Reducing Antimicrobial Resistance through Improved Sensitivity Testing

Problem: Antimicrobial resistance (AMR) presents major challenges for health systems and patients in Malawi and other LMICs because it threatens the ability to treat life-threatening infections, leading to increased healthcare costs and economic losses. In sub-Saharan Africa, in particular, there is a paucity of data regarding local AMR patterns due to lack of routine surveillance programs.

Limited laboratory resources and capacity are significant barriers to routine microbiology culture and susceptibility testing. Antibiotic susceptibility testing is important in guiding treatment so that patients receive appropriate antibiotics in a timely manner, and hospital antibiograms – a summary of local susceptibility and resistance patterns - provide guidance on appropriate empirical therapy, which has been shown to decrease mortality in patients with sepsis.

Solution: Working with Morgan Greenleaf and his team we have assisted in the creation of OpenAMR, a low-cost instrument to automatically read and interpret susceptibility testing results using computer vision. OpenAMR was designed, built, tested, and deployed locally with the goal of strengthening AMR surveillance.

Skills developed: 3D modeling, CNC routing, 3D printing, computer vision, system design, UI design, databases, cloud computing, & Python.

Project Partner: Morgan Greenleaf and colleagues

Improving In-patient Laboratory Specimen Collection in Hospitals

Problem: Efficient bedside collection of laboratory specimens is critical in reducing wastage and improving testing turn-around time. However, the process is hampered by challenges in coordinating the necessary equipment and supplies to collect the sample.

Solution: To improve specimen collection we have developed a mobile cart (trolley) that has space for all necessary equipment and can be taken to the bedside.

Skills developed: 3D modeling, CNC routing.

Project Partner: Kamuzu Central Hospital

Reducing Road Traffic Accidents through Radar-based Automated Speed Alerting to Drivers

Problem: Road traffic accidents are one of the leading causes of injury and death globally. In high-income countries, once common method used to alert drivers they are speeding is a road-side sign that displays the driver's speed. While this approach may be effective, it is cost-prohibitive for use in most low- and middle-income countries.

Solution: We have developed an inexpensive version of the speed warning system that can be locally manufactured at a fraction of the cost using open source hardware and software.

Skills developed: Electronic design, circuit board manufacturing

Project Partner: Oliver Gadabu

Increasing Local Capacity to Maintain Life-saving Medical Oxygen Machines

Problem: In most low- and middle-income countries medical oxygen is provided using oxygen concentrators. These devices extract nitrogen from room air, leaving oxygen concentrated to roughly 95%. While there are many possible modes of failure, one common problem is the failure of the electronic control unit. Replacement electronic control units are expensive and often difficult to procure.

Solution: To address this problem we have developed a replacement control board that can be manufactured locally and a fraction of the cost. We are currently in the process enhancing our design to add a telemetry module that will allow remote monitoring of the concentrators while in use, generating real-time alerts when an error occurs.

Skills developed: Electronic design, circuit board manufacturing

Project Partner: Kamuzu Central Hospital

Using Robust Mobile Clinical Workstation to Support Community Outreach in Rural Honduras

Problem: Many children in low- and middle-income countries do not have the opportunity to attend clinic for routine growth monitoring. Shoulder-to-Shoulder, an organization supporting a rural clinic in San Jose Honduras, conducts semi-annual site visits to 14 schools within a few hours travel from their clinic. Getting to these schools often requires walking the final mile or two. Once at the schools volunteer medical students and doctors conduct "well-visits", examining each child. Height and weight are recorded using a paper chart. However, since there is no record of these details from previous visits, it is difficult to know if a child's condition is improving.

Solution: To address this problem, we have developed a mobile data collection tool.

Skills developed: Electronic design, circuit board manufacturing, software development

Project Partner: Shoulder to Shoulder Foundation

Improving Building Security and Access Privileges Using RFID

Problem: Building security is essential. Locks and keys work, but do not provide flexibility to limit access to specific times of day or days of the week, easily revoke access, or to track access on an individual level.

Solution: To address this challenge our interns have locally developed a system using RFID. The system is currently in use at our Training Center.

Skills developed: Electronic design, circuit board manufacturing

Project Partner: Internal

Managing hypertension in rural settings using remote monitoring

Problem: Hypertension, or high blood pressure, is on the rise in LMIC settings. This can be managed with medication, however it requires regular monitoring. This is difficult for patients living in rural communities without easy access to healthcare.

Solution: We have developed a portable blood pressure instrument that sends blood pressure measurements to a central registry and receives recommendations based on analysis of current and past BP measurements for that patient.

Skills developed: Electronic design, circuit board manufacturing, 3D modeling, CNC routing

Project Partner: Community against Diabetes and Hypertension (CADH)

Extracting patient demographic details from health passport labels

Problem: Patients registered in many health centers and hospitals in Malawi are issued with a sticker to be affixed to their health passport displaying their demographic details. For more details see here. When patients visit a different health facility their demographic details may not be present in the local database due to connectivity challenges. This requires manually retyping the information into the local database.

Solution: We have developed a device that uses computer vision to capture text from a health passport label and store it in electronic format. A short demonstration video of the device in use can be viewed here. Not all information stored in the demographic record is shown on the label so additional details must be added manually.

Skills developed: Electronic design, circuit board manufacturing, 3D modeling, CNC routing, computer vision, optical character recognition.

Project Partner: Malawi National Registration Bureau