We have developed a healthcare delivery system for implementation of targeted treatment for genetic diseases in critically ill infants that is in use in 70 children's hospitals in the US. We have developed a test ordering system that can be integrated into the electronic health record. We receive a copy of the patient's electronic health record and a 0.5 mL blood sample. The latter is shipped overnight to our institute. We use natural language processing to extract all of the clinical features of a child's illness from their medical record. We have developed methods for genome sequencing in which DNA libraries are made directly from blood. We have built an LIMS to integrate all of the components, and automated pipelines for all of the analytic steps in genome sequence interpretation. We use artificial intelligence to go from sequence data and phenotype list to genetic disease diagnosis in 10 minutes. We have integrated ten information resources about genetic diseases and used artificial intelligence to curate the literature for 475 genetic diseases that cause critical illness in infants and for which there are effective treatments. The test output is both the genetic disease diagnosis and initial management guidance for the disorder. It takes 13.5 hours from receipt of blood sample to diagnosis and integrated management guidance. Typical rapid diagnostic genome or exome sequencing takes a week, is not integrated to the medical record and does not extend to management guidance.
It is the fastest diagnostic genome sequencing ever. It will end the "diagnostic odyssey" for patients with rare genetic diseases. It has reduced turnaround time from days/weeks to 13.5 hours. This is timely even for the most critically ill children and prevents physicians from having to treat patients empirically based on clinical likelihood. Since there are thousands of genetic diseases, empiric treatments are often incorrect or suboptimal. It is fully automated which enables scalability to meet the needs of tens of thousands of critically ill children. Current methods of rapid whole genome or whole exome sequencing are very manpower intensive, particularly at the steps of genome sequence interpretation and diagnostic reporting. In testing several thousand children over the last several years we have found that physicians receiving genomic test reports frequently have never before treated a patient with many of the conditions that we diagnose. Hence the need for integrated management guidance for the subset of diseases that have specific, effective treatments. Without such integrated, virtual management guidance the test results may not result in optimal changes in management and improved outcomes. It is the first comprehensive genetic test to include acute management guidance.
Our target customers are newborns, infants and children in intensive care units with diseases of unknown etiology at time of admission. We estimate that there are about 80,000 per year in the US. Our end-users are pediatricians who work in intensive care units - neonatologists, intensivists, hospitalists, and subspecialist consulting physicians, such as medical geneticists and neurologists. The "buyers" are hospitals with neonatal, pediatric and cardiovascular intensive care units, who are paid by private health insurance, Medicaid or self-pay.