Implementation of AI-Enabled Obstetric Ultrasound in Kenya: A Qualitative Study of Anticipated Usability, Feasibility, and Acceptability.

In many parts of the world, a pregnant woman's journey to a healthy birth is often interrupted by a simple lack of access to a specific kind of medical eye. Obstetric ultrasound is a standard tool in modern medicine that allows doctors to see inside the womb, checking the baby's growth, spotting potential complications early, and guiding critical decisions about care. However, in low-resource settings, the machines required for these scans are often too large, too expensive, or too dependent on a steady electrical grid to be useful. They sit in city hospitals while women in remote villages travel for hours, sometimes days, to reach them, often arriving too late for effective intervention.

To bridge this gap, researchers are developing a new kind of tool: an ultrasound device powered by artificial intelligence. This technology aims to be small enough to carry in a bag and smart enough to help a nurse or a midwife, rather than a highly specialized radiologist, take a clear picture and understand what it means. The idea is that if the machine can guide the user, the need for expensive, specialized experts in every village disappears. But before such a device can save lives, it must be trusted and used by the people it is meant to help. The success of this technology depends not just on how well the computer works, but on how well it fits into the complex reality of local hospitals, communities, and daily life.

A team of researchers from Kenya and Spain recently set out to understand exactly how this future technology would be received before it was even built. They did not test a physical machine on patients. Instead, they gathered a wide range of people—pregnant women, their partners, doctors, nurses, community leaders, and health officials—from both a busy city hospital in Nairobi and a rural district in Rabai. They asked these 84 individuals to imagine a world where this new, portable, AI-guided scanner was available. Through hundreds of conversations and group discussions, the researchers listened to what these people thought about the idea, what they feared, and what they hoped for.

The participants generally saw the potential in the device. They imagined a small, battery-powered machine that could be carried to a remote clinic or a home visit, eliminating the need for a woman to travel far to a city hospital. They believed it would be simple to use, perhaps requiring less technical skill than the large, complex machines currently in use. Many felt that if the device could help a nurse or a midwife take a scan and get a quick report, it would save time and reduce the strain on overworked staff. The promise of bringing a diagnostic tool to the doorstep of the community was a powerful vision that resonated across all the groups they spoke with.

However, this optimism was not unconditional. The people interviewed made it clear that a simple machine could not replace the need for human judgment and training. While they welcomed the idea of task-shifting—allowing non-specialists to perform scans—they worried that without proper training, the results could be wrong. They emphasized that a device might show a picture, but a trained person is still needed to explain what that picture means to the mother and to decide on the next steps in care. There was a strong consensus that the technology should support the healthcare worker, not replace their expertise. If the machine made a mistake, or if the person using it did not know how to interpret the data, the consequences could be serious.

Trust emerged as a central theme. In the rural areas, where people are less familiar with advanced technology, there were genuine concerns about safety. Some participants worried that the "artificial" nature of the device or the idea of radiation might harm the mother or the baby. Others wondered if a small, portable device could really see as clearly as the large, stationary machines in the city. This skepticism created a paradox: the very features that made the device attractive—its small size and portability—also made some people doubt its power. The researchers found that trust would not be given immediately; it would have to be earned over time through education, clear communication, and seeing the device work successfully in the hands of people they knew and respected.

The conversation also revealed that the technology does not exist in a vacuum. For the device to work, the entire health system around it must be ready. Participants pointed out that finding a problem with a scan is useless if there is no specialist available to treat it, or no clear path to refer the patient to a hospital. They warned that introducing a new tool without fixing the underlying gaps in the system—such as a lack of staff, poor roads, or insufficient funding—would lead to frustration. They stressed that the device should not be a separate, short-term project funded by outside donors, but rather something integrated into the government's health system with a long-term plan for funding and maintenance.

Social dynamics played a significant role as well. In many communities, decisions about healthcare are made collectively, often involving husbands, elders, or religious leaders. The researchers found that even if a pregnant woman wanted to use the new scanner, her partner or community leader might say no if they did not understand or trust it. Religious leaders, in particular, were seen as powerful gatekeepers whose approval could make or break the acceptance of the technology. The study suggested that for the device to be accepted, the community would need to be involved from the start, with trusted local figures helping to explain the benefits and address fears.

Ultimately, the study concluded that the path to success for AI-enabled ultrasound in Kenya is not a straight line of technological deployment. It is a complex journey that requires aligning the tool with the people who will use it and the systems that will support it. The device must be affordable, the users must be well-trained, and the community must trust that it is safe and accurate. The researchers found that while the technology holds great promise for improving maternal health, its value will only be realized if it is introduced with deep respect for the local context, the workforce, and the social fabric of the communities it serves. The future of this tool depends as much on human relationships and system readiness as it does on the intelligence of the machine itself.