191 Background: When a vesicant drug is administered to the peripheral vein, a gravity drip method rather than an peristaltic infusion pump is recommended. In addition, it is reported that the gravity based pump has less particle generation than the peristaltic pump that injects fluid through constant peristaltic movement. However, it is pointed out that the disadvantage of gravity based pump is that the rate of administration can easily change according to the change in the location of the injection site. We evaluate the accuracy of new gravity based pump (pressure-sensitive speed control fluid injection device, Accudrip made by Hanvit MD ) at various injection rates and hanging heights common to hospital settings. Methods: The accuracy and safety of the intravenous (IV) flow controller were evaluated for patients receiving IV chemotherapy in the oncology department. It was verified whether the accurate injection rate could be maintained even in various fluid heights, fluid amounts, and types of anticancer drugs that could affect the injection rate. Various anticancer drugs, including immune checkpoint inhibitor, were mixed with 250 ml/500 ml of normal saline and administered for 30 min/60 min, and glucose 500 ml/1000 ml was administered for 60 min/120 min to measure accuracy, each progressed 25 times. Medical staff were surveyed on the ease and limitations of medical practice for each peristaltic infusion pump and new gravity based pump. In addition, side effects occurring in patients during fluid administration were recorded. Results: The mean degree of error in the estimated end time of gravity based automatic infusion system was 3.09% (2.09 ~ 4.10). The error of the gravity based automatic infusion system was statistically significant depending on the patient age (p=.001) and the infusion time (p=.016), but the degree of error according to the type of fluids (p=.302) and fluid concentration (p=.844) was not statistically significant. Among the subjects of the study, nurses' satisfaction with the application of gravity based pump was 4.10 (±0.62) out of 5, and among the general characteristics of nurses, age ( p<.078), and working department (p<.91) did not show a statistically significant difference, but work experience ( p <.021) showed a statistically significant difference. Conclusions: The accuracy of Gravity based automatic infusion system was determined because the degree of error in the expected end time of infusion was not significant according to the infusion variable. In a survey of nurses who have used gravity based pump directly, it was found that the ease of use and satisfaction were similar to those of the conventional peristaltic pump. These findings will contribute greatly to the precise drug administration (over or under-administration) and minimization of adverse events in real-world hospital settings.