TL;DR
This paper introduces Sangoshthi, a low-cost mobile platform designed to train Community Health Workers (CHWs) in rural India, overcoming challenges of limited internet access and infrastructure. By combining internet and IVR technology, Sangoshthi facilitates interactive, peer-learning-based training sessions, demonstrating significant improvements in CHW knowledge and confidence.
What Problem Does This Paper Address?
In India, Community Health Workers (CHWs), known as Accredited Social Health Activists (ASHAs), play a crucial role in delivering maternal and child health services to rural populations. Despite their importance, traditional face-to-face training methods face significant hurdles: high costs, logistical complexities, and limited opportunities for ongoing capacity building. These challenges lead to suboptimal knowledge and skill sets among ASHAs, impacting their effectiveness. Furthermore, rural areas often suffer from intermittent internet access and a lack of smartphones, making conventional digital learning solutions impractical. This paper addresses the need for an accessible, scalable, and cost-effective training platform that can empower ASHAs in their local environments without requiring additional infrastructure.
Methods
Sangoshthi employs a client-server architecture, with an Android application for hosts and a server built on Freeswitch for call management. The core innovation lies in its hybrid approach: it leverages internet connectivity for the host's smartphone application to manage sessions and for non-voice data transfer, while utilizing Interactive Voice Response (IVR) technology for audio communication over the GSM cellular network. This allows CHWs with basic feature phones to participate in live, interactive "talk shows" hosted by experts. The system supports a conference call format where hosts can manage participant interaction, allowing listeners to express opinions by pressing "1" on their phone keypads. This design minimizes reliance on high-bandwidth internet, making it suitable for low-connectivity environments.
Datasets
The study utilized data collected during a four-week field deployment of Sangoshthi with 40 ASHAs in Haryana, India. The data collected included:
- Call Logs: Detailed records of call activities, including show call answer/drop timestamps, mute/unmute actions, and DTMF events.
- Questionnaire Data: Pre- and post-intervention questionnaires administered via telephone to assess knowledge levels.
- Interview Data: Transcriptions of telephonic interviews with ASHAs and face-to-face interviews with the expert and hosts, as well as transcriptions of 12 show call recordings.
No new benchmarks were introduced; the study focused on evaluating the existing system's effectiveness.
Key Insights and Ideas
- Bridging the Digital Divide: Sangoshthi effectively overcomes the limitations of low internet penetration and feature phone usage in rural India by integrating IVR with internet-based control.
- The Power of Peer Learning: The platform fosters a lively environment for peer learning, where ASHAs can share experiences and learn from each other's challenges and solutions.
- Complementary Training: Sangoshthi is not intended to replace traditional training but to complement it, providing in-situ learning opportunities without the need for additional infrastructure.
- Host Empowerment: The system is designed to be user-friendly for hosts, even those with no prior experience, by incorporating features like pre-recorded content and clear interaction management.
- Adaptability is Key: The system's ability to adapt to network issues (e.g., call drops) and user behavior (e.g., managing speaker priority) highlights the importance of robust design in challenging environments.
Key Results
- Knowledge Improvement: The treatment group, which used Sangoshthi, showed a significant 16% improvement in knowledge scores post-intervention, compared to a 5% improvement in the control group. Post-treatment scores were significantly higher for the treatment group (50%) than the control group (33%).
- Increased Confidence: ASHAs reported increased confidence in handling cases after participating in the Sangoshthi sessions.
- Positive User Experience: ASHAs, hosts, and the expert generally had positive experiences with Sangoshthi, finding the show duration appropriate and the sessions engaging.
- Interaction Metrics: An average show lasted over an hour, with 46 minutes dedicated to Q&A. On average, 12 unique speakers participated per show, and the system facilitated structured interaction without overlaps.
- Cost-Effectiveness: While call charges were the primary expense, the per-ASHA cost was estimated to be significantly lower than traditional face-to-face training, with potential for further reduction if ASHAs bear their own call charges.
Limitations and Future Work
- Limited Interaction Modes: The current version of Sangoshthi primarily supports instructor-learner interaction. Future work aims to incorporate learner-learner and learner-content interactions, potentially through quizzes and offline feedback mechanisms.
- Feature Phone Dependency: The current design is optimized for feature phones, limiting interaction to keypad inputs. Future work plans to extend the system to support smartphone users with richer interaction components.
- Parallel Sessions: The initial design did not allow for multiple parallel independent training sessions. Subsequent updates to the underlying Citizen Radio platform have addressed this.
- Content Availability: The authors plan to make show content available through regular IVR systems and explore ways to engage listeners beyond the one-hour duration.
- Network Robustness: While call drops were managed, persistent connectivity issues for some users were noted. Further improvements in network handling and user error mitigation are ongoing.
Summary updated