objective: to disclose the existence of rooming-in (ri) in public and government contracted private hospitals that offer obstetric beds, in the state of rio de janeiro, brazil, and to examine whether there is any association between ri and another quality care indicator which influences breastfeeding, namely the rate of cesarean section operations performed in these hospitals. method: a survey was made of the existence of ri through a questionnaire sent to the municipal health offices, the information collected being confirmed by telephone with each maternity hospital. the c-section rate data was obtained from the rio de janeiro state health office and divided into 2 groups: "below 40%" and "40% and above". the prevalence ratio was applied to the measurement of the association between the variables. results: a rooming-in rate of 65.2% was found for the state as a whole, with regional variations: a better situation in the capital (84.8%), an intermediate one in the interior (69.9 %), and a worse one in the metropolitan belt (44.2%). the public maternity hospitals revealed a higher rate (89.7%) than that of the government contracted private hospitals (53.3%). a direct relation between the practice of ri and low c-section rates was found in the hospitals. however, this association did not present the same weight in all regions of the state. the lowest proportion of hospitals adopting ri was verified in the metropolitan belt, whereas the highest relative number of hospitals with high c-section rates was observed in the interior. conclusion: it is concluded that to reverse the observed status, government authorities must fulfil their gerencial role within their own health system, as well as in the government contracted private hospitals.