Our time in Tanzania was nearing its end. By now, we had done an adequate amount of hospital visits and had identified problem areas that we could innovate within. It was time for us to do our final internship capstone project! Julia narrowed the projects down to two: a mechanism to prevent the backflow of fluid in suction pumps, and a modification to the Clean Machine to serve Amana Hospital’s needs. As Joel, Sadick and I had already worked on repairing and troubleshooting the suction pump from Amana, and we were assigned to the suction pump project.
After our experience repairing the suction pump, we had multiple ideas to address this problem. These ideas were varied- they ranged from a training and troubleshooting app to a completely revamped design of the pump. Our assigned project statement helped us narrow down these ideas and focus on the problem at hand. We knew that we had to focus on a mechanism to prevent, delay or limit the damage caused to the aspirators by fluid backflow. We narrowed down our ideas to those that explicitly addressed the problem.
During our visits to Amana and Muhimbili Hospitals, we had a chance to focus our observations and interviews on the suction pump problem. In addition to interviewing nurses and surgeons about the common problems faced with the suction pump, we got an opportunity to see broken suction pumps in the Muhimbili Hospital Graveyard, as well as speak to technicians who had to repair these suction pumps. From this, we learned about the problem with the current solution. The built-in valve mechanism was often removed by the users. The bacterial single-use hydrophobic filters were not always effective, expensive and had to be replaced often. Our solution had to serve the purpose of the valve as well as ensure that it was not removed by the users.
Although we had just about eight days till the end of the internship, our team decided to go through the entire engineering design process, doing focused research, forming design criteria and making screening and scoring matrices. Our design criteria focused on what we thought was the most important to Amana Hospital- requiring minimal clinician input, effectiveness, accuracy, durability, and low cost. After screening and scoring our ideas, we decided to prototype our top two ideas, to be able to further decide what was best suited to Amana Hospital’s needs. Our top two ideas were a multiple canister system- adding more canisters to the suction pump to delay the backflow of fluid, and a sensor system that would detect when the valve shut off the suction and alert the user that the canister needed to be emptied.
Using lunchboxes, modeling clay and pipes, we began by prototyping the multiple canister system. Although the prototype worked, we agreed that it was not a permanent solution to the problem. We could add multiple canisters to prevent the backflow, but it would take up precious space in a surgical theater. It was also a temporary solution- how many canisters could we keep adding before the backflow ruined the pump?
We decided to move on to our valve-sensor mechanism. It was a combination of a 3D printed valve which served as a reusable backflow prevention mechanism, coupled with a sensor system that would alert the user when the canister filled up. We hoped that the sensor system would let the user know exactly why the pump stopped working once the canister was filled, and hence prevent the removal of the valve. To prototype this idea, we needed a functioning suction pump to test our valve and our sensor mechanism. We had returned our repaired suction pump to Amana Hospital and did not think that they had one to spare for us to prototype on. Hence, in the spirit of ‘making do’, we fashioned our own suction pump out of a leak-proof lunch box, 3-D printed parts, pipe, and a hand-pump. We estimated the dimensions of the cage that would house the valve and 3-D printed it. Now, with a makeshift suction pump ready, we decided to work on our prototype.
We began by researching the shapes of existing valves. After some research and testing, we had a simple valve that fit in our cage, floated, and shut the suction off once the canister was full. Our next step was to prototype the sensor system. Here, we ran into some problems. We first attempted to use an IR sensor and emitter pair to detect the valve closing. Although this system worked in the beginning, we ran into some trouble making the system accurate. It would give us plenty of false positives and false negatives. We tried to fix this problem by using different materials and colors for the valves, as well as creating temporary ‘hoods’ to remove interferences to the system. But after multiple attempts were unsuccessful, we realized time was running out and we had to try a different system.
Our second attempt was a magnetic sensor system. This system was a little bulkier as it required the use of an Arduino board, as well as a bigger sensor and a magnet on the valve. Although we got the electronics to work, we ran out of time and couldn’t manage to design a valve that could float while handling the weight of the magnet. With just a couple of hours left until our presentations, we decided to create a functional prototype that would communicate the idea. For our final presentations to the DIT Faculty and Dr. Mafele from Amana Hospital, we presented our valve in our in-house suction pump and presented the sensor mechanism separately with an external magnet. These were enough to demonstrate our idea, and we also explained that the entire system could be customized to any make and brand of suction pump.
Our prototypes were well received by both the faculty and Dr. Mafele at Amana Hospital. Along with the positive comments, we also received multiple suggestions and additions that could make our system more effective and easier to operate. This experience reiterated to me how important it is to have constant clinician interaction and input while prototyping a device that will be used in a medical, surgical setting. Dr. Mafele and the professors at DIT brought up points that we hadn’t heard of during our interviews and that we hadn’t thought of during our brainstorm sessions.
Our project was far from a complete prototype, but we had done the work of developing the problem, homing in on a solution, eliminating options that would not work, and making the first attempt at a prototype. I felt myself get restless and frustrated when our ideas and their executions failed, and I was disappointed that we couldn’t make a complete, neat prototype. Julia, Matthew, and my team helped me put things in perspective and appreciate that working on any project is a slow, iterative process, and no amount of progress is too little progress. I’m sure that this project will continue to be worked on by engineers in DIT, and I plan to bring this project back to Rice so we can work on it here as well! In the meantime, I am happy to have learned so much more about CAD, 3D printing, Arduino, suction pumps, valves, sensors and patience in the process.