Symptomatic identification of malaria in the home and in the primary health care clinic

In endemic areas in the absence of microscopy, the WHO case definition of malaria is the presence or a history of fever without other obvious cause. Yet there is little empirical evidence on the accuracy, predictability and reliability of clinical signs and symptoms for diagnosing malaria within different endemic settings. Studying patients in endemic communities in the Philippines, we found that fever alone did not discriminate well for malaria. In contrast, a sequential occurrence of fever, chills and/or sweating, or a combination of all three symptoms was a good general predictor of the disease. However, the place of diagnosis and observation (home or clinic), age, and season affected the positive predictive values obtained. Specificities and positive predictive values were greatest (over 80%) for those at most risk - children under 9 years of age in highly endemic communities - and were most reliable when the diagnosis was made at home. Predictive values were also greatest during the season when childhood acute lower respiratory infections in the study area increase. The good predictability of clinical signs and symptoms for high-risk groups suggests that simple protocols can be developed for the management of malaria in endemic areas of the Philippines. Over the past 30 years, the mountainous area of Kalinga Apayao Province on Luzon Island in the Philippines was extensively deforested due to slash and burn farming. The malaria risk is reduced, but malaria is still endemic. During 1990-1992, morbidity surveys identified 614 malaria cases. Researchers wanted to determine the accuracy, predictability, and reliability of clinical signs and symptoms for diagnosing malaria. Most individuals (89%) claimed to have had fever, yet just 35.4% had a body temperature greater than 37.6 degrees Celsius. Only 51.8% of fever cases had parasitemia, indicating that the World Health Organization's recommended case definition of malaria (i.e., presence or history of fever) did not adequately identify malaria. Further, prior to this study, about 50% of the area's children were usually infected, but just 11.3% of children younger than 6 in this study had fever. A good general predictor of malaria included a sequential occurrence of fever, chills and/or sweating, or a combination of all 3 symptoms. The positive predictive values were: at-home observation and diagnosis (74-76% vs. 69-72% at the rural health clinic), age younger than 9 years (80-84% vs. 65-69% for = or 10 year olds), and presentation during November-January (94-100% vs. 74% for February-October). November-January was the season when the prevalence of acute lower respiratory infections was highest. These results demonstrate that health workers can develop simple algorithms with good predictability of clinical signs and symptoms for high-risk groups to manage malaria in endemic areas of the Philippines.