The following are samples of human factors, engineering, and usability projects that highlight my experience. These are a collection of examples from each stage of my career including current and on going human factors work, a biomedical engineering design project, and a usabil-a-thon consulting application.
The American Medical Association and MedStar Health’s National Center for Human Factors in Healthcare developed a framework, based on the science of user-centered design (UCD), to increase transparency of electronic health record (EHR) vendor usability processes. This framework uses a 15 point scale to examine EHR vendor use of UCD best practices.
As the research assistant on this project, I was responsible for employing many human factors and research skills. This included performing a large literature review and environmental scan of the current research, guidelines concerning EHR usability testing, and current evidence-based best practices in human factors and usability. This literature review was then used as a basis for a grading rubric that contained three dimensions: User-Centered Design Process, Summative Testing Methodology, and Summative Testing Results. The rubric was then used to grade and extract the relevant information from the EHR vendor's usability reports by myself and the research team. Additionally, I contributed to the final design and presentation of the results of the study that allowed for comparison between the EHR vendors in a simple and clear data visualization.
A complete methodology and results of this study can be found on the National Center for Human Factors in Healthcare website .
Led a team of five biomedical engineers to design, prototype, and characterize a shivering detection algorithm in Matlab in order to aid cardiac arrest patients. Collaborated with university faculty and physicians from Children's Hospital of Philadelphia, who are part of the center for simulation, advanced education, and innovation. Used interviews and focus groups to identify a design problem within the hospital. Received first place in the department of Biomedical Engineering and received the Faculty's Choice Award in the College of Engineering.
Abstract:
Targeted Temperature Management (TTM) is a recommended and commonly used method for cardiac arrest patients in order to slow metabolic rate, decrease intracranial inflammation and protect against reperfusion injury to prevent brain damage. Unfortunately, when body temperature is decreased, the body will try to maintain homeostasis causing a person to shiver and ultimately counteract the benefits of TTM. Currently, the Bedside Shivering Assessment Scale (BSAS) is used to diagnose shivering in these patients; however, it is subjective and time consuming for the health care professionals. Therefore there is a need for a device that can accurately detect the onset of shivering. The objective of this design was to create a non-invasive device that will detect and signal the early stages of shivering through utilization of electromyography (EMG) signals from the masseter and other muscle groups in order to quickly and accurately assess shivering. To achieve this, an algorithm was developed that processes EMG signals and identifies if shivering is occurring by accurately distinguishing a shivering EMG signal from noise and artifacts. The algorithm was loaded onto a computer and used to process raw EMG signals obtained from surface EMG electrodes located on the masseter muscles and thorax. Upon shivering detection, the device outputs both visual and audible alarms with silencing capabilities in order to appropriately alert healthcare personnel without interfering with other ICU practices. The design criteria for the device were set to shivering detection of two minutes or less, and a specificity of 80%. The algorithm was characterized in a manner to find true-positive rates that could be used to find maximum intervention time to compare to the two minute criteria that was set. Testing of the algorithm on known shivering and non-shivering data sets resulted in 65.40% true-positive rate and 100% specificity. With these specifications there is the potential that algorithm could be delayed 51.9 seconds in a two and a half minute EMG recording. Therefore, the algorithm met all the design criteria. Further development of the algorithm includes collecting a larger number of shivering data sets to investigate patient variability, optimize filtering and utilize additional muscle groups.
The final device will be compromised of a hardware prototype and a software algorithm that will be marketed as a stand-alone critical care patient-monitoring device. The target market end users for this device will be hospitals that currently use TTM, which are over 400 hospitals in the United States and currently increasing. This market is expected to grow as TTM becomes more accepted as the preferred treatment for out of hospital cardiac arrest patients. Currently there are no devices that are similar to the one proposed by our team, but there are patents that discuss similar technologies. Because the majority of the prototype is software based, the projected cost will be low. The only manufacturing required would be for the device housing including an LED and alarm.
This is currently an ongoing project for George Mason University's Physical Ergonomics class. This assignment includes identifying an occupation that is prone to common ergonomic issues, assessing the workplace ergonomics in the field, and creating a set of recommendations for the workers based on the observations.
Currently, we are scheduled to perform infield observations at the Falls Road Animal Hospital in Baltimore, Maryland. This assessment will include interviewing and surveying the current staff about typical job duties, injury records, and demographic data. We will also obtain anthropometric measurements of the veterinarians and their work space to determine how ergonomically friendly the current set up is. Additionally, video analysis will be performed to determine current and potential hazardous actions and postures.
This project is performed by myself and another graduate student, along with a professional ergonomist from JFA Associates, a renowned firm in the area. The ergonomic assessment will be completed by December of 2015.
In April 2015, I participated in the first annual Usabil-a-thon DC. This event was a Hack-a-thon style usability experience. My team of four human factors researchers and engineers had 12 hours to consult MetroStar Systems' usability and user experience designers to create an educational and healthy lifestyle application.
MetroStar Systems provided us with the app Woozy, that uses crowd sourcing to provide location information about sicknesses such as the Flu. The designers wanted to incorporate a gaming element to the app to appeal to the younger children of its possible users. This game was to be educational about healthy habits, easy to play, and addicting to users.
The human factors team employed a variety of principle usability research methods including competitor analysis, personas, design thinking sessions, think aloud protocols, storyboards, and low fidelity prototyping.
Many conclusions and recommendations made by the team have been incorporated into the app, which is currently available in the App Store on iTunes.
Previous project for George Mason University's Usability and Product Design class. The idea for this product is a mobile application aimed towards long-term, hospitalized, pediatric patients that allows users to maintain contact with other patients within the hospital. This is achieved via the mobile application and in person, through suggestions made by the application. Four main components will comprise the application including: (1) a game center, (2) a chat room, (3) a location check-in with an interactive map, and (4) an announcement bulletin that can be updated by the hospital staff. Each user will be prompted to create a simple profile indicating their name, location within the hospital (i.e. room), and a simple user symbol. This application is designed to address the boredom typically felt by patients during hospitalization, in addition to sadness or loneliness related to chronic and long-term medical ailments. The application is intended to provide a distraction through a positive and social outlet.
Many human factors and usability principles were employed during this project including personas, competitor analysis, context analysis, task analysis for various goals within the app, heuristic evaluation, design thinking protocols, and quantitative usability testing (all raw data for these methods can be found in the appendix). A complete presentation of testing and results can be found here, while slides to the left present the main testing, design, and recommendations for the project.
Led a team of two biomedical engineers in a redesign of a current cyclo-computer product due to ineffective user interface design. Heuristic evaluation along with a keystroke level GOMS analysis were used to evaluate the current interface. Based on these results, two novel prototype designs were constructed and evaluated with qualitative and quantitative testing methods. Constraints and criteria of the product were considered during the design of two novel interfaces. Results of this user testing were used to create design recommendations and improvements for future designs.
Abstract:
Even though the Cateye Strada Cadence Cyclo-Computer consists of a very simple user interface - three buttons and a large LCD display - the system still requires a convoluted, confusing series of button presses and holds to navigate through the structure of its operating system. To pinpoint the origin of the system’s design failure, a heuristic evaluation was performed, along with a keystroke level GOMS analysis to estimate execution time of common tasks. Based on this assessment along with user profiles in mind, two novel user interfaces were developed into prototypes, a touchscreen and multiple button option. To assess the success of the prototypes, a usability test and subjective survey were performed with six participants. Overall, the touchscreen prototype performed the most successfully however, both prototypes displayed decreased task times when compared with the original design. Future testing and design should focus on overcoming the prototyping limitations, the small participant group, and testing as a secondary test instead of primary task. In conclusion, the evaluation performed gave insight to further improvements of cyclo-computer user interface design.
