Identification of Jaundice in Newborns Using Smartphones

Neonatal jaundice is a common and most often harmless condition. However, when unrecognized it can be fatal or cause serious brain injury. Three quarters of these deaths are estimated to occur in the poorest regions of the world. The treatment of jaundice, phototherapy, is in most cases easy, low-cost and harmless. The crucial point in reducing the burden of disease is therefore to identify then children at risk. This results in the need for low-cost, reliable and easy-to-use diagnostic tools that can identify newborns with jaundice. Based on previous research on the bio-optics of jaundiced newborn skin, a prototype of a smartphone application was developed and tested in a pilot study and the application refined. This smartphone application will now be evaluated in a clinical trial set in two hospitals in Norway. The smartphone application gives immediate estimates of bilirubin values in newborns, and these estimates will be compared to the bilirubin levels measured in standard blood samples, as well as the results from ordinary transcutaneous measurement devices.

The smartphone jaundice app system works by taking pictures of newborns with a custom-made color calibration card placed on their chest. This makes it possible to measure the skin color precisely regardless of the specific light source that is used to illuminate the newborn. The measured skin color is then compared with items in a database of simulated newborn skin colors. These simulated newborn skin colors have been created using numerical simulations of how light moves through skin, with varying skin parameters including, but not limited to, skin thickness, blood concentration, melanin, and of course bilirubin, the pigment that causes jaundice. By comparing the measured skin color with the simulated skin colors that are most similar to it, the investigators can then estimate the bilirubin concentration in the newborn's skin by e.g. averaging the bilirubin concentrations used to create these simulated skin colors. In a group of 200 newborns with varying degree of jaundice, correlation between smartphone bilirubin estimates will be compared with total serum bilirubin and standard transcutaneous bilirubinometry.

Bilirubin estimates through color analysis of digital images obtained through smartphone application.

Sensitivity of the smartphone application to detect severe jaundice, defined as total serum bilirubin value > 250 umol/l.

Estimate of bilirubin levels from smartphone pictures, compared with bilirubin measurements in total serum. Correlation will be expressed by Pearson correlation coefficient, r.

Correlation between bilirubin estimates by smartphone pictures and by standard transcutaneous readings.

To evaluate the correlation between bilirubin estimates from a smartphone application and bilirubin estimates from a standard transcutaneous device in newborns with varying degree of jaundice. Correlation will be expressed by Pearson correlation coefficient, r.

- To evaluate the correlation between bilirubin estimates from a smartphone application and bilirubin estimates from visual inspection in newborns with varying degree of jaundice

Inclusion Criteria: Born in St.Olavs hospital, Trondheim or in Haugesund Hospital, Norway Born with gestational age >36+6 Birth weight ≥ 2500g and <4500g Age 1 - <15 days Are having a blood sample performed, as newborn screening or for jaundice assessment Exclusion criteria: Infants needing intensive treatment. This includes: Infants in the need for respiratory support Infants with conditions that could compromise skin circulation, as sepsis, heart failure or other Infants that have received phototherapy