Tesis de doctorado: Validación del diseño centrado en el usuario en un sistema notificador de interacciones farmacológicas
Archivos
Fecha
2016
Autores
Título de la revista
ISSN de la revista
Título del volumen
Editor
Resumen
"Es posible alertar a los médicos de la existencia de interacciones medicamentosas por medio de sistemas de soporte para la toma de decisiones. En el Hospital
Italiano de Buenos Aires se ha desarrollado un sistema de Historia Clínica Electrónica con un sistema notificador de interacciones farmacológicas promediando la década del 2000. El desarrollo de dicho sistema fue realizado mediante técnicas
de ingeniería de requerimientos e ingeniería de software tradicionales. Al observar que la tasa de anulación de las alertas era muy alta, se decidió rediseñar el sistema y sus alertas, con técnicas de Diseño Centrado en el Usuario.
El objetivo de la presente tesis es medir la eficiencia, efectividad y satisfacción (métricas de usabilidad) de las alertas desarrolladas con técnicas tradicionales
versus las desarrolladas con técnicas de diseño centrado en el usuario, mediante ciclos iterativos de prototipado (siguiendo la norma ISO 9241-210), en un ambiente controlado de laboratorio y validar dichos resultados en una prueba de terreno asistencial.
La metodología utilizada para la medición de la usabilidad de las alertas tradicionales versus las nuevas rediseñadas, por medio de un ambiente controlado por
casos clínicos (con interacciones reales según escenarios de atención) y exposición cruzada a ambos tipos de alertas; y luego validar dichos resultados en un estudio experimental controlado y randomizado a dos ramas en un escenario asistencial real.
El resultado del estudio evidenció que las alertas diseñadas según técnicas de diseño centrado en el usuario fueron más usables que las tradicionales, siendo más efectivas (menor tiempo empleado), eficientes (completadas con menos errores) y
satisfactorias, en las pruebas de laboratorio; y fueron más efectivas y satisfactorias en terreno asistencial, no así en el plano de la eficiencia.
Se concluye que la utilización de técnicas de diseño centrado en el usuario genera alertas más usables que el diseño tradicional, en pruebas de laboratorio y
parcialmente en terreno. Son necesarias futuras investigaciones para explicar esta diferencia."
"Physicians can be alerted when they prescribe a drug to a patient trough a decision support system, if this new prescription has be reported as potentially harm due to an interaction with another drug, this is known as drug-drug interaction. The Hospital Italiano de Buenos Aires has developed an Electronic Health Record system with a drug-drug interaction alert notifier module, since middle of 2000. The development of such system was generated using traditional methodology of software engineering requirements. Because was observed that the rejection ratio of the alerts was very high, then the alert system was redesigned with User Centered Design methodology. The objective of this thesis was to measure the efficiency, the effectiveness and the satisfaction (usability metrics) of the alerts developed with traditional methodology versus the alerts redesigned with User Centered Design methodology, by means of iterative cycles of prototyping (following standards ISO 9241-210), in a controlled laboratory environment, and later to validate the obtained results in a clinical environment. The methodology used to measure the usability of the traditional design alerts versus the redesigned alerts, was with real clinical interactions (according the medical service), and crossing the physicians to both type of alerts; finally the results of the previous study was validated with a controlled experimental study, randomized to two groups, in a real clinical environment. The results of the present studies showed that the redesigned alerts with User Centered Design were more usable than the alerts designed with traditional techniques, being more effective (low time consuming), efficient (completed with less errors), and satisfactory, all in the laboratory setup; about the clinical setup the alerts were more effective and satisfactory, but efficient. The conclusions are that using User Centered Design methods the alerts are more usable in the laboratory, and partially in the clinical setup. Future investigations are necessary to explain this difference."
"Physicians can be alerted when they prescribe a drug to a patient trough a decision support system, if this new prescription has be reported as potentially harm due to an interaction with another drug, this is known as drug-drug interaction. The Hospital Italiano de Buenos Aires has developed an Electronic Health Record system with a drug-drug interaction alert notifier module, since middle of 2000. The development of such system was generated using traditional methodology of software engineering requirements. Because was observed that the rejection ratio of the alerts was very high, then the alert system was redesigned with User Centered Design methodology. The objective of this thesis was to measure the efficiency, the effectiveness and the satisfaction (usability metrics) of the alerts developed with traditional methodology versus the alerts redesigned with User Centered Design methodology, by means of iterative cycles of prototyping (following standards ISO 9241-210), in a controlled laboratory environment, and later to validate the obtained results in a clinical environment. The methodology used to measure the usability of the traditional design alerts versus the redesigned alerts, was with real clinical interactions (according the medical service), and crossing the physicians to both type of alerts; finally the results of the previous study was validated with a controlled experimental study, randomized to two groups, in a real clinical environment. The results of the present studies showed that the redesigned alerts with User Centered Design were more usable than the alerts designed with traditional techniques, being more effective (low time consuming), efficient (completed with less errors), and satisfactory, all in the laboratory setup; about the clinical setup the alerts were more effective and satisfactory, but efficient. The conclusions are that using User Centered Design methods the alerts are more usable in the laboratory, and partially in the clinical setup. Future investigations are necessary to explain this difference."
Descripción
Palabras clave
SISTEMAS DE INFORMACION, TOMA DE DECISIONES, INGENIERIA DE SOFTWARE, HOSPITALES