Using Audit Results to Improve Data Completeness and Quality

Missouri Cancer Registry

The first audit of the Missouri Cancer Registry (MCR) was conducted in 2002 on 1999 diagnosis year (DY) data by CDC's National Program of Cancer Registries (NPCR) and Macro International. The second audit was conducted in 2008 on DY 2005 data. Results from the 2008 audit and a follow-up assessment of data quality were presented during an educational session at the 2010 Missouri State Tumor Registrars Association (MoSTRA) annual meeting in September 2010.

High-quality data are needed for public health surveillance and other activities. Ensuring that a central cancer registry (CCR) meets national standards for timeliness, completeness, and quality with limited resources is an issue of concern to all NPCR-funded CCRs.

After the first audit, MCR focused quality assurance (QA) and training efforts on errors and missed cases found during the audit to illustrate to reporting facility staff why cases need to be reported in a timely manner, how to improve the quality of data submitted to the central CCRs, the importance of providing text to support codes, and where to look for cases that might be missed. MCR also used the audit process, lessons learned, and contacts made with other CCRs to streamline its audit program.

MCR's overall estimated case completeness for the second audit was 96.7% (NPCR standard is 95%) compared with 95.6% in the first audit. The improvement in case completeness and quality of core data elements is attributed largely to NPCR funding that supported new MCR staff positions, increased training for reporting facility staff, reporting facility audits, and expanded non-hospital reporting.

MCR reduced errors in most elements identified in the first audit as being problematic. MCR's overall data accuracy for DY 2005 was only 94.7%, slightly below the 95% norm, compared with 96.9% in the first audit. The vast majority of discrepancies in DY 2005 were items introduced after DY 1999, many having been introduced in DY 2004. The fact that DY 2005 was only the second year for using the Collaborative Staging System and Collaborative Stage (CS) Site-Specific Factors, more required data elements, and reporting facility staff turnover largely explain but do not solve the problem.

To assess current data quality, MCR QA staff selected several areas of concern from the results of the second audit, and plan to run reports using more recent data to identify where data quality has improved and areas that still need improvement. For example, of the 301 discrepancies identified in the audit, CS extension accounted for 45 (15.0%) while Histology, CS Lymph Nodes, DX Metastasis at Diagnosed, and Derived SS2000 accounted for 19 (6.3%) each. The primary sites with the most discrepancies were respiratory system, followed by breast, digestive system, and male genital system.

MCR QA staff will use the results of their analysis of 2007 and/or 2008 data to review problem sites and fields with hospital registrars and other attendees at MoSTRA's annual meeting. A report also will be included in the MCR newsletter that is distributed electronically to all Missouri hospitals and many other CCRs, and posted on the MCR Web site. Depending on the outcome, MCR staff will either continue to use relevant portions of the DY 2005 audit results for future training sessions, or will focus primarily on new areas such as CSv2.