Résumé
This study presents some of the first statistical findings of the Severance Hospital discharge survey. The purpose of this study described in this report was twofold: 1) The underlying consideration was model building for the distribution of short-stay hospitalized patients; and 2) The immediate target was a better understanding of the impact of certain measurement probability discharged as cured present in the survey of medical records. In building up the duration-of-stay model, the conditional probabilities of discharge on a particular day were computed for the January-December, 1967 in order to obtain an appropriate theoretical distribution. The rise and fall of these conditional probabilities as duration of stay increased was characteristic of the log-normal distribution. Therefore, the log-normal distribution, ... was fitted to the available duratin-of-stay data. The parameters found for the distribution were estimated as mu=1.834 and delta=0.675 in 1967. As for measuring the probability of patient discharged as cured, life table model was adopted. Discharge as cured including improved patient and death including transferred to other hospital without improved are takan as the two attrition factors for decrement of a single population. The probabilities of being discharged as cured during the first week after hospitalization were found .as P0=0.41877 in 1967, P0=0.44113 in 1968 and P0=0.47387 in 1969. The average expectations of post discharge life were e0=2.048 weeks in 1967, e0=2.160 weeks in 1968 and e0=1.961 weeks in 1969. During the first fourteen days after hospitalization the probabilities as cured were 2P0=0.6814 in 1967, 2P0=0.7056 in 1968 and 2P0=0.7352 in 1969, which showed the rate has been steadly increased.