Automated TB Diagnostic

Franklin W. Olin College of Engineering, 2006 - $17,250

This E-Team is developing an automated tuberculosis (TB) tester for the developing world. The current method of TB diagnosis, acid fast bacilli (AFB) sputum microscopy, is slow and unreliable: after collecting the sample, technicians spend 20-30 minutes looking for TB on a recommended 300 fields on each slide. Technician fatigue, lack of training, technician shortages and human error make sputum microscopy, especially in the developing world, highly inaccurate. By automating the slide reading process and replacing error-prone technicians, the team believes the TB tester will make TB diagnosis faster and more consistent, reducing resources wasted on false positives and letting fewer false negatives slip by.