Background: About 830 pregnant women die every day due to preventable causes which often get missed in initial assessment. We assessed the quality of assessment for high-risk conditions or complications at the time of admission for childbirth in two referral hospitals from a southern state of India. Methods: We conducted a cross-sectional study among pregnant women admitted for childbirth in two hospitals (a district hospital and a medical college). We extracted information about high-risk assessment through history, examination and lab investigations and monitoring from case sheets and, prevalence of high-risk from registers. We interviewed the doctors and nurses, to understand the process of high-risk assessment for pregnant women at time of admission and monitoring thereafter. Results: Both the hospitals were lacking standard protocol for assessment and documentation in case-sheets was poor. Common investigations and basic examination was done for about two-thirds of cases but past and current history of common illness was elicited by only one-third of cases. Participants were insufficiently monitored during labour. 55% of the women had atleast one high-risk condition. Commonest high-risk was previous caesarean and complication noted was foetal distress. Most patients came in without referral records, and there was no mechanism for referral communication, or continuum of care across levels of care. Conclusions: A large proportion of pregnant women had any high-risk but a considerable proportion may be missed, or identified late. The documentation regarding assessment was poor. There is scope of improvement in high-risk assessment and monitoring of pregnant women admitted for childbirth.