iTriage was creates by two ER doctors to give users “quick access to a huge healthcare & medical database” (iTriage website). The content has been reviewed by Harvard Medical School and while over 12 million people have downloaded the iTriage health app, it still has some major usability problems. Therefore, before we begin a re-design of the application or engage in usability testing, we would need to complete some preliminary measures to evaluate the application.
A heuristic evaluation uses a small set of evaluators to judge compliance to a set of known usability principles. Typically this process involves anywhere from three to five evaluators who create written or verbal reports. In my report, I like to go through the interface twice (first for a general feel and second for details).
First, I had to determine a set of heuristics to use. Using the ten heuristics from Molich and Nielsen (1990) is a good place to begin. Then I’ve decided to add on two heuristics because the application will be used on a mobile platform. This requires compliance to a unique set of usability principles.
- Simple & natural dialog
- Speaks the user’s language
- Minimize user’s memory load
- Clearly marked exits
- Prevents errors
- Precise error messages
- Help and documentation
- Target size
- Screen size appropriateness
From there, I started to explore the app generally on my first time around and then I got into the details on a second look. Here is a short summation of my findings:
- Simple & natural dialog – This is inherently difficult for a medical application because you have to balance common terms with medical terminology. Some of the language used in the application is confusing (conditions vs symptoms). The language in the parentheses isn’t consistent on whether the medical name comes first or whether the common name comes first. There is also very little dialog to push the user along so there’s no natural flow of dialog present.
- Speaks the user’s language – There a distinction between messages and news which I would’ve considered to be the same things. The application does try to speak to novice users by defining terms in layman’s language while still having all the necessary medical terminology.
- Minimize user’s memory load – Currently there is no way to combine symptoms. Therefore the user would have to remember which ailments were brought up for one symptom and then go check another one. The user can save their ailments to “My iTriage” and keep track of them.
- Consistency – The drop downs were consistent in that they all looked similar and only one opened at a time. However, the user received clicking feedback when they couldn’t actually click on anything.
- Feedback – The critical indicator is a necessary, but scary feedback mechanism. When a user enters a symptom that may be linked to a critical condition, a red-orange pop-up appears and they are instructed that they should call 911. There’s good feedback when you generally press things because the application takes you to a new page or shows a visual change in the opening of pop-ups. However, when you try adding symptoms, there’s no feedback as to what you added or that it was saving symptoms.
- Clearly marked exits – Besides just exiting the application, there was no exit for the overview tour. While you are in the application pages, there’s no way to get back other than using your phone’s inherent back button
- Shortcuts – No shortcuts present
- Prevents errors – A medical application will have a large amount of errors as users will not be diagnosing themselves properly. This is a common phenomenon where you can enter a simple symptom but receive severe diagnoses such as cancer. However, I don’t know how much an application could do about getting a proper diagnosis. Then, as mentioned above, the critical indicator prevents errors in the application. You don’t want someone with a heart attack using iTriage to diagnose themselves and think that they have the common cold. Lastly, a user can edit emergency numbers to be whatever number they want. This can cause major errors because if they type another number than 911 into that spot and then use it to call 911 later, they won’t receive 911..
- Precise error messages – No error messages present
- Help and documentation – Some of the subpages do not push users along or give users help to figure out what they should do. For example, on the symptoms page, there’s no instruction to touch the body or what to do. There’s also no direct link to the glossary, in case someone wanted to look up a medical term.
- Target size – Nothing looks like a button in this application. The age slider has levels and an uneven range. The images on the page are too small. The spacing for definitions goes off the page and drop downs are only allowed to open one at a time.
- Screen size appropriateness – The way to display information is not best for the screen size at hand as much of it is block text. There’s also no easy way to scroll through lists.
This heuristic assessment was only for about an hour, so given more time, I’m sure there would be more to discover. Therefore, talking to other evaluators and taking more time to further explore the application in detail would give us a better idea of how to approach usability design and application re-design.