La dificultad de crear soluciones usables en sanidad
La productividad que un usuario puede conseguir con una herramienta depende, entre muchas otras cosas, de la usabilidad de la interfaz con la que trabaje. En el contexto de las aplicaciones sanitarias, las soluciones de diseño de interfaz que se aportan pueden llegar a suponer en determinadas ocasiones una serie de peligros que deriven en riesgos clínicos. ¿Es incompatible una solución productiva con una solución libre de riesgos clínicos? Con el ánimo de ofrecer una visión amplia de las problemáticas, en este artículo ofrecemos posibles soluciones atendiendo a los patrones de diseño más habituales usados para aumentar la productividad en las interfaces de usuario.
Para el análisis se proponen dos modos de uso concretos: la entrada de datos (interacción contra el sistema) y la salida de datos(representación de información e interacciones de manejo de datos) en un sistema sanitario.
Entrada de datos (garbage-in)
Ayuda para completar la información
- Problema: La ayuda puede estar descontextualizada
- Solución: La ayuda que se ofrezca al usuario no debe estar sólo contextualizada a su actividad habitual sino al contenido y la información que se requiere.
- Problema: La acción para autocompletar puede terminar en la selección de un valor no deseado
- Solución: Las acciones de autocompletar deben permitir al usuario final sobrescribir el valor y advertirlo de que está introduciendo un valor fuera del catálogo
Informar de errores de validación
- Problema: Los errores de los que se informan son parciales
- Solución: Informar al usuario de la posible existencia de más errores y de la necesidad de su revisión
Opciones por defecto
- Problema: Los valores por defecto pueden estar basados en suposiciones erróneas
- Solución: Informar de que existen valores por defecto e incluso ofrecer la posibilidad de eliminarlos
Atajos de teclado
- Problema: Pérdida de feedback visual sobre lo que se está haciendo
- Solución: Informar al usuario mediante un mensaje de la acción que se va ejecutar
Comandos (lanzar acciones, deshacer, acciones masivas)
- Problema: El comando puede provocar un fallo en la integridad de los datos
- Solución: Ejecutar comando en modo simulado hasta la completa confirmación por parte del usuario e informar de la vista local del resultado
Salida de datos (garbage-out)
Resumen de datos e índices
- Problema: Los datos no están actualizados
- Solución: informar al usuario de la última actualización de los datos y ofrecer la posibilidad de actualizarlos in-situ
- Problema: La búsqueda oculta resultados
- Solución: Informar de los resultados visibles y de la existencia de resultados ocultos ya sea por cuestiones de paginación como de coincidencia con la búsqueda
- Problema: Las notificaciones no están priorizadas
- Solución: Ofrecer un sistema de priorización y permitir la personalización de la prioridad de las alertas una a una o por tipologías
- Problema: Se ha considerado una alerta algo que no lo es
- Solución: Asociar las alertas a dominios clínicos estáticos e identificables de forma consistente
- Problema: Se ha asociado un icono poco intuitivo a una alerta
- Solución: Etiquetar las alertas con términos cortos y/o agruparlas bajo dominios predefinidos
El escenario sanitario
Debemos tener en cuenta que en el contexto sanitario, las tareas que un usuario debe realizar que requieran de un sistema informático pueden llegar a suponer hasta el 30% de su tiempo y por tanto existe una necesidad evidente de mejorar la eficiencia y productividad en su uso diario. Al mismo tiempo es importante considerar que durante la jornada laboral son múltiples y muy frecuentes los cambios de tareas, actividades y contextos, por ejemplo, un médico en la sala de urgencias en un intervalo de 10’ ha podido realizar el Triaje a dos pacientes, ser consultado por el equipo de enfermería 3 veces sobre asuntos de diferentes naturaleza, se ha podido levantar de su asiento 6 veces para diferentes comprobaciones de material, información, etc. Por lo tanto, si bien la propia tarea favorece y exige una alta capacidad de concentración, la capacidad de atención se reduce por el alto número de interrupciones. En este contexto, el usuario controla lo que hace pero está sometido a una carga alta de estrés y distracciones en casi cualquiera de sus actividades. Un error médico podría suponer un riesgo para la salud de sus pacientes.
Desde la visión más simplificada de un sistema informático como mera representación del modelo de información, la interfaz básica más directa consiste en la disposición de formularios para rellenar datos y vistas para mostrarlos. Incluso con ésta fórmula, la aplicación no estaría exentas de riesgos clínicos. Cualquier solución aplicable para la interfaz más sencilla no es incompatible con interfaces más ricas y usables. Por esto consideramos que no sólo no es incompatible ofrecer soluciones que aumenten la productividad y la usabilidad, sino que además es necesario.
In my previous post, I published a simple mock-up to introduce the idea of an Electronic Medical Record looking like Facebook social network.
We initially saw how taking advantage of this website could help to contribute for a better tool for doctors, nurses, and patients.
In this second part of the exercise, we can see as well the hi-fi prototype where the visual design will allow us to make new musings.
Notifications will lead the activity of users under specific contexts.This space would be used in a similar way, so we would expect to see there:
- New patient admissions
- New results
- Prescription modifications
- Activity of another physicians on my patient’s profile
- Events notifications
- Automatic clinical alerts
The search box will index patients, activities and tasks. This way, finiding a patient grouped under a particular list or jumping to the a functional module will be quick and simple.
The patient banner is now taking a considerable amount of space: Is is really needed? Well, If we follow the idea of having patients accessing to their own profiles, it would be nice to have this personlizable. This way, patients would create a human connection between them and their clinical staff. I think the balance between the clarity of a customizable banner and the used space make of this a valuable area.
Thumbnail area will show the teaser of each content type (clinical domains). This way the patient banner won’t be only an area to ensure the patient identification but to highñight a meaninful set of content related with the latest activity of the patient.
What about the ‘Like’ buttton?
What would ‘Like’ mean in a clinical context? It might be pretty controversial saying ‘Like’ under a patient diagnosis. However, there are other social actions which could provide a helpful support to the care activity.
Notifications could be triggered automatically by events or due to a professional opinion. A notification doesn’t express an emotion, but a objective idea. It’s not a bad idea having a mechanism to communicate subjective feelings but here we shouldn’t make an error by this Facebook ‘positive-thinking’.
About the prototype
The opinions, ideas and suggestions shared in this article and in the ‘Facebook-like EMR (Part I)’ one only represent me (Carmel Hassan).
This prototype has been made adapting the icons of Jigsoar. The shown data is fake and the girl in the picture is me! I know I know…
In 2008, Bob Watcher wrote an article called “Why the medical record needs to become more like Facebook” where lay down the idea of having the social network as the mirror for a new Electronic Medical Record (EMR) User Experience (UX). A collaboration and social framework to provide better care to patients while keeping useful information between physicians and nurses.
He was not the only one; already in 2007, Robert Nadler established a high-level model remarking out the core functionalities that it could have EMR software. However, it was not intended to get only a Facebook-like UI but a real social site to connect Patients and Doctors.
After the latest Facebook re-design I don’t know if those authors would still maintain the same opinion about this topic, anyway I do think it’s a very very interesting approach even assuming any usability issue that Facebook could have. So here are my two cents.
A Design Proposal
Users & Profiles
If a clinical solution would need to completely work like Facebook, every user (physician, nurse, etc.) would have their own user porfile. However, the analysis of having a EMR looking like Facebook timeline suggests that the profile page is planned to show only patient data. Why would we need to see a nurse profile, then?
Following this idea, the Home page would be reserved for (primary) users and the Profile pages for showing the electronic medical records of patients (secondary users).
A possible extension would be allowing the access to the EMRz by Patients, so they could also check their own EMR online by themselves. In this case, we could consider them as secondary users not as part of the network community, but just to contribute to their own medical history and keep a direct communication between them and they’re care providers.
Social activity will be generated by clinicians considering patient-centred documentation based on Profiles. As suggested before, there are two main social groups: the one created by clinical staff only, and the one where patients and clinicians would interact.
Social interaction is the key point of this proposal and it underpins the main usaibility goals:
- Meaningful use
Profiles: Patient Timeline
Facebook profile pages have been re-design to look more like a real timeline where any important event is chronologically displayed. In this example, right side will be used to show a summary of the most important event types associated to the patient medical history like Health issues, Allergies, Diagnosis, Requests and Results, Progress Notes, Prescriptions and any other clinical subject. On the left side, any user (doctor, nurse and the patient) could add comments anytime.
Privacy should have a strong presence here, since having restricted-access data is a valuable feature that doctors, nurses and patients. They all will need to control the visibility of the data they entered in a fashionable way. Although setting permission in Facebook is pretty hard and unclear to get control about which type of users see which type of patient data or personal comments is still a requirement. This could be done seeing users by their role, applications as domains, and groups as teams. Still, patients are a special kind of user which will have direct access to his own record.
There’s a weak line which separates the Facebook familiarity advantage into the most confusing UI for a productivity tool and it’s called “enjoyment”. The idea of this Facebook-like EMR was to promote some typical behavioural patter of users when interacting with a social network like:
- Safe Exploration
- News Stream
- Other people’s advice
- Personal Recommendations
However, there are some other existing activities that the clinical staff usually does and this UI is not oriented to
- Changes in Midstream
- Keyboard only
- Streamlined repetition activities
Nevertheless, the app integration approach of Facebook suggest me to become a good exercise to imitate outside the UX field.
The content is not about what people think or feels, but what physicians and nurses make, diagnose, treat or provide to patients. The language of the user interface should be adapted to the purpose in order to ease the kind of content expected to be entered.
Data entry is one of the most challenging features in healthcare apps. The smarter the application, the quicker the data is entered. The whole phylosiphy of Facebook puttin atention into every single details should be adopted also to create content such as vital signs, prescriptions, a soap note, or a discharge letter. As an example, below there’s a list of content with different natures which depends on the context of use, the user role and the business model.
- Task Oriented
- Lists of patients
- Lists of tasks
- Scheduled activities
- Content Oriented
- Patient history summary
- Patient evolution
- Patient current health status and diagnosis
- Clinical decisions
- Process Oriented
- Protocol based care planning
- Admissions and discharge
This is only the first part of the exercise. For the second part I’ll try more mock-ups and the high-fidelity prototype adding more details also in the content used in this sketch to get a better idea about how crazy (or not) is the proposal of being inspired by Facebook to create an electronic medical record.
There’s also an important gap where Facebook has no direct answer: clinical safety and episode-related information. We’ll talk about it also for the next post.
By now, how realistic do you think it is?
Information technology has made possible to improve healthcare services by increasing productivity, getting more accessible services, and automating daily care activities – among others. However, the software itself allows getting a better patient care at the same time that it opens a door to clinical risks. It is a responsability for designers to avoid hazards as well as design rich-interactive solutions.
As you know, it is part of the process of making design decisions: user centred design to ensure usable products. UCD, UX and Usability have a lot of meanings and aspects that influence in reducing the errros that clinical staff could make by only interacting with the software.
I’ll give some examples of how the main topics of usability could avoid clinical risks:
Visibility of system status
Working with the right patient in the right time. Keep always visible the patient information that will help users to identify it without any doubt.
Match between system and the real world
Use clinical language as much as you can to help to understand the purpose, the scope, and to support the clinical process. Clinical procedures has specific names which require clarity unambigously.
User control and freedom
Provide multiple visualization modes to ease the data understanding. Allow performing activities in different orders to get adapted both users and business process. The more natural the process is for the user the more confident s/he will feel.
Consistency and standards
Respect familiarities with the real world activities and with previous software solutions. Clinical staff is very hard-working with protocols and routines, consistency could be boring but effective.
Provide clear points of interaction and valuable data for decision-making. A correct labeling and control aesthetic will help to easily understand how the data can be managed, as well as a correct data presentation will avoid performing wrong actions.
Recognition rather than recall
Reduce the workload by creating recognisable solutions. Do not stress the user with a lot of information to be remembered.
Flexibility and efficiency of use
Provide shortcuts and give a chance to experts, the more efficient they’ll use the solution the more time they’ll spend providing care. Allow users workarounds.
Aesthetic and minimalist design
Minimal design will simplify the data presentation and a good aesthetic will increase the content meaning to facilitate the interpretation and the decision-making. Use aesthetic to provide meaning and avoid decorators that mix up the patient information.
Help users recognize, diagnose, and recover from errors
Allow undo actions for the most critical actions: prescribing medications, deleting patient data, managing patient history, etc. A mistake should be able to be recovered by a an undo action or a rollback process.
Help and documentation
Describe how every care activity can be achieved by your solution. Show main conflicts and risks if there are any dependency between them.
These are only some clues about how putting attention in usability can help to reduce the amount of clinical risks. Lately, every data that is presented in a screen and every allowed interaction between users and clinical software it’s a potential risk to provide a wrong care to patients. Undoubtlely, there are many other hazards non-UI-related that can end up in a catastrophic situation.
Nothing is free of hazards, but good design decisions can save time, costs, and even lifes.
At the heart of our strategy to unify healthcare technologies is our Bluebird technology. Bluebird is a cross functional and multi product environment supporting a modern user experience across functional areas and technology. Both modern design and excellent usability are important parts of the Bluebird approach.
Bluebird has its origins in the desire to offer a modern UI design and a good usability of healthcare applications. The ability to change the look and feel of the UI is important to underpin the aim of providing modern healthcare software using state of the art technology and design.
I’m happy to hear that as a member of such a great team.
There’s nothing new in a widget-based interface. The need of having a lot of information at a glance, in one single screen following a Dashboard pattern has been a frequent demand and a solution that we can see form many providers. So, what makes the difference?
The metaphor of a widget is translated in different systems sometimes as a small display area showing updated-information and other times as quick access to a specific application. Although we could absolutely ignore the important gaps between these two approaches, the key factors that are common and creates the most important distinction between widget-based interfaces in Healthcare are the Information Design, the Visual Design, and the Interaction Design.
The overall usability and utility in healthcare applications currently relies in the features, the performance, and the efficiency – as any other productivity tool – and is letting the User Experience (UX) aside. Understanding the most common misconceptions about this discipline, the UX has to recover its position, not to be the brightest star, but to be the leading one. In this sense, widgets – as part of a solution – are the interaction and meeting points between the user and the system, and they have to be designed to contribute to the global UX at the same time that cares about the user needs in a specific context.
Some example of solutions that make use of widgets can be found in the website Healthcare Scene (examples 1, 2, 3). I’ll let the readership to judge the real difference among them. I encourage you to pay attention to two elements: a widget (one single widget) and the ‘desktop’/screen as whole.
Is the Widget the right metaphor?
Is this the discussion? Maybe not. Widgets are a flexible and powerfully component. However, looking the design approach of Windows 8 Tile/Live Tile, I’m afraid that the simplification of a complex requirement into a read-only and low-interactive display areas is still beating the challenge of creating a full-productive tool for clinicians; the no-ui conversation is also present. However, I strongly believe that widgets is a perfect solution not for product requirements but also for user needs, and as said above, we only need to put special attention to the visual design of the UI components to get an elegant distinction, to the interactions with widgets to empower users and the right design of the information displayed in the widget content to let them make the right decisions.
…And do not forget it: may the technology be with you.
I’d also love to know if the NHS has got some progress on it and if there would be any possibility in this zero-investment philosophy comming from Spanish healthcare public sector for a Common User Interface to integrate the 17 different solutions that currently exist.
UX Designers for Healthcare should be placed in Hospitals, near users. They would only need time to observe, to ask, to help, and a table to sit and sketch.
… unfortunately, the most of the time, UX designers are in departments, with managers and technologists; not thinking in the problem, nor providing solutions, but offering workarounds, feasability analysis, and making real the statement: companies invest in UX.
There’s no discussion for me, you don’t need justify UX ROI anymore, the only investment you cannot afford as a company is to waste your resources and the real experience of your users.
I’m not suggesting to remove design departments; if they exist in your company that’s a great new, I have seen how much you can influence in your company’s values when there are more than one single “UI-guy”.
I’m not only telling that you have to fire your domain experts, if they exist in your company take advantage of them by creating a domain knowledge database.
However, the work methodologies and development processes require a transformation, not only in the techniques applied day by day in the office, but also in the way the worker participates and, as a consequence, the way the designer is involded in the problem s/he has to face.
Designers as a service
The mistake, imho has been consider that the only service a designer can provide is “Consultancy”. That’s good when there are few UX Experts in your company and you need to assess your solutions or solve certain problems.
However, a good User Experience is a long-term goal which requires continuity, not only during the developement stages.
My suggestion is stop sending “troops” of analysts to meet with your customer, stop asking for permission to record “a day in the life of…”, stop wondering when a user is available for testing your solution in your lab, but to consider UX Designers as the ideal co-workers and partners living with the problem.
UX Designers should be considered also a service for some periods of time in your hospitals and healthcare centre, UX Designers should be the insider observer of your users.