In hospitals and medical offices, phone lines and patient portals can quickly become overwhelmed with calls and messages.
Patients often call with questions about medications, appointment scheduling, or test results.
Doctors and nurses, already pressed for time, must manage these communications while ensuring high-quality care.<\/p>\n
AI chatbots help by automating routine interactions.
For example, the University of Pennsylvania\u2019s Abramson Cancer Center uses a system called Penny.
Penny is an AI-powered text messaging service that checks in daily with patients on oral chemotherapy for gastrointestinal cancers.
It asks patients about their medication plans and side effects, creating a continuous monitoring channel outside of clinical visits.
This system alerts clinicians if any responses suggest problems requiring attention.<\/p>\n
Northwell Health applies similar chatbot technology to a broader patient base, including postpartum women and chronic disease patients.
Their chatbots engage patients through tailored questions based on individual medical conditions.
The goal is to monitor health regularly after discharge and reduce the risk of hospital readmissions.<\/p>\n
At UC San Diego Health, the chatbot is integrated into the MyChart patient portal.
It scans incoming patient messages, drafts initial responses for non-emergency inquiries, and forwards these drafts for clinician review before sending.
This method helps clinicians manage communications more efficiently without losing the human touch necessary in patient interactions.<\/p>\n
Studies support the use of AI chatbots in patient communication.
In one, 78.6% of evaluations preferred chatbot responses over physician responses, noting greater empathy, tone, and thoroughness in the chatbot replies.
Patients who use these systems often prefer texting over calls because they can respond at their convenience.<\/p>\n
Despite many advantages, the use of AI chatbots in healthcare needs careful management.
Automated responses must be medically accurate and convey empathy to maintain trust between patients and providers.
Several experts say that AI cannot operate independently without human oversight.<\/p>\n
Christopher Longhurst, MD, explains it clearly: “A clinician absolutely has to remain in the loop and be engaged with the message.”
This requirement helps prevent mistakes and ensures messages do not sound overly robotic or impersonal.
When a chatbot produces a response, clinicians review and revise it as needed to confirm medical accuracy and adjust the tone so that patients feel heard and cared for.<\/p>\n
In cancer care, where medication plans are complex and risks high, the stakes are particularly significant.
Lawrence Shulman, MD, who oversees Penny at the University of Pennsylvania\u2019s Abramson Cancer Center, notes that patients on oral chemotherapy may go weeks between visits.
During that time, AI communication systems help monitor patients, but clinician review assures that concerns raised by patients are addressed promptly and accurately.<\/p>\n
Jeffrey Ferranti, MD from Northwell Health points out the pressure on physicians post-pandemic, describing them as “burned out and overburdened.”
AI systems help reduce workload, but clinician supervision is necessary to ensure quality and prevent misinformation, allowing doctors to focus on direct patient care.<\/p>\n
Maintaining patient trust also means being clear about how AI chatbots work.
Patients should be allowed to opt-in to chatbot communications and be informed about data privacy and how their information will be used.
This practice helps patients accept and use the technology.<\/p>\n
Healthcare AI often serves two main purposes:<\/p>\n
Because patients have different levels of health knowledge and conditions, chatbot questions are often customized.
Northwell Health\u2019s system adjusts questions based on patient risks and diagnoses.<\/p>\n
<\/p>\n