Objective: Published data on the observer variability with the transcranial doppler (TCD) ultrasound are limited by the use of the product moment correlation coefficient. This study was designed to quantify the intra and inter observer variability with the TCD in terms of the intra class correlation coefficient (ICC) and to assess the impact of lack of practice on the observer variability and the accuracy of data generated.
Methods: The study was performed in two phases. In phase I, three observers insonated the middle cerebral artery in 10 healthy volunteers and measured peak systolic and end diastolic cerebral blood flow velocities. In phase II, the same observers repeated the experiment on five healthy volunteers after an eight week break during which none of the observers were allowed to use the TCD system. The observers were blinded to the measurements obtained.
Results: The ICC for peak velocity measurement was 0.9 (0.81-0.99) in phase I and 0.85 (0.66-1.00) in phase II (p < 0.05). The ICCs for end diastolic velocity measurements for phase I and II were 0.8 (0.64-0.96) and 0.67 (0.33-1.00) respectively (p < 0.01).
Conclusions: A high level of observer agreement is possible with the TCD for measurement of cerebral blood flow velocities. Lack of regular practice with the system reduces the accuracy of measurements and impacts on both measured and calculated indices.