June 1, 2023
On Wednesday, May 17, 2023, the Global Handwashing Partnership and USAID MOMENTUM hosted a webinar entitled “Leveraging Human-Centered Design and Artificial Intelligence (AI) Systems for Behavior Change in Health.” The webinar shared an example from the SafeHands Project implemented under USAID MOMENTUM and included discussion focused on the replicability, sustainability, and ethics of using AI technology for behavior change.
The webinar was moderated by Aarin Palomares (Global Handwashing Partnership), who welcomed participants and presented an opening question to spur thinking for the open discussion portion of the call. The question was, “What would you keep in mind from an ethics perspective if this project was going to be implemented in your country?”
Following the short welcome, Nga Nguyen, a Senior Advisor for WASH and Behavior Change at USAID, provided the group with opening remarks. In her remarks, Ms. Nguyen highlighted current activities funded by USAID, including several behavior change activities focused on communication and the use of nudges to change hand hygiene behavior. Beyond USAID’s handwashing activities, the Agency also invests in infection and prevention control initiatives, including the SafeHands Project that is being presented today. She noted that the SafeHands project is an interesting and timely example for behavior change practitioners, as it brings together behavior change, human-centered design, technology (handwashing device), and ethics and compliance with the use of AI in various settings. She closed by stating that while many believe hand hygiene should be fundamentally promoted in many settings, this webinar investigates how we do that and the many questions on ethics and sustainability when it comes to AI as an emerging technology for development issues.
AI and Human-Centered Design for Behavior Change in Health
The Quicksand team presented findings from the SafeHands project. The team included:
Quicksand is a firm based in India that uses design methods and research to build people-centered products, services, systems, and strategies.
Ms. Shah started the presentation by providing some background on the project and the team’s approach. The SafeHands project aimed to use human-centered design to explore whether AI could be used to improve handwashing in the labor rooms of India’s public health facilities. This project was implemented across 6 hospitals and 3 states in India for 12 months. As part of this process, the team identified the individual and organizational factors affecting the uptake of the AI system and the core behavior of handwashing. The team’s approach followed human-centered design processes and is outlined below, first starting with qualitative research to understand people’s needs and preferences.
Ms. Baveja provided more detail on the team’s co-creation process to develop the final intervention. Following the initial steps of human-centered design (qualitative research to develop behavior insights), the team went through a co-creation process where they identified and tested multiple solutions. These solutions were given to participants on the ground to brainstorm and improve initial ideas to create new iterations and to inform the final intervention.
Based on this process, the team used both digital and non-digital products to design the final intervention. For the digital component, the team used an AI device called Varjahands 2.0, which has 2 components: 1) a display screen that is placed above a basin – the camera can recognize hand movements to ensure WHO’s 9 steps for handwashing are followed and provides live feedback of handwashing technique and 2) a dashboard with aggregated, anonymized data to help managerial staff monitor overall hand hygiene compliance within the facility. The team also proposed non-digital assets, including performance reports, orientations, peer learning sessions and patient feedback, as part of the intervention. The project is currently in Phase II, with the next phase focusing on tertiary facilities.
Finally, Mr. Sachdeva ended the team’s presentation with some thoughts on sustainability. He noted performance insights, with hand hygiene performance dropping. While the exact reasons are not known, he presented some considerations for this decrease in performance, including experiencing technological issues with the system, such as the practicalities of installing the device in public settings and connectivity to transfer data, as well as reduced on-ground support.
Following the presentation, the following questions were addressed through a question and answer session.
|Are there any publications available from your research?||No, the team has done additional presentations at various conferences but there is no academic output thus far. At the end of both phases, there may be a peer-reviewed manuscript developed.|
|Can this be linked to the Power BI dashboard or other technologies/systems?||As of now, it does not have this specific capacity. There is potential for the data to be integrated and compatible with other systems but would need to look at this specifically.|
|Did you get any willingness to pay feedback from hospital administrators? Is there maintenance cost?||The price is about $2,000 USD with one year warranty included. That price is reflective of research and development and the scale of the current operation. The price will go down with scale. Willingness to pay is a bit tricky because there have been issues on the ground with the technology, so it is a slightly complicated question at this stage. Right now, the price is too high for hospitals to pay and support from donors and project is still needed at this stage.|
|Is the compliance measured based on the WASH technique or is it compliance to hand hygiene based on opportunities?||The device measures quality not quantity.|
|What learnings are you taking to the next phase?||There are limitations to the device that will need to be further explored. Based on the scope of the device, there is no intention of using the device to measure quantity as that brings up a question of ethics and personal identification to track.|
Following the Q&A session, there was an open discussion to discuss these issues more broadly. Mr. Sachdeva started this discussion by making a note that administrators and staff had different perspectives of how the device should be used. Administrators had expressed interest in more personalized data, while staff reflect the realities of their daily lives. This also brings up the question of using anonymity within the data system – with aggregated, anonymized data, some staff may be blamed for poor compliance of other staff members. These contradicting perspectives are setting-specific and can be quite nuanced, which is why the human-centered design process is important to better understand each context.
The full webinar recording is available here.
If you would like to receive regular updates from the Global Handwashing Partnership, please subscribe to our mailing list or follow us online.
© 2017 The Global Handwashing Partnership (GHP).