Bayesian Estimation of Age-Specific Mortality and Life Expectancy for Small Areas With Defective Vital Records

Area of Science:

• Demography
• Biostatistics
• Public Health

Background:

• Small area demographic rate estimation is hindered by sampling variability and imperfect vital registration systems.
• Vital registration coverage quality often varies significantly across regions within countries.
• Accurate mortality schedules are crucial for public health planning and resource allocation.

Purpose of the Study:

• To develop a Bayesian regression model for small-area mortality schedules.
• To simultaneously address challenges of small local samples and underreporting of deaths.
• To improve the accuracy and reliability of demographic rate estimation in small areas.

Main Methods:

• Developed a Bayesian regression model combining a relational model for mortality schedules.
• Integrated probabilistic prior information on death registration coverage from demographic methods and field audits.
• Applied the model to small-area mortality data from Brazil.

Main Results:

• The model successfully accounted for underregistration of deaths and provided measures of uncertainty.
• Incorporating external estimates of vital registration coverage improved small-area mortality estimates.
• Bayesian estimates indicated that noise often dominates signal when comparing mortality levels in small areas.

Conclusions:

• The developed Bayesian model enhances small-area mortality estimation by addressing data limitations and registration issues.
• External data on vital registration coverage significantly improves demographic estimates and quantifies uncertainty.
• Local differences in life expectancy estimates are often smaller than uncertainty intervals, even in populous countries.