Testing for prostate and colorectal cancer: comparison of self-report and medical record audit

https://doi.org/10.1016/j.ypmed.2004.02.024Get rights and content

Abstract

Background. Self-reported data are often used to determine cancer screening test utilization, but self-report may be inaccurate.

Methods. We interviewed members of three health maintenance organizations and reviewed their medical records for information on digital rectal exam (DRE), prostate-specific antigen (PSA) test, fecal occult blood test (FOBT), sigmoidoscopy, and colonoscopy (response rate 65%). We calculated the sensitivity, specificity, concordance, and kappa statistic to compare the two sources for black men (n = 363), white and other men (n = 847), and women (n = 920) by study location.

Results. For DRE, FOBT, sigmoidoscopy, and colonoscopy, testing rates determined by self-report were higher than those in medical records. Kappa statistics showed fair to good agreement (0.40–0.80) for PSA, sigmoidoscopy, and colonoscopy among most subgroups. For DRE and FOBT, the agreement was poor except among participants from one HMO. Sensitivity was ≥80% for sigmoidoscopy among most subgroups, and ≥85% for endoscopy (sigmoidoscopy and colonoscopy), >75% for DRE, and ≥63% for PSA among all subgroups. Specificity exceeded 80% for FOBT and colonoscopy among all subgroups. Agreement was lower among older age groups. For all tests, agreement was poor between the reasons for testing.

Conclusion. Overreporting for some cancer tests should be considered when using self-reported data to evaluate progress towards reaching national goals for prevention behaviors.

Section snippets

Sample selection

Participants were randomly selected from member lists of Kaiser Permanente-Northern California (NC), Kaiser Permanente-Georgia (GA), and HealthPartners (HP), Minnesota. Men aged 45 years and older and women aged 55 years and older as of September 1, 1999 who had been enrolled in the plan for at least 5 years were eligible for selection. The study was approved by human subjects review boards at all participating sites and by the CDC.

We selected the study sample with stratification by sex and

Results

The demographic characteristics of the study populations (black men, white or other men, and women) are described in Table 2. In both groups of men, the largest proportion was aged 50 to 59 years. Most men were non-Hispanic, married, and employed; reported excellent or good health; and had more than a high school education. About 18% of black men and 10% of white or other men were current smokers. Most of the women were older than 60 years, non-Hispanic, and retired; reported excellent or good

Discussion

These results indicate that, for some cancer screening tests, agreement between information from self-reports and medical records is low. Tests more likely to have fair to good agreement between self-reported and medical record information included PSA, sigmoidoscopy, and colonoscopy. The sensitivities of self-reported data in our study groups are generally lower than those previously reported for colonoscopy [20]. The sensitivities for DRE [21], [23], [24], FOBT [20], [21], [22], [23], and

Acknowledgements

We thank Andy Nelson of HealthPartners for his support in planning and conducting this study.

References (29)

  • J.E. Oesterling et al.

    Serum prostate-specific antigen in a community-based population of healthy men: establishment of age-specific reference ranges

    J. Am. Med. Assoc.

    (1993)
  • P.H. Gann et al.

    A prospective evaluation of plasma prostate-specific antigen for detection of prostate cancer

    J. Am. Med. Assoc.

    (1995)
  • A.L. Potosky et al.

    The role of increasing detection in the rising incidence of prostate cancer

    J. Am. Med. Assoc.

    (1995)
  • S.J. Jacobsen et al.

    Incidence of prostate cancer diagnosis in the eras before and after serum prostate-specific antigen testing

    J. Am. Med. Assoc.

    (1995)
  • Cited by (103)

    • National Disparities in Colorectal Cancer Screening Among Obese Adults

      2017, American Journal of Preventive Medicine
      Citation Excerpt :

      First, the cross-sectional design prevents inferring a causal linkage between obesity and CRC screening utilization. Second, self-reported data may result in over-reporting of screening prevalence,46 inaccurate recall of physician recommendations, and BMI misclassification, though there is no reason to suspect differential reporting based on BMI that would account for the study findings. Third, the reasons for non-adherence investigated in this study were limited to the response options provided by NHIS; however, the results are consistent with the major barriers revealed in previous focus groups and qualitative work.39,40

    • The influence of physician recommendation on prostate-specific antigen screening

      2015, Urologic Oncology: Seminars and Original Investigations
    View all citing articles on Scopus
    View full text