2021.01.12

Establishment of systematic test measurement method

Research and Development by

Yoshitaka KIMURA, Tohoku University Graduate School of Medicine

Corresponding Research Area

Effective combination of PCR, antibody tests, and other tests necessary for improving test efficiency and reliability


Reduction of test burden by detecting patients with a low probability of testing positive

  • Collect and analyze trial data by mid-December
  • Start triage by distributing free app to limited area and using AI interview at time of fever consultation

Test app and PCR positive probability estimation method at time of COVID-19 diagnosis using hierarchical Bayesian AI

Summary
①Optimize medical resources by selecting pseudo cases: Verify app that extracts patients with low probability of testing positive for PCR test with just an interview out of the pseudo cases By verifying the prototype of the app being created and efficiently excluding those with a low probability for testing positive from testing, the burden of the medical site will be reduced.
②Standardize triage: Through hierarchical learning, pseudo cases that vary by region and season are excluded from the testing to improve the efficiency of the medical resources. Standardize triage and reduce bias between doctors and medical institutions.

Current progress
With conditional approval from the Ethics Committee of the Tohoku University Graduate School of Medicine, we are discussing clinical research procedures (data collection using questionnaires) with cooperating research facilities.
1. Hospital survey: Requested research cooperation from hospital facilities within Okinawa Prefecture and started discussions on research. We are coordinating with the staff of the infectious disease internal medicine and emergency departments to the end of collecting data from medical records of those who have been medically examined for pseudo symptoms.
2. Survey at testing spots in Tokyo: To prepare for a prospective study of COVID-19 pseudo symptoms (symptomatic patients), we are discussing the details of the work procedure with the director and staff of the Tokyo Health Center. We are expecting this to start soon.
3. Sendai City Testing Spot: We are going to collect data based on the progress of the two sites above.

Use Data


Basic information: 1. Age, 2. Gender, 3. Respiratory history, 4. Diabetes, heart disease, hypertension, chronic kidney disease, history of liver cirrhosis, 5. Other history, 6. Presence of close contact, 7. Presence of influenza or cold cases among family and surrounding people
Symptom items: 8. Presence of fever, 9. Body temperature at first visit 10. Sore throat, 11. Cough, 12. Mucus, 13. Nasal discharge, 14. Joint pain, 15. Respiratory distress, 16. General malaise, 17. Taste disorder, 18. Large fever temperature change between morning and evening, 19. Symptoms and location of symptoms vary on daily basis.

Courtesy of Tohoku University