Original articleSensitivity of self-reports of cancer in a population-based prospective study: JPHC Study Cohort I
Introduction
Many studies have relied upon self-reports of cancer, for example, to exclude cancer patients at the baseline in a cohort or in a case–control study, because there are usually no other methods to identify a person as being free of cancer. Thus, it is useful to determine whether self-reports of cancer are valid. There are two indicators commonly used for the validation. One is the percentage of actual cancer patients determinable by self-reports (i.e., sensitivity). The other is the percentage of self-reported cancer that is, in fact, cancer (i.e., the positive predictive value). The present study uses the first indicator.
The sensitivity of self-reports of cancer may differ with the given community or country 1, 2, 3, 4, 5, and by the time frame within a given community or country; the sensitivity was 0.33–0.61 in 1950s–70s 6, 7, 8, 9 and 0.71–0.93 in 1990s 1, 2, 3 in the US. In Japan, sensitivity is expected to increase, and telling a patient he or she has cancer used to be taboo, but is now under discussion 10, 11, 12.
The main purpose of the present study was to measure the sensitivity of self-reports of cancer in a population-based prospective study, which to our knowledge is for the first time outside North America or Europe. An additional aim was to analyze determinants that might affect the sensitivity; these include cancer site, gender, study area, smoking status, alcohol use, time between cancer diagnosis and the questionnaire, and years of formal education.
Section snippets
JPHC Study Cohort I
The present study was a part of the Japan Public Health Center-based prospective Study on cancer and cardiovascular diseases (the JPHC Study) Cohort I [13]. The subjects were drawn from the general public, either with or without disease, in four local areas: Ninohe Public Health Center (PHC) area in Iwate (Area A), Yokote PHC area in Akita (Area B), Saku PHC area in Nagano (Area C), and Ishikawa PHC area in Okinawa (Area D). All residents born from 1930 through 1949 were enrolled as of January
Results
Of 615 subjects, 221 self-reported their cancer in the questionnaire, for a sensitivity of 0.36 (95% CI 0.32–0.40) (Table 1). Sensitivity varied considerably by cancer site, with the values of 0.81, 0.42, 0.41, 0.26, 0.14 and 0.08 for breast, uterus, stomach, lung, colorectum and liver, respectively.
As to the cancer sites, sensitivity varied less by sex (Table 1), area (Table 2), smoking status, alcohol use, time between diagnosis and the questionnaire survey, and education (Table 3). The crude
Discussion
Sensitivity of self-reports of any cancer was 0.36 in the present study. This value is similar to the proportion of inpatients who were notified of their cancer in a hospital for the elderly [19]. This level is much lower than the 0.71 and 0.93 values from recent US studies 1, 2. Sensitivity was reported to be slightly lower in the age group of 70 years or older in a study [2]. However, this relationship did not explain the relatively lower sensitivity in the present study, because our
Acknowledgements
We express our appreciation to Mr. Michael Fahey, who critically reviewed interpretation and expression of the results; Ms. Yuko Iwatare, who was in charge of cancer registration; and other colleagues who have supported this project.
This work was supported by grants-in-aid for Cancer Research and for the Second-Term Comprehensive Ten-Year Strategy for Cancer Control from the Ministry of Heath and Welfare of Japan. Ako Yoshinaga is an Awardee of a Research Resident Fellowship from the Foundation
References (31)
- et al.
What do gastroenterologists in Europe tell cancer patients?
Lancet
(1993) - et al.
Accuracy of recall of hip fracture, heart attack, and cancera comparison of postal survey data and medical records
Am J Epidemiol
(1993) - et al.
Validity of self-reported cancers in a prospective cohort study in comparison with data from state cancer registries
Am J Epidemiol
(1998) - et al.
Comparison of active and cancer registry-based follow-up for breast cancer in a prospective cohort study
Am J Epidemiol
(1999) - et al.
Underreporting of cancer in medical surveysa source of systematic error in cancer research
Am J Epidemiol
(1976) - et al.
Validation of cancer prevalence data from a postal survey by comparison with cancer registry records
Am J Epidemiol
(1994) - et al.
Agreement between questionnaire data and medical records. The evidence for accuracy of recall
Am J Epidemiol
(1989) Measurement of prevalence of chronic disease by household interviews and clinical evaluations
Am J Public Health
(1957)- National Center for Health Statistics. Health interview responses compared with medical records. Vital and health...
- Madow MG. Net differences in interview data on chronic conditions and information derived from medical records. Vital...