Extent of thyroid surgery often miscoded in cancer registries

Efforts to improve the quality of health care critically depend on the accuracy of documentation that measure the effectiveness of quality improvement initiatives. Prior reviews of large databases have found substantial and pervasive deviations from standards of care in clinical practice.

Kiernan and colleagues hypothesized that some observed deviations were a result of documentation errors. As part of their study, researchers reviewed 918 thyroid cancer cases from the Tennessee Cancer Registry from 2004 to 2011. They found that of these cases, 369 were incorrectly coded, and most errors (85%) were attributable to facilities accredited by the American College of Surgeons’ Commission on Cancer (COC). This is an unexpected finding, since COC-accredited institutions are distinguished by their commitment to quality cancer care.

In reviewing the errors in documentation, the researchers found that, in fact, the correct procedures were done most of the time. Thus, there were errors in reporting, and not in practice. Facilities with COC accreditation were more likely to commit documentation errors that would reflect poorly on the quality of care delivered at their facilities. Although it is reassuring that documentation errors may be more common than deviations from the medical standard of care, this study highlights the need for quality assurance in data collection, as well as the substantial burden that such documentation requirements may represent for some institutions. We can wonder if billing records are subject to the same kinds of errors.

This study highlights important issues regarding the quality of data in hospital and state cancer registries, finding that in one state’s cancer registry, over 40% of thyroid cancer cases were incorrectly coded as thyroid lobectomy; in fact, 66% of those cases actually received a total thyroidectomy. The implication is that such coding errors may extend to other states and cancer diagnoses and undermine the interpretation of registry research studies.

Our previous publications on database and registry research on thyroid cancer summarize the advantages and disadvantages of various types of cancer registries used in the United States, and this study highlights one such disadvantage: quality control. Numerous studies have shown that cancer statistics and research conclusions can be greatly impacted by the quality of data used in registry research. In one of our publications, we present the advantages of a patient-centric, cloud-based disease management tool, the Thyroid Cancer Care Collaborative, integrated into physician workflow to improve the quality of thyroid cancer data used in research (Thyroid. 2014;doi:10.1089/thy.2013.0441 and Urken ML, et al. Head Neck. 2012;doi:10.1002/hed.23615).

While at first glance, this may seem like a study arguing for limited use of any registry data for large scale research, it is actually the contrary. Researchers must continue to use the wealth of information available in cancer registries (unmatched by any other disease process in the United States), while controlling for quality concerns in study design, and show cancer registrars that their daily work is useful and has meaning. Furthermore, hospitals should give tumor registries the resources needed to collect and report meaningful high-quality data — data that can be used by the hospitals themselves for internal quality control and process improvement. From 2000 to 2011, the number of new thyroid cancer cases per year reported to the National Cancer Database increased from 15,000 to over 35,000. Was there a related increase in resources devoted to tumor registries?