Pakpoom Thawong. The Accuracy of HIV/AIDS Reporting on Death Certificates. (). Mahidol University. Mahidol University Library and Knowledge Center. : , 2013.
The Accuracy of HIV/AIDS Reporting on Death Certificates
Abstract:
Background: Mortality data are important tools for public health
planning. Inaccuracy in the actual cause of death may wrongly
suggest differences in magnitude of health problems, which is
important for health planning and policy. HIV/AIDS is the leading
cause of death in Chiang MaL Studies on the accuracy of HIVI
AIDS reporting on death certificates in Thailand were not widely
available. ObJective: To quantify the accuracy of HIV/AIDS reporting
on death certificates. Design: A population-based cross-sectional
study. Setting: Four subdistricts of Mae Wang, Chiang MaL
Subjects and methods: Names and addresses of the decedents
on a death registration database between 1 st January 1997 and 31 st
December 1998 were searched for. The next of kin or the closest
relatives of those decedents were interviewed using a standard
and validated questionnaire, which included general and demographic
information. Past illnesses and any illnesses occurring 1
year before and at the time of death were discussed in detail using
an in-depth interview. The information was reviewed by a physician
review panel blinded to the information and independent of the actual
death certificates. Causes of death reported on death certificates
were sought from the demographic registration computer system
of the district. The cause of death on death certificates and the
cause of death agreed upon by the physician review panel were
compared. Consistency of the data was analysed using kappa
coefficients. Differences between kappa coefficients were analysed
using Fisher's ztransformation test. Results: A total of 201 deaths
were registered during the defined period. The death certificates
listed HIV/AIDS in only 42 (sensitivity = 65.6%) of 64 instances
where the physician review panel identified HIV/AIDS as the underlying
cause of death. HIV/AIDS was not recorded on the death
certificates of those who were likely to have died from/of HIV/AIDS
in 22 cases. The causes of death reported on the official death
certificates among these were recorded as disease of cardiovascular,
respiratory, gastrointestinal, neurological and urinary systems,
allergy and suicide. The death certificates, however had a high
specificity (1 00%). Overall observed agreement between the death
certificates and the physician review panel was 89.1 % (kappa =
0.72). The validity and agreement did not appear to be affected by
age, gender, presence of detailed history of illnesses from governmental
hospitals, location and time of death, or the persons who
gave opinion on the cause of death. Conclusions: An increase in
the HIV/AIDS related mortality trend is not likely to be caused by
false positive diagnoses of HIV/AIDS deaths because of its high
specificity. HIV/AIDS mortality rate, determined from death
certificate information, may underestimate the actual HIV/AIDS
related mortality in view of its low sensitivity. Results of the study
indicated that the death certification process needs some improvement
to ascertain that the mortality statistics of the nation are
accurate.
Mahidol University. Mahidol University Library and Knowledge Center