The aim of this study was to examine the possibility of predicting anxiety and depression 6 months after a cancer diagnosis on the basis of measures of anxiety, depression, coping and subjective distress associated with the diagnosis and to explore the possibility of identifying individual patients with high levels of delayed anxiety and depression associated with the diagnosis. A consecutive series of 159 patients with gastrointestinal cancer were interviewed in connection with the diagnosis, 3 months (non-cured patients only) and 6 months later. The interviews utilized structured questionnaires assessing anxiety and depression [Hospital Anxiety and Depression (HAD) scale], coping [Mental Adjustment to Cancer (MAC) scale] and subjective distress [Impact of Event (IES) scale]. Patient anxiety and depression close to the diagnosis were found to explain approximately 35% of the variance in anxiety and depression that was found 6 months later. The addition of coping and subjective distress measures did little to improve that prediction. A model using (standardized) cut-off scores of moderate to high anxiety, depression (HAD) and intrusive thoughts (IES subscale) close to the diagnosis to identify patients at risk for delayed anxiety and depression achieved a sensitivity of 75% and a specificity of 98%. Levels of anxiety and depression at diagnosis predicted a similar status 6 months later. The results also indicated that the HAD scale in combination with the IES intrusion subscale may be used as a tool for detecting patients at risk of delayed anxiety and depression.