About MindfulGarden

Described by nurses as the “digital crash cart for behaviors,” MindfulGarden is an evidence-based interactive behavior modification platform designed to arrest and de-escalate agitated behaviors associated with delirium. Using interactive digital media with calming visuals and audio, MindfulGarden takes the patients’ physiological inputs (motion, sound, and vitals) to drive the assets on screen; think of it as a video game the patients don’t know are playing. The graphical assets and sounds would match the patients at their level of anxiety, then bring them back to homeostasis so proper care can be administered by care staff.

The company was born under a personal experience with our founder’s mother, Esther (click here to see the founder’s story), who was misdiagnosed with dementia when she had developed post-surgical delirium. Although the physical symptoms of both conditions are very similar (irritability, physical and vocal agitation and violence, rambling, etc.), delirium is temporary and can be reduced with proper care. Unfortunately, Ester was ignored, left alone, and finally passed away due to a bowel infection from being unable to go to the washroom for 14 days. Thus, the goal for MindfulGarden is to create delirium management tools for care staff to quickly arrest and de-escalate the onset of delirium to calm the patients down so proper care can be administered.

My Role

As the Senior Product Designer at MindfulGarden, my primary role is to lead, manage, and execute research, design, development, and innovation of company products and technologies. I am also responsible for the industrial design of our first physical product, the delirium crash cart, and the UI/UX design of all interfaces (platform application, website, user experience research). In addition, I would have to manage and work with our developers and third-party facilities to realize my designs. Working with our VP of products, I would also collect feedback and data from our evaluation and research partners. After synthesizing the information using various design research methods, I would consolidate the findings and work with company executives to devise potential executable solutions.

Click here to see what I did with the industrial design.

Click here to see what I did with the interaction design (UI/UX)

Industrial Design

When I started working at MindfulGarden in 2017, the first iteration prototype was a large and heavy unit requiring constant power from a wall outlet. Powered by WinOS, the prototype had a sizeable 40-inch TV mounted on an articulating arm with a small PC driving the interactive experience. A very rudimentary interface is accessible on the same screen as the interactive experience via a mouse and keyboard.

After bringing the prototype digital crash cart to a partnering long-term-care home and watching the care staff’s interaction with the unit, we quickly discovered that the unit’s base was too large and heavy for daily transportation and the height of the base where the small PC is stored limited the placement of the unit. Furthermore, as the base of the unit was custom-made, each costing up to $3000 to manufacture, it was vital for me to recommend a redesign of the hardware configuration to make our product more scalable and efficient.

Apply lean design methodology and to keep our product agile, I proposed several modifications to the design of the digital crash cart.

First, I swapped the heavy and expensive custom-made cart base with a lighter, more economical, off-the-shelve TV mobile cart. The new body significantly increased the carts’ mobility while reducing production costs. Using off-the-shelve parts would also allow us to quickly test our designs and make modifications without wasting valuable assets. In addition, the new cart would be less straining for care staff to transport the units throughout the facility, increasing adoption and usage rate. Later designs added a horizontal handlebar to improve further usability and ergonomics based on care staff feedback.

Second, we replace the high-end small format PC (Alienware PC at the time) with cost-efficient mini PCs. Not only was the mini PC much cheaper, but it was also even smaller in physical format to help further slim down the design. As staff and patient safety from electrical components was a vital input, a 3D printed enclosure was made around the mini PC, which I later modified into a dedicated waterproof case mounted at the mid-stem of the cart.

Third, I added a long 20′ retractable extension cord to replace the standard extension cables used on the original cart. The extension cord helped versatility to place the carts as the locations of the electrical outlets no longer restrained them. We later added a large capacity battery in place of the extension cord as we observed that the long cable would cause a potential tripping hazard to care staff. In the latest version of the crash cart, we have upgraded the commercial-grade battery to a medical-grade hot-swappable battery system in compliance with device safety and regulations.

Finally, I added a small secondary touchscreen mounted on the back of the unit and moved the control interface for the care staff to the secondary screen. The rationale for a rear-mounted control panel was derived from observing how the care staff uses the crash cart during the patient care routine. For example, the care staff would have to use the keyboard and mouse to access sessions on the machine without appropriate resting trays. In addition, having the backend interface on the same screen as the interactive experience means that the patients’ viewing experience would be interrupted every time a setting needs to be changed. Furthermore, the safety of the care staff would be at risk while they have their back turned against a potentially violent patient experiencing hyperactive delirium while adjusting the settings. The placement of the second screen allows the care staff to face the patent while interacting with the interface while the patient experience remains uninterrupted. In a later design, I proposed replacing the rear-mounted touchscreen and the mini PC with a rear-mounted tablet, acting as both the control panel and the processor that drives the interactive experience. Using a tablet, not only would we allow the MindfulGarden unit to be even more scalable with the integration of either the Apple or the Android ecosystem, but it would also expand the hardware configuration to encompass more usage cases. Although I attempted to make the tablet configuration a reality in 2018, due to hardware and processing power restraints, the design did not start development until late 2021.

In all, my design approach for MindfulGarden was to keep the company lean and agile. As MindfulGarden is an early startup company with a limited budget, it is essential to keep the cost down, ensure we can be quick to scale, and always consider user input at every design cycle.

UX/UI

The first iteration prototype of the MindfulGarden crash cart had almost no user interface asides from a couple of sliders to adjust the intensity of the interactive assets on screen and simple numeric indicators of the sensor inputs. So when I first started working with MindfulGarden in 2017, I immediately focused on a few interaction shortfalls.

First, as mentioned in my works on the industrial design of the MindfulGarden crash carts, having the control panel be on the same screen as the patient experience poses several challenges, such as lack of care staff interface and safety and patient viewing experience interruptions. However, with the addition of the rear-mounted touchscreen control panel, I could design a dedicated user experience for the care staff. Using the original functions of the backend interface, I’ve revamped and created a more user-friendly interface layout.

Second, in the original interaction, the patient-facing interactive experience activates as soon as the unit is turned on, which gives the care staff zero time to prepare and position the unit for use. In the new interface, I have added a session start sequence to allow the care staff to power on the unit and ensure it is positioned correctly before launching the interactive patient experience. Having a proper session start sequence would also trigger the start of the patient data collection, allowing us to start building up the patient usage database and data analysis.

Third, the original system automatically has the patient-facing interactive experience following a natural circadian rhythm (day, dawn/dusk, night) based on the time of day. However, several feedbacks from the care staff indicated that they would like more control over the circadian stages depending on the patient and usage case. Hence, we have added a time of day selection function.

Lastly, as the old interface had all the functions crammed onto a single page, I mapped and organized the various operations and created separate interactions, making expansion to the control interface possible.