Quantifying the Contribution of Population-Period-Specific Information to Model-Based Estimates in Demography and Global Health

Leontine Alkema , University of Massachusetts Amherst
Guandong Yang, University of Massachusetts at Amherst
Krista Gile, University of Massachusetts at Amherst

Sophisticated statistical approaches are used to produce model-based estimates for demographic and health indicators even when data are limited, very uncertain or lacking. Examples in family planning and fertility-related research include the Family Planning Estimation Model (Cahill et al., 2018) that is used to estimate contraceptive use and unmet need for contraceptives worldwide, and a Bayesian accounting model (Bearak et al, 2020), used to estimate unintended pregnancies and abortions. Both models use hierarchical and temporal model structures such that data-rich population-periods can help to inform estimates in settings where data are limited, very uncertain or lacking. To facilitate interpretation and use of model-based estimates, we aim to provide a standardized approach to answer the question: To what extent is a model-based estimate of an indicator of interest informed by data for the relevant population-period as opposed to information supplied by other periods and populations and model assumptions? We propose a data weight measure to calculate the weight associated with population-period data set y relative to the model-based prior estimate obtained by fitting the model to all data excluding y. In addition, we propose a data-model accordance measure which quantifies how extreme the population-period data are relative to the prior model-based prediction. We illustrate the insights obtained from the combination of both measures in toy examples and for estimates produced by the family planning and accounting models.

See extended abstract

 Presented in Session 73. Issues in health data analysis