**Background:** An effective referral system is critical for timely access to emergency obstetric care (EmOC) and reducing maternal mortality. Despite India's progress in reducing maternal mortality from 398 to 103 per 100,000 live births (1997–2018), multiple referrals remain a major contributor to maternal deaths. Urban referral systems face unique challenges including inequitable spatial distribution of facilities, weak coordination, and overloaded tertiary hospitals. This study aimed to document patterns, reasons, and outcomes of obstetric referrals in public health facilities across urban Maharashtra.
**Methods:** This retrospective study used secondary data from municipal hospital records across 33 maternity (Level I), 16 peripheral (Level II), and 4 tertiary (Level III) hospitals in four municipal corporations (Brihanmumbai, Thane, Mira Bhyander, and Kalyan Dombivli) between 2016 and 2019. A standardized referral slip developed by SNEHA in collaboration with municipal corporations captured demographic information, referral reasons, clinical details, communication, and mode of transfer. Data were collected monthly by SNEHA investigators from duplicate referral slips and matched with "Received-In" data from receiving facilities to track outcomes. Inclusion criteria were pregnant women with ≥28 weeks gestation. Data were entered using CommCare on smartphones and analyzed using STATA V.14 with descriptive statistics.
**Key Results:** Of 199,888 deliveries at Level I and II facilities, 28,020 (14%) were referred, with 25,253 included after exclusions. Most referred women (74%) were under 30 years (mean age 24, SD 6), and 46% were primigravid. Referrals were higher from Level I (21%) than Level II (9%) facilities, with variation across corporations (MC 3: 32%, MC 2: 14%, MC 4: 13%, MC 1: 9%). The most common obstetric reasons for referral were pregnancy-induced hypertension/eclampsia (17%), previous caesarean section (12%), fetal distress (11%), and oligohydramnios (11%). Only 24% of referrals were exclusively for obstetric reasons. Non-medical reasons dominated: unavailable emergency operation theatre (47%), NICU (45%), anaesthetist (24%), paediatrician (22%), physician (20%), and obstetrician (12%). Referrals exclusively due to non-medical reasons accounted for 19% of cases. Phone-based communication between referring and receiving facilities occurred in only 47% of cases, and complete referral documentation was low at 37%. Only 45% of women were transferred by ambulance. Of all referrals, 60% (15,033) could be tracked to higher facilities, with 85% reaching the designated facility. Among tracked cases, 45% delivered by caesarean section, 96% resulted in live births, and 34% of newborns weighed <2,500 grams. No maternal deaths were reported.
**Clinical Implications:** The study reveals that a substantial proportion of obstetric referrals in urban Maharashtra are driven by health system deficiencies—inadequate infrastructure and staff shortages—rather than purely medical necessity. The low referral communication rate (47%) and incomplete documentation (37%) highlight critical process gaps that could delay care and increase risk. The high caesarean section rate (45%) among referred women reflects their higher obstetric risk profile. The findings support the need for: (1) formal communication and feedback systems between referring and receiving facilities, (2) upgrading health infrastructure (operation theatres, NICUs, MICUs) at lower-level facilities, (3) ensuring adequate staffing (anaesthetists, paediatricians, obstetricians), and (4) institutionalizing standardized referral protocols and documentation practices. The partnership model with SNEHA demonstrates potential for improving referral coordination, though the 40% loss to follow-up indicates ongoing challenges in tracking referred women through the system.