American Journal of Epidemiology Advance Access originally published online on August 2, 2005
American Journal of Epidemiology 2005 162(5):491-498; doi:10.1093/aje/kwi215
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
PRACTICE OF EPIDEMIOLOGY |
Validity of Self-Reports of Reasons for Hospitalization by Young Adults and Risk Factors for Discordance with Medical Records
The Coronary Artery Risk Development in Young Adults (CARDIA) Study
1 School of Medicine, University of Alabama at Birmingham, Birmingham, AL
2 School of Public Health, University of Alabama at Birmingham, Birmingham, AL
Correspondence to Dr. Jeffrey Roseman, Department of Epidemiology, School of Public Health, University of Alabama at Birmingham, 220 M Ryals Building, 1665 University Boulevard, Birmingham, AL 35294-0008 (e-mail: JRoseman{at}ms.soph.uab.edu).
This research focused on the validity of young adults' (mean age = 33 years; standard deviation, 3.9) self-reports of reasons for hospitalization and factors affecting validity in a longitudinal cohort study of over 5,000 young adults in four US cities (1985–1998). Self-reported reasons were considered discordant if they differed from those in medical records. Of the 321 self-reported hospitalizations, overall concordance was 92.5%; concordance ranged from 80% for infections to 100% for injuries/fractures and procedures/surgeries. There were no significant differences among mail, telephone, or face-to-face methods of collecting self-reports. In generalized estimating equations analyses, Black race (odds ratio = 4.23, 95% confidence interval: 1.72, 10.40; p = 0.002) and intravenous drug use (odds ratio = 6.06, 95% confidence interval: 1.17, 31.22; p = 0.03) were positively associated with discordance. Nonetheless, self-reports by Blacks were 90.0% concordant. Self-reports by Whites were 95.7% concordant. These results suggest that young adults' self-reported reasons for hospitalization are overwhelmingly concordant with medical records. This has important implications, since obtaining medical records has become more costly and logistically difficult.
cohort studies; data collection; hospitalization; longitudinal studies; medical records; reproducibility of results
Abbreviations: CARDIA, Coronary Artery Risk Development in Young Adults; CI, confidence interval; GEE, generalized estimating equations; ICD-9, International Classification of Diseases, Ninth Revision; OR, odds ratio