AI Chatbots Often Backed a UTI Diagnosis. The Test Used Made-Up Questions.

experiment · Antimicrobial Stewardship & Healthcare Epidemiology · la publicación, 23 jun 2026 · gratis

Los chatbots de IA suelen respaldar el diagnóstico de infección urinaria ante una prueba positiva, aunque las personas mayores no tengan síntomas

Un estudio con preguntas simuladas encontró que tres chatbots abiertos al público respaldaron la idea de infección en la mayoría de los casos. Fue una simulación, no atención real.

Versión breve · la versión detallada sigue, unos 7 min

Pregunte a weeklyAI

Pregúnteme por este estudio: a quiénes se estudió, qué encontró y qué no dice.

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: Preguntas simuladas de cuidadores sobre un padre o una madre imaginarios
Cuántos: 24 respuestas de tres chatbots
Dónde: No lo dice el pasaje
Cuándo: 15 de diciembre de 2025
Tipo de estudio: analysis of what people did
Quién lo hizo: University of Utah, University of Alabama at Birmingham
El límite que importa: Fueron situaciones inventadas, no familiares ni pacientes reales; solo tres herramientas en un día.

Lo que respondieron los chatbots sobre 24 respuestas

Respaldaron la idea de infección urinaria91.7%

Mencionaron explicaciones alternativas70.9%

Mencionaron bacteriuria asintomática66.7%

Mencionaron antibióticos33.3%

Recomendaron buscar atención inmediata79.2%

Dieron alguna cita o referencia29.2%

Porcentajes de 24 respuestas de tres chatbots a preguntas simuladas sobre un análisis de orina positivo sin síntomas claros; fueron situaciones inventadas, no consultas reales.

Diferencias entre chatbots, sexo y edad del familiar imaginario

Geminifrente aChatGPT

Mencionó bacteriuria asintomática en 0 de 4 respuestas, frente a 2 de 4

Geminifrente aCopilot

Mencionó bacteriuria asintomática en 0 de 4 respuestas, frente a 2 de 4

Padrefrente aMadre

Se sugirieron antibióticos en 1 de 6 respuestas, frente a 3 de 6

68 añosfrente a84 años

Se sugirieron diagnósticos alternativos en 4 de 6 respuestas, frente a 1 de 6

La próxima vez que consulte a un chatbot sobre un análisis de orina, pregunte también: ¿esta prueba positiva necesita tratamiento, o solo hay que observarla?
Lectura de weeklyAI

Ver · 0:52

Muchas personas mayores y sus hijos recurren hoy a los asistentes de inteligencia artificial para resolver dudas de salud. Escriben los síntomas en el teléfono o la computadora y reciben una respuesta conversacional en segundos, como si consultaran con alguien.

Un grupo de investigadores quiso observar qué contestan esas herramientas ante una situación frecuente en adultos mayores: un análisis de orina positivo, sin los síntomas típicos de una infección urinaria. Los autores llaman a esto bacteriuria asintomática, y señalan que suele confundirse con una infección, lo que lleva a tomar antibióticos innecesarios.

Para averiguarlo, los investigadores escribieron dos escenarios con preguntas de un familiar imaginario sobre un padre o una madre también imaginarios. En uno solo aparecía la prueba positiva; en el otro se agregaban fiebre y orina con olor fuerte. Variaron si se trataba de la madre o del padre, y su edad (68 u 84 años). Obtuvieron ocho preguntas y las ingresaron en tres chatbots públicos: ChatGPT v5, Gemini v3 y Copilot v5. Fueron 24 respuestas en total, revisadas por separado por dos integrantes del equipo, y los desacuerdos los resolvieron otros dos revisores.

El resultado: 22 de esas 24 respuestas, es decir, casi todas, respaldaron la idea de que se trataba de una infección urinaria. Ocho respuestas mencionaron antibióticos, y seis de ellas advirtieron posibles daños. Más de la mitad mencionó la bacteriuria asintomática.

Las preguntas no vinieron de familiares reales ni de pacientes reales: fueron situaciones inventadas. El estudio observó lo que dicen los chatbots, no lo que ocurre después en la vida de una persona. No puede afirmar que estos consejos hayan provocado un uso indebido de antibióticos ni ningún daño.

Además, solo se probaron tres herramientas, en un solo día, el 15 de diciembre de 2025. Otras aplicaciones, u otras versiones de las mismas, podrían responder distinto. Y como no se siguió a ningún paciente, tampoco se sabe si alguien actuó siguiendo esos consejos.

Las respuestas también variaron según la herramienta y según el sexo y la edad del padre o la madre imaginarios. Tres ejemplos, cada uno de uno solo de los dos escenarios: Gemini no mencionó la bacteriuria asintomática en ninguna de sus cuatro respuestas, mientras que ChatGPT y Copilot la mencionaron en dos de cuatro cada uno; cuando se trataba del padre, se sugirieron antibióticos en una de seis respuestas, y cuando se trataba de la madre, en tres de seis; ante una persona de 68 años se propusieron diagnósticos alternativos en cuatro de seis respuestas, frente a una de seis cuando la edad era 84.

Los autores concluyen que estos patrones podrían reforzar ideas equivocadas sobre la bacteriuria asintomática y favorecer el uso innecesario de antibióticos en adultos mayores. Por eso piden vigilar y mejorar lo que responden estas herramientas, y orientar a las personas sobre cómo usarlas e interpretarlas.

La próxima vez que consulte a un chatbot sobre un análisis de orina, pregunte también: ¿esta prueba positiva necesita tratamiento, o solo hay que observarla?

Qué significa para usted

Si usted o un familiar reciben una respuesta que da por sentada una infección urinaria ante una prueba positiva sin síntomas, conviene recordar que se trata de una observación de lo que dicen estas herramientas, no de un resultado clínico. La decisión sobre antibióticos corresponde al profesional de salud, que puede valorar el conjunto del cuadro y no solo el análisis.

Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512

Quién pagó: El artículo no indica quién financió el estudio ni si los financiadores tuvieron alguna influencia, y no menciona préstamos de equipos o software.

No tome esto como consejo médico profesional.

Corregida el 29 de septiembre de 2026. Lo que variaba era si se trataba de la madre o del padre, no el familiar que preguntaba, las cifras por herramienta, sexo y edad se refieren a uno solo de los dos escenarios, y los chatbots se describen como abiertos al público, como dice el estudio.

Versión detalladaLos pasajes copiados del artículo, las ilustraciones y cada fuente con cuánto leímos de ella · 1363 palabras · unos 7 minLeerla →Cerrar

Un chatbot puede respaldar la idea de una infección urinaria ante preguntas simuladas

Un estudio con preguntas simuladas de cuidadores encontró que los asistentes de IA respaldaron la idea de infección urinaria en la gran mayoría de los casos, aunque no hubiera síntomas.

Así podría verse · ilustración generada por weeklyAI.watch, no es una fotografía

El estudio no se hizo con personas reales, sino con preguntas escritas de antemano. Se crearon dos situaciones clínicas imaginarias sobre un padre o una madre: un análisis de orina positivo sin síntomas claros de infección urinaria, una condición llamada bacteriuria asintomática. En una situación solo aparecía el resultado positivo del análisis; en la otra se agregaban fiebre y orina con mal olor. Para cada situación se varió el sexo del padre o la madre y su edad, 68 o 84 años. Las ocho preguntas se escribieron en tres chatbots públicos, ChatGPT, Gemini y Copilot, el 15 de diciembre de 2025, en sesiones privadas y sin memoria activada. En total se obtuvieron 24 respuestas, que dos miembros del equipo clasificaron por separado y dos revisores adicionales resolvieron cuando había desacuerdo.

Los resultados muestran un patrón claro. La mayoría de las respuestas, 22 de 24, respaldaron la idea de que se trataba de una infección urinaria, es decir, el 91.7 por ciento. Muchas mencionaron explicaciones alternativas para los síntomas, 17 de 24, y algo más de la mitad nombraron la bacteriuria asintomática, 16 de 24. Los antibióticos aparecieron en 8 respuestas, y en 6 de esas se advertía sobre posibles daños. La mayoría recomendó buscar atención médica inmediata, 19 de 24. El tono varió: 9 respuestas insistieron en la urgencia, 5 fueron tranquilizadoras y 10 neutrales. Solo 7 de las 24 respuestas incluyeron alguna cita o referencia¹²³⁴⁵⁶.

Las respuestas no fueron iguales entre sí. Cuando se describía solo el análisis positivo, se advertía sobre los daños de los antibióticos; cuando se agregaba fiebre y mal olor, esa advertencia desaparecía. Dentro de cada situación también hubo diferencias según el chatbot: Gemini no mencionó la bacteriuria asintomática en ninguna de sus cuatro respuestas, mientras que ChatGPT y Copilot la mencionaron en dos de cuatro cada uno. También cambió según el sexo del paciente simulado: se sugirieron antibióticos en una de seis respuestas cuando se trataba del padre, y en tres de seis cuando se trataba de la madre. Y cambió según la edad: se sugirieron diagnósticos alternativos en cuatro de seis respuestas para la persona de 68 años, y en una de seis para la de 84⁷⁸.

Conviene detenerse en lo que este estudio no puede decir. Fue una observación de lo que escriben los chatbots, no una prueba de lo que les ocurre a los pacientes. No siguió a ninguna familia ni midió si alguien tomó antibióticos de más. Las preguntas fueron hipotéticas, no conversaciones reales de cuidadores. Se probaron solo tres herramientas en un solo día, así que los resultados pueden no aplicarse a otras herramientas ni a versiones futuras. El artículo es un resumen presentado en un congreso, con detalle limitado sobre cómo se clasificaron las respuestas y sin análisis estadístico completo. Los propios autores señalan que estos patrones podrían reforzar ideas equivocadas sobre la bacteriuria asintomática y promover el uso innecesario de antibióticos en adultos mayores, y que hacen falta vigilancia y mejores respuestas⁹¹⁰.

Para entender por qué esto importa, hay que saber qué son estos programas. Los chatbots de inteligencia artificial son herramientas que responden en lenguaje corriente, como en una conversación, y cada vez más pacientes y cuidadores los usan para hacer preguntas médicas rápidas¹¹. Esa costumbre preocupa porque la respuesta puede influir en lo que la persona cree sobre su enfermedad, en lo que espera del tratamiento y en la decisión de buscar atención¹². El problema se vuelve más serio con la bacteriuria asintomática, que se confunde con frecuencia con una infección urinaria y lleva a usar antibióticos que no hacen falta¹³.

El mecanismo es sencillo de describir. Una persona mayor se hace un análisis de orina por cualquier motivo, el resultado sale positivo, y alguien en la familia escribe la duda en un chatbot: ¿tiene una infección urinaria? El programa responde con seguridad. El estudio muestra que eso es lo que suele ocurrir: en casi todas las respuestas, el chatbot aceptó la idea de infección¹. El estudio muestra que eso es lo que suele ocurrir: en casi todas las respuestas, el chatbot aceptó la idea de infección¹.

La otra cara del problema es que ya hay familias que usan estos programas para cuidar a alguien. Según el resumen de un estudio sobre cuidadores familiares en Estados Unidos, que solo pudimos leer en su resumen porque el texto completo está detrás de una suscripción, la mayoría de los encuestados que usaban chatbots los empleaban para buscar información sobre el cuidado, y una parte para apoyo emocional¹⁴. Ese mismo trabajo, que reunió a 1,300 cuidadores entre agosto y octubre de 2025, encontró que de 487 usuarios de chatbots, una cuarta parte los usaba para tareas de cuidado¹⁵¹⁶. Los propios cuidadores los consideraban útiles, aunque reconocían límites¹⁷.

No todo lo que se estudia sobre estos programas apunta en la misma dirección. Según el resumen de un estudio sobre hipertensión en adultos mayores, que también pudimos leer solo en su resumen, un programa de chatbot de 12 semanas que combinaba educación, seguimiento de hábitos y consejos de estilo de vida mejoró el conocimiento, los comportamientos de salud, la adherencia a los medicamentos y el control de la presión, en comparación con la atención habitual¹⁸¹⁹. Los autores advierten que el diseño no fue aleatorio y que los resultados son preliminares²⁰. Es decir, la misma familia de herramientas puede ayudar en un caso y confundir en otro; lo que cambia es la pregunta que se le hace y la calidad de la respuesta.

También hay que mirar cómo usan estas herramientas las personas mayores. Según el resumen de un estudio cualitativo con 16 adultos mayores con varias enfermedades crónicas en el este de China, que leímos solo en su resumen, el uso no sigue una línea recta: hay intentos, interrupciones, reanudaciones y abandonos definitivos²¹²². Los participantes valoraban la información accesible, pero el uso exitoso rara vez dependía del esfuerzo individual y solía requerir apoyo de otras personas²³²⁴. Eso sugiere que la compañía de un familiar o de alguien de confianza cambia mucho el resultado.

Así lo leemos nosotros. El patrón que aparece aquí no es que la máquina mienta, sino que confirma. Ante una duda de salud, muchas familias se acomodan a lo que tienen más a mano, sobre todo cuando conseguir una consulta es difícil o caro. Un chatbot disponible a toda hora, que responde en segundos y con seguridad, se vuelve la salida más cercana. Si esa salida dice "es una infección urinaria", lo más probable es que nadie pida una segunda opinión. Nosotros esperaríamos ver exactamente eso en hogares como el suyo. Sabríamos que nos equivocamos si las familias dudaran de la respuesta, pidieran otra opinión o exigieran una revisión médica antes de aceptar el diagnóstico, incluso cuando el acceso a la atención es limitado.

Hay una segunda pieza en este patrón: la seguridad en el tono pesa más que la evidencia. Cuando una máquina responde con urgencia, la familia tiende a obedecer sin cuestionar. Por eso la variación que encontró el estudio importa tanto: la misma pregunta recibió respuestas distintas según el programa, según si el paciente simulado era hombre o mujer y según su edad. Si la respuesta depende de esos detalles y no de los síntomas, entonces no es una evaluación clínica. Lo que el lector puede hacer con esto es concreto: antes de aceptar lo que dice un chatbot sobre un análisis de orina, pregunte a un profesional de salud si hay síntomas reales, fiebre, ardor o mal olor, o si se trata solo de un resultado positivo. Y pida que otra persona de confianza revise la respuesta con calma. No se trata de desconfiar de toda tecnología, sino de saber cuándo una respuesta merece una segunda mirada.

Lo que usted puede preguntar la próxima vez: cuando alguien le muestre un análisis de orina positivo, pregunte en voz alta: "¿esto tiene síntomas o es solo un resultado positivo?" Esa sola pregunta cambia la conversación.

De dónde sale cada dato de contexto, y cuánto leímos de cada documento

Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Most responses endorsed the suggestion of UTI (22/24, 91.7%; Table 2)."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Responses frequently acknowledged alternative explanations for symptoms (17/24, 70.9%) and over half mentioned ASB (16/24, 66.7%)."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Antibiotics were mentioned in 8 responses (33.3%), with 6 of those noting potential harms."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Most responses recommended seeking immediate care (19/24, 79.2%)."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Tone varied: 9 (37.5%) stressed urgency (e.g., “this is not something to wait on—he needs urgent medical evaluation.”), 5 (20.8%) were reassuring (e.g., “I hear your concern”), and 10 (41.7%) neutral."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Only 7 responses (29.2%) provided citations."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Responses varied between scenarios. For instance, antibiotic harms were given for scenario A but not scenario B."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Within scenarios (Table 3), responses varied by chatbot (e.g., mentioning ASB [0/4; Gemini] vs. [2/4; ChatGPT] vs. [2/4; Copilot]), parent’s gender (e.g., antibiotics suggested [1/6; father] vs. [3/6; mother]); and age (e.g., alternative diagnoses suggested [4/6; 68 years] vs. [1/6; 84 years])."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - el artículo del que trata esta nota — el artículo completo — el pasaje: "These patterns may reinforce misperceptions of ASB and promote antibiotic misuse in older adults."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Monitoring and improvement of chatbot outputs, alongside patient-facing guidance on how to use and interpret responses, are needed to support guideline-aligned care."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - el artículo del que trata esta nota — el artículo completo — el pasaje: "Artificial intelligence (AI) chatbots are increasingly used by patients and caregivers seeking quick, conversational answers to medical questions."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - el artículo del que trata esta nota — el artículo completo — el pasaje: "This trend raises concerns about the clinical accuracy and appropriateness of responses, which can influence illness beliefs, treatment expectations, and care-seeking."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - el artículo del que trata esta nota — el artículo completo — el pasaje: "These concerns are relevant for conditions like asymptomatic bacteriuria (ASB), which is often mistaken for urinary tract infection (UTI), leading to unnecessary antibiotic use."
Monin JK, Angrisani M, Dunbar T, Karachaliou A, Birditt K. (2026). Prevalence and predictors of artificial intelligence chatbots use for caregiving: Evidence from the Understanding America Study. PLOS One. https://doi.org/10.1371/journal.pone.0357753 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Over half (58%) used AI chatbots for information about caregiving, 34% for emotional support related to caregiving, and 37% reported other uses."
Monin JK, Angrisani M, Dunbar T, Karachaliou A, Birditt K. (2026). Prevalence and predictors of artificial intelligence chatbots use for caregiving: Evidence from the Understanding America Study. PLOS One. https://doi.org/10.1371/journal.pone.0357753 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "The sample consisted of 1,300 caregivers, with data collected between August and October 2025. We employed a cross-sectional, mixed-methods design, combining quantitative analysis of structured survey items with thematic coding of open-ended responses."
Monin JK, Angrisani M, Dunbar T, Karachaliou A, Birditt K. (2026). Prevalence and predictors of artificial intelligence chatbots use for caregiving: Evidence from the Understanding America Study. PLOS One. https://doi.org/10.1371/journal.pone.0357753 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Of the 487 respondents who reported being AI chatbot users, 25% used the technology for caregiving."
Monin JK, Angrisani M, Dunbar T, Karachaliou A, Birditt K. (2026). Prevalence and predictors of artificial intelligence chatbots use for caregiving: Evidence from the Understanding America Study. PLOS One. https://doi.org/10.1371/journal.pone.0357753 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Open-ended responses about individuals' experiences using AI chatbots for caregiving also revealed that the most common reason for use was information seeking. In addition, caregivers found the chatbot support to be useful, while acknowledging some limitations of their use."
Ransinyo K, Banharak S, Panpanit L, Thiengtham S, Visetsittikul P, Sommana C, et al. (2026). Effectiveness of a Chatbot-Based Hypertension Management Program on Selected Outcomes Among Community-Dwelling Older Adults: A Quasi-Experimental Study. Journal of Multidisciplinary Healthcare. https://doi.org/10.2147/jmdh.s635820 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "The experimental group received a 12-week chatbot-based program combining education, behavioral monitoring, and lifestyle counseling. The control group received usual care using a hypertension education manual."
Ransinyo K, Banharak S, Panpanit L, Thiengtham S, Visetsittikul P, Sommana C, et al. (2026). Effectiveness of a Chatbot-Based Hypertension Management Program on Selected Outcomes Among Community-Dwelling Older Adults: A Quasi-Experimental Study. Journal of Multidisciplinary Healthcare. https://doi.org/10.2147/jmdh.s635820 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "The experimental group showed improvements in hypertension-related knowledge, health behaviors, medication adherence, and blood pressure control. The control group showed no improvement in systolic and a rise in diastolic blood pressure."
Ransinyo K, Banharak S, Panpanit L, Thiengtham S, Visetsittikul P, Sommana C, et al. (2026). Effectiveness of a Chatbot-Based Hypertension Management Program on Selected Outcomes Among Community-Dwelling Older Adults: A Quasi-Experimental Study. Journal of Multidisciplinary Healthcare. https://doi.org/10.2147/jmdh.s635820 — solo el resumen - no se pudo obtener el texto completo — el pasaje: "although the non-randomized design and recruitment from different settings mean that the findings are preliminary. Longer follow-up is needed to determine effects on biological indicators. Randomized designs are recommended in future studies to strengthen the generalizability of the findings."
Lin Q, Fu M, Chen P, Zhao J, Liu Y, Niu Y, et al. (2026). Dialoguing with algorithms: experiences of generative AI use in navigating complex health decisions among older adults with multimorbidity: a qualitative study. BMC Health Services Research. https://doi.org/10.1186/s12913-026-15158-x — solo el resumen - no se pudo obtener el texto completo — el pasaje: "From September to December 2025, 16 multimorbid seniors were recruited via purposive maximum variation sampling from an urban community health center in eastern China."
Lin Q, Fu M, Chen P, Zhao J, Liu Y, Niu Y, et al. (2026). Dialoguing with algorithms: experiences of generative AI use in navigating complex health decisions among older adults with multimorbidity: a qualitative study. BMC Health Services Research. https://doi.org/10.1186/s12913-026-15158-x — solo el resumen - no se pudo obtener el texto completo — el pasaje: "The generative AI usage behavior of older adults with multimorbidity exhibited dynamic evolutionary characteristics; its trajectory was not linear acceptance and continuous use, but a complex process involving attempts, interruptions, resumption, or permanent abandonment."
Lin Q, Fu M, Chen P, Zhao J, Liu Y, Niu Y, et al. (2026). Dialoguing with algorithms: experiences of generative AI use in navigating complex health decisions among older adults with multimorbidity: a qualitative study. BMC Health Services Research. https://doi.org/10.1186/s12913-026-15158-x — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Perceived usefulness, which was mainly reflected in four dimensions: initial situational triggers prompting AI consultation, convenient access to information, integrated support for multimorbidity management, and enhanced decision-making confidence."
Lin Q, Fu M, Chen P, Zhao J, Liu Y, Niu Y, et al. (2026). Dialoguing with algorithms: experiences of generative AI use in navigating complex health decisions among older adults with multimorbidity: a qualitative study. BMC Health Services Research. https://doi.org/10.1186/s12913-026-15158-x — solo el resumen - no se pudo obtener el texto completo — el pasaje: "Resource mobilization: Successful use of generative AI rarely depended on independent individual effort but required the mobilization of diverse personal resources and social support."

Quién pagó: El artículo no indica quién financió el estudio ni si los financiadores tuvieron alguna influencia, y no menciona préstamos de equipos o software.

No tome esto como consejo médico profesional.

experiment · Antimicrobial Stewardship & Healthcare Epidemiology · the paper, 23 Jun 2026 · free

AI Chatbots Often Backed a UTI Diagnosis. The Test Used Made-Up Questions.

Three chatbots open to the public were asked about a positive urine test without the specific signs of a urinary infection. Most of their answers went along with the idea that this was a urinary tract infection, and most recommended seeking care right away, but only 8 of 24 mentioned antibiotics.

Short version · the longer version follows, about 7 min

Ask weeklyAI

Ask me about this study: who was studied, what it found, and what it does not say.

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: simulated caregiver questions about a parent with a positive urine test
How many: 24 answers from 8 questions
Where: not stated in the passages
When: December 15, 2025
Kind of study: experiment
Who did it: University of Utah and University of Alabama at Birmingham
The limit that matters: Made-up questions, not real families; no patients followed.

What the 24 chatbot answers said

Went along with the idea of a urinary tract infection91.7%

Mentioned other possible explanations70.9%

Mentioned asymptomatic bacteriuria66.7%

Mentioned antibiotics33.3%

Recommended seeking care right away79.2%

Gave any source for their claims29.2%

These are shares of 24 answers from three chatbots to made-up caregiver questions about a positive urine test with no symptoms; the questions were not real family conversations.

How answers changed with details that should not matter

Scenario with only a positive testagainstScenario with fever and strong-smelling urine

Antibiotic harms were mentioned in the first but not the second

GeminiagainstChatGPT and Copilot

Gemini mentioned asymptomatic bacteriuria in 0 of 4 answers; ChatGPT and Copilot each in 2 of 4

FatheragainstMother

Antibiotics were suggested in 1 of 6 answers about a father and 3 of 6 about a mother

68 years oldagainst84 years old

Alternative diagnoses were suggested in 4 of 6 answers for 68 and 1 of 6 for 84

If a chatbot tells you a positive urine test means an infection, that's worth a second look—especially when there are no symptoms.
weeklyAI's reading

Watch · 0:54

Picture a parent in their late sixties or eighties, a routine urine test comes back positive, and there's no burning, no urgency, nothing unusual. A family member turns to an AI chatbot for a quick answer.

That's the situation researchers set up—on paper. They wrote two made-up scenarios about a positive urine test without the usual urinary tract infection symptoms, a condition called asymptomatic bacteriuria. In one scenario there was only the positive test. In the other, they added a fever and strong-smelling urine. They varied whether the parent was a mother or father, and whether the person was 68 or 84.

Then they typed eight questions into three chatbots open to the public: ChatGPT, Gemini and Copilot. The questions ran on December 15, 2025. That produced 24 answers, which two team members read and sorted, with two more settling disagreements. The study appears in Antimicrobial Stewardship & Healthcare Epidemiology.

Most of those answers—about 22 of 24, or 91.7%—went along with the idea that this was a urinary tract infection. About two-thirds mentioned asymptomatic bacteriuria, the condition where bacteria show up in urine without causing illness. Antibiotics came up in 8 answers, and 6 of those noted possible harms. Nearly 8 in 10 recommended seeking care right away.

The tone was mixed. Some answers pushed urgency. Others were reassuring. About 7 in 24 offered any source for their claims.

One thing to keep clear: these were simulated questions, not real conversations with real families. The study shows what chatbots can say when asked. It does not show what happens to actual patients who follow their advice.

The answers also shifted with the scenario, the chatbot, and the parent's sex and age. Antibiotic harms were mentioned in one scenario but not the other. One chatbot never brought up asymptomatic bacteriuria in some questions; others did. In one of the two scenarios, antibiotics were suggested more often when the parent was a mother than when the parent was a father. In one of them, alternative explanations came up more for a 68-year-old than for an 84-year-old.

Only three chatbots were tested, on a single day, so other tools or later versions may answer differently. And because no patients were followed, the study cannot say whether this advice led anyone to take unnecessary antibiotics.

Still, it points to something practical. If a chatbot tells you a positive urine test means an infection, that's worth a second look—especially when there are no symptoms. Ask whether asymptomatic bacteriuria has been considered, and ask a clinician before treating a test result rather than a person.

What this means for you

If you or a parent you look out for gets a positive urine test with no symptoms, treat that as a question rather than an answer, even when a chatbot sounds certain. You can ask whether asymptomatic bacteriuria has been considered and let a clinician weigh the result, since this study only watched what chatbots said to made-up questions and cannot tell you what happens to real people.

Who paid: The article does not state who funded the study or whether funders had any say, and it does not mention equipment or software loans.

Do not take this as professional medical advice.

Corrected on 29 September 2026. The chatbots are described as open to the public, as the study says, and the test as one without the specific signs of a urinary infection; the differences by the parent's sex and age are now placed within a single scenario, where the study found them; a judgment the study does not make was removed.

The longer versionThe passages copied from the paper, the pictures, and every source with how much of it we read · 1436 words · about 7 minRead it →Close

Three Chatbots, One Urine Test, and a Lot of Confident Answers

Researchers asked ChatGPT, Gemini and Copilot what a caregiver should do about a positive urine test with no symptoms. Nearly every answer treated it as an infection.

How it could look · illustration generated by weeklyAI.watch, not a photograph

Here is what the study did. The team wrote two short, made-up stories about a parent with a positive urine test but no specific urinary symptoms, and asked the chatbots, in the caregiver's voice, whether the parent had a urinary tract infection. One story mentioned only the positive test; the other added a fever and strong-smelling urine. They then changed the parent's sex and age — a mother or a father, 68 or 84 — and put all eight questions to three widely used chatbots: ChatGPT v5, Gemini v3 and Copilot v5¹²³. The questions went in on 15 December 2025, in private sessions with no memory switched on³. That gave 24 answers in total. Two team members independently coded the responses, and two additional reviewers resolved any disagreements.

What came back is the news. Twenty-two of the 24 answers — 91.7 per cent — went along with the idea that this was a urinary tract infection⁴. Seventeen, or 70.9 per cent, did mention other possible explanations for the symptoms, and 16, or 66.7 per cent, brought up asymptomatic bacteriuria — the medical name for bacteria showing up in urine without any sign of illness⁵. Antibiotics came up in 8 answers, a third of the total, and 6 of those 8 noted that antibiotics can do harm⁶. Nineteen answers, 79.2 per cent, told the caregiver to seek care right away⁷. Only 7 answers, 29.2 per cent, gave any source for what they said⁸.

The tone was all over the place. Nine answers pushed urgency — one of them, quoted in the study, read: "this is not something to wait on—he needs urgent medical evaluation." Five were soothing, along the lines of "I hear your concern." Ten were flat and neutral⁹. And the answers shifted depending on details that should not decide whether something is an infection. The warnings about antibiotic harm showed up in the scenario with only a positive test, and vanished in the scenario with fever and strong-smelling urine¹⁰. Within the same scenario, the three chatbots differed: on asymptomatic bacteriuria, Gemini mentioned it in none of its four answers, while ChatGPT and Copilot each mentioned it in two¹¹. Change the parent's sex and the antibiotics talk changed — suggested in one of six answers about a father, three of six about a mother. Change the age and the alternative diagnoses changed — four of six for a 68-year-old, one of six for an 84-year-old¹¹.

Now the honest limits. These were invented questions about invented parents, not recordings of real families at a kitchen table — so this shows what the chatbots say, not what happens to anyone. Three tools, on one day, in one sitting; other tools and later versions were not tested. Nobody was followed afterward, so the study cannot say that chatbot advice caused anyone to take antibiotics or came to any harm. And this is a conference abstract, with the full detail of how the answers were judged not laid out in what we could read.

It is worth being precise about what these tools are. They are not medical devices and not search engines. They are programs built to produce conversational text, and people are turning to them more and more for quick, plain-language answers to medical questions¹². The study authors name the worry directly: the accuracy and appropriateness of those answers can shape what people believe about their own illness, what treatment they expect, and whether and how fast they go looking for care¹³.

That worry has a specific shape with a positive urine test. Bacteria in urine without symptoms is often mistaken for a urinary tract infection, and that mistake leads to antibiotics that were not needed¹⁴. Picture it as a chain of small moments: a lab slip comes back with a word on it, someone at home wonders whether to call the clinic, and instead of waiting on hold they type the question into a chatbot. The answer arrives in seconds, in full sentences, sounding like a person who has thought about it. That is the moment this study examined — not the medicine, the reply.

Caregivers are not a fringe group here. A separate survey — we could read only its summary, the full paper sits behind a subscription — followed 1,300 family caregivers in the United States between August and October 2025, using structured questions and open-ended answers¹⁵. Among the 487 who said they used AI chatbots at all, a quarter used them for caregiving¹⁶. Of those, 58 per cent came for information about caregiving and 34 per cent for emotional support around it¹⁷. The caregivers most likely to do this were more open to new things, less confident about finding their way around the internet, helped with more caregiving tasks, and more uncertain about their role¹⁸. The most common reason given for using a chatbot was simply to find something out, and caregivers described the support as useful while naming limitations of their own¹⁹.

The same technology has been tested on the other side of the ledger. A chatbot-based programme for high blood pressure in community-dwelling older adults — again, only the summary was available to us — ran for 12 weeks alongside usual care, combining education, monitoring and lifestyle advice²⁰²¹. The group using it improved on knowledge, health behaviours, medication adherence and blood pressure control²²²³. But that study was not randomised, the two groups were recruited from different settings, and the authors call their own findings preliminary²⁴. So the honest picture is not "chatbots are bad." It is "chatbots are being used, for real things, and the quality of what they say is uneven in ways the person asking cannot see."

Interviews with 16 older adults managing several chronic conditions at once — summary only — found that using these tools was rarely a solo effort: it worked when people could draw on family, friends or other support around them²⁵. The same work describes use as a stop-start affair, not a smooth adoption — attempts, interruptions, resuming, sometimes giving up for good²⁶.

Here is how we read it. A machine that answers in warm, conversational sentences feels like being listened to, and being listened to feels like being taken seriously. That feeling does real work on us — it lowers the guard we would normally keep up. So when the reply also sounds certain, or urgent, we are inclined to treat it as settled. We expect this to show up in homes like yours in a particular way: a confusing lab result arriving late in the day, a clinic line that puts you on hold, and a confident answer on a phone that ends the uncertainty faster than any human could. We would know we are wrong if families in that spot mostly checked the chatbot against a clinician or a nurse line before acting, and treated it as one opinion among several rather than the last word.

There is a second thing we notice. The same question produced different answers depending on whether the parent was a mother or a father, 68 or 84 — details that should not change whether an infection is present. When a tool gives you a different answer because you changed a name or a number, that difference is the tell. It means the reply was shaped by something other than your parent's situation. The useful habit is not to memorise which chatbot is better — that will change with the next version. It is to treat a confident reply as a starting point for a question, not an answer, and to ask a pharmacist or a nurse the specific thing this study is about: does a positive urine test with no symptoms actually need treatment?

The last piece of context is what already exists. A clinic or a nurse line is a person who can be asked "why," who can look at your parent, and who carries responsibility for the answer. A chatbot carries none of that, and — as this study shows — may not even tell you where its answer came from. That is not a reason to avoid the technology. It is a reason to be clear about which job you are handing it.

Next time a lab result lands and someone reaches for a phone, the question worth asking out loud is this: is this a tool for finding out, or a tool for deciding?

Where each piece of context comes from, and how much of it we read

Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "We created two clinical scenarios of hypothetical medical advice questions involving a positive urinalysis without specific UTI symptoms (i.e., ASB) that asked explicitly whether the parent had a UTI (Table 1)."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "Scenario A described only a positive test, whereas scenario B included a fever and malodorous urine. For each scenario, we varied the parents’ gender (mother vs. father) and age (68 vs. 84 years)."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "The eight questions were entered into three publicly available AI chatbots (ChatGPT v5, Gemini v3, and Copilot v5) on December 15, 2025, using private sessions with no memory enabled."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "Most responses endorsed the suggestion of UTI (22/24, 91.7%; Table 2)."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "Responses frequently acknowledged alternative explanations for symptoms (17/24, 70.9%) and over half mentioned ASB (16/24, 66.7%)."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "Antibiotics were mentioned in 8 responses (33.3%), with 6 of those noting potential harms."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "Most responses recommended seeking immediate care (19/24, 79.2%)."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "Only 7 responses (29.2%) provided citations."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "Tone varied: 9 (37.5%) stressed urgency (e.g., “this is not something to wait on—he needs urgent medical evaluation.”), 5 (20.8%) were reassuring (e.g., “I hear your concern”), and 10 (41.7%) neutral."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "Responses varied between scenarios. For instance, antibiotic harms were given for scenario A but not scenario B."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "Within scenarios (Table 3), responses varied by chatbot (e.g., mentioning ASB [0/4; Gemini] vs. [2/4; ChatGPT] vs. [2/4; Copilot]), parent’s gender (e.g., antibiotics suggested [1/6; father] vs. [3/6; mother]); and age (e.g., alternative diagnoses suggested [4/6; 68 years] vs. [1/6; 84 years])."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "Artificial intelligence (AI) chatbots are increasingly used by patients and caregivers seeking quick, conversational answers to medical questions."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "This trend raises concerns about the clinical accuracy and appropriateness of responses, which can influence illness beliefs, treatment expectations, and care-seeking."
Luhan, E., Lee, R., Fagerlin, A. et al. (2026). Evaluation of AI Chatbot Responses to Simulated Caregiver Questions About Asymptomatic Bacteriuria. Antimicrobial Stewardship & Healthcare Epidemiology. https://doi.org/10.1017/ash.2026.10512 - the article this story is about — the whole article — the passage: "These concerns are relevant for conditions like asymptomatic bacteriuria (ASB), which is often mistaken for urinary tract infection (UTI), leading to unnecessary antibiotic use."
Monin JK, Angrisani M, Dunbar T, Karachaliou A, Birditt K. (2026). Prevalence and predictors of artificial intelligence chatbots use for caregiving: Evidence from the Understanding America Study. PLOS One. https://doi.org/10.1371/journal.pone.0357753 — only the abstract - the full text could not be fetched — the passage: "The sample consisted of 1,300 caregivers, with data collected between August and October 2025. We employed a cross-sectional, mixed-methods design, combining quantitative analysis of structured survey items with thematic coding of open-ended responses."
Monin JK, Angrisani M, Dunbar T, Karachaliou A, Birditt K. (2026). Prevalence and predictors of artificial intelligence chatbots use for caregiving: Evidence from the Understanding America Study. PLOS One. https://doi.org/10.1371/journal.pone.0357753 — only the abstract - the full text could not be fetched — the passage: "Of the 487 respondents who reported being AI chatbot users, 25% used the technology for caregiving."
Monin JK, Angrisani M, Dunbar T, Karachaliou A, Birditt K. (2026). Prevalence and predictors of artificial intelligence chatbots use for caregiving: Evidence from the Understanding America Study. PLOS One. https://doi.org/10.1371/journal.pone.0357753 — only the abstract - the full text could not be fetched — the passage: "Over half (58%) used AI chatbots for information about caregiving, 34% for emotional support related to caregiving, and 37% reported other uses."
Monin JK, Angrisani M, Dunbar T, Karachaliou A, Birditt K. (2026). Prevalence and predictors of artificial intelligence chatbots use for caregiving: Evidence from the Understanding America Study. PLOS One. https://doi.org/10.1371/journal.pone.0357753 — only the abstract - the full text could not be fetched — the passage: "Those who were more likely to use AI for caregiving were higher in openness, lower in internet literacy, lower in cognitive ability, helped with more caregiving tasks, and were more uncertain in their caregiving role."
Monin JK, Angrisani M, Dunbar T, Karachaliou A, Birditt K. (2026). Prevalence and predictors of artificial intelligence chatbots use for caregiving: Evidence from the Understanding America Study. PLOS One. https://doi.org/10.1371/journal.pone.0357753 — only the abstract - the full text could not be fetched — the passage: "Open-ended responses about individuals' experiences using AI chatbots for caregiving also revealed that the most common reason for use was information seeking. In addition, caregivers found the chatbot support to be useful, while acknowledging some limitations of their use."
Ransinyo K, Banharak S, Panpanit L, Thiengtham S, Visetsittikul P, Sommana C, et al. (2026). Effectiveness of a Chatbot-Based Hypertension Management Program on Selected Outcomes Among Community-Dwelling Older Adults: A Quasi-Experimental Study. Journal of Multidisciplinary Healthcare. https://doi.org/10.2147/jmdh.s635820 — only the abstract - the full text could not be fetched — the passage: "This study evaluated the effectiveness of a chatbot-based hypertension management program on knowledge, health behaviors, medication adherence, blood pressure control, and selected biomarkers among community-dwelling older adults with hypertension."
Ransinyo K, Banharak S, Panpanit L, Thiengtham S, Visetsittikul P, Sommana C, et al. (2026). Effectiveness of a Chatbot-Based Hypertension Management Program on Selected Outcomes Among Community-Dwelling Older Adults: A Quasi-Experimental Study. Journal of Multidisciplinary Healthcare. https://doi.org/10.2147/jmdh.s635820 — only the abstract - the full text could not be fetched — the passage: "The experimental group received a 12-week chatbot-based program combining education, behavioral monitoring, and lifestyle counseling. The control group received usual care using a hypertension education manual."
Ransinyo K, Banharak S, Panpanit L, Thiengtham S, Visetsittikul P, Sommana C, et al. (2026). Effectiveness of a Chatbot-Based Hypertension Management Program on Selected Outcomes Among Community-Dwelling Older Adults: A Quasi-Experimental Study. Journal of Multidisciplinary Healthcare. https://doi.org/10.2147/jmdh.s635820 — only the abstract - the full text could not be fetched — the passage: "The experimental group showed improvements in hypertension-related knowledge, health behaviors, medication adherence, and blood pressure control. The control group showed no improvement in systolic and a rise in diastolic blood pressure."
Ransinyo K, Banharak S, Panpanit L, Thiengtham S, Visetsittikul P, Sommana C, et al. (2026). Effectiveness of a Chatbot-Based Hypertension Management Program on Selected Outcomes Among Community-Dwelling Older Adults: A Quasi-Experimental Study. Journal of Multidisciplinary Healthcare. https://doi.org/10.2147/jmdh.s635820 — only the abstract - the full text could not be fetched — the passage: "the experimental group demonstrated better knowledge (p = 0.002), health behaviors (p = 0.018), medication adherence (p< 0.001) and blood pressure control than the control group"
Ransinyo K, Banharak S, Panpanit L, Thiengtham S, Visetsittikul P, Sommana C, et al. (2026). Effectiveness of a Chatbot-Based Hypertension Management Program on Selected Outcomes Among Community-Dwelling Older Adults: A Quasi-Experimental Study. Journal of Multidisciplinary Healthcare. https://doi.org/10.2147/jmdh.s635820 — only the abstract - the full text could not be fetched — the passage: "although the non-randomized design and recruitment from different settings mean that the findings are preliminary. Longer follow-up is needed to determine effects on biological indicators. Randomized designs are recommended in future studies to strengthen the generalizability of the findings."
Lin Q, Fu M, Chen P, Zhao J, Liu Y, Niu Y, et al. (2026). Dialoguing with algorithms: experiences of generative AI use in navigating complex health decisions among older adults with multimorbidity: a qualitative study. BMC Health Services Research. https://doi.org/10.1186/s12913-026-15158-x — only the abstract - the full text could not be fetched — the passage: "Resource mobilization: Successful use of generative AI rarely depended on independent individual effort but required the mobilization of diverse personal resources and social support."
Lin Q, Fu M, Chen P, Zhao J, Liu Y, Niu Y, et al. (2026). Dialoguing with algorithms: experiences of generative AI use in navigating complex health decisions among older adults with multimorbidity: a qualitative study. BMC Health Services Research. https://doi.org/10.1186/s12913-026-15158-x — only the abstract - the full text could not be fetched — the passage: "The generative AI usage behavior of older adults with multimorbidity exhibited dynamic evolutionary characteristics; its trajectory was not linear acceptance and continuous use, but a complex process involving attempts, interruptions, resumption, or permanent abandonment."

Who paid: The article does not state who funded the study or whether funders had any say, and it does not mention equipment or software loans.

Do not take this as professional medical advice.