BACKGROUND: Medical patients are often screened for distress in the clinic using a questionnaire such as the Hospital Anxiety and Depression Scale (HADS) while awaiting their consultation. However, might the context of the clinic artificially inflate the distress score ? To address this question we aimed to determine whether those who scored high on the HADS in the clinic remained high scorers when reassessed later at home. METHOD: We analysed data collected by a distress and depression screening service for cancer out-patients. All patients had completed the HADS in the clinic (on computer or on paper) prior to their consultation. For a period, patients with a high score (total of > or = 15) also completed the HADS again at home (over the telephone) 1 week later. We used these data to determine what proportion remained high scorers and the mean change in their scores. We estimated the effect of ‘ regression to the mean’ on the observed change. RESULTS: Of the 218 high scorers in the clinic, most [158 (72.5 %), 95% confidence interval (CI) 66.6–78.4] scored high at reassessment. The mean fall in the HADS total score was 1.74 (95% CI 1.09–2.39), much of which could be attributed to the estimated change over time (regression to the mean) rather than the context. CONCLUSIONS: Pre-consultation distress screening in clinic is widely used. Reassuringly, it only modestly overestimates distress measured later at home and consequently would result in a small proportion of unnecessary further assessments. We conclude it is a reasonable and convenient strategy.