The United Nations is committed to reducing global infant mortality rate (IMR) and maternal mortality ratio (MMR) compared to the 1990 figures by 67% and 75% respectively by the year 2015.
My journey to help achieve this goal began in 2009 with a visit to Santhemarahalli located in the Chamrajnagar District of rural Karnataka, India. Spurred by the National Rural Healthcare Mission (NRHM), an Indian Government initiative, the Chamrajnagar District has developed an extensive array of Primary Healthcare Centers throughout the region. The National Rural Healthcare Mission in India has identified that one of the ways that they can help to reduce infant mortality rates and maternal mortality ratios is through institutional delivery of newborns. An average of about eighty babies were annually born In the Santhemarahalli Government Hospital in the recent years. The hospital did not have a blood bank or an anesthetist on-site, both of which are critical if infant delivery requires an operative process. There were several reported instances of maternal death in the region and official statistics indicated a maternal mortality ratio of 250/100,000. The institutional delivery system at Santhemarahalli was in need of some intervention.
A team of colleagues from GE and I met with a doctor who had been working at the Santhemarahalli Government Hospital for four months. This doctor has begun working at this particular hospital primarily in order to meet the practice requirements for nomination to the Royal Society of Obstetrics and Gynecology. At the same time, an Ultrasound scanner had been sourced through the Karuna Trust in a unique Public-Private Partnership. This partnership also compensated the healthcare worker for facilitating institutional delivery and the patient for hospitalization charges.
The availability of the ultrasound machine and the doctor introduced the women and built a relationship between them and the hospital several weeks prior to the delivery. In a transformation that was remarkable, institutional delivery rose to an average of 80 per month (from an average of 80 per year) within three months of the doctor and ultrasound machine being available at the hospital.
The Santhemarahalli visit was an incredible demonstration of how access to ultrasound could make a positive difference in maternal and fetal health.
A conventional ultrasound scan by an expert sonographer involves navigating the probe to the clinically valid 2D plane to acquire an image, correcting the image brightness and contrast for perceptual needs, manually identifying the region of interest on an image and using the mouse-like caliper to trace contours and make measurements. Automating this process through computer-assisted methods could revolutionize the reach of ultrasound to primary care physicians stationed in remote areas worldwide.
Obstetric ultrasound images vary greatly due to differences in the patients, fetal anatomy and size, the user and the scanner. While the fetal femur is a long bone that usually offers a good contrast, the fetal head is an elliptical bone that appears with discontinuous boundaries in ultrasound. The soft tissue in the fetal abdomen wall has a low contrast and an under-visualized boundary. Our team at Global Research in Bangalore has devised algorithms to withstand these challenges and measure each feature in about 3 seconds. Application specialists at GE Healthcare are currently evaluating the algorithms for possible integration into ultrasound systems.
The automated measurements promise a future in which ultrasound can be used to date pregnancy, monitor fetal growth, signal the need for early intervention for assisted delivery thereby help in reducing fetal distress and maternal death during childbirth in areas such as rural Karnataka, India.