Patient wellness and preventative care are increasingly becoming a concern for many patients, employers, and healthcare professionals. The federal government has increased spending for wellness alongside new legislation which gives employers and insurance providers some new tools for encouraging preventative care. Not all preventative care and wellness programs have a net positive savings however. Our research attempts to create a patient wellness score which integrates many lifestyle components and a holistic patient prospective. Using a large comprehensive survey conducted by the Centers for Disease Control and Prevention, models are built combining both medical professional input and machine learning algorithms.

Models are compared and 8 out of 9 models are shown to have a statistically significant (p = 0.05) increase in area under the receiver operating characteristic when using the hybrid approach when compared to expert-only models. Models are then aggregated and linearly transformed for patient-friendly output. The resulting predictive models provide patients and healthcare providers a comprehensive numerical assessment of a patient’s health, which may be used to track patient wellness so at to help maintain or improve their current condition.