"Providing access to reliable health information for health workers in developing countries is potentially the single most cost effective and achievable strategy for sustainable improvement in health care."
Pakenham-Walsh N, Priestly C, Smith R
"Meeting the information needs of health workers in developing countries"
British Medical Journal 1997; 314:90.


New: Watch the video to learn about our latest work.

HealthLine Case Study


Community health workers in developing countries are often trained only for a few months before they begin providing health services, and have a great need to access updated and reliable health information. In the HealthLine project, we're designing an information access system specifically for such health workers, which they can call any time, toll-free, from any cell phone or landline. Through a spoken conversation with the system in their native language, they would be able to learn more and/or fill in any health information gaps that they may have. This would in turn enable them to provide better health services to their community.

HealthLine is a collaborative research project involving Carnegie Mellon University (Pittsburgh, PA), Aga Khan University (Karachi, Pakistan) and the Health & Nutrition Development Society (Karachi, Pakistan).

HealthLine is supported by:


As of January 2009, we have completed a large user study with health workers in rural Sindh comparing two automated information access systems in the Sindhi language. In one system, health workers would use the telephone keypad to select menu choices, while in the other they would speak instead. Most users preferred the spoken interface, though there are still significant challenges for the least literate users. Such users had significant difficulty using either system, and we are now looking into ways of streamlining the interaction based on their needs.

As of May 2008, we have completed a series of user studies with health workers in rural Sindh to understand their preferences with respect to speech-based information access as compared to text-based information access. One major finding is that our initial choice of Urdu (Pakistan's lingua franca) as the target language was sub-optimal, as community health workers often are not always fluent in Urdu -- even though they may have attended school up till 5th grade. We are now re-building our prototypes in Sindhi, and will be conducting user studies with health workers of varying literacy levels to evaluate the prototypes through objective metrics (such as task success rate, task completion time) and subjective metrics (user ratings).

As of September 2007, we have completed a series of field-based interviews with various health workers to understand the context in which they work, their literacy and comprehension skills, as well as their baseline health information needs. We have identified appropriate health information texts, and have completed a successful prototype test in August. We are now creating a system with more information, utilizing voice-based search, and will be evaluating it with a wide range of community health workers, with varying literacy levels.

Research questions:

  • What are the principles of speech interface design for a voice-only information access mechanism that provides large amounts of detailed information? Since low literate users have no other options for automated information access, all details must be provided through the voice channel.
  • What role can voice-based search play in such interfaces?
  • How do low-literate users view voice-based search? Can they easily learn how to use it?
  • Most existing content is designed for visual access, and is optimized for reading. How must this content be adapted so that it is optimized for hearing?

(You'll need Flash if streaming, XviD or VLC if downloading)

Detailed Overview (English, 34 mins)
| Google Video | Download

This video gives a detailed overview of HealthLine's goals and methodology.

Discussion Facilitation (Urdu, 13 mins) Stream | Google Video | Download

This video is meant as a tool to be used in focus group meetings with community health workers, to elicit their needs, opinions and suggestions on the project.

Concept Demonstrations (XviD required)
Download Demo 1, Demo 2 and Demo 3

These concept demonstrations show three different interaction styles that could potentially be useful for health workers. They are only meant to give a flavor of what is possible with speech technologies.


Carnegie Mellon

Jahanzeb Sherwani
Roni Rosenfeld
Rahul Tongia

Cornell University

Nosheen Ali

Aga Khan University

Yousuf Memon
Mehtab Karim

Health & Nutrition Development Society

Tanveer Ahmed
Sarwat Mirza
Anjum Fatma



Jahanzeb Sherwani, Nosheen Ali, Carolyn Rose, Roni Rosenfeld. Orality-Grounded HCID: Understanding the Oral User. To appear in Information Technologies & International Development, Special Issue on Human Computer Interaction and Development, December 2009.

Jahanzeb Sherwani, Sooraj Palijo, Sarwat Mirza, Tanveer Ahmed, Nosheen Ali, Roni Rosenfeld. Speech vs. Touch-tone: Telephony Interfaces for Information Access by Low Literate Users. In Proc. Information & Communications Technologies and Development, Doha, Qatar, April 2009.

Jahanzeb Sherwani, Nosheen Ali, Sarwat Mirza, Anjum Fatma, Yousuf Memon, Mehtab Karim, Rahul Tongia, Roni Rosenfeld. HealthLine: Speech-based Access to Health Information by Low-literate Users. In Proc. IEEE/ACM Int'l Conference on Information and Communication Technologies and Development, Bangalore, India, December 2007.

Jahanzeb Sherwani, Nosheen Ali, Rahul Tongia, Roni Rosenfeld, Yousuf Memon, Mehtab Karim, Gregory Pappas. HealthLine: Towards Speech-based Access to Health Information by Semi-literate Users. In Proc. Speech in Mobile and Pervasive Environments, Singapore, September 2007.

Jahanzeb Sherwani, Dong Yu, Tim Paek, Mary Czerwinski, Yun-Chen Ju, Alex Acero. VoicePedia: Towards Speech-based Access to Unstructured Information. In Proc. Interspeech, Antwerp , Belgium

Jahanzeb Sherwani, Rahul Tongia, Roni Rosenfeld, Yousuf Memon, Mehtab Karim, Gregory Pappas. Towards Speech Interfaces for Health Information Access by Semi-literate Users. In Proc. AI in ICT for Development, Hyderabad , India


Join our mailing list and receive updates on the project: