280 likes | 285 Views
This article discusses the implementation of visual editing standards in a data management system to ensure high quality and accurate cancer data. It also explores the historical perspective of the standards and introduces a sampling plan for visual editing. Additionally, the article presents an audit module for recoding and provides insights into the issues and considerations associated with implementing these changes.
E N D
Is Your Data Management System Flexible for Quality Control Activities? Winny Roshala, CTR Data Standards and Quality Control Unit NAACCR: June 13-19, 2009, San Diego, CA
CCR Visual Editing Standards: Accuracy Rates • Implemented January 1, 2000 with 100% visual editing on 13 data items • Automated software was developed to calculate accuracy rates • Accuracy Rate Standard: 97%
CCR Visual Editing Standards: Purpose • Assure high quality data for analysis • Provide consistency in the visual editing process • Quantify the accuracy of cancer data from cancer reporting facilities • Standardize accuracy rates • Standardize format for reporting rates to registrars/facilities
CCR Visual Editing Discrepancies • Defined as the quality or state of being discrepant, i.e., disagreeing, being at variance • A discrepancy arises when a more appropriate code should have been selected for a data item based on submitted documentation
CCR Visual Editing Discrepancies • Discrepancies are counted prior to cases being linked or consolidated • Each data item is considered one potential discrepancy with the following exceptions: • Site/subsite • LN’s Pos/Examined • Site Specific Factor fields
CCR Visual Editing Standards: Calculation of Accuracy Rates • Percent Discrepant: Number of discrepancies divided by the number of abstracts, multiplied by the number of data items • Accuracy Rate: 100% less the percent discrepant
CCR Visual Editing Standards:Historical Perspective • In December 2005, in order to reduce a backlog, admissions from abstractors with an accuracy rate of 99% were no longer visually edited • This “push through” represented approximately 64% of admissions
CCR Visual Editing Standards:Historical Perspective • Due to budget cuts, this percentage was reduced again by adding abstractors with an accuracy rate of 98% to the admissions no longer visually edited
CCR Visual Editing Standards:Historical Perspective • In February 2008, due to a further reduction in state funding, the CCR changed it’s approach to reducing the proportion of cancer registry abstracts that are visually edited • Instead of focusing on individuals, the CCR went to a random sampling of cases for visual editing, reducing it from 100 percent to 40 percent
CCR Visual Editing Standards • Quality for the remaining 60% of abstracts will be monitored by targeted visual editing and through recoding and reabstracting audits • Hospital registrars continued to receive monthly Discrepancy Reports
Visually Edited Data Items • County of Residence at Diagnosis • Sex • Race • Spanish/Hispanic Origin • Date of Diagnosis • Diagnostic Confirmation • Site/Subsite* • Laterality (Only paired sites listed in Volume I) • Histology • Grade • CS Tumor Size • CS Extension • CS Lymph Nodes • Number of Regional Nodes Positive/Examined* • CS Metastasis at Diagnosis • CS Site Specific Factors 1-6* • Class of Case * Counted as one discrepancy
VISUAL EDITING SAMPLING PLAN • Run the edits against the admission • Set a flag indicating whether any edit errors exist • Check the admission as to whether it qualifies for required review and set a flag indicating true/false • Check the site to determine whether it is one of the sites that require 100% visual editing and set a flag indicating true/false
The following sites had a high discrepancy rate and will continue to undergo 100% visual editing: • Lip • Nasal Cavity & Middle Ear • Accessory Sinuses • Thymus • Heart, Mediastinum, and Pleura • Retroperitoneum and Peritoneum • Adrenal Glands • Other Endocrine Glands • Other Ill-Defined Sites • Unknown Primary Site
VISUAL EDITING SAMPLING PLAN • If the site is determined not to require 100% VE, using the system function Random, generate a number between 0 and 99 • Set a flag indicating VE Required to true if the number generated is 37 or less, else set the flag to false • Once all of the checks are complete, and if all flags are set to false, the admission will bypass Visual Editing
Issues to Consider • Percent of cases randomly selected for visual editing • List of sites which require 100% visual editing • New review tasks that need to be added to the database
Issues to Consider • Programming changes may require little time, however the deployment of programming changes may require a full build • Deployment may be delayed to comply with a scheduled release
What About the 60% of Cases Bypassing Visual Editing? Eureka Recoding Audit Module (RAM)
RAM Features • Ability to select data items and text fields • Accessibility • Audit current data • Automatically sends cases from the primary auditor to the secondary auditor • Generates reports • Multi-purpose tool
Summary • With diminishing resources, changes in the CCR visual editing practices were necessary • Due to the flexibility in our data management system, Eureka, we have the ability to quickly refine the sampling plan for visual editing as needed • This has resulted in our ability to redirect resources while carefully monitoring quality control on our data
Summary • The development of Eureka RAM has enabled us to focus on the cases bypassing the visual editing process • RAM is instrumental in quickly identifying problem areas in coding and instruction • Training efforts can be mobilized earlier • Future uses for RAM may include training of new staff, targeted visual editing and special studies
Acknowledgements • Nancy Schlag, CCR Operations Section Chief • Andrew Sutliff, Eureka Programmer Analyst • Kyle Ziegler, Audit Coordinator, Quality Control Specialist • Vic Belen, Administrative Assistant
Contact Information • Winny Roshala, BA, CTR • California Cancer Registry • 1825 Bell St., Suite 102 • Sacramento, CA 95825 • Phone: (916) 779-0313 • Email: wroshala@ccr.ca.gov