California Cancer Registry
Data Standards and Quality Control
Breast Recoding Audit
Final Report
June 2006
Introduction and Sampling Methodology
In July 2006, the Data Standards and Quality Control (DSQC) Unit of the California Cancer Registry (CCR)
initiated an audit of breast cancer cases. A total of 50 cases were randomly selected from each region,
including Region 4. Regions 1 and 8 and Regions 7 and 10 were combined and sampled as single regions,
respectively.
The criteria for cases sampled for this audit were as follows:
•
•
•
•
•
•
•
•

•

Diagnosed between July 1, 2004 and June 30, 2005
Female patients only
Class of Case 0, 1, and 2
Invasive and in situ malignancies
Diagnosed in California
Resident within reporting region only
Reporting facility within resident region
Type of reporting source 1 (hospital)
All ACTUR cases were removed

The audit included 41 data items as follows:
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•

•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•

County at DX
Sex
Race 1
Spanish/Hispanic Origin
Date of Diagnosis
Primary Site including Subsite
Laterality
Histologic Type ICD-O3
Behavior Code ICD-O3
Grade
Diagnostic Confirmation
CS Tumor Size
CS Extension
CS Lymph Nodes
CS Mets at DX
CS Site Specific Factor 1
CS Site Specific Factor 2
CS Site Specific Factor 3
CS Site Specific Factor 4
CS Site Specific Factor 5
CS Site Specific Factor 6

Regional Nodes Positive / Regional Nodes Examined
Date of Primary Surgery
Surgery of Primary Site - Summary
Scope Regional LN Surgery - Summary
Surgery Other Regional/Distant Sites - Summary
Date Radiation
Regional Radiation Modality
Radiation Boost Modality
Date Chemotherapy
Chemotherapy - Summary
Date Hormone
Hormone - Summary
Date Immunotherapy (BRM)
Immunotherapy (BRM) - Summary
Date Transplant Endocrine Therapy
Transplant/Endocrine - Summary
Date Other Treatment
Other Treatment - Summary
Sequence Number - Hospital
Class of Case

Note: The data items Primary Site including Subsite (CXX.X) and Regional Nodes Positive / Regional Nodes
Examined are counted as one discrepancy, respectively.
California Cancer Registry

Page 1

9/26/2008

 California Cancer Registry
Data Standards and Quality Control
Breast Recoding Audit
Final Report
June 2006
Audit Calculations
Fifty cases were randomly sampled for each region to ensure that there were 40 cases recoded for each region.
The audit results were calculated as follows: The number of cases recoded (40) was multiplied by the number
of data items (41), resulting in 1,640 possible discrepancies per region. This number was then multiplied by
the number of regions audited (seven), resulting in a total of 11,480 possible discrepancies. There were only
34 cases that could be recoded for Region 4 in the 50 cases sampled because 16 of the cases belonged to
another region. Using the above method (34 multiplied by 41), there were a total of 1,394 possible
discrepancies for Region 4. This number was then added to the 11,480 from the other seven regions, resulting
in a total of 12,874 possible discrepancies.

Audit Results
There were 160 discrepancies identified, which resulted in a discrepancy rate of 1.24%. Therefore, the
accuracy rate for this audit is 98.8%.

Discrepancies by Regional Registry
Region 3 and 7/10 had the least number of discrepancies with 12 (7.5%). Region 8 had the highest number,
with 43 (26.9%) discrepancies. The remaining regions varied with the number of discrepancies. Table 1
illustrates the distribution of the discrepancies across the regional registries.
Number of Discrepancies by Region
43

45

40

35

30
24
22

25

20
20
14
15

13

12

12

10

5

0
Region 2

Region 3

Region 4

Region 5

Region 6

Region 8

Region 9

Region 10

Table 1
California Cancer Registry

Page 2

9/26/2008

 California Cancer Registry
Data Standards and Quality Control
Breast Recoding Audit
Final Report
June 2006
Discrepancies by Data Item
Of the 160 discrepancies identified in this audit, the highest number of discrepancies occurred in Scope of
Regional Lymph Node Surgery. There were 18 discrepancies in this data item. The data item Surgery of the
Primary Site was the second highest with 15 discrepancies identified. CS Site Specific Factor 3 had the third
highest number with 13 discrepancies found. Data item discrepancies identified in this audit are illustrated on
Table 2.

Distribution of Discrepancies

20
18
16
14
12
10
8
6
4
2
Class of Case

Sequence Number – Hospital

RX Summ – Other

RX Date – Other

RX Summ – Transplt/Endocr

Date_Transp_Endo

RX Summ – BRM

RX Date – BRM

RX Summ – Hormone

RX Date – Hormone

RX Summ – Chemo

RX Date – Chemo

Rad – Boost RX Modality

Rad – Regional RX – Modality

RX Date – Radiation

RX Summ – Surg Oth Reg/Dis

RX Summ – Scope Reg LN Surgery

RX Summ – Surg Prim Site

RX Date – Surgery

Regional Nodes Positive / Exam

CS Site Specific Factor 6

CS Site Specific Factor 5

CS Site Specific Factor 4

CS Site Specific Factor 3

CS Site Specific Factor 2

CS Site Specific Factor 1

CS Mets at DX

CS Lymph Nodes

CS Extension

CS Tumor Size

Diagnostic Confirmation

Grade

Behavior Code ICD-03

Histologic Type ICD-03

Laterality

Primary Site incl Subsite (CXX.X)

Date of Diagnosis

Spanish/Hispanic Origin

Race 1

Sex

County at DX

0

Table 2
Scope of Regional Lymph Node Surgery discrepancies can be attributed to the timing of the surgical events
and whether or not the “sentinel” lymph node was involved. More discussion of these discrepancies can be
found in the Discrepancies in Treatment Fields section of this report.

California Cancer Registry

Page 3

9/26/2008

 California Cancer Registry
Data Standards and Quality Control
Breast Recoding Audit
Final Report
June 2006

Surgery of the Primary Site comprised 9.4% of all discrepancies. A detailed discussion of these discrepancies
can be found in the Discrepancies in Treatment Fields section of this report.
The discrepancies in Site Specific Factor #3, which had 13 (8.1%) discrepancies, were related to the number
of positive ipsilateral regional lymph nodes. A discussion of these discrepancies can be found in the
Discrepancies in Staging Fields section of this report.
There were no discrepancies identified in the data items County of Residence at Diagnosis, Sex, Behavior,
Diagnostic Confirmation, Surgery Other Regional/Distant Sites Summary, Date Immunotherapy (BRM),
Immunotherapy (BRM) Summary, Date Transplant/Endocrine, Transplant/Endocrine Summary, Date Other
Treatment, and Other Treatment Summary, and Class of Case.

Discrepancies in Staging Fields
There were 49 discrepancies in staging data items, representing 30.6% of all discrepancies identified as
follows:
•
•
•
•
•
•
•

•

CS Tumor Size -- 6 (12.2%)
CS Extension -- 1 (2.0%)
CS Lymph Nodes -- 9 (18.4%)
CS Mets at Diagnosis -- 2 (4.1%)
CS Site Specific Factor 3 -- 13 (26.5%)
CS Site Specific Factor 4 -- 8 (16.3%)
CS Site Specific Factor 5 -- 4 (8.2%)
Regional Lymph Nodes Positive/ Regional Lymph Nodes Examined -- 6 (12.2%)

These discrepancies are illustrated in Table 3.

California Cancer Registry

Page 4

9/26/2008

 California Cancer Registry
Data Standards and Quality Control
Breast Recoding Audit
Final Report
June 2006

Discrepancies in Staging Fields
14

13

12

10

9
8

8
6

6

6
4
4
2
2

1

0
CS Tumor
Size

CS
Extension

CS Lymph
Nodes

CS Mets at
Dx

CS SSF #3

CS SSF #4

CS SSF #5

Lymph
Nodes
Pos/Exam

Table 3

The highest number of discrepancies occurred in Site Specific Factor #3, which had 13 (26.5%) discrepancies.
The Site Specific Factor #3 field represents the number of positive ipsilateral axillary lymph nodes. The most
common problem identified was that registrars coded 000 when there were no lymph nodes removed. The
definition of code 000 is “All lymph nodes removed are negative.”1 The appropriate code is 098, which is
defined as “No lymph nodes removed or examined.”1 See page 463, CS Coding and Staging Manual, Note #3.
The CS Lymph Node field had the second highest number of discrepancies with nine (18.4%). The
discrepancies in this field were due to a misunderstanding of the differences in IHC staining versus H&E
staining and coding this field when the registrar has limited information.

1

Collaborative Staging Manual, Page 463
California Cancer Registry

Page 5

9/26/2008

 California Cancer Registry
Data Standards and Quality Control
Breast Recoding Audit
Final Report
June 2006
Discrepancies in Treatment Fields
Fifty nine discrepancies were identified in treatment fields, representing 36.9% of all discrepancies identified
as follows:
•
•
•
•
•
•
•
•
•
•

Date of Primary Surgery -- 4 (6.8%)
Surgery Primary Site Summary -- 15 (25.4%)
Scope Regional LN Surgery Summary --18 (30.5%)
Date Radiation -- 2 (3.4%)
Regional Radiation Modality -- 5 (8.5%)
Radiation Boost Modality -- 5 (8.5%)
Date Chemotherapy -- 1 (1.7%)
Chemotherapy Summary -- 1 (1.7%)
Date Hormone -- 1 (1.7%)
Hormone Summary -- 7 (11.9%)

These discrepancies are illustrated in Table 4.

Number of Discrepancies in Treatment Related Fields
20

18
15

15
10
7
5

5

4

5

2
1

0

1

1

0
Hormone Summary

Date Hormone

Chemotherapy
Summary

Date Chemotherapy

Radiation Boost
Modality

Regional Radiation
Modality

Date Radiation

Surgery Other
Regional/Distant Sites
Summary

Scope Regional LN
Sur Summary

Surgery Primary Site
Summary

Date Primary Surgery

Table 4

California Cancer Registry

Page 6

9/26/2008

 California Cancer Registry
Data Standards and Quality Control
Breast Recoding Audit
Final Report
June 2006
Scope of Regional Lymph Node Surgery is defined as “removal, biopsy, or aspiration of regional lymph
nodes that took place, either at the time of the primary surgery or during a separate surgical event.2
Miscounting the number of lymph nodes removed and whether or not one of those lymph nodes was or was
not the “sentinel” lymph node(s) were common issues identified.
The discrepancies identified in Surgery of the Primary Site were due to a code being used to indicate a
specific type of mastectomy without supporting text or s text indicating a specific type of surgical procedure
but coded to an “NOS” or nonspecific code.

Conclusions
The results of this audit were very good with an overall accuracy rate of 98.8%. Nevertheless, there were
problems identified that need to be addressed. The following recommendations are made:
1) Create training modules and/or conduct training for abstractors and visual editors on:
a. Surgery of the Primary Site
b. Scope of Regional Lymph Node Surgery
c. CS Lymph node fields:
i. CS Lymph Nodes
ii. CS Site Specific Factor #3
d. Tumor Size
e. Grade
f. Subsite
g. Ambiguous Terminology
2) Perform audit of High Quality Abstractors
3) Perform a recoding audit of specific data items to evaluate improvement:
a. CS Lymph Nodes
b. Site Specific Factor #3
c. Scope Regional Lymph Node Surgery
d. Surgery Primary Site
Note: High Quality Abstractors (HQA’s) were identified by the regions during the reconciliation process. All
calculations and graphs within this report include discrepancies made by HQA’s. Discrepancies made by
HQA’s are discussed separately in the accompanying report titled “Breast Recoding Audit Final Report
Supplement A – High Quality Abstractors.”
Auditors:
Cheryl M. Tatum, CTR
Winny Roshala, CTR

2

Dawn Leytem, CTR
Kyle L. Ziegler, CTR

Commission on Cancer, American College of Surgeons, FORDS manual, page 138-139
California Cancer Registry

Page 7

9/26/2008