An Evaluation of an Early Warning Alert and Response Network (EWARN) in Darfur, Sudan
/ 0
- Details
- Category: Vol. 81, June 2013
- Hits: 1311
An Evaluation of an Early Warning Alert and Response Network (EWARN) in Darfur, Sudan,DIANE F. MOROF, ALAA ABOU-ZEID and MUIREANN BRENNAN
Abstract
Objective: To conduct a field-based evaluation of an Early Warning Alert and Response Network (EWARN) in Darfur, Sudan.
Methods: Using adapted surveillance evaluation guide-lines, evaluators reviewed EWARN documents and conducted semi-structured in-depth interviews and group discussions with key informantsat national, state, and local levels. Eval-uators conducted visits at 18 purposively sampled clinics in all Darfur states. Observers examined morbidity reporting, laboratory functions, and disease controland nutrition data. Qualitative and quantitative analysis identified common themes and examinedkey variable frequencies.
Results: All clinicians described EWARN as useful; most indicated that its primary usefulness was early outbreak detection. Between January and October 2009, there were a total of 30 alerts with 10 confirmed outbreaks, 16 negative results, and four results with pending laboratory tests. Of the 26 alerts with investigation results, 10 were confirmed (positive predictive value [PPV] = 38%). The sensitivity of the outbreak detection system could not be determined on the basis of available data. Lack of clarity and variations in the application of case definitions and laboratory testing led to differences in reporting of specific conditions and rendered trend data less reliable. Collecting data on non-epidemicprone diseases at every site was burdensome. Few deaths were reported at the clinic level.
Conclusions: EWARN is a useful system for outbreak detection. Refining, standardizing, and increasing training frequency on case definitions, expanding laboratory capacity, and focusing data collection on epidemic-prone diseases would greatly improve the system's outbreak and surveillance ca-pacity. Mortality reporting from outpatient clinic data should be eliminated.