The term “ubiquitous computing” tends to invoke images of young folks flaunting their fancy iPhones and iPads, but in reality this concept can be leveraged to help an entirely new demographic, one that needs it now more than ever. Margret Morris, Jay Lundell, Eric Dishman and Brad Needham of Intel conducted a series of ethnographic studies with elders with various degrees of cognitive deficits to better understand their needs and design more appropriate technologies.
This enterprise is not a new one. Researchers at Carnegie Mellon University created “Nursebot”:
Nursebot helps elders accomplish their daily tasks by guiding them through their environment. The primary goal of this project was to develop a mechanism that allows elderly people with cognitive defects a chance at a normal life. Pearl (the nursebot) lives with the patient and assists them by performing tasks such as:
- reminding the elderly patients to take their medication or to go see the doctor
- acting as a proxy for the professional caregivers by allowing the professionals to have direct interaction with the patient despite being physically far away
- collecting data about the patient, which can be very useful for surveilling the daily health of the patient
- assisting the elderly if they fall or need to manipulate some physical object such as lifting a heavy object or taking something out of the microwave
- allowing the patient to engage in some social interaction
As helpful as this robot seems, the implementation and execution would be difficult because it is a technology all too foreign to the patients. If the goal is to help them lead more independent lives, the priority should be on leveraging the technologies they are already comfortable with. The nursebot technology, and other such attempts to help those with cognitive deficits live richer lives, has a steep learning curve and an inherent unfamiliarity that makes most elderly patients uneasy. Morris et al. suggest that designing technology to help this demographic should be centered around a certain set of requirements that is more true to what these patients want, and would benefit from the most.
After conducting a total of 45 interviews, Morris et al. concluded with four principles that ought to guide any ubiquitous computing for those with cognitive disabilities.
1. The technology should support early detection.
Those with cognitive deficiencies often encounter a stage of denial where they cannot come to terms with the fact that there might be something wrong with them. The technology should be aware of such a phenomenon.
This chart shows that due to denial and uncertainty, the patient might have a level of perceived functionality that is incongruent to their real level of functionality. Morris claims that this phenomenon of denial can be avoided by using the “embedded assessment” strategy – essentially embedding small cognitive tests and tasks into the patient’s environment and daily schedule. This method incorporates cognitive tests in such a way that they provide practical value to the daily lives of the patients. Morris claims that this would help the patient overcome denial and encourage early detection of cognitive defects. The most effective and well received example of embedded assessment for patients with cognitive decline is some kind of rehearsal tool to recognize names and faces. Morris et al. devised a prototype to address this particular issue:
This demographic of cognitively declining elders require the most user-centered designs possible, because dealing with the technology shouldn’t be a chore in itself. Morris et al. believe that such examples of embedded assessment satisfies both the overall objective of early detection and denial avoidance and is a simple and intuitive task that the patients would perform anyway.
2. The technology should adapt to support patients with varying levels of cognitive deficiencies.
This principle seems fairly obvious since there are naturally varying degrees of cognitive deficiencies in the elderly patients observed. A tool that only looks at one subset of patients would be too chauvinistic in its functionality, and wouldn’t scale well for individual variability. In fact, patients with cognitive deficiencies often experience varying degrees of decline within the same day, so the technology should be able to adapt to the variability within and across patients. They suggest a design that would adjust the level of support based on data received from sensors and cognitive tests embedded into daily activities (puzzles, games, etc.).
3. The technology should support or catalyze social relations.
Social interactions are an important component of cognitive functionality. We are inherently social beings, and being deprived of social interaction (especially for those already cognitively declining) might be catastrophic. The interviews Morris et al. performed supports this hypothesis – the socially isolated elders seemed less well cognitively, emotionally, and physically. The folks behind Nursebot might chime in here and claim their robot passes this test with flying colors. However, Morris claims that the system designed for elders ought to catalyze or spawn human interaction, and not replace it (which is what Nursebot sought to do). They prototyped an activity tracking system that helped “Heidi” that sensed, tracked, and databased her movement throughout the house, then signals her with the opportune moment to contact others who also have the same tracking system in their homes.
This system allows Heidi and others like Heidi to seek out contact with each other “organically”, with a little help from the system. This prototype prompts elders to seek social interaction with others, rather that sit at home and talk to a robot.
4. The technology should leverage familiar interfaces.
This is perhaps the most important principle that ought to drive design of technologies for elders with cognitive deficiencies. Morris reports that of the interviews conducted, the elders tended to avoid dedicated “computer rooms”, instead preferred to sit at the kitchen table, or their favorite chair. So technology for them should be accessible in interfaces they know and use already. The familiarity of the object may itself be a powerful recall cue, and will have the longest usability.
Morris et al. distilled these principles from 45 observational interviews conducted over long periods of time. Such a rich ethnographic study has the power to yield extremely valuable results, as explained above. If we truly want to build tools to help cognitively declining elders, we have to put their requirements front and center, and not get sidetracked by designing the coolest or most advanced piece of technology. Though Nursebot and SmartHouses are impressive in their own right, they can’t serve their intended purpose if elders don’t use them! Making interfaces easy, intuitive, and fun are the most important aspects of designing technologies for this demographic.