This is the link to access the roadmap, we currently have 5 active squads. Below are the updates from the various working groups:
CHT Plugins - the team has completed the ideation phase and is moving into the technical design stage. They are working towards an MVP that will allow one to add custom in-app tabs which will show static or dynamic content, work offline and interact with CHT data models.
Hosting TCO reduction V2 - the group is exploring reducing spaces required for upgrades by splitting large ddocs and making changes to the upgrade process to allow more control of what is reindexed.
CHT multi-agent system - a hierarchial multiagent system has been built which has multiple agents that covers research, QA and development and works together to automate the workflows and improve the CHT development process. The * MCP server has been set up using kapa.ai. Please reply to this * forum post if you would to be involved in this work
Task prioritization - the overdue task bubble functionality has been built and will be released in v5.1.0
Display of previous month targets- the team is finalizing two remaining tasks and this improvement is likely to be available in the next CHT release.
2. Community priorities overview.
This is the list of priorities that have been raised by the community:
Ability to pause on a form
Client and task visibility across catchment areas
Support for uploading attachments in contact forms
option to hide target goal past the goal
Offline synchronization support.
Task filter
The community is exploring ways to organize and list these priorities to make it easier to track, discuss and act on them collectively.
1.Massive Online Open Validation Evaluation (MOOVE) Initiative
This intervention is being implemented in Kenya by Strathmore University, Kenyatta National Hospital and the University of Nairobi and it aims to evaluate the performance of large language models (LLMs) in supporting clinical decision support by comparing their outputs to local clinical guidelines assessing their usability in primary care settings
The study has 3 phases hypMOOVE
in this stage, the accuracy of LLMs is evaluated using clinical scenarios without patients involvement
LLMs outputs are reviewed retrospectively by Expert Clinicians SilentMOOVE
De-identified data will be captured during routine patient visits. Clinicians will review the LLMs outputs after patient encounter and provide feedback
The Implementation will take place in 3 counties in Kenya (Kajiado, Nairobi and Migori)
2. Squad updates
Display of previous month targets - the team is finalizing this feature and it will be available in CHT v5.1.
Task filters have been added to CHT which will enable users to filter tasks based on due date, task type and place (for users who manage many places). These changes will be available in the next CHT release.
The focus of our March 2026 Community Meetup discussion was the CHT hosting total cost of ownership and the ongoing initiatives aimed at reducing the hosting costs.
Why does this matter? Disk space is the main cost driver, as CouchDB creates massive index files which consume disk space far beyond the actual data stored. For some projects, the disk costs are significant which might pose a challenge to the long term sustainability of the projects.
TCO phase 1 - What was achieved?
CouchDB Nouveau was introduced to replace the existing CouchDB search which has enabled disk savings in the excess of 30% for large instances. These improvements were released in v5.0. For partners who plan to upgrade, there are a few important considerations, including:
You will need 3-5 times disk space during the upgrade.
Kubernetes deployments need additional manual steps.
Early 5.0 upgrades for production CHT deployments have delivered 5- 25 % disk savings, depending on the instance size.
The TCO V2: What’s next
TCO phase 2 aims to split views into separate design documents, which would ensure faster upgrades and further reduce the disk space needed for the upgrade.
How to get involved: we encourage you to join the weekly TCO V2 community squad calls held every Tuesday from 3:15 - 4:15 pm GMT to contribute and share your perspectives. Meeting link.
Capturing GPS Data in CHT
GPS data is captured automatically in CHT reports, with every form edit storing an additional location entry. To enable geolocation capture, users must grant location permission to the Android app. If permission is denied twice, users are directed to device settings to manually enable it.
In addition to automatic capture, a number of projects use the map widget which is embedded directly into forms as a more deliberate and structured way of collecting geolocation data.
Analysis of GPS submissions.
A review of geolocation data across various CHT projects from multiple geographies revealed a GPS capture success rate ranging from 30% to 70%. The primary driver of failures was geo timeouts, with additional cases attributed to users not having granted location permission.
Other challenges shared by deployments included low GPS accuracy in remote and rural areas and instances where users turned off location services to conserve battery life.
Using GPS data to deliver responsive care.
Several projects are already leveraging geolocation insights to strengthen community health service delivery. Notable examples include the use of geospatial data to optimize health service delivery in Uganda and the Datharm team’s work using GPS data to deliver targeted routine immunisation services across three states in Nigeria.