Author:
|
Fernandes, Fabiola; Castillo, Paola; Bassat Orellana, Quique; Quintó, Llorenç; Hurtado, Juan Carlos; Martínez, Miguel J.; Lovane, Lucilia; Jordao, Dercio; Bene, Rosa; Nhampossa, Tacilta; Ritchie, Paula Santos; Bandeira, Sónia; Sambo, Calvino; Chicamba, Valeria; Mocumbi, Sibone; Jaze, Zara; Mabota, Flora; Ismail, Mamudo R.; Lorenzoni, Cesaltina; Sanz, Ariadna; Rakislova, Natalia; Marimon, Lorena; Cossa, Anelsio; Mandomando, Inácio; Vila Estapé, Jordi; Maixenchs, Maria; Munguambe, Khátia; Macete, Eusébio; Alonso, Pedro; Menéndez, Clara; Ordi i Majà, Jaume; Carrilho, Carla
|
Abstract:
|
Although autopsy diagnosis includes routinely, a thorough
evaluation of all available pathological results and also of any
available clinical data, the contribution of this clinical
information to the diagnostic yield of the autopsy has not been
analyzed. We aimed to determine to which degree the use of
clinical data improves the diagnostic accuracy of the complete
diagnostic autopsy (CDA) and the minimally invasive autopsy
(MIA), a simplified pathological postmortem procedure designed
for low-income sites. 264 coupled MIA and CDA procedures (112
adults, 57 maternal deaths, 54 children and 41 neonates) were
performed at the Maputo hospital, Mozambique. We compared the
diagnoses obtained by the MIA blind to clinical data (MIAb), the
MIA adding the clinical information (MIAc), and the CDA blind to
clinical information (CDAb), with the results of the gold
standard, the CDA with clinical data, by comparing the ICD-10
codes and the main diagnostic classes obtained with each
evaluation strategy (MIAb, MIAc, CDAb, CDAc). The clinical data
increased diagnostic coincidence to the MIAb with the gold
standard in 30/264 (11%) cases and modified the CDAb diagnosis
in 20/264 (8%) cases. The increase in concordance between MIAb
and MIAc with the gold standard was significant in neonatal
deaths (kappa increasing from 0.404 to 0.618, P=.0271), adult
deaths (kappa increasing from 0.732 to 0.813, P=.0221) and
maternal deaths (kappa increasing from 0.485 to 0.836,
P<.0001). In conclusion, the use of clinical information
increases the precision of MIA and CDA and may strengthen the
performance of the MIA in resource-limited settings. |