Can a smartphone application be used to identify newborns with neonatal jaundice?

In this preliminary study, a camera-based phone application effectively identified newborns with risky levels of serum bilirubin. We need more research to see if this will work in real-world situations. And, of course, we have to ask the most important question: Will payers include it in healthcare coverage plans? (LOE = 2b)

Taylor JA, Stout JW, de Greef L, et al. Use of a smartphone app to assess neonatal jaundice. Pediatrics 2017;140(3):e20170312.  [PMID:28842403]

Cohort (prospective)

Foundation

Other

These investigators enrolled 530 newborns younger than 7 days who were at least 35 weeks' gestation (20.8% African American, 26.3% Hispanic, and 21.2% Asian American). The newborns were recruited from 7 centers across the United States. None of the babies had undergone phototherapy though some of them may have been enrolled because they were clinically jaundiced. As a result, the authors obtained a wide range of serum bilirubin levels, from 2 mg/dL to 25 mg/dL (.0342 mmol/L to .427 mmol/L). At the same time as the blood draw the smartphone application was used to obtain a total of 6 images at 3 distances, using a special color calibration card placed on the child's sternum. All children had both serum testing and photos taken. The overall correlation between the smartphone application results and the serum bilirubin was 0.91 (95% CI .89 to .92) and was similar across all ethnic groups. The app had 100% sensitivity to identify a serum bilirubin of 17 mg/dL (.291 mmol/L) or higher (negative predictive value 100%) and had a sensitivity of 84.6% to identify a serum bilirubin level in the high-risk zone on the Bhutani nomogram (negative predictive value 97.2%). This was an efficacy study that evaluated the use of the application by trained investigators. We will need further research to determine whether clinicians or parents can effectively use it in the office or at home.