What Researchers Did
Researchers checked how well hospital diagnosis codes for carbon monoxide poisoning matched what doctors found in medical records, using Taiwan's health insurance claims data.
What They Found
In this validation study of Taiwan claims data, diagnosis codes alone did not accurately identify carbon monoxide poisoning, but accuracy improved when hyperbaric oxygen therapy during hospitalization was added to the code. The team reviewed a random sample of 557 cases from 5,571 patients who newly had a carboxyhemoglobin blood test at four hospitals from 2011 to 2020. For ICD-9-CM code 986, sensitivity was 90.9% and positive predictive value was 72%. For ICD-10-CM code T58, sensitivity was 94.9% and positive predictive value was 76.6%. Positive predictive value was higher when poisoning was the primary diagnosis, and it could reach 100% when patients with either code also received hyperbaric oxygen therapy during hospitalization.
What This Means for Canadian Patients
This study is about the accuracy of billing and diagnosis codes, not about how well any treatment works. It does not report patient outcomes after hyperbaric oxygen therapy. It shows that when researchers use claims data to find carbon monoxide poisoning cases, receipt of hyperbaric oxygen therapy during hospitalization was one data point that made case identification more accurate.
Canadian Relevance
No direct Canadian connection identified.
Study Limitations
The study used a 10% sample from four hospitals in one country, and only patients who had a carboxyhemoglobin test were included, so the results may not apply to other settings or to poisoning cases that were never tested.