The reliability of telephone interviews compared with in-person interviews using memory aids

PURPOSE: Information obtained by telephone interviews and in-person interviews is generally considered comparable, but it is unclear if extensive memory aids can be used effectively with telephone interviews. We compared a telephone interview to an in-person interview using the same questionnaire and memory aids in both. METHODS: A convenience sample of 103 women, aged 25 to 69 years, completed a telephone interview and at least four weeks later, completed an in-person interview. Memory aids included a life events calendar, cue lists, and worksheets. RESULTS: Agreement values, measured by kappa/weighted kappa, were as follows: parity (1.00), age at menarche (0.76), menopausal status (0.95), a history of reproductive organ surgery (0.98) or tubal ligation (0.91), self-reported infertility (0.76), and a first degree family history of breast/ovarian cancer (0.90). Agreement values for duration variables, measured by the intraclass correlation, were as follows: lactation (0.96), oral contraceptive use (0.98), any hormone replacement therapy (0.98), exclusive estrogen and progesterone therapy (0.83), and exclusive estrogen therapy (0.99). CONCLUSIONS: The good to excellent level of agreement found in this study indicates that telephone administration of our questionnaire with extensive memory aids is a reliable method of obtaining detailed exposure information relative to in-person interviews. Ann Epidemiol 2003; 13:495-501. (C) 2003 Elsevier Inc. All rights reserved.