Iteration is a benefit of easily adaptable software
Once you release a version of your data collection app into the field, you may learn new ways in which it can run better. Even if you tested your application with users prior to release, having them use it in real life scenarios may highlight areas where the application needs improvement. Or maybe the government releases new standards of care after launch, and your forms are out of date. How do you optimize your app after it has already been built and deployed?
Digital tools can be easier to adapt while still in use, but there are some tips to keep in mind for a smoother process. First, let’s unpack what we mean by optimization.
3 Steps to optimization
How can you learn from the real-life running of the program? Look for opportunities to optimize your app, if your circumstances do not already lead you in this direction.
1. Look at the data
Now that there is a program collecting data digitally, this is a valuable resource going forward. You can start to analyze and use the available data collected by the system, to improve the system and adapt it further.
In CommCare, you can see which forms are being submitted by each user under the “Submissions by Form” report. Find out whether there are any forms being used less than expected, and start compiling questions to take back to the users.
For example, in one project, hardly anyone was using the “Register Pregnancy” form, and the program had to find out why – were fewer women falling pregnant? Were the users aware they needed to register with that particular form? Were women unwilling to have their data entered in the phone? By following up, you can then tailor a solution to the problem.
Program performance indicators
By exporting the form data or case data, you can analyze your program according to its performance indicators. You will know what to look for, but some examples to get you thinking are:
- How many households received follow up on time?
- Which is the preferred method of family planning for married women?
- How many people know where their closest health facility is?
- How many children accessed counseling?
- How many farmers did not have the right knowledge to plant their crops on time?
Analyzing the program might give you ideas for optimization, such as a new counseling module specifically for children, with multimedia and cartoon images, because they were not accessing the standard counseling form very much. Or if most women are using condoms, do they know about the other family planning methods? Do you need to include more educational content to broaden their options? Do you need to add dates and icons to remind users to follow up with households on time?
Worker performance metrics
In CommCare, there are seven generic worker activity reports to give you insight into when your users are completing and submitting forms, and what their interaction level is with each case (a case could be a mother, a farm, or whatever else you are tracking over time).
This research can identify gaps in the users’ knowledge or performance that could be addressed through additional content, such as a training module or better UX design.
2. Understand from the user perspective
After analyzing the data, take any questions back to the users to try and understand the why behind some of the findings.
- Why are most follow up visits completed late? Are the households unavailable? Should a “visit attempt” be captured instead? Is it unclear in the app that a follow-up is due? Are users confused by the workflow?
- Why do most users complete all their forms by Thursday? Does something happen on Fridays to prevent data collection?
It is also good to get user feedback now that they have had a chance to use the application alongside their real-life workflows. Check in with them to see if any workflows can be improved going forward. Is anything hindering them from collecting data? What has been revealed that testing did not pick up?
3. Combine the insights into additional content and workflows
Take what you have learned from analyzing the data and talking to the users, and see how this translates into new content or tweaks to existing content in your app. Involve your users in brainstorm sessions, or host some focus groups with beneficiaries to understand what to optimize.
Once you have an idea of changes to be made, start mapping them out. Questions to ask yourself at this stage are:
- Will the changes add to the current workflow or replace it?
- Are you adding new forms or modules?
- What media can be added to enhance the experience for beneficiaries and users?
- Do you need to improve the reminders, icons, and alerts for follow up or regular tasks?
Two examples to learn from
Data-driven decisions: adding new screening criteria, based on good data
In one of our projects, a paper process screened children for common illnesses and then they were either referred to or treated at the village clinic. When we digitized the tool, we were able to collect additional data at the household level, such as whether the adults had been exposed to tuberculosis, which is a variable that used to be collected separately.
Because the app facilitated real-time data collection, it made it easier to analyze the combined trends and gave the program directors the idea to start flagging children who were more at risk of having TB symptoms. For example, determining how many children displayed a cough and fast breathing and how many lived with an adult who had TB.
Once the algorithm for catching children at increased risk was defined and the new workflow mapped out, it was built into the existing form and deployed to the field. Because it was an addition to the end of existing workflows, it did not affect previously collected data. It did, however, enable those screened in Version 2 to also be flagged for TB screening where necessary.
User experience: small tweak to accommodate user preference
In the same form, the binary ‘fast breathing’ variable was calculated by the children’s breaths per minute. On paper, this process was done with a 60-second timer, as the health workers counted the child’s breaths. In the digital version, we introduced a breath counter app, which integrated with the CommCare form, making counting the breaths much easier (and more accurate). Health workers would tap for each breath counted in a 60-second window before the result was submitted. During initial user testing, the group of participants expressed how much they enjoyed the new technology.
But once the app was rolled out to all users, the program started getting some feedback about users who preferred the old, analog method of breath counting, which some of them had been using for the last eight years or more. Some veterans were unhappy that they were forced to use the digital version of the breath counter.
Therefore, for Version 2 of the app, we decided to let the users choose which version of the breath counting method they wanted to use. This small branch in workflow helped users to feel like their voices were heard, and catered for two ways of getting the same data.
Applying these lessons
There are just a few ideas and tools to help you start optimizing your mobile applications that are already in use. Unlike paper-based forms, which cost a lot to reprint and distribute, the beauty of digital tools is that they can be adapted as you learn new ways of doing things. With careful analysis of current usage, a complete understanding of the app’s workflows, and a clear plan for implementing the new content, you can ensure your app is optimized to suit the users and beneficiaries in your program.