experiment · JAMA network open · la publicación, 1 oct 2026 · gratis
Un chatbot ayudó a jóvenes con un gen de cáncer hereditario a entender su riesgo antes de los 30
En cuatro centros académicos de Estados Unidos, el conocimiento subió del 52% al 78% después de la visita. El estudio no prueba que el aparato calme a nadie, ni que funcione fuera de esas clínicas.
Versión breve · la versión detallada sigue, unos 9 min
Pregunte al chat de weeklyAI
Pregúnteme sobre la IA en su vida: le señalo las notas que weeklyAI ha publicado sobre el tema, esta semana y antes.
Las conversaciones se guardan mientras exista weeklyAI, para mejorar la publicación. Se responde en el idioma en que usted escribe.
- El estudio, de un vistazo
- Quiénes
- Jóvenes de 12 a 24 años con un síndrome hereditario de predisposición al cáncer
- Cuántos
- 106 jóvenes
- Dónde
- Cuatro centros académicos de Estados Unidos
- Cuándo
- No lo dice el pasaje
- Tipo de estudio
- experiment
- Quién lo hizo
- Dana-Farber Cancer Institute, Emory, Universidad de Chicago y Huntsman de Utah
- El límite que importa
- Se hizo en cuatro centros académicos, con población sin diversidad racial ni étnica
Consulta sola frente a consulta más chat electrónico
El conocimiento correcto del riesgo antes de los 30 años pasó de 38% a 37% sin chat, y de 52% a 78% con chat
La angustia no aumentó con el chat; la baja observada no fue concluyente
Un temor razonable era que saber más angustiara a los jóvenes. No ocurrió: la angustia no subió en el grupo del chat.
En una familia donde circula un gen de cáncer hereditario, muchos jóvenes salen de la consulta sin que la visita resuelva todas sus dudas. El artículo describe un estudio anterior en el que buena parte de la conversación pasa entre el médico y los padres, y a veces en la visita se dan resultados de estudios de detección. Un equipo de cuatro centros de cáncer de Estados Unidos probó un complemento: un chat electrónico llamado AYA-RISE, que se abre después de la consulta.
El chat permite leer la información al nivel de detalle que cada quien elija. Sobre el riesgo, por ejemplo, puede decir "riesgo alto de cáncer en el futuro" o "más de 5 personas de cada 10 desarrollarán cáncer". También tiene secciones sobre qué estudios hacer según la edad, cómo hablar con amigos y dónde buscar apoyo. Lo desarrolló el equipo de investigación, sin usar inteligencia artificial generativa, y la interfaz la hizo Clear Genetics, hoy parte de Invitae/Labcorp.
Los investigadores repartieron al azar a 106 jóvenes de entre 12 y 24 años, con uno de 9 síndromes hereditarios, en dos grupos: visita sola o visita más chat. La edad promedio era de 18 años, y el 64% eran niñas o mujeres.
Antes de la visita, poco más de la mitad del grupo con chat (52%) sabía bien cuál era su riesgo de cáncer antes de los 30 años. Después, lo sabía el 78%. En el grupo de la visita sola, el conocimiento se quedó igual: 38% antes y 37% después. Esa diferencia fue medida y no parece obra del azar.
El estudio tiene bordes claros. Se hizo en cuatro centros académicos con clínicas de genética ya establecidas, y el 93% de los participantes eran blancos. De cada cuatro jóvenes invitados, sólo uno aceptó participar, y los planes originales de reclutar 300 se recortaron a 106. El tamaño real del beneficio queda por confirmar. Los participantes tenían que hablar inglés.
Un temor razonable era que saber más angustiara a los jóvenes. No ocurrió: la angustia no subió en el grupo del chat. Bajó un poco, pero esa baja no alcanzó a distinguirse del azar, así que el aparato no está probado como calmante.
Si en su familia hay un joven con un gen de cáncer hereditario, vale la pena preguntar en la clínica de genética si ofrecen algún material de seguimiento digital después de la consulta, y en qué idioma está disponible.
Qué significa para usted
Si en su familia hay un joven con un gen de cáncer hereditario, pregunte en la clínica de genética si ofrecen algún material de seguimiento por escrito después de la consulta y en qué idioma está. El estudio midió el beneficio en cuatro clínicas académicas de Estados Unidos, con jóvenes de habla inglesa, así que en su país puede que ese material no exista todavía. Vale la pena preguntar y observar si lo ofrecen.
Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794
Quién pagó: El estudio fue financiado por la subvención 1U01CA243688 del Instituto Nacional del Cáncer (Dra. Mack); el financiador no tuvo ningún papel en el diseño, la ejecución, el análisis ni la decisión de publicar, y la interfaz del chatbot fue desarrollada por Clear Genetics, ahora Invitae/Labcorp.
No tome esto como consejo médico profesional.
Los hallazgos de otros estudios que aquí se mencionan provienen de este documento; esos estudios no se consultaron uno por uno.
Versión detalladaLos pasajes copiados del artículo, las ilustraciones y cada fuente con cuánto leímos de ella · 1738 palabras · unos 9 minLeerla →Cerrar
Un chat después de la consulta ayudó a jóvenes con riesgo hereditario de cáncer a entender su riesgo antes de los 30
En un ensayo con 106 pacientes de 12 a 24 años en cuatro centros académicos de Estados Unidos, quienes usaron la herramienta después de ver a su equipo médico mejoraron su conocimiento; el estudio no prueba que reduzca la angustia ni que funcione fuera de esos centros.
Imagine la escena: una joven de 17 años sale de su consulta de genética en un hospital de Boston con una carpeta bajo el brazo. Adentro viene el nombre de un síndrome hereditario, una lista de estudios de detección y una cifra que nadie le explicó del todo. Lo que este ensayo probó fue sencillo: darle a esa joven, después de la visita, una herramienta electrónica con la que pudiera preguntar a su ritmo. El resultado fue que más de ellas entendieron su riesgo de cáncer antes de los 30 años.
El estudio se hizo en cuatro centros académicos de Estados Unidos: el Dana-Farber Cancer Institute en Boston, Emory University con Children's Healthcare of Atlanta, la Universidad de Chicago y el Huntsman Cancer Institute de la Universidad de Utah. Participaron 106 jóvenes de 12 a 24 años, con edad promedio de 18.3 años; 64 por ciento eran niñas o mujeres. Todos tenían uno de nueve síndromes hereditarios de predisposición al cáncer, entre ellos Li-Fraumeni, DICER1 y Lynch. Se les asignó al azar a dos grupos: uno solo tuvo su consulta habitual y el otro tuvo la consulta más la herramienta, llamada AYA-RISE. Se les preguntó antes y después.
La pregunta central era si sabían cuál es el riesgo promedio de tener cáncer antes de los 30 años para alguien con su síndrome. Al empezar, 52 por ciento del grupo con chat (28 de 54) respondía bien; al terminar, 78 por ciento (42 de 54). En el grupo que solo tuvo la consulta, la proporción casi no se movió: de 38 por ciento (20 de 52) a 37 por ciento (19 de 52)1. El equipo calculó que la ventaja del chat era clara y no producto del azar1.
Hubo otros resultados que el estudio mismo no puede dar por buenos. Sobre el riesgo a lo largo de toda la vida, las diferencias entre grupos no fueron concluyentes: la precisión pasó de 33 a 52 por ciento con el chat y de 44 a 50 por ciento sin él2. Sobre el conocimiento de los estudios de detección recomendados, la ventaja también favoreció al chat pero no alcanzó a ser concluyente. Y sobre la angustia: el equipo no vio un aumento del malestar psicológico después de la intervención3. La reducción de angustia que se observó tampoco fue concluyente, y los propios autores escriben que el estudio tenía pocos participantes para detectar un efecto en el bienestar4.
¿Qué es exactamente AYA-RISE? No es un chat que inventa respuestas. Todo el contenido lo escribió el equipo de investigación, sin materiales generados por inteligencia artificial5. La pantalla de inicio ofrecía seis caminos: conversar, aprender, planear, compartir, conectar y crecer6. El de conversar abría un chat interactivo que explicaba el riesgo de cáncer y los estudios recomendados, con páginas de apoyo para hablar con amigos y sobrellevar la incertidumbre7. Lo interesante es cómo manejaba las cifras: a quien quería lo general se le decía "alto riesgo de cáncer en el futuro"; a quien quería números, "más de 5 personas de cada 10 desarrollarán cáncer en el futuro", y cada quien elegía cuánto detalle quería8.
Piense en lo que cubre una consulta de genética. El equipo médico habla con el joven, muchas veces con los padres presentes, revisa resultados de estudios y explica lo urgente. Lo que no siempre queda es el espacio para la pregunta que da vergüenza, la que se piensa en el camión de regreso a casa. La herramienta se diseñó justo para eso: para usarse después de la visita, cuando el joven ya está solo con su teléfono o su tableta y puede explorar a su manera7. No se probó en lugar de la consulta, sino después de ella.
El problema que atiende es grande. Más de 10 por ciento de los adolescentes y adultos jóvenes con cáncer tienen una predisposición hereditaria, lo que eleva su riesgo de nuevos cánceres a lo largo de la vida9. La detección temprana permite encontrar tumores cuando son más tratables10. Pero los modelos de comunicación pensados para esta edad todavía están en pañales, y hacen falta nuevas formas de ayudarles a entender su riesgo y a aceptar los estudios de detección11. En otras palabras: la medicina ya sabe qué vigilar; lo que faltaba era cómo contarlo.
Vale la pena comparar los dos brazos del estudio, porque la diferencia es toda la historia. Los del grupo de solo consulta entraron con 38 por ciento de respuestas correctas y salieron con 37: la visita, por sí sola, no movió la aguja en este grupo. Los del chat entraron con 52 y salieron con 781. Conviene notar que el grupo del chat ya sabía más al empezar, y aun así fue el que mejoró. Lo que el estudio no puede decir es por qué: si fue el chat, si fue el tiempo extra frente a la pantalla o si fue simplemente haber tenido más tiempo para pensar.
El estudio tiene límites que conviene tener presentes. De 798 jóvenes evaluados, 549 resultaron elegibles y 106 aceptaron participar; los autores escriben que se inscribió 25 por ciento de los que se acercaron y eran elegibles, y que tuvieron que reducir la meta de participantes para terminar dentro del plazo de financiamiento12. También señalan que el estudio se hizo en centros académicos con clínicas de predisposición al cáncer ya establecidas, y que la población carecía de diversidad racial y étnica13. Además, los cuestionarios iniciales se aplicaron después de la asignación al azar, lo que pudo introducir sesgos de expectativa. Y la herramienta tendrá que actualizarse conforme cambien las guías de detección14.
¿Quién pagó y quién participó? El estudio lo financió el Instituto Nacional del Cáncer de Estados Unidos con una subvención a la doctora Jennifer Mack; el financiador no tuvo papel en el diseño, la conducción ni la publicación. La interfaz la desarrolló Clear Genetics, hoy parte de Invitae/Labcorp, y varios autores declararon vínculos con la industria, entre ellos empleos y acciones en empresas de genética y farmacéuticas. Los autores escriben que ven posibilidades de llevar la herramienta más allá de los centros académicos, a la comunidad, donde muchos jóvenes reciben atención y donde faltan recursos para enfermedades raras15. Es una promesa razonable, pero todavía es eso: una promesa.
Así lo leemos nosotros. Hay un patrón que se repite en la medicina: cuando alguien con autoridad decide qué información recibe el paciente y en qué forma, lo que se aprende termina siendo lo que esa autoridad contó, no lo que la persona preguntó por su cuenta. En una consulta cargada, el joven suele escuchar mientras los adultos hablan de él. El chat cambia quién controla el detalle: el joven elige si quiere la frase general o la cifra. Si eso es lo que está pasando, esperaríamos que en casa se notara la diferencia: un joven que sale de la consulta sabiendo explicar su riesgo con sus propias palabras, no solo repitiendo lo que dijo el médico. Sabríamos que nos equivocamos si los del grupo de solo consulta mejoraran igual, o si los del chat no eligieran niveles distintos ni preguntaran cosas propias. Lo que usted puede hacer es sencillo: si un joven de su familia sale de una consulta de genética, pregúntele qué entendió, con sus palabras, y anímelo a preguntar lo que quiera, en vez de dar por hecho que el médico ya lo explicó todo.
También conviene ver la herramienta por lo que es: un complemento, no un sustituto. El chat funcionó porque venía después de hablar con el equipo, no en lugar de hacerlo. Si en su clínica o en la escuela de su hijo aparece una herramienta así, pregunte si viene acompañada de una persona con quien hablar y si el joven puede elegir qué temas explorar. Y antes de confiar en cualquier herramienta de salud, pregunte de dónde salen sus datos y quién la revisó.
Otros trabajos que hemos leído sobre chatbots de salud ayudan a poner esto en perspectiva, aunque no dicen nada sobre este estudio. Un ensayo con chatbots de lenguaje natural para promover la vacuna del VPH encontró que, a los 45 días, ninguno de los chatbots aumentó la intención de vacunar, mientras que los materiales de salud pública mantuvieron un efecto modesto16. Otro estudio probó un chatbot que aplicaba entrevistas motivacionales en atención primaria y reportó mejoras sostenidas en actividad física y alimentación [C2.3, C2.5]. Son resultados distintos en poblaciones distintas: lo que sugieren es que un chatbot no es mágico por sí solo, y que el diseño y el contexto importan tanto como la tecnología.
El estudio deja una pregunta abierta y la responde en parte. ¿La herramienta reduce la angustia? No se puede afirmar: la reducción observada no fue concluyente y el estudio tenía pocos participantes para medir bienestar4. Lo que sí puede decirse es que no aumentó el malestar3. Esa distinción importa para las familias que temen que hablar de cáncer con un joven lo asuste más.
Volvamos a la joven de Boston con la carpeta bajo el brazo. Lo que este ensayo muestra es acotado y valioso a la vez: en cuatro centros académicos, con 106 jóvenes, una herramienta electrónica sencilla y bien hecha, usada después de la consulta, logró que más de ellos entendieran su riesgo antes de los 30 sin que aumentara su angustia [M8, M10]. Lo que no muestra es que funcione en una clínica comunitaria, en español, con jóvenes de otras razas o en quienes no tienen un síndrome hereditario. Tampoco muestra que reemplace la conversación con el médico.
Lo que esto hace posible para usted es concreto. Si en su familia hay un joven con un síndrome hereditario, ya sabe qué pedir: no solo la consulta, sino también material que pueda revisar después, a su ritmo, y la oportunidad de preguntar lo que no preguntó frente al médico. La pregunta que puede llevar a la próxima cita es esta: "¿Hay algún material o herramienta que mi hijo pueda revisar en casa después de esta consulta, y con quién puede hablar si le queda alguna duda?"
De dónde sale cada dato de contexto, y cuánto leímos de cada documento
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Following the clinic visit, 37% (19 of 52) of AYAs in the control group and 78% (42 of 54) in the chatbot-based intervention group had accurate knowledge, reflecting greater improvement in knowledge of cancer risk by age 30 years (OR, 3.50; 95% CI, 1.45-9.19; P = .005)."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Findings for lifetime cancer risk were not statistically different between groups, with baseline accuracy in 44% of AYAs (23) in the control group and 33% of AYAs (18) in the intervention group, and after-visit accuracy in 50% of AYAs (26) in the control and 52% (28) in the intervention group."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "We also assessed psychological experiences, including problems related to cancer predisposition and overall distress, to ensure that enhanced knowledge did not come at the expense of psychological well-being. Fortunately, we did not see increased psychosocial distress after the intervention."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Findings were similar for change in distress thermometer ratings, with a non–statistically significant reduction in distress in intervention relative to control (β = −0.50; 95% CI, −1.09 to 0.08; P = .09)."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The research team developed all material; no material was derived from generative artificial intelligence."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The home page included hyperlinks to 6 options: chat, learn, plan, share, connect, and grow."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The intervention uses an interactive chat to educate AYAs about cancer risk, with resource pages focused on issues of importance to AYAs, such as talking with peers and coping with uncertainty."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "The chat format allowed AYAs to obtain information in their preferred level of detail. For example, information about cancer risk was offered in general terms (“high risk of cancer in the future”) and in numbers (“more than 5 people out of 10 will develop cancer in the future”); a branching pattern allowed patients to choose their preferred level of detail."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "More than 10% of adolescents and young adults (AYAs) with cancer have a hereditary cancer predisposition, resulting in heightened lifetime risk of new cancers."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Screening allows for lifesaving early detection when cancer is more readily treatable."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "AYA-specific models for cancer risk communication and care are still emerging; new models are needed to help AYAs understand risk and embrace lifesaving screening interventions."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Only 25% of AYAs approached for this study enrolled, and we revised our enrollment goals during the study to match our funding timeline. Limited enrollment may have affected the results."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Our study was conducted at academic centers with established cancer predisposition clinics, and our population lacked racial and ethnic diversity."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "As guidelines evolve, the intervention will need to be updated to accurately reflect current standard of care recommendations and support evolving practices of clinical care teams."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - el artículo del que trata esta nota — el artículo completo — el pasaje: "These findings offer promise for dissemination beyond academic settings and into the community, where many AYAs receive care and where dedicated resources for rare diseases are often lacking; reducing burdens of study procedures may help to increase uptake."
- Sehgal NKR, Rai S, Tonneau M, Agarwal AK, Cappella J, Kornides ML, et al. (2026). Large Language Model Chatbot Conversations vs Public Health Materials and Parental HPV Vaccination Intentions. JAMA Network Open. 10.1001/jamanetworkopen.2026.16822 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "At 45 days, neither chatbot increased intent relative to controls, whereas public health materials maintained modest effects."
Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794
Quién pagó: El estudio fue financiado por la subvención 1U01CA243688 del Instituto Nacional del Cáncer (Dra. Mack); el financiador no tuvo ningún papel en el diseño, la ejecución, el análisis ni la decisión de publicar, y la interfaz del chatbot fue desarrollada por Clear Genetics, ahora Invitae/Labcorp.
No tome esto como consejo médico profesional.
experiment · JAMA network open · the paper, 1 Oct 2026 · free
A Chatbot After the Clinic Visit, and What It Did for Young People With an Inherited Cancer Risk
In a trial at four US cancer centers, a chat tool helped young people learn their risk of cancer before age 30. The trial was small, mostly White, and tested only after a specialist visit.
Short version · the longer version follows, about 8 min
Ask the weeklyAI chat
Ask me about AI in your life: I point you to the stories weeklyAI has published on it, this week and before.
Conversations are saved for as long as weeklyAI exists, to improve the publication. Answers come in the language you write in.
- The study at a glance
- Who
- Young people aged 12 to 24 with an inherited cancer risk syndrome
- How many
- 106
- Where
- Four US cancer centers
- When
- Published October 7, 2026
- Kind of study
- experiment
- Who did it
- Dana-Farber Cancer Institute, with Emory, University of Chicago and Huntsman Cancer Institute
- The limit that matters
- Small, mostly White, only at four academic clinics, only after a specialist visit
These are the shares of two small groups in one US trial; the chatbot group started out knowing more, and the study cannot say the tool would work in other clinics or languages.
Clinic visit alone versus clinic visit plus the chatbot
37 percent knew their risk before age 30, against 78 percent with the chatbot
Distress did not rise with the chatbot; the small drop was not statistically significant
Treat this as a promising result, not a guaranteed one.
In some families, a parent or a sibling has had cancer, and a test has shown that the same gene runs in the family. The young people in those families, some of them teenagers, know they may face cancer early. What they understand about that risk is often shaped in one place: a visit to a cancer genetics clinic.
A research team wrote the chatbot's information for exactly that moment, and a genetics company built the screen it runs on. The four US cancer centers took part by enrolling young people in the study. It is called AYA-RISE, and it comes after the clinic visit, on a phone, tablet or computer. It offers a chat, plus pages to learn more, plan screening by age, share with peers, connect to support groups and cope. It is written for each of nine inherited cancer syndromes, and it lets a young person choose how much detail to see: "high risk of cancer in the future," or "more than 5 people out of 10."
They tested it on 106 people aged 12 to 24 who have one of those syndromes. Half were assigned to see the clinic team alone. The other half saw the team and then used the chatbot. Everyone answered questions before and after.
Knowledge of the risk of cancer before age 30 rose among those who used the chatbot, from a little over half to 78 percent of them. Among those who had the visit alone, it stayed flat, at about 37 percent. The difference was large enough that chance is an unlikely explanation.
The study cannot say the tool works beyond those four academic clinics, or in Spanish, or for syndromes it was not built for. Only about one in four of the young people approached joined, and the planned group was cut from 300 to 106. Treat this as a promising result, not a guaranteed one.
Did the chatbot upset anyone? It did not appear to. Distress did not rise. The small drop in distress was not statistically significant, so the tool looks safe in this trial but is not proven to calm anyone.
What to look for: whether the genetics clinic near you offers a follow-up digital tool like this one, and whether it exists in your language.
What this means for you
The trial was run at four academic cancer centers, on 106 young people, nine in ten of them White, so a clinic near you may not offer anything like AYA-RISE yet. What you can ask is whether your genetics clinic gives young patients a follow-up tool of their own, in their own language. A specialist is the one to ask about any risk in your family.
Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794
Who paid: The study was funded by National Cancer Institute grant 1U01CA243688 (Dr Mack); the funder had no role in design, conduct, analysis, or the decision to publish, and the chatbot interface was developed by Clear Genetics, now Invitae/Labcorp.
Do not take this as professional medical advice.
The longer versionThe passages copied from the paper, the pictures, and every source with how much of it we read · 1675 words · about 8 minRead it →Close
A Chatbot After the Clinic Visit Helped Young People With Inherited Cancer Risk Understand It. Distress Did Not Rise.
A randomized trial at four US cancer centers tested a take-home chat tool for people aged 12 to 24. The gain was in their grasp of short-term risk, and the study is small.
Picture the end of a long appointment. A teenager has just heard a genetics team talk about a gene in the family, about screening that starts young, about a risk that may arrive before age 30. The parents have asked most of the questions. The young person nods. Then everyone goes home.
That gap between what is said in a room and what is understood afterward is what a randomized trial published October 7, 2026, in JAMA Network Open set out to measure. The authors are a large team led from the Dana-Farber Cancer Institute in Boston, with colleagues at Emory University and Children's Healthcare of Atlanta, the University of Chicago, and the University of Utah's Huntsman Cancer Institute. The work was paid for by a National Cancer Institute grant, and the funder had no role in the design, the analysis or the decision to publish.
The trial enrolled 106 young people aged 12 to 24, each with one of nine inherited conditions that raise cancer risk, including Li-Fraumeni syndrome, DICER1, Lynch syndrome and BRCA1- and BRCA2-associated breast and ovarian cancer. The average age was 18.3 years, and 64 percent were girls or women. Half were assigned to a standard clinic visit; the other half got the visit plus a tool called AYA-RISE, for AYA Risk Information and Screening Education, which they used afterward.
The scale of the problem is not small. The article describes a study in which more than 10 percent of adolescents and young adults with cancer carry an inherited predisposition, which raises their lifetime risk of new cancers1. The article describes a study in which screening is what turns that risk into something medicine can act on, because it allows cancers to be found early, when they are easier to treat2. Yet the authors write that ways of talking about risk built specifically for this age group are still emerging, and that new ones are needed to help young people understand risk and take up screening3.
Consider what the clinic visit covers. It is built around the whole family, and often around a parent's questions, and it may arrive on a day when screening results are being handed over. The authors note that the article describes a study in which young people with inherited cancer risk feel clinic visits do not meet all their needs, partly because so much of the conversation runs between clinicians and parents. A private screen, with no parent and no doctor looking over a shoulder, is a different room entirely.
Here is how the tool worked, step by step. It opened through a secure link on a phone, tablet or computer, and the home page offered six paths: chat, learn, plan, share, connect and grow4. The chat answered questions about cancer risk, recommended screening and prevention, with content written for each specific syndrome. The format let a young person choose how much detail they wanted: risk could be described in general terms, as a "high risk of cancer in the future," or in numbers, as "more than 5 people out of 10 will develop cancer in the future," and a branching path let them pick5. The other pages carried a summary of the chat, screening recommended by age, guidance on talking with peers about risk, links to support organizations, and help with coping and bereavement6.
One detail matters for anyone worried about machines making things up. The research team wrote all of the material, and none of it was generated by artificial intelligence7. The interface itself was built by Clear Genetics, now part of Invitae and Labcorp, and several of the authors reported ties to industry, including to those companies.
Before the visit, 52 percent of the young people in the intervention group (28 of 54) and 38 percent in the control group (20 of 52) correctly estimated their risk of cancer before age 30. After the visit, 78 percent of the intervention group (42 of 54) had it right, against 37 percent of the control group (19 of 52). The difference favored the group that used the chatbot, and the authors report it as statistically significant8.
The picture narrows when you look at the other questions. For lifetime risk, accuracy went from 33 percent to 52 percent in the intervention group and from 44 percent to 50 percent in the control group, a difference the authors report as not statistically significant9. Knowledge of recommended screening also leaned toward the tool but did not reach the threshold the authors set, and most participants already knew their screening recommendations at the start.
On distress, the authors looked carefully for harm. They write that they did not see increased psychosocial distress after the intervention10. The measured reduction in distress was not statistically significant11.
Two other studies we read only in their summaries show why this result should be held with care. In a large online trial among parents in the US, Canada and the UK, a chatbot built on a general language model raised parents' stated intent to vaccinate a child against HPV right after the conversation, but at 45 days neither chatbot version still did, while ordinary public health materials held a modest effect; no intervention increased actual vaccination1213. The summary concludes that well-designed public health messaging may match or exceed short chatbot conversations for this purpose14. A second trial, in routine primary care, tested an automated chatbot that used motivational interviewing to change lifestyle habits and reported improvements in physical activity and diet at 12 weeks, sustained at nine months, with no serious adverse events1516. A third, a test of whether a general chatbot could rewrite medical reports in plainer language, found readers understood simplified reports better and no medical information was lost, under the conditions of that study1718.
The limits are stated by the authors themselves. Only 25 percent of the young people approached enrolled, and the team revised its enrollment goal during the study to fit the funding timeline, which the authors say may have affected the results19. The trial ran at academic centers with established cancer genetics clinics, and the participants lacked racial and ethnic diversity20; 93 percent identified as White. Baseline surveys were given after randomization, which the authors say may have introduced expectancy bias. The main question also required comfort with numbers, and the study did not systematically record why eligible people declined.
What follows from the finding is maintenance as much as celebration. The authors write that as screening guidelines evolve, the tool will need updating to reflect current standards of care21. They also see promise in spreading this kind of tool beyond academic centers and into communities, where many young people receive care and where resources for rare diseases are often thin, and they suggest that reducing the burden of study procedures could help more people take part22.
Here is how we read it. A young person of 16 or 20 is still learning to weigh consequences, and a skill that is growing that fast takes hold when it is practiced, not when it is heard once. That is the plain reason a single conversation, however good, tends to fade for someone this age, and why a tool they work through themselves, at their own pace, right after the visit, would be expected to stick better. If we are wrong, the young people who used the chat should have done no better than those who only saw the clinic team, or the gain should have appeared only among the oldest. That is not what happened here.
There is a second reason the private screen may have worked, and it has nothing to do with software. In a clinic room, a young person sits between a doctor who knows and a parent who worries, and the polite thing is to nod. A chat has no authority in the room, so the questions a person would not ask aloud get asked. We would expect that to show up as better understanding without more worry, which is roughly the shape of the result. If it were wrong, distress would have risen, or the users would have gained nothing.
We also read other work that keeps us sober about chat tools in general. One thing to watch for is warmth: a program that echoes a person's words and keeps asking for more can feel like being listened to, and that feeling comes from the format, not from understanding. Another is staying power: a conversation that lifts intentions can lose its effect within weeks, while a plain leaflet holds. Treat a friendly health chat as a place to gather information and rehearse questions, and check any number it gives against what the clinic told you.
And a boundary to name, since it is about carrying over rather than about this result: this trial does not show that a tool like this would work in a community clinic, in Spanish, or for a family outside the nine syndromes it was built for. The place to look for that is the next study, not this one.
What a family can take from this today is narrower and more useful than a headline. The tool was tested as an addition to a clinic visit, never as a replacement for one, and it was tested only in people aged 12 to 24 with a known inherited syndrome. It did not make distress worse, which is the reassurance a parent needs before suggesting it.
The appointment ends. The young person walks out with a phone, a link and a question they were too polite to ask in the room. Ask the genetics team whether anything like this exists for your family — and then let the young person, not the adults, be the one who opens it.
Where each piece of context comes from, and how much of it we read
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "More than 10% of adolescents and young adults (AYAs) with cancer have a hereditary cancer predisposition, resulting in heightened lifetime risk of new cancers."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "Screening allows for lifesaving early detection when cancer is more readily treatable."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "AYA-specific models for cancer risk communication and care are still emerging; new models are needed to help AYAs understand risk and embrace lifesaving screening interventions."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "The home page included hyperlinks to 6 options: chat, learn, plan, share, connect, and grow."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "The chat format allowed AYAs to obtain information in their preferred level of detail. For example, information about cancer risk was offered in general terms (“high risk of cancer in the future”) and in numbers (“more than 5 people out of 10 will develop cancer in the future”); a branching pattern allowed patients to choose their preferred level of detail."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "The intervention uses an interactive chat to educate AYAs about cancer risk, with resource pages focused on issues of importance to AYAs, such as talking with peers and coping with uncertainty."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "The research team developed all material; no material was derived from generative artificial intelligence."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "Following the clinic visit, 37% (19 of 52) of AYAs in the control group and 78% (42 of 54) in the chatbot-based intervention group had accurate knowledge, reflecting greater improvement in knowledge of cancer risk by age 30 years (OR, 3.50; 95% CI, 1.45-9.19; P = .005)."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "Findings for lifetime cancer risk were not statistically different between groups, with baseline accuracy in 44% of AYAs (23) in the control group and 33% of AYAs (18) in the intervention group, and after-visit accuracy in 50% of AYAs (26) in the control and 52% (28) in the intervention group."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "We also assessed psychological experiences, including problems related to cancer predisposition and overall distress, to ensure that enhanced knowledge did not come at the expense of psychological well-being. Fortunately, we did not see increased psychosocial distress after the intervention."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "Findings were similar for change in distress thermometer ratings, with a non–statistically significant reduction in distress in intervention relative to control (β = −0.50; 95% CI, −1.09 to 0.08; P = .09)."
- Sehgal NKR, Rai S, Tonneau M, Agarwal AK, Cappella J, Kornides ML, et al. (2026). Large Language Model Chatbot Conversations vs Public Health Materials and Parental HPV Vaccination Intentions. JAMA Network Open. 10.1001/jamanetworkopen.2026.16822 — only the abstract - the full text could not be fetched — the passage: "At 45 days, neither chatbot increased intent relative to controls, whereas public health materials maintained modest effects."
- Sehgal NKR, Rai S, Tonneau M, Agarwal AK, Cappella J, Kornides ML, et al. (2026). Large Language Model Chatbot Conversations vs Public Health Materials and Parental HPV Vaccination Intentions. JAMA Network Open. 10.1001/jamanetworkopen.2026.16822 — only the abstract - the full text could not be fetched — the passage: "No intervention increased self-reported vaccination uptake."
- Sehgal NKR, Rai S, Tonneau M, Agarwal AK, Cappella J, Kornides ML, et al. (2026). Large Language Model Chatbot Conversations vs Public Health Materials and Parental HPV Vaccination Intentions. JAMA Network Open. 10.1001/jamanetworkopen.2026.16822 — only the abstract - the full text could not be fetched — the passage: "The findings of this randomized clinical trial suggest that well-designed public health messaging may match or exceed the impact of short chatbot conversations for HPV vaccine promotion."
- Wong CHY, Wong MNK, Hou WK, Ting FST, Ng YM, Che Hin Chan C. (2026). Motivational interviewing chatbot improves lifestyle in primary healthcare settings in a pragmatic randomised controlled trial. npj Digital Medicine. 10.1038/s41746-026-02728-w — only the abstract - the full text could not be fetched — the passage: "Modified intention-to-treat analysis (n = 460) showed significant improvements in the intervention group compared to usual care in physical activity (576 MET-min/week), fruit/vegetable intake (0.27 portions/day), and committed action (0.95 a.u.) at 12 weeks."
- Wong CHY, Wong MNK, Hou WK, Ting FST, Ng YM, Che Hin Chan C. (2026). Motivational interviewing chatbot improves lifestyle in primary healthcare settings in a pragmatic randomised controlled trial. npj Digital Medicine. 10.1038/s41746-026-02728-w — only the abstract - the full text could not be fetched — the passage: "Effects were sustained at 9-month follow-up. No serious adverse events were reported."
- Stephan D, Schumacher S, Al-Nawas B, Kämmerer PW, Thiem DGE. (2026). Improving patient communication and understanding of medical reports through text simplification with ChatGPT: a comparative study. Oral and Maxillofacial Surgery. 10.1007/s10006-026-01643-4 — only the abstract - the full text could not be fetched — the passage: "Across all document types, patients rated simplified texts significantly higher in clarity, structure, and usefulness (p < 0.001), with significantly improved comprehension accuracy (e.g., MRI 80.3% vs. 53.7%; p < 0.001)."
- Stephan D, Schumacher S, Al-Nawas B, Kämmerer PW, Thiem DGE. (2026). Improving patient communication and understanding of medical reports through text simplification with ChatGPT: a comparative study. Oral and Maxillofacial Surgery. 10.1007/s10006-026-01643-4 — only the abstract - the full text could not be fetched — the passage: "No loss of medical information was observed."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "Only 25% of AYAs approached for this study enrolled, and we revised our enrollment goals during the study to match our funding timeline. Limited enrollment may have affected the results."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "Our study was conducted at academic centers with established cancer predisposition clinics, and our population lacked racial and ethnic diversity."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "As guidelines evolve, the intervention will need to be updated to accurately reflect current standard of care recommendations and support evolving practices of clinical care teams."
- Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794 - the article this story is about — the whole article — the passage: "These findings offer promise for dissemination beyond academic settings and into the community, where many AYAs receive care and where dedicated resources for rare diseases are often lacking; reducing burdens of study procedures may help to increase uptake."
Kamihara, J., Cronin, A. M., Uno, H. et al. (2026). Chatbot Intervention and Cancer Risk Knowledge in Adolescents and Young Adults With Cancer Predisposition. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2026.35794
Who paid: The study was funded by National Cancer Institute grant 1U01CA243688 (Dr Mack); the funder had no role in design, conduct, analysis, or the decision to publish, and the chatbot interface was developed by Clear Genetics, now Invitae/Labcorp.
Do not take this as professional medical advice.
The findings of other studies mentioned here come from this document; those studies were not opened one by one.

