[en] Radon (²²²Rn) present in the spa facilities has been identified as an agent of additional radiation burden both for bathers and the working personnel. Although some researchers have been concentrated on the health impact of radon activity transient increases due to water flow of domestically used water, limited work has been reported on this effect in spa facilities. The aim of this work was to study the exposure both of bathers and working personnel due to water flow under working conditions in some thermal spas in Greece. The spas studied were built at two locations in Greece; Lesvos Island and the city of Loutra- Edipsou. The study was performed between 1999 and 2003. Measurements involved determination of radon concentration in the waters of the thermal spas, monitoring of radon and short-lived decay product activity (both in attached and unattached form) and coarse particles (> 500 nm). Water radon concentration was measured by Alpha Guard PQ2000Pro of Genitron GmbH equipped with an appropriate unit (Aquakit). Indoor radon activity was monitored by Alpha Guard PQ2000Pro and EQF3023 of Sarad GmbH which was also used for monitoring of radon short-lived decay product activity. Coarse particle concentration measurements were performed by a GRIMM 1.104 Portable Dust Monitor. Measurements indoors were performed in the treatment room and in the reception room, if existing, under working conditions. Several procedures for bath filling and pouring out of the treatment room, during the measurements, were studied. Radon concentrations of thermal waters were found to lie in the range 1 Bq.L^-1 and 304 Bq.L^-1. Some of these concentrations may be considered as elevated. Elevated (about 2-6 kBq.m^-3) indoor radon concentrations were measured in the treatment rooms of the spas of Polichnitos, Eftalou and Loutra Edipsou. This found to be in accordance to the high radon potential of these spas, as detected out of the measurements of their thermal waters. Concentration peaks both for radon, radon daughter and coarse particles, were found to appear during filling of baths in the treatment process. According to the bath filling and pouring out procedure followed, these peaks presented variations. The peaks were attributed to radon degassing. Peaks were also detected for coarse particle concentration (droplets) at the same time periods which were related to peaks in radon and daughter activity. A time delay which reaches up to 4 hours between radon peak and daughter peak concentration has been systematically observed for each progeny nuclei. The unattached fraction of radon progenies found to be below 0.1 % for each daughter nuclei. Thus, almost all progeny nuclei attach to coarse particles at their generation. Doses were estimated using proposed models. The doses delivered to the bathers during treatment were in the range of 0.00670 mSv per year to 0.1279 mSv per year, while the doses delivered to personnel were in the range 0.4 and 4.5 mSv per year. The peaks in concentrations measurements were related with an additional short term health impact both for bathers and working personnel. (author)