Objective: No study on mutations in the K-ras oncogene and cancer of the exocrine pancreas or cancer of the biliary system has analyzed the reliability of clinical and epidemiological information.
Methods: Agreement between patient and surrogate on factors potentially related to both tumours was evaluated within a multicentre prospective study. Interviews were personally administered to both patient and surrogate (N = 110 pairs). Agreement was examined via the simple kappa index (k), the weighted kappa index (kw), the percentage of simple agreement, and the percentages of positive and negative agreement.
Results: Agreement for medical history was excellent (k between 0.89 and 0.76), as it was for tobacco consumption (k = 0.98). Agreement was moderate for coffee consumption (k = 0.68), frequencies of food groups (kw from 0.66 to 0.38), and consumption of alcoholic drinks (k from 0.66 to 0.32). Surrogates indicated a higher consumption of alcohol than patients.
Conclusion: Surrogates can be an alternative source of information when patients cannot be interviewed, but information on alcohol consumption should be treated with caution.