Stage Data Capture in Ontario February 23, 2010 Agenda Refresher: Ontario s Stage Capture Project Collaborative Staging and Population Stage Reporting in Ontario Use of Stage Data in System Performance Measurement Collaborative Staging Data Quality Program Accessing CCO data Appendices (other project updates) 2 1
Ontario s Stage Capture Project 3 Ontario s Stage Capture Project A provincial initiative that aims to improve the quality & completeness of cancer staging g data in Ontario To enable semi-automated capture of cancer stage at diagnosis for 90% of all eligible new cases, using the Pan Canadianendorsed Collaborative Staging (CS) minimum data set. 4 The Project is part of a broader national initiative led by the Canadian Partnership Against Cancer 4 2
Population stage rate doubled when provincial stage capture project started in 2007 cer cases t stageable (as per AJCC 6 th edition) canc % of Ontario incident with reported stage Percent of Ontario incident cancer cases where stage at diagnosis was reported to CCO 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 2005 cases 2006 cases 2007 (half year) cases 2007 cases as measured Mar 2006 as measured Mar 2007 as measured Mar 2008 as measured Nov 2009 Data Source: CCO Cancer Registry and Collaborative Staging Database; Cancer staging for eligible stageable cancer cases by specified diagnosis year; Extracted on dates specified. Ontario cancer centres have expanded TNM stage reporting to all patients 3
Significant efforts from Regional Cancer Centres submitting stage data for cancer centre and host hospital cases Percentage of regional cancer centre hospital cases (cancer centre cases plus surgery-only cases in host hospitals ) with valid stage reported (2006/2007 to 2008/2009) 7 CS data collection initiated in 71 hospitals for top 4 disease sites CS Wave 1 CS Wave 2 CS Wave 3 Mar 2008: Initiated data collection in 4 CSA* hospitals in LHIN 7 Jun 2008: Added 7 CSA* hospitals in LHIN 4, 5, 6 Mar 2009: Added 24 CSA* hospitals in LHINs 1, 2, 3, 8, 9, 11, 13 Oct 2009: Added 36 non- CSA hospitals with at least 25 cases/year, all LHINs Final Wave: 2010 Feb 2010: Begin data collection in 14 regional cancer centres (RCCs were previously submitting stage data to CCO on cancer centre and host hospital cases) Cancer Surgery Agreement (CSA) hospitals excluding Regional Cancer Centres Top 4 sites: breast, lung, prostate, colorectal 4
Collaborative Staging and Population Stage Reporting in Ontario 9 Collaborative Staging data included in these reports Source: Ontario Cancer Registry (OCR) Passive registry OCR does not actively find and register cancer cases; cases are created from data inputs (DAD, NACRS, pathology reports, etc.) Assumptions on hospital records CCO CS abstractors rely solely on information in hospital EHRs No data quality to be performed on hospital records Hierarchy in identifying primary hospital for a case Case exclusions: Previously staged cases, Regional Cancer Centre cases and surgical cases from host hospitals of RCCs (until 2010) 10 5
TNM Stage data included in these reports Source: RCCs submit stage data monthly to CCO s Data Book on all stageable cancer cases Seen at the cancer centre Diagnosed at the host hospital Stage data reported by HIM professionals at RCCs 11 Comprehensiveness of provincial stage data by LHIN for all disease sites 12 6
Comprehensiveness of stage capture by LHIN for top 4 disease sites 13 Comprehensiveness of stage capture by disease site 100% Percentage of Ontario Cancer Cases Diagnosed in 2007 with Valid Stage by Disease Site As Measured in November 2009 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 14 7
Comprehensiveness of stage capture by treatment centre type 15 Use of Stage Data in System Performance Measurement Incidence and Referral Patterns Relative Survival Resection Rates Guideline Concordance Chemo/ Radiotherapy Treatment Rates 16 8
Incidence Rates by Stage 17 Distribution of cancer stage at diagnosis for top 4 disease sites 100% 90% 80% 70% 60% Distribution of Incident Cancer Cases by Stage Top 4 Disease Site Cases Diagnosed in 2007 5% 7% 14% 39% 20% 30% 48% 10% 50% 40% 30% 20% 43% 30% 27% 6% 82% IV III II I 10% 21% 18% 0% Breast Colorectal Lung Prostate Staged Cases*= 6,927 5,941 4,999 6,728 % Staged: 90% 84% 78% 81% 1% 18 9
Distribution of cancer stage at diagnosis across LHINs for colorectal cancer cases 19 Distribution of cancer stage at diagnosis across LHINs for breast cancer cases Sorted in descending % of Stage I and II 20 10
Distribution of cancer stage at diagnosis across LHINs for lung cancer cases Sorted in descending % of Stage I and II 21 Resection Rates by Stage 22 11
Variation of resection rates across LHINs for non-small cell lung cancer cases 23 Variation of resection rates by stage at diagnosis for non-small cell lung cancer cases 24 12
Treatment Rates by Stage 25 Variation in treatments across LHINs for non-small cell lung cancer cases 26 13
Variation in treatments across LHINs for stage IIIA/B non-small cell lung cancer cases 27 Colon Guideline Chart Review - Results Treatment Practice Patterns of Fully Resected Stage III Colon Cancer Cases - Analysis of CCO Data Only for 2007/08 (based on 8 RCCs*) N=376 Not Treated with Guideline Adjuvant Chemo 180 48% Treated with Guideline Adjuvant Chemo (includes clinical trials) 196 52% * Grand River, Hamilton, Kingston, London, PMH, Sudbury, Thunder Bay 28 Based on data available from CCO ALR Databook for 2007/08, from 8 RCCs 14
Colon Guideline Chart Review - Results Treatment Practice Patterns of Fully Resected Stage III Colon Cancer Cases - Analysis of CCO Data and Chart Reviews from Eight RCCs* for 2007/08 N=376 Not Treated with Guideline Adjuvant Chemo 132 35% Treated with Guideline Adjuvant Chemo 244 65% * Grand River, Hamilton, Kingston, London, PMH, Sudbury, Thunder Bay 29 Based on data available from CCO ALR Databook for 2007/08, and Chart Review from 8 RCCs Colon Guideline Chart Review - Results Treatment Practice Patterns of Fully Resected Stage III Colon Cancer Cases - Analysis of CCO Data and Chart Reviews from Eight RCCs* for 2007/08 Referred Elsewhere N=376 16 Not treated: Other 32 9% Not treated: Patient Choice 36 10% Not treated: Age/Condition 48 13% 4% 5-FU plus LV / FOLFOX 173 45% Xeloda 47 13% Clinical Trials 24 6% * Grand River, Hamilton, Kingston, London, PMH, Sudbury, Sunnybrook, Thunder Bay 30 15
Survival Rates by Stage 31 Variation in survival rates by stage at diagnosis and sex for lung cancer cases 32 16
Collaborative Staging Data Quality Program 33 Robust data quality program will ensure CS data is of high quality Processes in place to ensure high data quality EDITS software, Registry Plus, visual review, re-abstraction, training, year-end end cleanup Comprehensive array of DQ measures implemented to assess and monitor multiple dimensions of data quality for 2007 CS cases Reliability, Timeliness, Completeness, Validity, Usability Participated in national inter-rater reliability study led by the Canadian Partnership Against Cancer (CPAC) Extensive ongoing training for CCO abstractors with support from the Public Health Agency of Canada Executive Summary of DQ report available on CCO web portal 17
Accessing CCO data and reports 35 CCO s Secure Web Portal https://external.cancercare.on.ca Collaborative Staging folder Hospital specific subfolders RCCs: TNM Staging or Data Book Production CS data quality report (Executive Summary) 2007 CS DQ Assessment Executive Summary_Final.pdf CS chart level and hospital aggregate report (Excel) All CS data elements collected from your hospital and derived with CS abstraction software Hosp 2007 Report.xls CS LHIN and ON aggregate report (Excel) LHIN X 2007 report.xls Description of CS data elements and codes (Word) CS implementation - data variables and rationale for hospitals.doc Sample slides for optional regional presentation by Dr. Brierley (focus: clinical indicators and other guideline concordant analysis) LHIN Stage Presentation - 2010-01-26.pdf 36 18
Requesting other data from CCO http://www.cancercare.on.ca/toolbox/system info/requestccodata/ 3 ways to request other data (including chart level data from outside your hospital for research) Apply for access to iport Order SEER Stat CDs Request other data 37 Upcoming Activities Complete staging for 2008 diagnosis year by Spring 2010 Requires secondary abstractor access Follow-up with reports and presentation Write paper on colon guideline concordance findings for publication Produce range of stage based indicators for Lung Cancer Journey Continue efforts to ensure high quality stage data is available on a go forward basis 38 19
Questions / Discussion 39 Appendices 40 20
Appendix A: Collaborative Staging automation 41 Leveraging automated data capture from electronic clinical reports for CS data collection Synoptic Clinical Notes Synoptic Imaging Data Synoptic Lab Data Ontario Laboratory Information System (OLIS) Synoptic Surgical Data Canadian Cancer Registry Electronic Transfer of Collaborative Staging Data Elements Cancer Registry and Data Warehouse Synoptic Pathology Data Pathology Information Management System (PIMS) Clinicians Administrators and Planner Researchers and Epidemiologists 21
Automating capture of CS pathology data from synoptic reports and site specific factor lab data from OLIS CS minimum data set 1. CS Tumour Size 2. CS Extension 3. CS TS/Ext Evaluation Method 4. CS Lymph Nodes 5. CS Regional Node Evaluation Method 6. Regional Lymph Nodes Positive 7. Regional Lymph Node Examined 8. CS Metastases at Diagnosis 9. CS Metastases Evaluation Method 10. CS Site Specific Factor 1-25 Recent study demonstrated feasibility and benefit of automatically collecting CS required lab test data from OLIS Structured Lab Test Results from OLIS Synoptic Pathology Reports 2007 trial with Lakeridge, UHN and CVH revealed that prepopulating CS abstracts with synoptic pathology report data resulted in time saving without compromising data quality As of Dec 2009, 23% of all pathology resection reports received in discrete synoptic format Since May 2008, 6,335 path reports from 46 hospitals received in discrete synoptic format 25.0% % reports submitte ed in Discrete Synoptic Format 20.0% 15.0% 10.0% 5.0% 0.0% May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov 2008 2009 Overall Synoptic Reporting 0.0% 2.4% 3.9% 4.3% 5.5% 7.5% 8.5% 10.3% 9.5% 12.0% 14.5% 15.9% 18.6% 18.5% 19.6% 19.6% 20.1% 21.1% 23.1% Specimen Received Month Data Source: CCO PIMS Database; Reports received by month of date of surgery; from May 08 to Nov 09, as of Dec 24/09. 22
As of Dec 2009, 97% of all discrete synoptic pathology reports complete against CAP checklist 100.0% of Discrete Synoptic Reports Complete for CAP % 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 90% target 0.0% May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov 2008 2009 5 Common Cancers 0.0% 92.3% 95.0% 96.0% 92.9% 87.4% 88.5% 89.7% 92.0% 95.6% 92.9% 94.4% 94.4% 93.1% 94.0% 94.5% 96.8% 97.1% 97.0% Specimen Received Month Data Source: CCO PIMS Database; Reports received by month of date of surgery; from May 08 to Nov 09, as of Dec 24/09 OLIS data can increase completeness rates for CS lab data 100% 90% 80% 70% 60% 3772 cases 50% 40% 30% 204 cases 20% 10% 0% 3290 cases Prostate Colorectal Ovarian Disease site Jan 2007 Mar 2008 Collaborative Staging data 23
OLIS provides 65% more PSA results on prostate cancer cases than hospital record abstraction alone 100% n=69) % prostate cancer patients cluded in comparability analysis, n (inc 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% OLIS data does not match manual abstraction OLIS data matches manual abstraction OLIS provides PSA result and not found from manual abstraction Synoptic Reporting in Imaging and Surgery e-surgery AB and ON working together on content standards - Developing template for endometrial cancer and hope to be able to incorporate synoptic path reports ecancer ovarian in use in ON (WebSMR in AB) - 2 uses: operative notes, outpatient clinic notes for colposcopy clinic - 140 reports to date at UHN - London and Hamilton to implement soon e-imaging Standards being led by Radiological Society of North America - 13 template development subcommittees - 70 templates launched at RSNA 2009 annual conference - Available at: http://reportingwiki.rsna.org/index.php?title=templates Ontario pilot on synoptic MRI report for rectal cancers focusing on content standards for reporting 48 24
Appendix B: Available stage data 49 Current Stage Data Available 2006 cases TNM only from cancer centres Approximately 50% completeness in provincial stage capture for all sites 2007 cases TNM + CS - for top 4* sites - 80%+ stage capture (depending on site) TNM only - for all other sites, limited to cancer centres cases N.B. In early 2010, additional CS staging will be undertaken in cancer centres to ensure 90% completeness for 2007 CRC cases Data prior to 2006 only TNM; <30% completeness *Breast, colorectal, lung, prostate cancer 50 25
Future Stage Data Availability 2008 and 2009 cases TNM + CS @ 90% stage capture for top 4 sites* TNM only - for all other sites, limited to cancer centres cases 2010 and 2011 cases CS for top 4* sites @ 90% stage capture TNM only - for all other sites, limited to cancer centres cases 2012 cases CS for all sites @ 90% stage capture *Breast, colorectal, lung, prostate cancer 51 Chart level reports include all data variables collected in Registry Plus Demographic First name Middle name Last name Date of birth Sex HIN Other Abstract ID Reporting hospital (hospital @ abstraction) Hospital chart # 1st contact date Abstract completion date OCR group # Class of case Tumour Primary site (ICD-O3) Laterality Histology (ICD-O3) Behaviour (ICD-O3) Histologic grade Age @ Dx Dx date Diagnostic confirmation Staging at Diagnosis CS Tumour size CS extension CS Tumour size/extension i eval CS lymph nodes (LN) CS lymph nodes eval # Regional LN positive # Regional LN examined CS Metastasis CS Metastasis eval Site specific factors ( up to 6) CS Version Derived T and T descriptor Derived N and N descriptor Derived M and M descriptor Derived Stage group Derived SEER 2000 Derived SEER 1977 AJCC reported* Clinical T, N, M, Group AJCC reported* Pathologic T, N, M, Group *AJCC reported: whatever data is available in the electronic hospital patient chart 26