Colorectal cancer is the second leading cause of cancer deaths in the country. Finding the disease early by screening, like with the fecal immunochemical test (FIT), helps lower deaths. Even though screening is helpful, some groups get screened less often. For example, Hispanic and Latino people have a screening rate of 53.4%, while non-Hispanic white people have a rate of 70.4%. This difference causes more deaths and is a challenge for healthcare systems.
Language and cultural differences are big reasons why Spanish-speaking and other non-English-speaking groups do not get screened as much. Usual ways to reach patients, like calls from hospital staff or letters in the mail, do not work well for people who don’t speak English well or don’t use online health systems. This makes it hard for doctors to offer preventive care to all patients.
New technology in artificial intelligence (AI), especially agents that talk in many languages, can help improve how patients get information and increase screening rates. This article looks at how multilingual AI care agents help get more people screened for colorectal cancer in places like Pennsylvania and Maryland. It also talks about how these AI tools can help medical offices run more smoothly in the U.S.
Hippocratic AI worked with WellSpan Health to create an AI care agent called Ana. Ana speaks both English and Spanish. She helps solve language and communication problems around colorectal cancer testing. The AI talks on the phone with patients, explains why screening is important, and helps them set up to get FIT test kits.
In September 2024, a study was done with WellSpan Health in Pennsylvania and Maryland. It tested how well Ana reached 1,878 patients who needed colorectal cancer screening but did not have active online health accounts. These patients included 517 Spanish speakers and 1,361 English speakers.
The study found a big difference in how many patients answered the calls. Almost 9 out of 10 Spanish speakers (89%) picked up, while just over half of English speakers (53%) did. Spanish speakers talked on the phone longer, about 6 minutes, compared to 4 minutes for English speakers. This extra time shows they talked more and were more interested, probably because they could use their own language.
Most importantly, 18% of Spanish speakers agreed to get the FIT test kit. That is more than two and a half times higher than the 7% of English speakers who agreed. This shows that when people get information in their own language, they are more likely to participate in preventive care.
Spanish-speaking patients in the study were on average younger (57 years) than English speakers (61 years). There were more women in the Spanish-speaking group (49.1%) compared to the English-speaking group (38.4%). But even when considering age, sex, and call length, speaking Spanish doubled the chance of agreeing to the FIT test.
These results show that AI that speaks the patient’s language helps reduce health differences. By breaking language barriers, the AI not only teaches but also builds trust for groups that usually get less care.
Helping Hispanic and Latino groups get preventive care like cancer screening is very important. Low screening leads to higher death rates from cancer in these communities. Language differences, culture, and less knowledge about healthcare often make these patients avoid care.
Ana, the AI, talks naturally and kindly in Spanish. This helps patients understand and feel comfortable getting information. Unlike regular calls by humans, which can be limited by time, staff, and language skills, AI lets patients ask questions and get answers without feeling rushed. The longer call times with Spanish speakers show how well this works.
This kind of AI outreach also helps patients who don’t use online health services. It fills in gaps where emails or online messages don’t reach, helping medical offices connect with patients who might have been missed.
Experts like Dr. Meenesh Bhimani of Hippocratic AI say that AI in many languages can close health gaps instead of making them worse. Dr. R. Hal Baker from WellSpan Health adds that Ana makes sure patient talks stay kind and steady, especially for Spanish speakers.
Doctors, office managers, and IT staff can use AI to do more than just talk to patients. AI can help with office tasks, save staff time, make patients happier, and improve screening rates.
Generative AI voice agents are not simple chatbots. They use large language models that can have smooth, clear conversations. They change what they say based on the patient’s needs. They also use language and tone that match the patient’s culture and reading level.
In colorectal cancer screening calls, these AI agents give medical advice that is very accurate. Studies with over 300,000 fake patient talks found that more than 99% of the advice was correct with no serious harm. This is good news for doctors thinking about using AI.
Besides calls, these AI agents can help check on chronic diseases, decide how bad symptoms are, and make sure patients take their medicine. They do this in many languages. For medical offices, this tech helps speed up answering patients and reach many people fast.
But problems remain. Sometimes the AI waits too long or talks over the patient. AI must be watched and updated often to follow safety rules and laws. Healthcare groups using these must train staff and check quality carefully.
The study from WellSpan Health and Hippocratic AI shows that AI care agents speaking many languages help get more Spanish-speaking patients screened for colorectal cancer. More widely, AI voice tools in healthcare can make work run smoother, reduce health differences, and help medical teams give better preventive care. Using this technology carefully will be important for healthcare offices wanting to improve care and health for patients who do not speak English well in the United States.
The primary objective was to evaluate the effectiveness of a multilingual AI care agent in engaging Spanish-speaking patients for colorectal cancer screening compared to English-speaking patients.
The study included 1878 patients eligible for colorectal cancer screening; 517 were Spanish-speaking and 1361 were English-speaking patients without active web-based health profiles.
The AI agent made personalized telephone calls in the patient’s preferred language, provided education about colorectal cancer screening, and facilitated fecal immunochemical test (FIT) kit requests.
The primary outcome was the fecal immunochemical test (FIT) opt-in rate to gauge patient engagement with colorectal cancer screening.
Spanish-speaking patients had significantly higher engagement: FIT opt-in rates were 18.2% versus 7.1%, connect rates were 69.6% versus 53.0%, and call durations averaged 6.05 minutes versus 4.03 minutes for English speakers.
Yes, Spanish language preference was an independent predictor of FIT test opt-in with an adjusted odds ratio of 2.012, meaning Spanish speakers were twice as likely to opt-in after controlling for demographic factors and call duration.
Spanish-speaking patients were younger (mean age 57 vs 61 years) and more likely to be female (49.1% vs 38.4%) compared to English-speaking patients.
The findings suggest that language-concordant AI outreach can reduce longstanding disparities in preventive care access by significantly increasing engagement among non-English-speaking populations.
Limitations included being conducted in a single healthcare system, a short study duration, and the absence of follow-up data on whether patients completed screenings after opting in.
Future research should focus on assessing long-term adherence to screenings and determine whether increased engagement with AI outreach translates into improved clinical outcomes for patients.
Simbo is using AI agents to automate all phone related workflows. Connect now to know more.
Schedule AI Agents Demo Now© Since 2019, Simbo AI.
