Background In Tanzania the perinatal mortality rate is estimated at 69 per 1000 births.1 In Tanzania perinatal death reviews are recommended as a strategy to improve quality of clinical care for pregnant women and newborn babies. The study was conducted to analyse all perinatal death reviews reported to the national level from January 2004 – December 2007 in order to describe factors contributing to perinatal mortality in Tanzania, completeness of reporting and assess quality of perinatal death reviews. Methods This cross sectional descriptive study analysed 418 perinatal death reviews that were notified over the four year period. Standard procedures of double data entry and consistency checks were carried out using Epi Data Version 3.1, while STATA Version 8 and Microsoft Excel were used for analysis. Frequency and distribution tables and charts were used for analysis of demographic characteristics, antenatal problems, intrapartum complications, maternal, foetal and health facility factors contributing to the perinatal deaths. The Chi squared test was calculated to see whether the differences among the early neonatal deaths, fresh and macerated stillbirths were due to chance when compared for gestational age, birth weight, place of delivery, mode of delivery ,assistant at delivery, parity, intrapartum complications and avoidability of perinatal death. Action plans developed by perinatal death review teams were assessed by two independent assessors looking at components of the plan; problem identification, action, timeframe and outcome indicator. The quality of action plans was assessed by scoring each component of the plan as (1) unsatisfactory, (2) good and (3) very good. Results: The 418 reviewed perinatal deaths over the four years imply a perinatal death review rate of 1.4/1000. Of 418 perinatal deaths, 191(46%) were early neonatal deaths, 130(31%) fresh stillbirths and 97(23%) macerated stillbirths. Reports of perinatal death reviews were from 15 (71%) regions and 32,(23%) districts. None of 42 reporting health facilities reported consistently over the four years. The reported cases were mainly form hospitals 97%. One third of perinatal deaths occurred with mothers who were primigravida (32%) and a third of mothers were in risky age groups (below 20, > 35 years). The majority 82% had only primary level education and were farmers. Most mothers 99% attended ANC and complications were detected in 34% of cases. Deliveries took place mainly at health facilities, 94% with assistance of skilled attendant. Macerated stillbirths were more commonly delivered by SVD (93%) compared to early neonatal deaths (67%) and fresh stillbirths (48%, p<0.001). Intrapartum complications occurred in 62% of cases. Low birth weight was present in 55% of all perinatal deaths,, 46% among early neonatal death cases compared to 13% of fresh stillbirths and 36% of macerated stillbirths (p<0.003). Fifty seven percent of reported cases were preterm. More fresh stillbirths 51% were term compared to 38% and 39% for early neonatal deaths and macerated stillbirths (p <0.001). .Severe asphyxia was present in 45% of cases. Asphyxia ranked highest as a cause of probable cause of death in both fresh stillbirths (42%) and early neonatal deaths (41%), whereas untreated syphilis ranked highest for macerated stillbirths (18%). Over half (54%) of the deaths were considered avoidable by the review teams. Identified contributing factors at health facility level were among others poor quality of ANC services (38%), delay in diagnosing intrapartum complications (13%), poor monitoring of labour (8%) and delayed decision making at health facility level (8%). Key contributing maternal factors were poor compliance to ANC and delivery services (50%) and maternal conditions or complications (32%). The key contributing newborn factors were asphyxia and prematurity or low birth weight. The quality of action plans was poor with 75% graded as unsatisfactory. Conclusions: There is a need to increase coverage and quality of perinatal death reviews including a systematic way of providing feedback to reporting facilities. Continuous quality improvement of ANC, delivery and newborn care services, implementing maternal and newborn community interventions including strengthening linkages between health facilities and communities are key steps to reducing perinatal deaths.