ANALYSIS OF PRIMARY-CARE STAFF LANGUAGE ABOUT AGGRESSION AT WORK USING CONCORDANCING TECHNIQUES TO IDENTIFY THEMES

Objective. We aimed to describe features of language used during inter views about the extent of aggression and violence at work and their ef fect on primary care staff. Method. Forty-four primary health care tea m members in the West Midlands were interviewed, and interviews were r ecorded on videotape. The language content of these interviews was ana lysed using Cobuild concordancing software. Outcome measures used were word frequency, collocation and mutual information (MI) scores for la nguage use. Results. A total of 17 517 words spoken by interviewees we re analysed. Violence in this sample was perceived as occurring princi pally in connection with unmet demands for such things as prescription s and referrals. Only patients were perceived as violent: health care workers used other terms to describe their own feelings and responses. Sixty-eight specific incidents of violence were recounted, features p erceived as salient being drink, youth and to a lesser extent mental i llness. Conclusions. Concordancing software can be successfully used i n the qualitative examination of videotaped interviews. In this study, the technique rapidly identified a number of perceived training needs among a variety of primary care staff.