Diagnostic tests are usually evaluated for their informational content and effect on health outcomes. Yet patient attitudes toward taking risks and the morbid sequelae associated with the diagnostic test or its consequences are seldom considered. A prototypical model is presented which incorporates patient attitudes into the evaluative process. Based on traditional indices of test efficacy, preoperative searches for occult metastases in patients with presumably operable bronchogenic carcinoma should never be performed. If patient attitudes toward perioperative death are considered, many should have preoperative staging tests. The exact percentage of patients benefiting from testing varies with test sensitivity and specificity. This integrative approach is appropriate for oncologic patients who frequently undergo treatments with high morbidity and mortality rates.