Reminder: GBD 2019 Preliminary Demographics Comments due Today
Received On :
28.06.2019 18:11
Dear Collaborators, This is a reminder that comments on preliminary population, mortality, and fertility results for GBD 2019are due by midnight Pacific Time onFriday, June 28th.
Please see below for details on how to access these results, and key questions to consider when reviewing them. We invite you to review the countries of your interest or expertise as part of the opportunities available exclusively to GBD Collaborators,and to submit any comments using the form linked below.
I. Reviewing the Preliminary Results
To enable you to share feedback as efficiently and easily as possible, we have framed four critical questions regarding these results. We would greatly appreciate your feedback on these questions for the country or countries that you have expert knowledge.
The questions are:
Are there cases where the level or trend of population, mortality, or fertility does not match the data or the demographic profile of that country or location?
Are there sources that we have included that should be marked as outliers (due to unreliable data collection, unrepresentative samples, or other concerns)?
Additionally, looking towards GBD 2020, are there additional sources that are not currently represented in the results that are important to incorporate?
Lastly, we are making changes to how we determine completeness estimation of death registration systems in GBD 2019 by adding two more point results based on the method used in estimating population in GBD. For this purpose, we would ask that you please review your country's completeness results.After reviewing your country's completeness results, are there any issues we should be aware of?
II. Accessing the Preliminary Results
These preliminary results are shared via excel tables and PDFs. In order to access the results, please visit:
Included in the cloud folder for your country are the following:
Excel tables showing population, total fertility rate, total number of deaths, life expectancy at birth, under-5 mortality rate (5q0), and adult mortality rate (45q15).
Graphs with a variety of outputs, each labeled. These graphs also show the input data points, and model diagnostics for the estimation process.
Data sources included in these results can be found at the link below:
In GBD 2019, we have made considerable changes in the methodologies used in fertility and mortality estimation. These changes are manifested in the changes in our preliminary estimates. Here is a brief list of the changes we have made:
Implementation of the GBD standard location list in the modeling process:
To ensure consistency and continuity of the GBD estimates, we are systematically implementing the standard location list to ensure stability in estimated effect sizes of key covariates. The GBD 2019 standard location list includes all national locations included in the GBD analytical hierarchy, and subnational units from major populations including China, India, and the United States. In a regression model, we first run the model by using only the empirical data from the standard locations. Then, the fixed effects of any employed covariate are obtained. These coefficients are then used, and kept constant, in a re-run of the same model where data from all locations in GBD 2019 are included. Estimates for locations that are not in the GBD standard location list are then obtained. As most analytical models in our demographic estimation process involve a first stage model that links key covariates to the outcome variable of interests, implementation of the standard location list has an important impact on our estimates.
Improvement in the evaluation of completeness of death registration:
In GBD 2019, we are further improving our assessment of completeness of vital registration systems which include Vital Registration (VR), Disease Surveillance Points (DSP), and Sample Registration System (SRS). There are three traditional formal demographic methods developed to assess completeness. They are called the Death Distribution Methods which include Generalized Growth Balance method (GGB), Synthetic Extinct Generation method (SEG), and the combined method (GGBSEG). In GBD 2019, we developed two new methods based on the same Bayesian Hierarchical Population Estimation model used in GBD 2017 to provide an additional two point estimates of completeness between pairs of censuses. These two methods take advantage of the Bayesian analytical framework we developed for population estimation that can take into account uncertainty in reported deaths and coverage of the associated censuses. With the five point estimates of completeness for each pair of censuses and the VR in between, we outlier the two points estimates with the highest and lowest values, and proceed to the same completeness data synthesis model that has been used in the past rounds of GBD. We invite you to pay close attention to the location-specific completeness assessment graphs and would appreciate the comments you have on these new estimates.
Please kindly review the preliminary results by midnight Pacific Time, June 28, 2019 and respond to the questions at the link provided: https://ihmeuw.wufoo.com/forms/rco9zr0a6kkwh/
On that date, we will close access to the password-protected site in order to incorporate suggestions and produce new results. Due to the annual GBD timeline, we will be unable to incorporate any changes requested after this date. Regardless, we will review all comments in detail, and suggestions will be taken into consideration for GBD 2020 and beyond.
Lastly, please remember that you are receiving privileged access to these preliminary results as a GBD collaborator; these results are confidential and should not be cited or circulated.
Thank you again for your collaboration. We look forward to improving these results with your help.
GBD Secretariat
Institute for Health Metrics and Evaluation | University of Washington
2301 5th Avenue, Suite 600 | Seattle, WA 98121 | USA gbdsec@uw.edu | http://healthdata.org