Using information technology to improve the quality and completeness of cancer incidence data

Background: All US states have laws requiring facilities to report new cancer cases to a central cancer registry (CCR). Among challenges facing CCRs are: obtaining accurate race/ethnicity information; capturing cases diagnosed in physician offices or small-caseload hospitals; and obtaining complete treatment information. Missing or inaccurate race/ethnicity data complicates the difficult task of reducing cancer disparities in minority populations. Underreporting of cases is largely due to lack of human and financial resources. Information technology (IT) offers the possibility of improving data quality and case completeness. Objectives: 1) Describe how the Missouri Cancer Registry (MCR) used IT to improve race/ethnicity data quality, and 2) obtain previously unreported cases and treatment information. Methods: We analyzed existing data to identify facilities with higher percentages of missing race/ethnicity data. We conducted linkages with other databases (hospital discharge, Medicare, Medicaid, etc.) to obtain missing information and identify incorrect information. We developed an educational program to train reporting facility staff. We targeted specific oncology practices and partnered with the Missouri Health Information Technology (MO HIT) Assistance Center to identify electronic health record (EHR) vendors and physician offices implementing approved vendor EHRs. Results: We obtained missing race/ethnicity data and corrected inaccurate data. We found that some facilities set a default value of unknown or non-Hispanic for ethnicity. We adapted and expanded an on-line training program to include educating vendors and facility IT staff as well as registry and admissions staff. We are bringing previously unreported cancer cases directly into MCR from physician office EHRs and a data repository. In late 2011 the project will be expanded to include acute care hospitals. Results: The percent of cases with unknown or incorrect race/ethnicity data is increasing while the number of previously unreported cases is decreasing. Conclusion: IT can improve data quality and completeness and benefit minority and underserved populations.