Step 1 of 8
Undiagnosed No More finds patients whose invisible illness is missing from the health record, for nine conditions from Long COVID, ME/CFS, and Lyme disease to lupus, endometriosis, and depression. This tour follows the tools and the evidence behind them.
In the NIH All of Us Research Program, only 24% of people in care for chronic fatigue have its code, against 85% for hypertension. All of Us records and surveys, with CDC, Census, and CMS data, show where nine invisible illnesses go uncoded.
Nine open models share one scorer. Each was fit outside the All of Us Workbench from released aggregate counts, and eight learn from patients' own reports of a diagnosis, so they find people whose diagnosis was never coded. The About page has the data, the models, and the four papers.
The clinician view opens inside the health record through SMART on FHIR. For each condition it shows the estimate, each dated finding with how much it moves it, the patient's answers, the criteria to check, and the next step.
Patients and advocates in the TOPx sessions described rebuilding their history for every clinician and wanting to approve what is shared. The app asks validated questions in English or Spanish, keeps answers on the device, and ends with a summary the patient prints or downloads for the visit.
The Atlas, for health systems and policy: pick a condition to see where and for whom it goes uncoded, what else is in the record, and what finding patients earlier costs. Its numbers are released aggregate estimates, not patient records.
The tools speak HL7 FHIR, SMART on FHIR, and CDS Hooks, and read USCDI data. A new condition is a new model file, not new software.
Transcript and captions.
Watch the walkthrough on YouTube (opens youtube.com).
The patient app opens on Maria's synthetic answers, and the clinician view starts on her chart.
Open the patient app Open the clinician view Watch the demo video
Every patient on this site is synthetic.
For clinicians
For policymakers
For patients
For developers