Table 6 Voice-to-Text Clinical Evaluation Scale (V2T-CES) Assessment Framework

From: Enhancing clinical documentation with voice processing and large language models: a study on the LAOS system

V2T-CES Assessment Items

Evaluation Criteria

Rating Scale

I. Accuracy Dimension

1. How well does the system recognize specialized ophthalmological terminology compared to general medical terms?

 

1

2

3

4

5

2. How effectively does the system automatically correct errors caused by accent variations or speech rate changes?

 

1

2

3

4

5

3. How accurately does the system transcribe speech in noisy clinical environments (operating rooms, patient areas)?

 

1

2

3

4

5

II. Efficiency Dimension

4. Does the system achieve at least 50% time savings compared to manual keyboard input for documentation? (≥ 50% time savings vs. manual input)

 

1

2

3

4

5

5. How much does the system reduce the need for manual confirmation and correction of transcribed text?

 

1

2

3

4

5

6. How well does the system automatically populate structured EMR fields from voice input?

 

1

2

3

4

5

III. System Performance

7. How effectively does the system use clinical context to correct transcription errors and improve accuracy?

 

1

2

3

4

5

8. How seamlessly does the system integrate with existing EMR workflows without disrupting clinical operations?

 

1

2

3

4

5

9. How well does the system support hands-free operation during patient examinations and procedures?

 

1

2

3

4

5

IV. Overall Assessment

10. Overall, how much does this system optimize your clinical documentation workflow compared to current methods?

 

1

2

3

4

5

  1. Scoring Metrics:
  2. • Accuracy Index = (Q1 + Q2 + Q3)/15 × 100.
  3. • Efficiency Index = (Q4 + Q5 + Q6)/15 × 100.
  4. • System Compatibility Index = (Q7 + Q8 + Q9)/15 × 100.
  5. Threshold Values: Individual ≥ 80: Clinically viable45; Composite ≥ 85: Replacement potential.