Abstract: The invention refers to a system for delivering a machine-learning based behavioral intervention for the care of withdraw symptoms such as Neonatal Abstinence Syndrome in neonates. The system obtains biosensor or behavioral information about a patient from a wearable device on the patient and makes a determinative recommendation or takes appropriate action based on its evaluation. The biosensor and behavioral information is collected by way of a wearable device over progressive periods of time. When the data is indicative of a need for treatment because the patient is exhibiting symptoms or indicating relapse traits, this information is sent to the system for evaluation. When the data is indicative of a need for treatment, action is taken in the form of a recommendation or when the device is configured to support direct treatment, direct action can be taken to modify alleviate the patient.