|
Ongoing CRGC Unknown Race Mitigation Project
Kyle L. Ziegler, BS, ODS
Director, Data Management
Cancer Registry of Greater California
Each year the Cancer Registry of Greater California (CRGC) works diligently to maintain an unknown race rate below the NAACCR standard of 3.0%. Over the past several years, meeting this benchmark has become more challenging. The CRGC has developed a four-pronged approach to achieve this metric. The four approaches utilized are: a MD office directed followback, review of modified records, a text-to-code verification in the statewide data base Eureka, and a hospital follow back on patients abstracted and submitted with an unknown race.
Ascertaining a <3.0% of unknown race on patients is becoming increasingly difficult to maintain. It is becoming more common for physician offices to not collect the patient's race; pathology reports often do not have a race, which is an important factor when many more cases are abstracted directly from the pathology report (Path Only Cases); and hospitals may inadvertently submit cases with unknown race coded but documentation within the text indicates a known race.
The CRGC developed several strategies targeting the “point of entry” of a patient’s race; that is the source where the patient encounter would result in the capture of the patient’s race. The first of these approaches is the MD office follow back. The MD office directed follow back consists of identifying patients reported by MD offices with an unknown race. With assistance of the CRGC Business Analyst, who identifies the patients by performing a SQL query on a copy of the CRGC database and creates MD office packets in PDF format, the Data Collection staff fax out the packets to the MD offices. When the packets are returned, the results of the follow back are evaluated and when a known race is provided, the patient’s race is updated in the statewide database. To assist physician offices in understanding why we need this information, the CRGC created a letter that is included in the physician office packet describing the importance of a patient’s race. Click here to see letter.
This special project is performed at minimum three (3) times per year. It is preferred to be conducted quarterly, four times per year; however, it may not be beneficial (number of cases needing updated), staff resources and submission deliverables also impede the quarterly routine.
The review of unprocessed information is another method we have developed in an effort to mitigate the increasing number of unknown race in the statewide database. This review starts with running a series of SQL scripts to compare unprocessed Admissions and Modified Records for known race codes where the associated Consolidated Patient in the statewide database has an unknown race. Up to four different lists are created and provided to the Visual Editing team in the Data Collection Unit. The Visual Editors update the patients race for the patients in the listing. This process occurs monthly.
The third strategy to mitigate unknown race is Data Miner Query performed in the statewide database and conducted by the Data Quality Control Unit. The query produces a spreadsheet that identifies patients in the database with an unknown race code in Race 1. The list also has each patient’s reported race (99) and associated text fields, specifically the history, final diagnosis, and remarks fields. Oncology Data Specialist (ODS) formerly Certified Tumor Registrars (CTR’s), are instructed to document the patients race in the text fields when initially abstracting the case. The text-based statement is usually “WM” for white male, “BF” for black female, or a statement such as “pt is Asian” or any variation thereof. By performing this review, the CRGC often finds cases in the database where the CTR did not code the patients race according to the text the abstractor provided.
The fourth strategy CRGC utilizes is a hospital follow back. We piloted this approach in the fall of 2023. This process includes creating a list of patients that were reported with an unknown race which is then sent back to the hospital and with a request to review the medical record and verify the patient’s race. Lists were sent to 72 reporting facilities on 789 patients, resulting in 456 (57%) Race updates from an Unknown to a Known Race value. This approach proved to be very beneficial.
RESULTS: Over the last year, the CRGC performed three separate MD office follow back to collect race information. The CRGC contacted 609 MD offices inquiring about the race of 3,397 patients. We were able to update 723 (21.3%) patients with a known race because of this effort. The hospital follow-back project resulted in 456 (57%) updates, while the remaining other two efforts resulted in the updating of approximately 100 patients. These efforts resulted in approximately 1300 updates to one of the most important data items registrars collect.
SUSTAINING SUCCESS: The CRGC uses every opportunity to update an unknown race to a known race. These documented strategies described here have yielded positive results to date and are now performed routinely by the CRGC. These strategies are time-consuming and require staff time to produce, perform and update patient data sets for this one single data item. Although it is an important data item for research and for proper classification, these projects are taxing on the already stretched resources. Sustained long-term success will require a more automated approach to these strategies and developing linkages at the central registry level to reduce the current manual work efforts.
| 
