Accurate geospatial health data for better precision public health

Post author: Mourice Barasa, Research Officer for Monitoring & Data Systems

Precision public health is an emerging practice to have more granularity in prediction and understanding of public health risks while customizing treatments for more specific and homogeneous subpopulations, often using new data, technologies, and methods. It is about delivering “the right intervention at the right time, every time to the right population”. Accurate geolocation data is key in identifying when and where to intervene. The Community Health Toolkit (CHT) has the ability to collect geolocation data as health workers provide care in the community. 

As technical stewards of the CHT, Medic works with our partners to integrate the use of geospatial data to inform care delivery. Previous CHT deployments collected geolocation data passively with no health worker involvement in the process.

In early 2021, Google revised the Google Play store compliance requirements for all developers to explicitly declare app permissions and how the data collected is being used. Among other things, developers are required to indicate how they use location information to ensure end users are fully aware and consent for such use. 

Involving health workers in data collection 

With the efforts to promote openness and improve geolocation data accuracy, Medic is working with our partners to shift to active involvement of the health worker in geolocation data capture. In the new approach, the health worker is informed that location data is being captured and would confirm when recording the data. With Partners in Health in Malawi, we implemented a GPS data collection form where health workers are actively prompted to capture their current location data while also capturing the GPS location data in the background (passively). An initial attempt to map all households served by the health workers yielded only 60% of submitted forms where data was only passively captured even with a note reminding the health worker to turn on their location. However, when the health worker was prompted to record the geo-coordinates, 100% coverage was realized of which less than 2% were outside the catchment area or appeared problematic.

Photo credit: Partners in Health, Malawi, 2018

A comparison of active and passive data collected simultaneously in the same form identified some differences. Passive capture of geo-coordinates resulted in coordinates missing from 8% of submitted forms as compared to no cases of missing coordinates from actively collected geopoints. The coordinates collected with the health worker’s involvement were observed to be better distributed as compared to the geo-coordinates collected passively, which were clustered around specific areas. With regards to distance, about 81% of the geolocated points were within 50 meters from each other, 9% between 50-100 meters, while the remaining 10% were more than 100 meters from each other. 

These observations are important in informing how we can improve on data accuracy, especially the geolocation data which is important in the quest for improving precision public health. 

Achieving data accuracy through open-source communities

Enketo, the form engine which the CHT uses for web forms, has continued to improve on its functionalities, including improving geospatial data capture. However, the need to improve geolocation data accuracy has been identified as an issue worthy of more attention, and more users are seeking to improve on it. Learnings from the data collected through the CHT and other platforms using Enketo will be helpful in identifying ways to improve accuracy and reliability of collected geolocation data, and in the long term, contribute to improved precision public health. It also increases openness, as involving users in the collection of geolocation data is a possible enhancement to data quality and data volumes as opposed to general fears that drive passive geo-data collection.

Improving geospatial data accuracy is a journey for Medic and other members of the CHT community. Together, we are making efforts towards achieving data accuracy, as this is key in understanding and expanding universal health coverage.

Scroll to Top