Task-flow, Sketching, Asset Library, Wireframing, Prototyping

About this project

One of the greatest risks to Patient safety in drug delivery is human error. Doses must be calculated using a myriad of variables and a decimal point can be the difference between life and death. This project was focussed on improving safety  through managing the drug and it’s clinical uses across all levels of a healthcare organization. From pharmacist to clinician, the drug is controlled by an intelligent software that provides faster, safer delivery of drugs from a drug library to the infusion pump. Due to the confidentiality of this project, deliverables and product images are not able to be shown.

My Involvement

I was brought in as a lead designer and to lead a team in the design of desktop software that would support safety critical decision across multiple users and use cases. I worked with three different development teams that encompassed both the desktop software as well as hardware components related to the delivery of drugs. I met with subject matter experts in key roles of leading pharmaceutical companies to facilitate requirements gathering as well as review designs in various stages.  

Task-flow creation and analysis

Working closely with Product Managers, I distilled hundreds of individual requirements and their dependencies into task-flow diagrams. This visually mapped the logic in each use case and revealed conditions, dependencies and errors that were incorporated into requirement revisions. Once this was resolved, it paved the way for a faster design approach that would be able to satisfy safety needs as well as desired functionality. With annotated references to supporting requirements, these artifacts were a critical part of the traceability of design considerations for internal and federal review.

Wireframes

Asset Library

Using the task-flow diagrams, design work on the interface could begin. First on the list was to create a custom asset library of widgets. Each was documented with it’s own native styling and behavior then grouped into classes. This provided a key foundation for consistency and reuse throughout designs and development teams would benefit by limiting on-off widget coding. 

Typography

Typography and a Style Guide were defined for all widgets and text elements as this project would be heavily using forms. Safety aspects along with conditional dependencies imposed constraints on the layout of content. For that reason, the typography was handled up front so that we could design to scale. And as with the widget library itself, consistency and reuse were other key factors driving this effort.

Wireframes

Now that the base elements for UI design had been created, wireframing ensued. Key subsystems of the main application were each geared towards specific users and their tasks. As such, I took the approach to align individuals, including myself, with each subsystem. In addition, I managed the design and review process. We started off sketching designs to quickly iterate on concepts and refine the needs for the interface and a content strategy.

Using the team edition of Axure, we were able to check out / in our respective work into a SVN repository that would allow efficient and visible work. Custom report templates were created in Axure to allow us to quickly generate output documentation on a moments notice for key stakeholders which was also managed under version control. The documentation included the details for interactive behavior beyond what was native to the respective widget as well as the styling applied as defined in the style guide. 

Prototype Creation

Due to the complexity of conditional behavior and dependencies, it was decided that prototyping would be based off a duplicate of the wireframe pages. This allowed us to impose whatever means necessary to render the interactive behavior desired without tainting the integrity of the wireframes themselves. By using an interactive prototype, we were able to iterate through the review process faster with both internal and external stakeholders. The prototypes were eventually used in usability testing by a dedicated Human Factors team and allowed data to be collected without imposing on development teams progress on approved UI designs.