Computational algorithms to improve medical scheduling

Current approaches to patient and staff scheduling negatively impact patient care by distracting providers from clinical work. This project focused on creating an advanced model for automated large-scale scheduling of healthcare providers. The schedule found a Pareto optimal solution – thus balancing constraints for all staff members. This empowered administrators and clinicians to have more control over their schedules, so that clinicians can prioritize patient care.

Our team interviewed over 40 personnel (providers, administrators, assistants) involved in medical clinic operations throughout the country to assess the need for increased efficiency in administrative processes using technology. Interviews showed a disconnect between schedulers and administrators about the time it takes to schedule providers throughout clinics, and demonstrated a need due to the cost this additional time creates.  We then successfully developed use-cases for a process to intake scheduling constraints from a medical clinic within the Baltimore area, and created a schedules that showed less violations of scheduling constraints, and far-reduced the time (from weeks to seconds) to create multiple schedules showcasing constraint trade-offs.

The project is ongoing, and is now conducted through the Malone Center for Engineering in Healthcare at the Johns Hopkins University.  The current portfolio of work has expanded to broader problems in healthcare operations, including therapist location-optimization, outpatient no-show prediction, and provider forecasting.

Automated RGBD to C-arm calibration

Orthopedic surgery often requires the difficult placement of many tools. To ensure the correct placement, surgeons take multiple X-Rays, which often increases the procedure time and radiation exposure for the patient. This project centered on utilizing computer vision techniques to create an automated calibration algorithm that overlaid a Cone Beam Computer Tomography (CBCT) image with a live Red-Green-Blue-Depth Camera (RGBD) image. This overlay created a mixed-reality visualization that could be utilized by surgeons to improve navigation and surgical outcomes. The project was a collaboration with the Johns Hopkins Computer Aided Medical Procedures (CAMP) lab.

How might we improve the educational experiences for new immigrants?

This project was conducted through the IDEO Course for Human Centered Design. Our group focused on building better educational opportunities for Hispanic immigrants in Baltimore. To gain insight into the different hispanic Baltimore communities, we conducted in-person primary source interviews with various community members. Following interviews, we created a plan to build better social relationships between Hispanic community members to better serve the interests of the focus population and existing Baltimore community.

Subcutaneous injection of therapeutic monoclonal antibody injections

The aim of this project was to develop a subcutaneous injection device for monoclonal antibody (mAb) suspensions. The project was a collaboration between the Center for Bioengineering, Innovation and Design of Johns Hopkins University and Janssen Pharmaceuticals.