Completeness and Quality of Kampala Mortuary Dataset

Authors: Kavi Bhalla, Jerry P Abraham and Michael Lipnick

Date: July 7, 2010

Objective: To do a crude test of the completeness and quality of the Kampala Mortuary dataset

Method: The number of deaths in the mortuary records was compared with an estimate of the number of deaths expected in the Kampala metropolitan area

Findings:

- The mortuary dataset captures a substantial fraction of injury deaths in Kampala

- The quality of cause of death attribution of the dataset is low with a substantial number of deaths assigned to unspecified causes. Only four external causes (RTA, Fall, Burn, and Assault) can be distinguished.

Background

- Data Source Analyzed: Kampala Mortuary records (Kampala city mortuary and Mulago Hospital mortuary), for six months (July - Dec 2007).

- Analysis in this report should be repeated with Census 2007 data:

o The analysis in this report uses crude estimates of population and mortality in Kampala. Typically these have been extrapolated from estimates available for the country in international databases. Accurate data for population and mortality in Kampala should be available from the 2002 Population Census. This analysis should be updated when these estimates are available. (Note that the 2002 Census includes a question on household deaths but unlike some other censuses does not ask if the death was due to injury/violence)

- Injury cases comprise 100% of the deaths in the mortuary dataset analyzed here.

Assessment of Quality of Injury Coding

o Only four specified forms of injuries (Road Traffic, Fall, Burn, Assault) can be distinguished from the dataset. Key external causes that are not included at all include drowning, suicide, poisoning and mechanisms of RTA and assault.

o Unknown external cause comprised 22% of all injuries

TABLE 1: Distribution of cases in the Kampala mortuary dataset