1.3 Creating the database with murder
Microdata provided by the Department of Health of Mato Grosso State selected data only on homicides, according to the ICD-10 codes shown in the table above. [Note: There are no tables or graphics on this page.] This database includes the municipalities of Mato Grosso State, and the record of all reported homicides and deaths that occurred in legal interventions against residents in the municipality. For deaths with unknown intentionality, only a proportion of cases is considered homicide, so a random selection procedure within each one of the regions was conducted, to choose these cases according to the defined proportion by the proposed methodology.
For example, only 10% of deaths with unknown intent caused by other causes should be considered homicides according to the methodology adopted. So, if a particular region had 50 cases classified in the category causa mortis, five of these cases would be randomly selected to compose the homicide database, along with the declared homicides and legal interventions.
Regarding deaths with unknown intentionality caused by sharp instrument or firearm, the percentage of intentional homicides among deaths was obtained independently for each region and used to calculate the number of cases to be randomly selected. The proportion of selected cases by regional was obtained from the ratio given in the table above.
Data from the State Health Department are available in two ways: according to the place of occurrence of deaths and according to the place of residence of the victims. The latter type of record has been adopted for several reasons. Among them, the fact that the population used in the denominator of the homicide rate corresponds to the resident population — therefore the homicides in the numerator should also follow the same concept. Moreover, many deaths occur in hospital, so the place of occurrence of death is not always relevant in terms of estimation of the risk. It is worth remembering that the choice means that you are measuring the risk to residents in a particular location, regardless the place where the homicide occurred.
