Objectives. Determine the quality and variability of perinatal mortality (PM) records in Lima and Callao. Materials and methods. Key informants were interview to evaluate the processes. The reference for foetal deaths was taken from the obstetric registry and for early neonatal deaths as the neonatal registry. These were compared with records at a diverse range of levels in hospitals, Regional Directorates of Health (DISA) and Civil Registries (CR). The filling quality of death certificates and necropsy registries was also reviewed. Results. There were different and not well defined recording processes. The registration processes varied and were not well-defined. In the statistics offices of the hospitals there were recorded 2% more of fetal deaths. There was no discrepancy in foetal mortality registration in 3 hospitals using the Perinatal Information System; data for early neonatal death was only found in these hospitals and in the Instituto Especializado Materno Perinatal. Maternal perinatal section of the DISA and five hopsitals statistics offices recorded less than 9% fetal deaths and 23% of early neonatal deaths. The statistics office of the DISA only registered foetal deaths from 3 hospitals and did not register any early neonatal deaths. An underreported of 50% (-100% to +11.1%) of foetal deaths and 33.3% (-89.2 to +66.7%) of early neonatal deaths was found in the CR. In 70% of foetal death certificates the cause of death was recorded as “unknown”. 29% of foetal deaths and 24,9% of early neonatal deaths were autopsied. Conclusions. There is an important discrepancy in the basic records between the DISA and CR, which suggests that the processes and tools must be standarized, establish a surveillance system to incorporate the use of Perinatal Information System and feedback information to reduce the unrecorded perinatal mortality.

Objetivos. Determinar la calidad y variaciones en los registros de mortalidad perinatal (MP) en Lima y Callao. Materiales y métodos. Se encuestó a informantes clave para evaluar el proceso de registro. Se compararon los registros de MP de diversos niveles en hospitales y las respectivas direcciones de salud (DISA) y registros civiles (RC) tomando como referencia para muertes fetales (MF) el registro de los servicios de obstetricia y para muertes neonatales precoces (MNP) los de neonatología. Se revisó la calidad del llenado de los certificados defunción y registros de necropsias. Resultados. Los procesos de registro son variados y no bien definidos. En la oficina de estadística de los hospitales se encontró un registro mayor de MF (+2%). En los tres hospitales que usaron el Sistema Informático Perinatal (SIP) no hubo discrepancia en los registros de MF; para MNP sólo se encontró datos en cuatro (tres usaban SIP) de los once hospitales evaluados. En las DISA el componente materno perinatal , registra menos de 9% de MF en cinco hospitales, y para MNP menos de 23%. La oficina de estadística de la DISA sólo registra MF en tres hospitales y no existen datos de MNP. En los RC se encontró un subregistro de 50% (-100% a +11,1%) para MF y de 33,3% (-89,2% a +66,7) para MNP. 70% de los certificados de defunción de MF registraron “desconocida” como causa de muerte. Sólo 29% de MF y 24,9% de MNP fueron necropsiadas. Conclusiones. Existe discrepancia en los registros de mortalidad perinatal entre las DISA y registros civiles, lo que sugiere la necesidad de estandarizar procesos e instrumentos, establecer un sistema de vigilancia, incorporar el SIP y retroalimentar la información para disminuir el subregistro de la mortalidad perinatal.