Frequently Asked Questions

How do I get in contact and what should I do if I find a data error or problem?

We encourage everyone to get in contact and provide feedback on the database. The database is being launched in "beta" form precisely to give civil society time to become acquainted with it while also continuing to refine the database and add features per any feedback we receive.

If you have a suggestion for improving the user interface, a problem with any of the charts, or a general comment on the database, we would love to hear it. If you are reporting a data error in the database related to a specific mechanism or funder, we encourage you to check the original COP documents beforehand to see if the problem is a coding issue there before alerting us.

All enquiries and comments can be sent to dbinfo@amfar.org.

What is the difference between the PEPFAR Dashboards and the amfAR database?

amfAR's database collects data directly from the publicly released COP documents—the same data that appears on PEPFAR's own dashboard under "Planned Funding". However, amfAR's database goes further by capturing allocations data that aren't available on PEPFAR's Dashboard, including:

How can stakeholders best use this data?

By making this data available, amfAR hopes that anyone who is interested in PEPFAR can use the database as a vehicle to gain a greater understanding of PEPFAR programs within and across countries.

The primary audience for amfAR's PEPFAR database are civil society, academics and implementers.

This is a beta version of the database. As such, we welcome all feedback as we continue to update the database as new information becomes available and refine the user interface, as needed, across various platforms.

What cautions do you have for people in analyzing the data?

Critically, the database contains only PEPFAR COP allocations – not actual expenditures. Only limited expenditure data is available from PEPFAR at present. As such, the database cannot be utilized for cost-effectiveness analysis. Additionally, PEPFAR represents only one part of all HIV spending in a given country or region. The spending environment into which PEPFAR invests its resources matter and should not be looked at independently of the contributions of other funding entities such as domestic country funds, Global Fund grants, private resources, and other sources of spending on health in a given country.

Additionally, caution should be taken when drawing conclusions from searched narratives. Many narratives over the years have been redacted or simply had no narrative section available in the COP. This means that searches of these mechanisms will return no results, even if the programs ultimately performed work relevant to the search. Additionally, the narratives are complex and often times a phrase will be used in only in passing and thus may not signify significant program activities. For these reasons, searches should be limited to performing qualitative assessments of PEPFAR programs rather than anything quantitative.

What is the Not Available/Unknown Funding and why is there so much of it?

The COP/ROP data is the planned funding that gets approved before the relevant fiscal year begins. These data include planned bilateral funding initially approved for countries that submit PEPFAR COP/ROPs and subject to change due to updates and other actions approved by the U.S. Global AIDS Coordinator. As such, in some instances, publicly available allocations data may not represent the total amount of PEPFAR funding that was ultimately allocated to a specific program area or country in a given fiscal year. For instance, much of the 'Not Available'/'Unknown' allocation data were subsequently awarded either during the same COP/ROP year or in subsequent years as part of 'pipeline' funding. However, because what implementing partners and mechanisms were ultimately awarded in some grants/contracts have not been publicly released, they are presented here as 'Not Available' and 'Unknown' as our database cannot account for the individual mechanism details for those awards. If these data are ever updated or made publicly available, we will endeavor to incorporate the appropriate information into our database.

How does the search function work?

The search function merely looks for a specific phrase or phrases within the mechanism narratives. For certain key words, the system will perform "mesh searches" based on methodology sourced from PubMed. For instance, a search for MSM will also search the narratives for gay men, men who have sex with men, homosexual, and bisexual. When this is done, an asterisk will appear next to the search term in the title.

To separate search terms utilize "and". So, if you wish to search for commercial sex work programs in Gauteng province of South Africa, search for "FSW and Gauteng" which will return narrative sections that contain both terms.

What methodology did you use to create the database?

COPs for all countries were downloaded from PEPFAR's website. Where available .rtf or .doc files were utilized and .pdf files in other cases. Documents were then converted to .txt utilizing standard open source tools (unrtf or pdftotext). Python scripts were coded for each COP year to extract both headline data and mechanism specific data and enter it into a MySQL database. For certain years, scripts for both .rtf and .pdf files had to be coded separately as not all COPs were released in consistent formats. (Note: Data for India in 2013 is pulled from a combination of the published COP and data from PEPFAR's Dashboard as the COP does not contain any mechanism data. Additionally, Management and Operations allocations for 2013 are enhanced with additional data from PEPFAR's Dashboard that isn't available from the COP documents themselves. These amounts are reflected in mechanism numbers 90000 and above.)

After extracting all data, inconsistent partner name coding for primary recipients were harmonized and amended accordingly. A web interface was developed using Laravel, php and javascript (utilizing C3js and Google Charts) to graph the data. Php scripts were coded to detect inconsistencies in the dataset by strategic area, budget code, country, year, mechanism, organizational type/primary partner, and funding agency and all attempts were made to root out any errors. Additionally, all data from PEPFAR's Dashboard were incorporated to the database and Php scripts were used to detect inconsistencies between the two datasets.

Any amounts that remained unallocated either to individual partners or mechanisms were deemed "Not Available" and assigned a mechanism number of '1'. Overall, this represents about 18.8% of total PEPFAR planned funding.