Introduction: Early detection and case management remain the main strategies in malaria control programme (MCP) in a low endemic area such as in Java Island, Indonesia. These strategies require an understanding of the community’s care-seeking behaviour in relation to the various health services. Since most malaria cases in Java are diagnosed at home by the village malaria workers (VMWs), this study aimed to examine the user-provider interface in early detection and case management of malaria, particularly the interaction between the clinical malaria patients and the VMWs. Methods: The number of blood slides examined and the laboratory results over a 20-year period were retrieved from the routine malaria surveillance system. The population at risk of malaria and the rainfall data were obtained from secondary sources. In addition, age, sex, malaria species, types of drugs, drug and dosage and time lapse between slides taken and examined were recorded from the malaria registers at the three endemic health centres from 1994-1998. The quality of diagnosis was examined by re-reading 153 slides at the Faculty of Medicine, Gadjah Mada University. Prior to using rapid assessment procedures (RAP), we proposed 11 criteria and applied these to 15 published RAP studies in health. For each of the papers, two authors assessed the adequacy of information provided independently. Using the criteria as a guideline for developing a protocol, a RAP study of malaria was thereafter carried out. Data were retrieved from 38 free-listings, 28 in-depth interviews, seven focus group discussions and unstructured observation. A qualitative thematic content analysis was applied. Finally, based on the RAP results, a one-year longitudinal study of care-seeking behaviour of all clinical malaria cases treated by the VMWs was conducted in Mlonggo II area. Age, sex, daily actions and date of consultations were recorded by all VMWs in a diary that covered prior all contacts between the patients and the VMWs. Also, 24 interactions between the VMWs and the patients were audiotaped. Results: The incidence of malaria during the 1990s fluctuated, albeit at the lower level than that of the 1980s, and it reached a peak (3.5/1000 population) during an outbreak in 1996-97. There was no clear association between the El Nino phenomenon and incidence of malaria. The incidence was almost twice as high in children <15 years than in adults (15+ years). In <5 year old children the risk of P. vivax malaria was higher than the risk of P. falciparum. Comparisons between active and passive case detection (ACD and PCD) showed that almost 60% of 10, 493 confirmed malaria cases in Jepara were identified from ACD. ACD also detected significantly higher percentage of P. falciparum gametocyte infections than PCD (14.7% vs. 5.7%). The duration between slides taken to examination was however longer for ACD than for PCD (2.3 vs. 1.1 days). Applying the criteria to the published RAP studies, it was found that information was limited to address subjectivity, staff and ethics criteria. In Jepara, malaria (known as katisen or panas tis) was considered a common but minor illness. This was also reflected by the most common action taken by the patients, i.e. not doing anything. However, when the illness was perceived as important, the community had a good access to different health care providers. Eighty seven percent of cases had been treated by the VMWs on day four of the illness period. On day two, the proportion not treated was significantly higher in male than in female cases (60.7% and 54.6%; p 0.01) and in those <15 years of age compared with those 15+ years (71.3% and 56.9%; p<0.001). Insufficient understanding of malaria signs and symptoms likely leads to delay in illness recognition and treatment. Interactions between the VMWs and the patients were mostly focused on medical tasks, and low compliance with treatment was a common concern of the VMWs in the interaction. Conclusion: El Nino phenomenon was not associated with an epidemic in Jepara. A possible association between age and the risk of P. vivax malaria needs further investigation. In this decentralised health care system, ACD should be continued in a focus endemic area and therefore, efforts to retain the VMWs should be considered. This research showed similar findings between the RAP study and the longitudinal study. A consistent gap was found between the common understanding and the biomedical description of malaria. The performance of the VMWs supports the MCP through early contact with clinical malaria patients. Visits of VMWs within four days of symptom recognition appear to be the ideal situation for both the programme and the community. If case management continues to be the main strategy in MCP in this low endemic area, the emic perspective of the people must be well integrated to improve home treatment. Likewise, simple interventions to strengthen the role of VMWs in home management should be conducted.