The correct diagnosis of brain disorders is tricky; error rates fluctuate from 18% to 70% among specialists. In a recent study funded by insurers, medical complications from the misdiagnosis of brain disorders were estimated to cost $150B per year in the U.S., alone. That means that 1% of U.S. GDP was spent on complications from botched neurological and psychiatric diagnoses.
Brain disorders are notorious shape-shifters and talented mimics. Like a virus, their symptoms morph over time, and they regularly appear disguised as entirely different disorders from what they actually are. To make matters worse, brain disorders cannot be definitively diagnosed until autopsy. So, there's zero quality control.
The common approach to identifying a brain disorder is through what's known as 'differential diagnosis'. This includes throwing all possible diagnoses that may produce a given patient's symptoms on a board and observing which one(s) holds up through the evaluation and testing process. Disorders are known to band together, ganging up to make detection of any one member all the more difficult, and the patient's life all the more miserable. Accurate differentiation of one disorder from another, especially in mild stages when symptoms are weak and vague, is next to impossible. This is what makes Miro Health's recently released study results so compelling.
Miro Health's first Concurrent Validity and Test Retest Reliability study has concluded with unexpected preliminary findings. The Concurrent Validity study pitted MiroMIND's assessments against the current standard of care to determine the likeness of their measurement properties. The Test Retest Reliability study followed MiroMIND's results over three time points: If a patient is assessed with Miro on a Monday, a Wednesday, and a Friday, how likely are the results to be stable? The results of both studies were impressive.
However, it's the unintended preliminary findings of disorder differentiation that is most promising. Miro's machine learning approach successfully clustered discriminating features of patients with Alzheimer's disease (memory impairment), aphasia (language impairment), Parkinson's (motor impairment), and frontal stroke (executive impairment). Not only do Miro's findings advocate for symptom-specific treatments, but more importantly, MiroMIND can successfully measure the effectiveness of those treatments.