MAKERERE UNIVERSITY IGANGA-MAYUGE HDSS CONSISTENCY OF PHYSICIAN VERBAL AUTOPSY CODING IN IGANGA/MAYUGE DSS 2009 INDEPTH AGM By Aloysius Mutebi (VASA coordinator),

MAKERERE UNIVERSITYIGANGA-MAYUGE HDSS

CONSISTENCY OF PHYSICIAN VERBAL AUTOPSY CODING IN

IGANGA/MAYUGE DSS

2009 INDEPTH AGM

By

Aloysius Mutebi (VASA coordinator), Karin Kallander, Doreen Nabukalu, Daniel Kadobera, Edward Galiwango, George Pariyo

Introduction• VASA is a method of obtaining as much information as

possible about a deceased person by asking questions from the next of kin

• Principle of VA: A caretaker or next of kin is interviewed for symptoms and actions preceding the death, and a likely cause-of-death assigned.

• Cause specific mortality data is used to generate the burden of disease profile for a specific population.

• WHO/In-Depth verbal autopsy tool merged with the Bangladesh Matlab Social autopsy => “VASA” but now we have a new WHO/INDEPTH/SAVVY tool adapted since January 2009.

Background of Iganga/Mayuge DSS

• Iganga/Mayuge DSS started in 2004 and conducted its first baseline census in March 2005

• It is located in the mid Eastern part of Uganda, 118Km from Kampala, along the East African Highway

• The DSA covers an area of 3,931sq.Km and currently follows up a population of 68,300 people two times a year.

• The DSS has so far collected six rounds of data since it started

Objectives

• To find out how well the physician coders agree when coding the possible cause of death using verbal autopsy data.

• To find out where they mostly disagree in terms of disease codes, and for which age groups of the deceased

Methodology• Study Site

– Iganga/Mayuge DSS with 68,000 people in 12, 000 households

• Study Population– Neonates: 0 – 28 days, children: 29 days – 14 years,

adolescents and adults: 15 years+• DSS scouts report both death and birth using Death and

birth report forms respectively• Since 2009 we have adapted the revised

WHO/In-depth/SAVVY tool for VASA data collection• Death is also captured during the routine round of data

collection by the field assistants

Methodology cont’d• Three physicians code for both the most possible

immediate and underlying causes of death using the ICD10

• Where at least 2 physicians have the same code that is taken as the most possible cause of death

• Where all the three disagree, the form is put aside and reviewed by all the three as a team

Conclusion • The physicians are doing a good job considering the

78% agreement of at least 2.• The disagreement is mainly among the adolescents

and adults group

Recommendations

• We recommend that the physicians continue the fact that they are available and doing well

• We also recommend continuous refresher training for the coders

Acknowledgement • InDepth Network• Makerere University• Host Districts: Iganga & Mayuge

• Funders– Sida/SAREC – The Rockefeller Foundation

• Iganga/Mayuge DSS Management• Iganga/Mayuge DSS Staff• Iganga/Mayuge DSA communities

The End

Thank you