California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

The Data Standards and Quality Control Unit (DSQC) of the California Cancer Registry (CCR)
initiated a recoding audit of prostate cancer cases in an effort to determine data quality among
these cases.

Audit Sampling and Performance Methodology
All regions were involved in this audit. Due to limited available resources, the sampling
reflected the three California SEER registries. Therefore, the regions that make up the Cancer
Registry of Greater California (CRGC) were audited as one region. Regions 2, 3, 4, 5, 6, and 10
are considered the CRGC, and each region in this group had 10 cases sampled. Therefore, there
were a total of 60 cases sampled for CRGC. Region 8 and Region 9 each had 60 cases sampled,
for a total of 180 cases audited. The audit was based on cases diagnosed in 2011.
Each case was recoded based on the text documentation submitted on the abstract. The original
code must match the recode by the auditor. A discrepancy resulted if these codes did not match.
The audit was performed utilizing the peer review method. Each case was audited blindly and
independently by two separate auditors. When both the primary and secondary auditors
completed their assigned region, the primary auditor then compared the discrepancies between
the auditors. The two auditors then reconciled their differences and discussed any issues
discovered during the audit process.

Audit Findings
The auditors were instructed to apply CS Version 02.03, for this audit since the audit consisted of
cases diagnosed in 2011.
There were 186 discrepancies noted on this audit. The distribution of the top nine data items
with discrepancies were:
•
•
•
•
•
•
•

CS Extension
CS Site Specific Factor 13 (Number of Cores Examined)
CS Mets at DX (Bone, Brain, Liver, Lung)
CS Site Specific Factor 1
CS Site Specific Factor 11
CS Site Specific Factor 12
CS Site Specific Factor 3

32 (17.2%)
21 (11.2%)
20 (10.8%)
17 (9.1%)
17 (9.1%)
14 (7.5%)
12 (6.5%)

A detailed distribution of all discrepancies among regions and data items can be found in Table
1.
June 15, 2012

1

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

Detailed Discrepancy Analysis
CS Extension
The data item CS Extension- Clinical Extension had the highest number of discrepancies
identified on this audit. There were 32 (17.2%) discrepancies identified.
•

20 (62.5%) were recoded to code 150 (Tumor identified by needle biopsy (clinically
inapparent) Stated as cT1b with no other information on clinical extension)
o 8 recoded to code 150 from code 210
o 4 recoded to code 150 from code 230
o 3 recoded to code 150 from code 220
o 3 recoded to code 150 from code 300
o 1 recoded to code 150 from code 410
o 1 recoded to code 150 from code 999

•

4 (12.5%) were recoded to code 240 (Clinically apparent tumor confined to prostate,
NOS. Stated as cT2 [NOS] with no other information on clinical extension.)
o 1 recoded to code 240 from code 150
o 1 recoded to code 240 from code 200
o 1 recoded to code 240 from code 220
o 1 recoded to code 240 from code 490

•

4 (12.5%) were recoded to code 300 (Localized, NOS. Confined to prostate, NOS.
Intracapsular involvement only. Not stated if T1 or T2, clinically apparent or inapparent)
o 1 recoded to code 300 from code 150
o 1 recoded to code 300 from code 200
o 1 recoded to code 300 from code 230
o 1 recoded to code 300 from code 440

•

3 (9.4%) were recoded to code 200 (Involvement in one lobe/side, NOS. Clinically
apparent on physical exam or imaging, do NOT use information from biopsy to
determine extent of involvement.(See Note 3c))
o 1 recoded to code 200 from code 220
o 1 recoded to code 200 from code 230
o 1 recoded to code 200 from code 300

•

1 (3.1%) were recoded to code 230 (Involves both lobes/sides. Clinically apparent on
physical exam or imaging, do NOT use information from biopsy to determine extent of
involvement. (See Note 3c) Stated as cT2b with no other information on clinical
extension)
o 1 recoded to code 230 from code 210

June 15, 2012

2

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

There were 18 (56.3%) discrepancies in the CS Extension – Clinical Extension recoded from
either code 200, 210, 220 or 230. All of these codes have a specific instruction listed in the
definition of the code and in the data item specific notes, listed at the top of the page of the CS
Stage Manual for this data item. Note 3c for this data item states “Codes 200 to 240 are used
only for clinically/radiographically apparent tumor/nodule/mass which is palpable or visible by
imaging. To decide among codes 200-240, use only physical exam or imaging information, and
not biopsy information.” In each of the 18 discrepancies, the abstractor used information from
the biopsy to code this field.
A related data item is CS Site Specific Factor #3 (CS Extension – Pathologic Extension)
which had 12 (6.4%) discrepancies.
•

4 (33.3%) were recoded to 970 (No prostatectomy done within first course of treatment)
from code 960 (Unknown if prostatectomy done)
o In all of these cases the abstractor coded the surgery fields to 00 (No surgery) and
in three of the cases there was supporting documentation that there was no
surgery performed.

•

4 (33.3%) were recoded to a higher, more extensive extension code
o 1 recoded to code 300 from code 230
o 1 recoded to code 350 from code 230
o 1 recoded to code 350 from code 320
o 1 recoded to code 480 from code 415
Two primary issues were demonstrated in these four cases. Involvement of bilateral
lobes of the prostate was the issue in three of the cases, while in the fourth case, there was
extracapsular extension and specific margins involved which are documented in Note 6 in
the instructions for coding this data item.

•

4 (33.3%) were recoded to a lower, less extensive extension code
o 1 recoded to code 230 from code 350
o 1 recoded to code 230 from code 406
o 1 recoded to code 400 from code 406
o 1 recoded to code 485 from code 490
In two of the cases noted above, it was documented that both lobes were involved with
negative prostatic capsule and margins. However, the original codes used in these cases
are codes that indicate specific margins were involved. In one case, the original code
indicated bilateral lobe involvement; however, the pathology text clearly stated that both
lobes were not involved. In the remaining case, the original code indicated there was
microscopic bladder neck involvement (code 490) however; the text clearly stated the
seminal vesicle was involved (code 485) and not the bladder neck.

June 15, 2012

3

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

Site Specific Factors
There were a total of 107 discrepancies noted in the CS Site Specific Factor fields. This
accounted for 57.5% of all discrepancies identified on this audit. All Site Specific Factors will be
discussed together with the exception of CS Site Specific Factor #3 (CS Extension – Pathologic)
which was discussed above, in relation to the data item, CS Extension.
CS Site Specific Factor # 13 (Number of Cores Examined) had the second highest number of
discrepancies on this audit. There were 21 (11.3%) discrepancies noted in this field.
Discrepancy Detail
•

•

•

•

11 (52.3%) known value was changed to 991 (Biopsy cores examined, number unknown)
o The text stated there was positive cores and stated the number of positive cores
but did not state the number of cores examined. In some instances, it was obvious
that the number coded in this field was the number of positive cores and not the
number of cores examined. While it may be possible that all cores sampled were
positive, the text documentation did not support this.
3 cases (14.3%) were recoded from 999 (Unknown or no information) to 991 (Biopsy
cores examined, number unknown).
o In each case there were clear statements indicating positive cores but number was
unknown.
2 cases (9.5%) were recoded from 991 (Biopsy cores examined, number unknown) to a
known value
o Text documentation clearly stated a number.
5 cases (23.8%) were recoded from a known value to another known value
o This is due to the abstractor not counting the number of cores correctly when
there were several number of cores documented from different locations in the
prostate.

CS Site Specific Factor # 1 (Prostatic Specific Antigen (PSA) Lab Value) had 17 (9.1%)
discrepancies identified.
Discrepancy Detail
•

11 (64.7%) discrepancies were due to not rounding the PSA value up to the nearest tenth
in nanograms/milliliter.
o 060 recoded to 061
o 078 recoded to 079

June 15, 2012

4

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

•

6 (35.3%) discrepancies were due to not following rules correctly
o Example #1: PSA documented as >100. Abstractor coded this field as 100. Rules
state to code PSA values greater than 98.0 ng/ml to code 980
o Example #2: PSA documented as 8.39 ng/ml. Abstractor coded this field to 839.
The appropriate code for this field is 084

CS Site Specific Factor # 11 (Gleason’s Tertiary Pattern Value on Prostatectomy/Autopsy)
had 17 (9.1%) discrepancies identified.
Discrepancy Detail
•
•

•

9 (52.9%) discrepancies were recoded from 999 (Unknown or no information) to 998 (No
prostatectomy/autopsy performed).
o The text and codes clearly documented no prostatectomy was performed.
5 (29.4%) discrepancies were recoded from 998 (No prostatectomy/autopsy performed)
to 999 (Unknown or no information)
o These cases did have prostatectomies but there was no tertiary pattern noted by
the pathologists.
2 discrepancies (17.6%) were recoded from a known value to 999 (Unknown or no
information)
o There were no statements documenting a Gleason’s tertiary pattern in the text to
support these codes.

CS Site Specific Factor # 12 (Number of Cores Positive) had 14 (7.5%) discrepancies
identified.
Discrepancy Detail
•

•

6 (42.9%) recoded to 991 (Biopsy cores positive, number unknown) from a known value
o Statements such as “core biopsies positive” or “positive core biopsy, specific
number positive unknown.”
Abstractors appeared to code the number of cores examined in this field
and in Site Specific Factor 13 (Number of Cores Examined), which is the
appropriate place to code the number of cores examined.
4 (28.6%) recoded to 991 (Biopsy cores positive, number unknown) to 999 (Unknown or
no information)

June 15, 2012

5

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

•
•

3 (21.4%) recoded to a known value from another known value
o Text documentation clearly stated specific number of cores positive. These
discrepancies were most likely due to miscounting the number of cores positive.
1 (7.1%) recoded to a known value from 991 (Biopsy cores positive, number unknown)
o Text clearly stated specific number of cores positive

CS Site Specific Factor # 9 (Gleason’s Primary Pattern and Secondary Pattern on
Prostatectomy/Autopsy) had nine (4.8%) discrepancies identified.
Discrepancy Detail
•
•
•
•

4 (44.4%) discrepancies were recoded from 999 (Unknown or no information) to 998 (No
prostatectomy/autopsy performed)
o The text and codes clearly supported there was no prostatectomy performed
2 (22.2%) discrepancies were recoded from a known value to 998 (No
prostatectomy/autopsy performed)
o Text and codes document there was no prostatectomy performed
2 (22.2%) discrepancies were recoded from a known value to 099 (Primary pattern
unknown, secondary pattern unknown)
o The Gleason’s patterns were not documented
1 (11.1%) discrepancy was recoded from one specific value to another specific value
o This recode was specifically 034 recoded to 035. This was most likely due to a
key stroke error.

CS Site Specific Factor # 7 (Gleason’s Primary Pattern and Secondary Pattern on Needle
Core Biopsy/Transurethral Resection of Prostate (TURP)) had 7 (3.8%) discrepancies
identified.
Discrepancy Detail
•

•

6 (85.7%) discrepancies were recoded from one specific code to another specific code
o All were due to not capturing the primary and secondary patterns correctly
Example #2: Text states Gleason 4+3. Abstractor coded 034. Correct code
is 043.
1 (14.3%) discrepancy was recoded from 999 (unknown, or no information) to 998 (No
needle core biopsy/ TURP performed)

June 15, 2012

6

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

CS Site Specific Factor # 8 (Gleason’s Score on Needle Core Biopsy/Transurethral
Resection of Prostate (TURP)) had 4 (2.2%) discrepancies identified.
Discrepancy Detail
•

All of these discrepancies were due to discrepancies identified in Site Specific Factor #7,
described above. All were clearly documented in the text.
o 2 (50%) were recoded from 034 to 043
o 1 (25%) was recoded from 044 to 043
o 1 (25%) was recoded from 053 to 054

Histology
There were six (3.2%) discrepancies identified in Histology. The Multiple Primaries and
Histology rules that abstractors use to code histology addresses the issue identified on this audit.
Rule H10 for Other Sites Histology Coding Rules found in the Multiple Primary and Histology
rules state “Code 8140 (adenocarcinoma, NOS) for prostate primaries when the diagnosis is
acinar (adeno) carcinoma.”
•
•

5 (83.3%) were recoded from 8550 (Acinar) carcinoma to 8140 (Adenocarcinoma)
1 (16.7%) was recoded from 8101 (Carcinoma NOS) to 8000 (Malignancy, NOS)

Grade
Grade had six (3.2%) discrepancies noted. Volume I, Section V.3.5.6 states Gleason’s score 2,
3, and 4 are to be coded to grade 1 (well differentiated); Gleason’s score 5 and 6 are to be coded
to Grade 2 (moderately differentiated); and Gleason’s score 7, 8, 9, and 10 are to be coded to
Grade 3 (poorly differentiated).
•
•
•

3 (50%) discrepancies were recoded from code 1 (Well differentiated) to 2 (Moderately
differentiated)
o All cases were Gleason’s score 6
2 (33.3%) discrepancies were recoded from 2 (Moderately differentiated) to 3 (Poorly
differentiated)
o Both cases were Gleason’s score 7
1 (16.7%) discrepancies were recoded from code 3 (Poorly differentiated) to code 2
(Moderately differentiated)
o Text documented a Gleason’s score 6

June 15, 2012

7

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

CS Metastasis at Diagnosis
There were five (2.7%) discrepancies noted in the data field CS Mets at Diagnosis.
Discrepancy Detail
•
•

4 (80%) were recoded from 40 (Distant metastasis, other than distant lymph nodes or
bone (s); Carcinomatosis) to code 30 (Metastasis in bone (s))
o All 4 cases documented bone involvement as the only metastatic site
1 (20%) was recoded from code 99 (Unknown, distant metastasis not stated) to 00
o Text documentation includes statement “No metastatic disease identified.”

CS Lymph Nodes
There were four (2.2%) discrepancies noted in CS Lymph Nodes. This field is to be used to code
regional lymph node involvement only. Distant lymph nodes are captured in another field.
Discrepancy Detail
•

2 (66.7%) discrepancies were recoded from 999 (Unknown; regional lymph nodes not
stated, cannot be accessed, or not documented) to 000 (No regional lymph node
involvement)
All discrepancies were due to abstractors not following the rule as stated in the CS
Manual, Part I regarding coding “Inaccessible Lymph Nodes.” This rule states “The
Collaborative Stage Data Collection System allows data collectors to record regional
lymph nodes as code 000 negative (based on clinical evaluation) rather than 999
unknown when three conditions are met:
o
o
o

•

There is no mention of regional lymph node involvement in the physical examination,
pretreatment diagnostic testing or surgical exploration.
The patient has clinically low stage (T1, T2, or localized) disease.
The patient receives what would be usual treatment to the primary site (treatment
appropriate to the stage of disease as determined by the physician) (or patient is offered
usual treatment but refuses it).”

2 (33.3%) discrepancies were recoded from 100 (Specific regional lymph node chains) to
999 (Unknown; regional lymph nodes not stated, cannot be accessed, or not documented)

June 15, 2012

8

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

CS Mets at Dx – Metastatic Sites Fields
There were a total of 20 (10.8%) discrepancies identified in the field CS Mets at DX – Metastatic
Sites Fields. There are four independent fields in this group of data items and are directly related
to the field CS Mets at Diagnosis. The four CS Mets at Dx – Metastatic Sites Fields are Bone,
Breast, Liver, and Lung. These discrepancies are demonstrated in Graph 1.
The instruction in the CS Manual for each of these data items state “If CS Mets at Dx is coded to
00 (No metastatic disease), this field must be coded to 0 (No metastatic disease).
In all 20 discrepancies, the scenarios were identical. In each case, the CS Mets at DX was
coded to 00 (no metastasis) and each of the CS Mets at Dx – Metastatic Site Fields were coded
according to whether or not there was a specific procedure performed to evaluate that site.
As an example, if a patient had a bone scan and a CT of the chest and abdomen, the abstractor
would code the fields:
•
•
•
•

CS Mets at Dx – Bone
CS Mets at Dx – Brain
CS Mets at Dx – Liver
CS Mets at Dx – Lung

0
9
0
0

The fields CS Mets at Dx Bone, CS Mets at Dx Lung, and CS Mets at Dx Liver were coded
based on clinical exams performed. The field CS Mets at Dx – brain was coded to 9 (Unknown
if involved) because there was no CT or MRI scan of the head. While this may be logical to the
abstractor, instruction for coding this data item is not being followed.

Resource Allocation, Benefit, and Core Team
•
•

Projected Hours to Complete Project: Estimated at 500 hours from project kick-off to
Final Report submission.
Actual Hours to Complete Project: Approximately 720 hours.

There were significant delays beginning this audit. The Recoding Audit Module (RAM) had a
significant number of program enhancements performed to upgrade its functionality. The RAM
enhancements did not roll into the changed Eureka environment as expected and required
additional programming changes. The additional programming changes and testing caused a
significant delay in the initiation of the audit.
The core team included Kyle L. Ziegler, CTR, Cheryl Moody, CTR, Taina Valone, CTR,
Marianne Schlecht, CTR, Katheryne Vance, CTR, and Lois Inferrera, CTR
June 15, 2012

9

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

Conclusions
One hundred twenty nine (69.4%) of the discrepancies identified on this audit were the result of
not following the coding directions which are clearly stated in the CS Manual. Instructions such
as rounding up to the nearest tenth in nanograms/milliliter for the PSA value in Site Specific
Factor #1, using code 998 when no prostatectomy was performed in Site Specific Factor #11,
coding adenocarcinoma (8140) when the pathology report states acinar carcinoma per the
Multiple Primary and Histology rules, and not following the “inaccessible lymph node” rule in
the CS Staging Manual, Part 1, when coding the data item CS Lymph Nodes. All of the 111
discrepancies could have been eliminated if the abstractors had followed the instructions in the
CS Manual.
Of the 32 discrepancies noted in CS Extension (Clinical Extension), only eight (25%)
discrepancies resulted in a change in the stage group. Seven of the changes in stage grouping
were down-staged and one change in CS Extension resulted in the case being unstageable.
Table 2 demonstrates the changes in stage grouping as a result of the discrepancies in CS
Extension. The remaining 23 (71.9%) discrepancies had no impact on the stage grouping and
most of these discrepancies did not impact the “T” category of the T-N-M stage. In most of the
discrepancies, the CS Extension code may have changed but the stage grouping and “T” category
remained the same. For example, eight of the 23 discrepancies did not change stage groupings,
but the “T” category changed from T2b, T2c, or T2NOS to T2NOS.
In the related field, CS Site Specific Factor #3 (CS Extension – Pathologic Extension), none of
the 12 discrepancies resulted in a change in stage or “T” category.
The discrepancies noted in the data items CS Mets at DX – Metastatic Sites
(Bone/Brain/Liver/Lung) are easily resolved with training. On May 15th, an e-mail request was
submitted to the NAACCR Edits Committee, to develop an edit which would prevent the coding
of CS Mets at Dx- Bone, Brain, Liver and Lung to anything but 0 if the data item CS Mets at Dx
was coded to 0. The NAACCR Edits Committee agreed with the recommendation from the
California Cancer Registry and an edit is currently being developed to be included in the next
NAACCR edit metafile release. Through the findings in this audit, the inclusion of this new edit
will assist to improving data quality nationwide.

June 15, 2012

10

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

Recommendations
The results of this audit should be published in the CCR newsletter “The Criterion.” A majority
of the discrepancies noted on this audit can be attributed to not applying documented rules such
as demonstrated in CS Mets at Dx – Metastatic Sites or Histology.
Training efforts should be focused on CS Extension, the Site Specific Factor fields, and
Histology. There were simple, straightforward guidelines and rules that were not being
followed. Since almost 70% of the discrepancies were the result of abstractors not following
rules clearly documented, reminders and training would greatly reduce the errors that have been
identified on this audit.

June 15, 2012

11

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

Prostate Recoding Audit Discrepancies by Data Item
N=186
Data Item
CS Extension
CS Site Specific Factor # 13 (Number of Cores Examined)
CS Site Specific Factor # 1 (PSA Lab Value)
CS Site Specific Factor # 11 (Gleason's Tertiary Pattern Value on Prostatectomy/Autopsy)
CS Site Specific Factor # 12 (Number of Cores Positive)
CS Site Specific Factor # 3 (CS Extension - Pathologic)
CS Site Specific Factor # 9 (Gleason's Primary Pattern and Secondary Pattern on Prostatectomy/Autopsy)
CS Mets at Diagnosis Brain
CS Site Specific Factor # 7 (Gleason’s Primary Pattern and Secondary Pattern Value on Needle Core Biopsy/TURP)
Grade
Histology Type
CS Mets at Diagnosis
CS Mets at Diagnosis Liver
CS Site Specific Factor # 10 (Gleason’s Score on Prostatectomy/Autopsy)
CS Lymph Nodes
CS Mets at Diagnosis Bone
CS Mets at Diagnosis Lung
CS Site Specific Factor # 8 (Gleason’s Score on Needle Core Biopsy/TURP)
CS Tumor Size
Number of Lymph Nodes Examined
CS Site Specific Factor # 2 (PSA Interpretation)
Date of Diagnosis

Table 1
June 15, 2012

12

Number of Discrepancies
32
21
17
17
14
12
9
7
7
6
6
5
5
5
4
4
4
4
3
2
1
1

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

Distribution of Stage Grouping Changes
N=8
Stage Grouping Change
From Stage
To Stage
IIA
I
III
IIB
III
I
I
Unstageable

Number
Changed
4
2
1
1

Table 2

June 15, 2012

13

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

Distribution of Discrepancies in CS Mets at Dx Metastatic Sites
n=20
Bone; 20% (n=4)
Liver; 25% (n=5)

Lung; 20% (n=4)
CS Mets at Diagnosis Brain
CS Mets at Diagnosis Liver
CS Mets at Diagnosis Bone
CS Mets at Diagnosis Lung

Brain; 35% (n=7)

Graph 1
June 15, 2012

14

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

CS Extension Code Definitions

June 15, 2012

15

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

CS Extension Code Definitions - continued

June 15, 2012

16

 California Cancer Registry
Prostate Recoding Audit
CCR Project ID #300.73
Final Report
Prepared by Kyle L. Ziegler, CTR

Site Specific Factor #3 (CS Extension – Pathologic Extension) Code Definitions
Note 6: When the apical margin, distal urethral margin, bladder base margin, or bladder neck margin is involved and there is no
extracapsular extension, use code 400.

June 15, 2012

17