Automated methods indicate specialty disparities in follow-up time for cancer screening tests within one system

Abstract

IntroductionA study utilizing automated algorithms was used determine if abnormal results for cancer screening tests could be identified and if appropriate follow-up took place. We examined our system in order to determine whether follow-up varied by sociodemographic characteristics and/or clinical specialty. MethodsA retrospective, observational study was conducted to assess the time to follow-up for each abnormal cancer screening test result at Brigham and Women's Hospital, a tertiary academic medical center, from June 30, 2008 to June 29, 2010. Time to follow-up was evaluated using cumulative distribution function curves for mammograms, Pap smears, and prostate-specific antigen (PSA) tests. ResultsOut of 110,000 total cancer screening tests, we identified 4,378 abnormal mammograms, 1,066 abnormal pap smears, and 1,114 abnormal PSA tests. Mammograms were followed-up at the fastest speed, followed by PSA tests, then Pap smears. Mammograms ordered by OB/GYN (HR=1.154, p<0.001) or Hematology/Oncology (HR=1.400, p<0.001) were followed-up faster than those ordered by General Internal Medicine. Pap smears ordered by OB/GYN were followed-up slower compared to General Internal Medicine (HR=0.685, p=0.001). PSA test orders by Hematology/Oncology were followed-up at twice the speed of tests ordered by General Internal Medicine (HR=2.213, p=0.003); tests ordered by other specialties were followed-up at half the speed of General Internal Medicine (HR=0.465, p<0.001). ConclusionsThe disparities in time to follow-up for cancer screening tests within one health system indicate that best practices are not employed uniformly. Health IT methods may be a viable option to help bridge this gap.