Why Physicians Resist Using Algorithm-Based Apps

Algorithm-based technologies, including AI applications, in health care can provide incredibly accurate predictions and diagnoses for a plethora of diseases and can educate physicians in real time about treatments they might not be familiar with. For example, recent research that I conducted with colleagues shows that physicians can use mobile apps for timely, high-speed access to online repositories of information about infrequently used drugs.

But our interviews with 32 physicians (across 12 specialties) that work in the United Kingdom’s National Health Service (NHS) found that doctors are reluctant to use algorithms to help them make diagnoses. All 32 of the people were hesitant even though they recognized the value. There were several reasons:

— They lacked guidance or support on how to use technology. “You know absolutely nothing, and yet “you’re not told anything,” one physician said.

— The technology was not formally integrated into their workflows. “It’s all a bit niche right now,” one doctor named Paul said. Another complained that the technology was incompatible with his role.

— They felt that they needed to behave “professionally” in the presence of patients and worried about how patients, colleagues, and senior management would perceive their use of the technology. One physician explained that “I think looking at your phone might be perceived by onlookers or patients as something social rather than actually work-related and assume [that I am] not doing work.”

— They feared that if they heavily depended on technology to make diagnoses and come up with treatment plans, their own expertise would deteriorate. One physician explained: “I think what you learn as a doctor in training you need to nurture … because if you don’t … you’re going to lose the skill of how to talk to patients [and] how to come up with a treatment plan. And if you become too dependent on things like [technology], you’re potentially de-skilling yourself in other more important ways.”

Physicians can cope with these issues by taking certain measures:

1. They should communicate clearly to patients that the technology complements, but does not replace, their expertise and, as a result, can lead to a fuller understanding of their conditions and better health care outcomes.
2. When there is no clear guidance from senior management about the use of a particular technology, physicians can obtain information about its appropriate use by observing their colleagues or respected role models in their organization.

One such role model we interviewed told us: “I’ve recommended [mobile apps] to junior doctors that are on call. [I told them that] I found them very useful when I was in their position [because] the app gives you treatment options [and tells you] what are the indications [that the patient requires] an operation.”

3. Physicians should make sure that other health care providers and patients are using the technology appropriately. For example, several physicians we interviewed discussed how they use mobile apps that employ AI (e.g., PostOp) after treatment to help guide patients in their recovery. Such apps explain to patients what their recovery journey is going to look like.

Greg, a surgeon, told us how he uses the app to set milestones that patients who have had knee surgery need to achieve in physical therapy. “Normally, when I see patients in clinic I can explain to them what that is and [how] the app has a little bit more information. After the operation, they have to have physiotherapy, and each hospital is slightly different in what regime they use, but [the app] gives the patient the idea. They can gauge their progress [by] looking at the app. Most of them even discuss the similar guidelines and milestones [provided by the app] with the physiotherapists to make sure that they’re achieving their targets.”

4. Health care organizations, as well as specialty societies and medical associations, should take the lead in encouraging doctors to use the technology — for example, by providing detailed guidelines or instructions for how to use the technology in particular kinds of care.

Medical tools based on algorithms, including AI, are proliferating and have the potential to improve health outcomes. But to a significant extent, whether they do have an impact depends on physicians’ willingness to use them and their ability to integrate them effectively into care.