Analyzing tracker audit data in R

hi community!

With DHIS2 tracker being implemented in more countries than ever as a point-of-care data collection tool, it’s important to know when care providers are actually capturing data during their work day. Many EHR software providers find that system usage outside work hours can duplicate reporting work and overburden health care providers, in which case the tracker system creates as many problems as it solves.

Unfortunately its somewhat difficult to know when DHIS2 Tracker users are most active, what they work on, and for how long. To our team at e-Registries, this finding was a bit frustrating for our maternal and child health tracker projects. Luckily we discovered the trackedentitydatavalue audit log could be a potential resource for analyzing tracker usage patterns.

We produced a few trial scripts in R to pull the audit log through API, and render the analysis outputs in an RMarkdown document. As part of our ongoing commitment to the community, we have generifed the analysis for any DHIS2 instance and any user with access to the audit log. To anyone who is familiar with R for statistical analysis, it is very “plug and play” tool with minimal configuration required; the output html has some client-side interactivity. but it requires no “Shiny” server or advanced programming knowledge to set up.

A version of this analysis has been run for two large production systems in Palestine and Bangladesh, and we are now seeking additional feedback. You are encouraged to download, explore, and repurpose as you find useful!

The sample analysis output is found here.

The code is freely available here:

If you have any questions or suggestions for future improvements, please get in touch. Our hope is that such analyses interacting with the API can provide proof-of-concept for new native apps on the DHIS2 platform, specifically oriented to supporting tracker program managers and researchers.