Distribution and spatial analysis of public healthcare in Indonesia are rarely discussed. Most of the research tell about quantity and quality which do not mention the spatial aspect. This research purposes to know the relationship between health facilities selection and geographic conditions by combining medical facilites data, number of doctor and spatial analysis approach. The selected study areas are referral health facilities chosen by DHF (Dengue Hemorrhagic Fever) patients in Purwosari District, Gunungkidul Regency. Information was collected by interviews and field measurements (altitude and location plotting) and also secondary data collection (hospital accreditation data, number of doctors, and number of beds). GPS, Altimeter, ArcGIS, and SPSS used for data collection and analysis. The spatial analysis method in this research were kernel density, buffer, ruler, and altitude of each facility. Meanwhile, the medical analysis approach used Pearsons' correlation and multinomial logistic regression in SPSS. The results of the spatial analysis show that the DHF patients in Purwosari dominantly choose health facilities in Bantul Regency because of having closer distance, relatively short travel time, lower elevation, and many health facilities centers. At the same time, Puskesmas Purwosari (Sub-district level Health center) was only chosen as an early reference and a place for treating mild dengue patients. In addition, distance, duration of travel time, and the number of doctors had a strong correlation based on the Pearson's correlation. These three variables have the greatest linear relationship with the response variable. Meanwhile, some variables; elevation, number of doctors, accreditation, and number of beds were considered by patients in choosing a hospital based on multinomial logistic regression analysis.