The cancode interaction analysis system in the oncological setting: reliability and validity of video and audio tape coding

Abstract

Cancode is a computerized interaction analysis system developed for cancer consultations. This paper assesses its reliability and validity, and compares the use of audio versus video tape; by assessing 30 consultations between an actor and 10 oncologists. Weighted Kappa inter- and intra-rater scores ranged from 0.5 to 1.0 and 0.58–1.0, respectively, and use of video tape did not alter verbal coding. Factor analysis of verbal codes revealed two factors, 'verbal control' and 'verbal support'. Verbal and non-verbal doctor behavior differed by patient type ('verbal support' P=0.007, 'verbal control' P=0.004, 'Responsiveness' P=0.000, and 'Immediacy' P=0.000). Inter-doctor variation was noted for 'verbal support' (P=0.000) and 'Relaxation' (P=0.000). 'Responsiveness' was negatively correlated with 'verbal support' (−0.58) and 'verbal control' (−0.65). Cancode is reliable, valid and sensitive to doctors behavioral changes. For a more passive patient, the doctor may switch from a 'cure' to 'care' oriented consult, responding to psycho-social instead of informational needs.