A methodology for segregating rural and urban mortality
This study involved the design and testing of the Rural Urban Mortality Measurement (RUMM) technique. The technique generates independent estimates of rural and urban mortality for all age-groups by segregating death registration data into areas of similar characteristics to urban and rural areas. These areas are referred to as inferred urban and inferred rural populations in the study.
In order to assess the reliability and validity of the RUMM technique, it was applied to the Philippine death registration data of 1975 and 1980, and to the 1980 death registration data . for Thailand and Peninsular Malaysia. This application followed several procedural stages involving evaluation and assessment of the reliability and completeness of death and population data.
Application of the RUMM technique to Thailand and Peninsular Malaysia allowed the comparison of inferred urban arid inferred rural mortality estimates to the estimates generated for urban and rural areas. This is because Thailand and Peninsular Malaysia have rural and urban data on death registration.
The assessment procedure which compared the closeness of the patterns and levels of mortality between inferred urban and urban areas, and between inferred rural and rural areas, showed no difference. The differences in survival ratios for each age-group and the mean differences were found to be close. to zero. This led to the conclusion that the mortality estimates for inferred urban and inferred rural populations are valid representations of· levels and patterns of mortality found in urban and rural areas. Therefore, in cases where rural and urban tabulations of deaths do not exist, RUMM technique provides a valid method for calculation of mortality estimates.
This study also presented the strengths and weaknesses of the technique especially when applied to sub-national populations. Mainly, weaknesses result from using the Brass Growth Balance Equation to assess completeness of death registration. Substitution of alternative estimates of death registration completeness tends to strengthen the technique.
Finally, this study showed the robustness of the RUMM technique as well as its non-dependence on any specific index of urbanization arid on any technique of assessing completeness-of death registration.