemic early warning. However, for effective intervention this approach needs to be coupled with a surveillance system that can locate incipient epidemics in space and time in a sufficiently timely manner. In developing such a system, a balance needs to be reached in which it is possible to collect sufficiently specific malaria data on a weekly basis without over-burdening already stretched health facilities with increased administration and form-filling. In East Africa, the HIMAL project - a collaboration between the Kenyan and Ugandan Ministries of Health (www.HIMAL.uk.net) is currently testing a system for epidemic early detection (based on a network of sentinel health facilities) and a prediction system in a number of epidemic prone districts. Results from HIMAL and similar studies should help shed more light on precipitating factors for epidemics in African highlands and allow us to improve on the rather rudimentary approaches currently being touted.
|
|
|
| Articles will be freely available for a promotional period until 31st March 2003. |