We often get asked how to develop a Graphical User Interface for a medical device?
Maddison's guiding principles: For more insight download our medical interface design guide below.
1. First, do no harm. The overarching principle for developing a medical device is to use a process and practices that as far as possible reduce the risk of harm. hence the design process you must legally follow and the resulting documentation and testing is it driven by this principle.
2. Define the governing regulatory requirements. This is based on the adage: ‘Start with the end in mind’. Establish which standards your are designing to and what testing is required. Aligning a to the wrong requirements can be fatally expensive.
3. Involve users and test the device with them early. When users are clinical specialists with expert knowledge, who have been trained on and are accustomed to using predicate or analogous devices. Therefore, they come to your interface with existing heuristics (mental models of likely interactions, operating methods and responses.) Testing with users during all development phases, starting simply and adding complexity ensures that users’ existing mental models are harnessed to make your interface safe and intuitive.
4. Understand and use existing medical conventions. The medic about to use your interface for the first time in an emergency has probably been trained on or use done on a similar machine. They will assume that your interface works in the same way. The existing similar interface may not be intuitive to you but it is to the medic.
5. Develop the interface with successive build–test–learn loops. Use an iterative testing process, supporting users to test your interface – starting with a simple prototype that can be changed easily. As your confidence increases through testing, refine the prototype and its features.
6. Design the interface to be intuitive. An intuitive interface reacts in the way that a user would expect. For a user to anticipate a reaction, they hold a expectation of how the interface will react- a heuristic. In the early phases of design, you must discover and map their heuristics. Mimicking the user’s heuristic will make your interface intuitive. Where possible, make use of natural mappings – that is, natural arrangements of button or controls carry with them an intrinsic explanation of how they work or interact with a user. An intuitive interface reduces the need for any information from a user’s memory to perform a task. (This is described as reducing cognitive load.)
7. Use poka-yoke principles. Mistake-proofing, that is, the addition of a design feature that makes it nearly impossible for a user to misuse the device. This principle can be illustrated in simple terms using a children’s toy in which shaped blocks must be pushed through matching holes. It is impossible to put the wrong block in the wrong hole.
8. Design in redundancy. You should assume that your interface will be used in a busy, noisy environment and that it is probably not your user’s primary focus. Use multiple simulations to design ways in which your interface communicates with the user, so ensuring that it delivers the feedback needed by the user. Design alternative systems for using your interface, thus supporting users with different expectations, especially where there is a legacy system or common competitor. Providing more than one way to complete a task will help your audience achieve their own goals.
9. Design in forgiveness. Humans make mistakes – especially when under pressure, as medics often are – no matter how well your user interface is designed, so you need to use approaches that support users to prevent or recover from errors. This approach is generally called ‘forgiveness,’ the key being to prompt or pre-empt a user when an action they area bout to take may result in unexpected, irrevocable, or critical actions. Predict and explain the resulting action to the user, requiring them to confirm that the expected outcome before it happens. In addition, give them the ability to undo or stop the action if they realise they have made an error.