OBJECTIVE: When using non-patients as respondents in health state preference measurement, it is necessary to develop accurate descriptions of treatment alternatives and health states patients may be likely to experience. We developed descriptions of commonly used treatment modalities for colorectal cancer (CRC) and ensuing health states for use in subsequent preference measurement studies. METHODS: Following a literature review and clinical expert input, we identified four commonly used treatment modalities for CRC. Clinical data for each of the modalities were abstracted from the literature, and reviewed by oncologists to confirm if the literature descriptions of the modalities were consistent with experience in clinical practice. Draft descriptions were then prepared, and critically reviewed by patients from each of the treatment modalities for not only accuracy and relevance but also comprehensibility. Patients were also specifically instructed in these qualitative interviews to describe the impact of their treatment modalities on their everyday lives, to ensure that the ensuing descriptions captured relevant patient outcomes and were phrased in language used by patients. Finally, interviews were carried out with lay persons to review the descriptions for clarity and comprehensibility. RESULTS: The three commonly used treatment modalities for CRC were Modified de Gramont, Mayo, and Xeloda. It was decided to describe each modality in terms of efficacy, adverse events, administration procedures, and patient outcomes. For ease of comprehension, patients and lay persons suggested presenting the information in separate sections. Subjects were satisfied that the ensuing descriptions accurately captured the relevant issues and were comprehensible. CONCLUSION: We have successfully prepared accurate and comprehensible descriptions of treatment modalities and ensuing health states in CRC. These descriptions may be useful in health state preference measurement studies using non-patients as respondents.