Background: Idiopathic sudden sensorineural hearing loss (ISSNHL) is an emergency disease, and its pathogenesis is still largely unknown, making it difficult to diagnose and develop a therapeutic strategy. To predict the treatment outcomes and further customize the treatment strategy, we used a physician decision support system that incorporates complex information from electronic health records. We first developed the infrastructure of the physician decision support system, including an integrated data warehouse and an automatic data de-identification workflow.
Methods: We next conducted a cohort study to evaluate the treatment outcomes of 757 ISSNHL patients using the modified Siegel's criteria. The complete recovery (<25 dB) as a hearing outcome for ISSNHL patients was compared based on pretreatment hearing grades and disease onset with adjusted for age and sex after treatment initiation.
Results: The results showed that a complete recovery hearing outcome based on pretreatment hearing grades and disease onset in consideration of age and sex was associated with a low risk of pretreatment hearing Grade 2 (26-45 dB) (adjusted odds ratio 12.3, 95% confidence interval [CI]: 4.8-31.3) and disease onset ≤7 days (adjusted odds ratio 13.9, 95% CI: 4.2-45.8), respectively. Hearing recovery outcomes among complete recovery and noncomplete recovery (>25 dB) subjects according to pretreatment hearing grades were 32.9% (Grade 2 or 26-45 dB HL), 25.4% (Grade 3 or 46-75 dB HL), 31.1% (Grade 4 or 76-90 dB), and 4.5% (Grade 5, or >90 dB HL) (p < 0.0001). Patients with pretreatment hearing Grade 2 who received treatment within ≤7 days of disease onset had the highest rate of complete recovery (32.9%, 23/70).
Conclusion: In summary, using the physician decision support system, we successfully identified two predictors, the pretreatment hearing Grade 2 (26-45 dB) and treatment within ≤7 days of disease onset, associated with the highest odds of achieving complete recovery (<25 dB) of hearing in patients with ISSNHL.
Download full-text PDF |
Source |
---|---|
http://dx.doi.org/10.1097/JCMA.0000000000000450 | DOI Listing |
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!