Subject: Data Standards

Missouri Cancer Registry

Policies and Procedures – Data Quality

Subject: Data Standards
Function: Data Quality / Revised
11-06-06

OBJECTIVE: To ensure that MCR data meets required data standards set forth by NPCR, NAACCR and other relevant standard setters

DEFINITIONS:

POLICY: MCR QA staff and Project Manager will develop procedures to ensure that data meets all required standards.

NPCR STANDARDS: Within 12 months of the close of the diagnosis year, 97% of the cases will pass an NPCR-prescribed set of standard edits.

Within 24 months of the close of the diagnosis year, the data should meet the following NPCR criteria:

Data are 95% complete based on observed-to-expected cases.
There is a 1 per 1,000 or fewer unresolved duplicate rate.
There are 3% or fewer death-certificate-only cases.
The percent missing for critical data elements are:
2% or fewer age
2% or fewer sex
2% or fewer race
2% or fewer county
97% pass an NPCR-prescribed set of standard edits.

The central cancer registry will have an overall program of quality assurance that is defined in the registry operations policy and procedure manual. The quality assurance program will consist of, but is not limited to:

A designated certified tumor registrar (CTR) or CTR-eligible person will be responsible for the quality assurance program.
Qualified, experienced CTR(s) or CTR-eligible personnel conduct quality assurance activities.
At least once every 5 years, case-finding and re-abstracting audits from source documents will be conducted at each reporting facility.
Data consolidation procedures will be performed according to an accepted protocol.
Procedures will be documented for follow-back to reporting facilities on quality issues. These procedures will include rules for identifying when action or further investigation is needed.
Detailed documentation of all quality assurance procedures will be maintained.
The central cancer registry will have a designated education/training coordinator who provides training to the central cancer registry staff and reporting sources to assure high quality data.

Subject: Quality Assurance Plan
Function: Data Quality / Revised 1/06;
10/31/06

OBJECTIVE: To have complete, high quality data.

DEFINITIONS:

POLICY: MCR staff will perform established quality assurance/control procedures regularly.

PROCEDURE:

QA staff will consist of one or more CTRs.
The Quality Assurance Coordinator will be a CTR with 3 – 5 years of experience.
QA staff will maintain documentation for the “MCR Abstract Code Manual” which is updated and published annually on the website.
QA staff will utilize a computerized Missouri-specific edit set to enable them to focus on activities requiring human judgment, analysis or interaction.
MCR’s central registry software will maintain data integrity as a function of data quality.
The central registry software will provide process controls.
QA staff will provide training to central registry staff to ensure data collection is accurate, consistent and complete.
Describe activities for monitoring quality
Checks on case ascertainment
Checks on data quality
The following quality reviews will be performed on the completed data set prior to sending to Research Analyst for analysis:
Run program to verify sex based on common name table
Review sex data item for unknown or invalid codes
Review race data items for unknown or invalid codes
Review age at diagnosis date item for unknown and ages over 110
Review blank Tobacco Years. If Tobacco History = 0 then Tobacco Years should = 00, else Tobacco Years should = 99.
Review breast cases for sex=male. Review based upon first names and contract hospitals regarding questionable cases.
Review ages <15 for all published sites.
Other finalization steps:

Load Death Clearance cases.
Run database against itself in AutoMatch to check for duplicates

Run duplicate protocol and resolve possible duplicates.
Run GenEdits and resolve errors.
Run Interrecord edits and resolve errors.
Review final file for required fields.

Quality control procedures will be documented.
QA staff will document rules and guidelines used to reconcile edits and/or discrepancies (SEER for multiple primaries, FORDS coding rules, etc.)

REFERENCES:

Other Policies?

NPCR Program Standards?

NAACCR Standards for Cancer Registries Volume III, 3.1., Data Quality

Approved:

______

Subject: Initial Abstract Processing
Function: Data Quality / Revised
7-27-04
01-31-06

OBJECTIVE: To ensure all cases are properly processed and reviewed for quality assurance when needed.

DEFINITIONS:

POLICY: Registry staff will follow the data receipt procedures policy and perform routine QA functions on incoming data.

PROCEDURE:

Data Receipt policy will be followed for initial handling of data.
QA staff will review each case placed in a “pending” file by the central registry software. Reasons for case going to Pending:
Failure to pass edits
Need for consolidation review
For each case in the “pending” file:
Demographics will be reviewed to determine if all fields are complete.
Personal History will be reviewed for cases with sequence numbers 01-59 and 61-88 for documentation of other cancers.
Class of case, diagnosis date and dates of first contact will be reviewed for accuracy and relationship
Site and histology codes will be compared with text documentation to verify accuracy.
Collaborative stage data items will be compared with text to verify accuracy.
Laterality and diagnostic confirmation fields will be compared with text to verify accuracy.
Treatment codes will be compared with text to verify accuracy and to ensure that coded treatment is first course and not subsequent.
All MCR required fields will be reviewed for completeness (see Required Data Items List).
Vital/cancer status and any death information if the patient has expired.
All death codes should be in ICD-10 codes.
Minor discrepancies will be corrected based on text documentation.

QA feedback will be given to reporting facilities for each major discrepancy. List of major and minor discrepancies is being developed (10/31/06).

An email will be sent to the reporting facility to confirm that the data have been reviewed.

REFERENCES:

MCR Data Receipt Policy

NPCR Program Standards?

Approved:

______

Subject: Computer Edits
Function: Data Quality / Revised
1-31-06; 11-06-06

OBJECTIVE: To develop and maintain a Missouri-specific computerized edits for QA review.

DEFINITIONS:

Edits: Data edits are logical rules, typically embodied in a computer algorithm, that evaluate to “true,” “false,” or “maybe,” for any value(s) of (a) data item(s). Central registry edits are applied to all records to check for item validity, internal consistency, and inter-record consistency. Data edits may involve a single field, multiple fields in a single record, multiple fields in different records within one database, or multiple fields in multiple databases. (NAACCR Volume III)

POLICY: MCR will use the central registry software NPCR edit set until a customized Missouri edit set is finalized.

PROCEDURE:

The applicable edit metafile will be chosen when loading cases into Previewer.
The applicable metafile will also be run during import and as needed during the QA process.
Edit reports will be forwarded to the reporting facility whenever necessary.
MCR will utilize the designated GenEdits and Interrecord edits prior to any Call for Data.
Information from edit procedures will be analyzed as time allows,identifying areas for improvement (i.e., data sources, coder, or clarity of instructions in the manual).
MCR staff will create and maintain a customized version to be utilized with the central registry software, other abstracting software and eventually for hospitals to use prior to submitting data to MCR.

REFERENCES:

Other Policies?

NPCR Program Standards?

NAACCR Standards for Cancer Registries, Volume III, 3.1.4., Edits and Data Processing Capabilities for Data Quality

NAACCR Standards for Cancer Registries, Volume III, 3.2.3., Standards for Data Edits

NAACCR Standards for Cancer Registries, Volume III, 5.8, Edits

Approved:

______

Subject: Visual Editing
Function: Data Quality / Revised
11-06-06

OBJECTIVE: To ensure complete, high quality data.

DEFINITIONS:

POLICY: MCR will follow specific, standardized visual editing procedures.

PROCEDURE:

These procedures are in addition to policies outlined in “Initial Abstract Processing.”

Selection criteria for cases reviewed may be based on:

Edit checks
Rule and guideline changes
Reporter experience
Previous reporting history

Some edits will be resolved by automatic correction rules based on software settings (replace unknown or blank codes, etc.)
Some edits will require notification of the reporting facility.
For conflicting information, QA staff will contact reporting facilities by telephone.
Error reports are generated upon import????

Details about edit reports may be found in the “initial abstract processing” policy.

REFERENCES:

Initial Abstracting Processing Policy

NPCR Program Standards?

NAACCR Standards for Cancer Registries, Volume III, Section II.B.3: Quality Control Activities

Approved:

______

Subject: Guidelines for Consolidation
Function: Data Quality / Revised
11-06-06

OBJECTIVE: To provide guidelines in the resolution of discrepancies during the consolidation process.

DEFINITIONS:

Consolidation: the process of reconciling or compiling data obtained from more than one source on the same person or tumor.

POLICY: The MCR data management system performs automatic consolidation whenever possible. When the system cannot determine a correct value, QA staff will follow standardized process when resolving errors/discrepancies. Technical aspects of the consolidation process are outlined in the policy: Internal Matching and Case Consolidation.

PROCEDURE:

QA staff will check information provided in text fields to resolve errors.
If there is no text, or if the text information does not resolve the error, QA staff will contact both registrars for further information to resolve the problem.

REFERENCES:

MCR: Internal Matching and Case Consolidation

NPCR Program Standards?

NAACCR Standards for Cancer Registries, Volume III, 5.9., Record Consolidation

Approved:

______

Subject: Re-abstracting Audits
Function: Data Quality / Revised
02-10-06

OBJECTIVE: To audit the quality of data being submitted by hospitals.

POLICY: MCR will perform 6-8 re-abstracting audits annually.

PROCEDURE:

Facility Selection- Priority is given to those with specific concerns identified by quality assurance staff, followed by random sampling. The primary focus will be on facilities that:

Have consistent excessive errors on required fields from recent submissions.
Have new registrars or frequent registrar turnover

Data Selection - Review and re-abstraction of ten percent of the hospital caseload for facilities with two hundred or fewer cases will be performed. A maximum of twenty cases will be reviewed and re-abstracted for all other facilities. Patients will be randomly selected from the audited year or from recent transmissions if there has been registrar turnover. Cases from the top four primary sites or other specific primary sites will be reviewed. Eighteen required data items will be re-abstracted.

Accuracy Rate - The acceptable accuracy rate for re-abstracting audits is 95 – 100%.

Resolution - After re-abstraction has been completed, a summary and resolution forms identifying any abstracting discrepancies are prepared and returned to the registrar(s) for review. A telephone conference is held and the discrepancies are collaboratively resolved, with discussion and reinforcement of MCR coding rules and guidelines. Opportunity is given for registrars to provide more information or additional documentation that may support the original coding.

Final Letter/Summary (post-site visit)

Following resolution of discrepancies, a final letter and audit packet with detailed reports and statistical summary is sent to the hospital.
Additional training may be scheduled, if warranted.

6. Abstracts are revised in the MCR database for cases found to have discrepancies as a result of the audit.

REFERENCES:

MCR: Facility Audits, Data Processing Operations

NPCR Program Standards?

NAACCR Standards for Cancer Registries Volume III, 2.3.5, Case finding Audit Results

NAACCR Standards for Cancer Registries Volume III, 3.3.1, Reabstracting and Recoding Audits

Approved:

______

Subject: Management Reports
Function: / Revised
11-06-06

OBJECTIVE: To monitor the registry database and operations of the registry through the use of management reports.

DEFINITIONS:

POLICY: MCR staff will follow established guidelines for data submission to NAACCR and NPCR.

The central cancer registry annually submits a data file to the NPCR-Cancer Surveillance System (CSS) that meets the reporting requirements outlined in the NPCR-CSS Submission Specifications document and meets criteria for publication in United States Cancer Statistics.

PROCEDURE:

REFERENCES:

Policy

NPCR Program Standards II-B

NPCR Program Standards II

NPCR Program Standards VIII

Approved:

______

Building NameMissouri Cancer RegistrySection VII Date Revised: 11/2006