Background: Few studies in developing countries have examined posthospital mortality and little is known about the magnitude of posthospital mortality and risk factors for long-term survival. A better understanding of the determinants of posthospital mortality could help improve discharge policies and interventions with implications for overall childhood mortality.
Study population: In the period from 1991 to 1996 all paediatric admissions coming from the Bandim Health Project's area were registered at the National Hospital in Bissau, Guinea-Bissau. Posthospitalization information from a population-based surveillance system was available for 4153 admissions contributed by 3373 individuals having between 1 and 8 admissions during the period. Three thousand six hundred forty seven (3647) admissions by 2950 children resulted in live discharges. Postdischarge mortality included all deaths during 1 year following live discharge.
Results: Among the 221 children who died during the first year after discharge, 170 died in the community and 51 children died during a subsequent hospitalization; thirty-eight died on the day of discharge and almost one third had died within the first 2 weeks. The overall in-hospital and 12-month posthospital mortality was 20%. Compared to the mortality level in the community and controlled for other determinants of childhood mortality, children discharged from hospital had 12 times higher risk of dying during the first 2 weeks after discharge. The mortality rate ratio (MR) was 6.2 (95% confidence interval 3.8-10.2) times higher when we excluded those who died at the day of discharge. For the period 30-91 days after discharge the MR ratio was 3.7 (2.5-5.5), and in the period 3-6 months after discharge, the risk estimate was still 2.5 (1.6-3.9) times higher than community mortality. In a multivariate analysis, the all-dominating risk factor was discharge status as 'fled' in the sense of nonmedical discharge, the MRs being 18.6 (9.5-36.6) in the first 2 weeks after discharge and 4.0 (2.0-8.3) in the remaining part of the first year. Other significant risk factors for postdischarge mortality included ethnic group, housing quality and maternal education, and were similar to risk factors for community mortality. The same diagnoses that had high acute mortality, including anaemia, diarrhoea and 'other', were also associated with high postdischarge mortality.
Conclusion: There was a marked increase in mortality after hospitalization, the effect being particularly strong for children who fled the hospital. Improved discharge and follow-up policies might have an important impact on survival after paediatric hospitalization. Studies on the effect of focused intervention at discharge are needed.