Originally Published MDDI October 2005
By Erik Swain
Using doctors as design partners, conducting usability testing, determining user needs and constraints early on, and devoting appropriate resources to human factors are among the keys to a successful product design. Those principles and more were in evidence in the case studies presented at a June conference on human factors.
All of the products cited at the AAMI Conference on Human Factors, Ergonomics, and Patient Safety for Medical Devices had successful launches and gained market share. Three of them—the Diego powered dissector from Gyrus ENT LLC (Bartlett, TN), the Partner rhythm assistant from Guidant Corp. (Indianapolis), and the SonoSite 180 hand-carried ultrasound system from SonoSite, Inc. (Bothell, WA)— had also won Medical Design Excellence Awards.
The Diego dissector turned out well in part because the design team convened a surgeon panel to evaluate designs and test whether they felt right, said Amy Schwartz, director of user research for the health practice of IDEO (Palo Alto, CA), which worked on the design with Gyrus.
The panel was small, but included "extreme" users, she said. "We needed to include surgeons in the middle of nowhere who have no support, because they need to have things that other surgeons may not," she said.
Before design of the device, used in ear, nose, and throat surgery, began, the team observed ENT surgeons and discovered what problems they were having with their existing devices. "Workarounds provide evidence of a strong user need and may also suggest design solutions," Schwartz said. "They are the gold standard of observational research."
The Partner rhythm assistant, made by Guidant Corp. (Indianapolis), enables the patient to activate atrial defibrillation. It is a handheld device used with Guidant's
Vitality AVT implanted defibrillator.
From that, the team learned what sorts of designs caused strain on a surgeon's hand, caused cords to tangle, and made a device too heavy. It then engaged in a participatory design session with the surgeon panel. The surgeons were given prototyping materials and asked to design the ideal handpiece, with help from IDEO designers. This enabled them to see what could and couldn't be done.
The next step was usability testing, which found that the main problem was the cord was too stiff. That led to the final design, which had a bundled cord, a swivel to rotate the blade in a way that wouldn't tire the surgeon's hand, and a simplified handpiece design.
Guidant began designing its patient-controlled atrial defibrillation device by reviewing the literature and finding an outside design firm, Worrell Inc. (Eden Prairie, MN), with expertise in the area. The next step was to interview patients and clinicians about what features were most important, said Richard Stein, advising system design engineer at Guidant.
Guidant used that information for the initial designs. Key attributes included size, shape, weight, cost, battery longevity, input control characteristics, and use environments. "The device had to be able to tell the users what to do, and how they were doing," said Stein. "It also had to be easy to use by patients with weak vision, hearing loss, loss of memory, and a variant cognitive ability."
The company brought four designs to the first round of user testing. One emerged as the favorite, but users disagreed about what sorts of icons were best. Guidant designed a larger study with 150 patients to address that issue further.
Then it was time for the final usability study, designed by Guidant with FDA input. It assessed whether users understood what to do under certain conditions, and how to recognize when they were doing a function improperly. The 50 patients understood how to use the device. After a few more minor tweaks, Guidant submitted the design for FDA approval.
When SonoSite spun off of a larger company to pursue handheld ultrasound, it hired three full-time human factors engineers that were involved in all phases of product design and development, said Stephanie Barnes, human factors engineering manager. "Our CEO always appreciated human factors and understood that you can't design this kind of product without significant human factors input," she explained.
The first step was to determine corporate, customer, and project needs, which would define the level of investment needed. That determined which components of the system would require a large-scale human factors program, and which ones would require a small-scale one.
The level of scale determined how many people would be put on the project, how many tools would be needed, and how many iterations would be necessary. The handheld component underwent 17 usability tests, two major ergonomic studies, and four input- and display-related tests. It needed four rounds of physical design and three rounds of user-interface design, Barnes said. But management knew it needed to provide that level of resources because the requirements had already been determined.
Copyright ©2005 Medical Device & Diagnostic Industry