A cross-sectional analytical study of the decision to use health care services in Khon-kaen Hospital was conducted in 272 out-patients at Medicine Department during two weeks of April 1991 to determine the general socio-demographic characteristics of the patients and four sets of factors adapted from Andersens Health Behavioral Model affecting the variation of time taken for the decision (0-90 days). Most of the patients are native Buddhists, educated for upto four years, and working in agriculture sector. Most of them were married, came from the average household size, and identified as sufficient income earning households. Of the above socio-demographic characteristics female patients compared with male patients seem to delay seeking the hospital services (P < 0.05). The mean time taken for the decision to use hospital services is approximately two weeks (14.34 + 20.38 days) which cover more than three-fourths (78.7%) of total cases of patients. More than a half (58.8%) decided within one week. Almost all of the fifteen selected factors in the framework including distance between the hospital and where the patients stay, number of stops of transportation to the hospital, self affordability to the service cost, acceptability of the services in three aspects : ease of gaining services, service providers manner, and hospital facilities

illness-related attitude on etiology and self treatment, perceived seriousness of illness in three aspects : danger to life, disability, distress, and chronicity

consultation and referral, previously exposed multiple treatment number and options are significantly related to the mean time taken for the decision by univariate analysis (t-test and ANOVA) (P < 0.05) except illness etiology attitude (P > 0.05). The final linear regression model of relationship between time taken for the decision and selected significant factors reveals that number of previously exposed treatment and perceived seriousness of illness are the relatively most important factors affecting variation of time taken for the decision, "multiple" number of treatment compared with "never treat" could delay time taken for the decision for approximately 23 days, "mild" degree of perceived seriousness of illnes compared with "severe" could delay time taken for the decision for approximately 19 days after controlling other factors (P < 0.05). For the patients who have same characteristics, ones who stay in every 100 kilometer farther from the hospital would delay to use the hospital for approximately 7 days (P < 0.05). All the above three factors according to the final model can explain the variation between 0-90 days of time taken for the decision 42.5% in linear basis (P < 0.05). Health care providers should recognize the previously exposed umltiple treatment of the patients as one of the major medical history investigation issues before providing the services. Health care policy decision makers should base organization system of health care services upon not only health care provider side but health care recipients perception. The operational studies on health care service location to determine the most appropriate geographic distance between communities and the outlets should be conducted.