CANCER PREVENTION INSTITUTE ,Yv or CALIFORNIA Preventing Cancer. Promoting Life. GREATER BAY AREA CANCER REGISTRY SEER FINAL REPORT August 1, 2003 - July 31, 2010 CONTRACT N01-PC-35136 Submitted: July 31, 2010 Sally L. Glaser, Principal Investigator Kathleen Davidson-Allen, CTR, Director of Registry Operations SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 SEER FINAL REPORT AUGUST 1, 2003 - JULY 31,2010 TABLE OF CONTENTS (1) WORK DONE AND RESULTS ACHIEVED ..1 REGISTRY OPERATIONS .. 1 1. Cancer Case Volume 1 2. Case Finding .. 2 a. Out-of?area pathology laboratories .. 3 b. Electronic pathology (e-path) reporting .. 3 c. Casefinding challenges .. 3 3. Data Collection .. 5 a. Acute care facilities and outpatient departments 6 b. Multi-specialty group practices .. 6 c. Other reporting entities (non-hospitals) .. 6 d. Resident cases diagnosed outside of our SEER region .. 8 e. Data Collection challenges .. 8 4. Quality Control .. 9 a. Visual editing .. 9 b. Data quality improvement activities performed Quality control activities with Surveillance Research 10 d. Special quality control activities .. 10 e. Quality Control challenges .. 13 5. Education and Training .. 14 a. Ongoing education and training opportunities .. 14 b. Registry Compliance Program .. 15 6. Monitoring Completeness and Timeliness of Reporting .. 15 a. Completeness of 15 b. Timeliness of reporting .. 16 c. Completeness and timeliness challenges 16 7. Follow-up Methods (Passive and Active) 17 a. Passivefollow?up ..17 b. Active follow-up ..20 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-35136 c. Follow-up . . .. 21 8. Registry Information Management Systems .. 21 a. Data entry system for casefinding 22 b. Pathology data management system 23 0. Electronic pathology reporting .. 23 d. Web portal (RegWeb) .. 24 e. GBACR doctor database .. 24 f. SEER Registry Operations Cost Study (ROCS) reporting tool .. 24 9. Production data 24 h. Casefinding system .. 25 i. Eureka and GBACR casefinding systems .. .. .. 25 SURVEILLANCE RESEARCH .. . 25 1. Dissemination of GBACR and Related Data 26 a. Production and dissemination of annual reports 26 b. Fact sheets .. 27 c. Data access/communication .. 27 2. Surveillance Research by Surveillance Research Staff .. 27 a. Improving the utility of SEER data through research .. 28 b. Other work initiated by Surveillance Research investigators .. 29 OTHER ACTIVITIES OF NOTE . . 31 1. Registry Leadership .. 31 2. Legal/Regulatory Issues .. 32 3. Biorepository .. 33 a. California Biorepository Research 34 b. Change to California state .. .. 34 4. Process Management/Continuous Process Improvement .. 35 3. Overview of methodology .. 35 b. Process Management/Continuous Process Improvement activities 36 c. Other Process Improvement activities .. 37 (2) LIST OF STUDIES .. 37 (3) CURRENT STATUS AND FUTURE PLANS .. 38 1. Organization Chart and Current staffing 38 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-351 36 a. Organization chart .. 38 b. Staffing .. 38 Rules for Reporting to the Registry .. 39 Methods Used to Obtain Data and Percent of Records from Each Methods .. 39 Follow-up Methods (including percent passive follow-up by age group) .. 39 Key Relationships with Other Registries and Organizations .. 39 a. CCR and its regional registries 39 b. Stanford .. 39 c. UCSF .. 39 d. American Cancer Society 40 e. NAACCR and SEER .. 40 Current Process Management Methodologies and Plans for the Future .. 40 Future Plans, Challenges, 40 a. Registry Information Systems .. 40 b. Physician self-reporting .. 40 c. Process Management/Continuous Process Improvement 41 d. Electronic health records and cloud computing .. 41 9. Education and training 42 f. Improved internet?based dissemination of cancer statistics A. 42 e. Biorepositories .. 42 SEER Final Report: August 1. 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 (1) WORK DONE AND RESULTS ACHIEVED During the SEER contract period 2003-2010, the Greater Bay Area Cancer Registry (GBACR), operated by the Cancer Prevention Institute of California (CPIC) (formerly the Northern California Cancer Center has accomplished its contract deliverables by registering regional cancer cases from its nine catchment counties to a high level of quality and completeness, and by using and distributing cancer registry data extensively for research and education. We are proud of our work, as the GBACR data contributed to SEER are of importance given the size and diversity of the Greater Bay Area population, and as our registry?s data have helped in the monitoring and understanding of cancer trends and facilitated important epidemiology and cancer control studies relevant to cancer control both locally and nationally. An overview of our work and results during the contract period, and (in italics) in the past contract year (August 2009 July 2010), follows. Within categories of this report, we separate our work and results into three categories: ?Registry Operations? to describe our data collection and quality control activities, ?Surveillance Research? to describe our data use and dissemination activities, and ?Other Activities of Note? to describe our work in legal/regulatory compliance, biorepositories, and other strategic efforts. REGISTRY OPERATIONS Throughout the contract period, the GBACR Registry Operations Unit (ROU) identified and registered regional cancer cases by consistently performing several core functions including casefinding, data collection, quality control, training, and monitoring case completeness and timeliness; many of these activities were executed under information systems data management. Over the contract years, we have focused on refining processes, procedures and execution of once cumbersome paper-based, manual tasks activities through streamlining and automation to bring greater efficiency and productivity. An overview of our work and results follows. 1. Cancer Case Volume Characteristics During this contract, the GBACR contributed over 200,000 incident cancer cases, diagnosed from 2001 to 2007, to the SEER program. The attached tables show the case counts of in situ, malignant, and all cancers reported to our registry for every calendar year during the contract period. Appendices 1?6 show case counts according to race for the years 2001-2007 for the San Francisco?Oakland and San Jose-Monterey SEER regions, respectively. Appendices 7-9 show case counts by Hispanic ethnicity for the years 2001-2007 for the San Francisco-Oakland SEER region; Appendices 10-12 show counts for the years 2001-2007 for the San Jose- Monterey SEER region. Over time, the proportions of our case volume diagnosed in Hispanic, Chinese, and Filipino patients have grown substantially in both regions, while the proportions of African-American and Japanese patients have remained stable. Proportions of non-Hispanic white patients have declined over time. These changes likely reflect regional demographic trends as well as changes in cancer incidence patterns among specific groups. During the contract period, the annual case volume (2003 through 2007) grew by approximately 8.6% in the San Francisco?Oakland region and about 6.5% in the San Jose-Monterey area. In recent years, the total annual GBACR case volume has been about 32,000 newly diagnosed cases, with these individual tumors identified and consolidated from over 47,000 individual reports to the registry, some from multiple reporting sources. The proportion of cancer cases 1 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 diagnosed or treated outside of hospital settings at free-standing pathology laboratories, radiation centers, nursing homes, or private physician offices) has remained constant, representing about 5% of our total malignant case volume. higher proportions of in situ cancers were diagnosed outside of hospital settings, especially in the San Jose-Monterey region. Appendices 13-15 show case counts according to cancer type (malignant, in situ, all) and type of reporting source for the years 2001-2007 for the San Francisco-Oakland SEER region; Appendices 16-18 show these counts for the San Jose-Monterey SEER region. Appendix 19 shows the ten cancers most frequently diagnosed among men and women in our GBACR region in the period 2003 through 2007. Prostate cancer was the most common cancer in men, and breast cancer was the most common in women. Lung cancer ranked as the second most common cancer diagnosed in both men and women. Melanoma was the fourth most common cancer diagnosed in men and the fifth in women, and was diagnosed with increased frequency in our region over time, as was thyroid cancer, the eighth most common cancer diagnosed in women in our region. Whereas the HIV-associated cancer Kaposi?s sarcoma was once the fifth most common cancer diagnosed among men in our region, it ceased being even among the ten most commonly diagnosed malignancies. 2. Casefinding As the contract period ended, the GBACR region was home to 63 medical facilities, which included community and university-based hospitals, a 12-hospital health maintenance organization (HMO) system (Kaiser Permanente), several children?s specialty hospitals, a rehabilitation hospital, and two Veterans Administration (VA) hospitals. In addition, our region housed three multi-specialty medical practices, which were subjected to the same casefinding practices as the medical facilities, and ten free-standing pathology laboratories that include seven ?multi-purpose? laboratories, two dermatopathology laboratories, and one laboratory specializing in urinary system pathology. Casefinding in these facilities has been undertaken using highly trained staff and detailed, tailored protocols for locating cancers. These methods have been developed and field-tested over the many years of GBACR operation, and we have worked constantly to make them more efficient. Our approach has been based in having a strong, well-trained staff; in staying abreast of all facilities in which cancers are diagnosed; and in keeping close tabs on how cancers are reported within facilities. Throughout the contract years, the GBACR has used a PC-based, stand-alone data entry application developed in-house to collect (abstract and code) pathology information from hospital pathology departments and free-standing pathology laboratories and to identify cases for our rapid case ascertainment studies. This data entry system is called for its two source documents: 1) the Confidential Morbidity Report (CMR) report, the mechanism used by physicians to report non-hospital cancer cases to the GBACR, and 2) the pathology reports. In the latter half of the contract period, this pathology report data entry system was redesigned, programmed, tested and released for use in December, 2008. The redesigned application is more efficient because it is web-based and includes additional fields to collect more patient, tumor, and physician information. This additional information facilitates linkage of these reports to cases submitted by reporting facilities and follow?back to the appropriate non?hospital reporting entity. This application was used in facilities not reporting electronically using the Artificial Intelligence in Medicine?s (AIM) electronic pathology (e-path) reporting software. 2 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract NO1-PC-35136 Annually, GBACR Data Collection staff manually review over 400,000 pathology reports in our medical facilities and free-standing pathology laboratories; approximately 33,000 prove to be reportable cancer diagnoses and are entered into our casefinding application. a. Out-of?area pathology laboratories Out-of-area pathology laboratories include those in other parts of California and elsewhere in the United States. Paper copies of pathology reports from GBACR residents diagnosed in other areas of the state have been sent to the GBACR, entered into our application, and transmitted to our case?nding system for follow-back. Paper pathology reports for out-of-area residents diagnosed in GBACR laboratories similarly have been sent to the other California regional registries. During the contract period, we received paper reports from four out-of-state pathology laboratories, and the California Cancer Registry (CCR) received electronic reports for three national laboratories, which it routed to GBACR using the AIM e-path software. The third national lab (US Labs) recently began sending electronic data. Appendix 20 includes a list of the GBACR free-standing laboratories and the out-of-state pathology laboratories. b. Electronic pathology (e-path) reporting Over the course of the contract, we have had a steady increase in the number of e-path installations. At the end of this period, we have 10 e-path installations that cover 23 reporting facilities and free-standing pathology laboratories. These include several high-volume hospital pathology laboratories (Kaiser, University of California at San Francisco) and the four largest GBACR-region free-standing path labs. Combined, these facilities generate approximately 550,000 pathology reports each year; with electronic reporting, we have been able to reduce the number of relevant reports that are screened by 80%. With the capture of location information (ordering facility and ordering physician) in both the AIM e-path software and our application, the follow-back process has become more efficient, as we have been able to accurately target the reporting entity for follow-back and thus greatly reduce the number of second or third follow-back requests. During the last contract year, AIM e-path software also was installed in two hospital pathology laboratories and one free-standing pathology laboratory. 0. Case?nding challenges Beginning in 2002, we began encountering considerable difficulty in gaining access for casefinding activities in several medical facilities in our reporting area. This was a due to misinterpretation by these facilities of the state cancer reporting law, particularly as it applied to our rapid case ascertainment activities for research studies and the HIPAA Privacy Rule. With the guidance of Dr. Robert McLaughlin, the CPIC Legal and Regulatory Affairs Officer, who oversees issues of compliance, the GBACR was able to restore access to all casefinding documents. Since then, we have had no problems with access. One of the challenges of implementing electronic reporting in regional facilities with cancer registries has been the requirement of the facility registry to have the pathology data go directly into its reporting software as well as to the GBACR. While the AIM software provides a mechanism to review data that are sent to the GBACR, it lacks a mechanism to incorporate the 3 SEER Final Report: August 1, 2003-July 31. 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 data into the facility?s cancer reporting software (a CCR product). does have an electronic reporting module called ?Cancer Alert System? (CAS), but that software has not been widely used in California due to limited resources at the OCR. In specific, this has been an issue for the installation of e-path at Stanford, a well-known treatment and referral center. This installation has been a long-standing GBACR goal, as casefinding and follow-back processing of Stanford?s high volume of regular and consultative pathology reports have been very labor? intensive for GBACR staff. In the last contract year, we worked successfully with a team comprising Stanford, AIM, and CCR staff on a combined installation projected to go live on August 24, 2010. Once we have this installation up and running, it not only will introduce considerable efficiency based on electronic transmission of Stanford pathology but also will pave the way for installations in other hospitals with high-caseload cancer registries that are not yet submitting their pathology data electronically. Keeping abreast of the changes in healthcare practices in the Bay Area has also been challenging. However, the GBACR has employed numerous methods to track changes in our reporting facilities, including new medical facilities, hospital mergers, and new physicians in the regions. One method is to subscribe to and circulated two relevant publications to appropriate GBACR Data Collection, Quality Control, management, and administrative staff. The San Jose Business Journal, recognized as the region?s business and industry standard, has been useful for tracking new hospitals, outpatient treatment facilities, hospital mergers, activities, and overall changes to health care patterns. The California Healthline, a daily electronic publication of the California Healthcare Foundation, reports on news, policy and opinion in the health-care arena in California and nationwide, and its California Hospital News Roundup for the Week section has proven to be extremely useful, as it provides very detailed information on changes in health-care facilities. Another approach to keeping up with regional changes has been to request, every two years throughout the contract period, current listings of surgery centers, clinics and pathology laboratories from the California Department of Consumer Affairs, the state agency responsible for licensing facilities and laboratories. We reviewed the listings, contacted new facilities and laboratories, and, when applicable, established a case reporting method. This ensured that the GBACR remained active in performing casefinding from all possible reporting sources within the region. In addition, we have identified new physicians primarily through our casefinding and data collection activities. The California physician license number is a required data element in the CCR DM8 and is collected on all casefinding and data collection documents. In general, to track individual physicians who diagnose and treat cancers, the GBACR developed and maintained a comprehensive database of doctors, which it updated regularly to ensure that the most current contact information was available for data collection and casefinding purposes. This database also was used to identify physicians for other activities, such as data collection for Patterns of Care and other SEER studies. To be sure we were up-to-date in our knowledge about non-hospital facilities, the GBACR conducted a survey in 2008 of the 89 surgery centers and endoscopy centers in our catchment area. We learned that 17 in fact did not diagnose or treat cancer, and that four were not free- standing surgery centers but rather were af?liated with reporting facilities. The survey also inquired about where each center sent its pathology specimens to be read, so that we were able to validate that we were reviewing pathology reports at those primary laboratories. 4 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-351 36 Changes to reporting facilities and physicians that we identified by these various means, including the addition of new reporting sources, were added to the statewide database and immediately communicated to all staff. 3. Data Collection As required by SEER, the GBACR collects all newly diagnosed cancer cases with the behavior code (in situ) or (malignant) according to the International Classi?cation of Diseases for Oncology Per SEER guidelines, the following sites are not reported: neoplasms, malignant, NOS of the skin; epithelial carcinomas of the skin; papillary and squamous cell carcinomas of the skin; and basal cell carcinomas of any site except genital sites. Carcinoma of the cervix in situ has not been reportable since January 1, 1996 and prostatic intraepithelial neoplasia (PIN of the prostate since January 1, 2001. The GBACR reports cases with site and histology coded using the current version of the ICD-O, as mandated by SEER. During the contract period, newly diagnosed cases were coded using the updated ICD-O 3rd edition. During the contract period, SEER implemented several new requirements relating to the collection and coding of reportable cases, as follows: 0 Collection and reporting of primary intracranial and central nervous system tumors with a behavior code of (benign) or (borderline) according to ICD-O-3, effective with cases diagnosed January 1, 2004. 0 Collaborative Staging System (CS), effective with cases diagnosed January 1, 2004. Cases diagnosed prior to January 1, 2004 are coded to the coding system that was in effect at the time of diagnosis. 0 SEER Program Coding and Staging Manual 2004, effective with cases diagnosed January 1, 2006 through December 31, 2006. Casefinding List, effective October 1, 2006. 0 Multiple Primary and Histology Coding Rules, effective with cases diagnosed January 1, 2007. SEER Program Coding and Staging Manual 2007 with 2008 revisions, effective with cases diagnosed January 1, 2007 and forward. Casefinding List, effective January 1, 2009. Hematopoietic Neoplasm reportability instructions and data collection rules, effective with cases diagnosed January 1, 2010 and forward. 0 Casefinding List, effective January 1, 2010. 0 Collaborative Stage Data Collection System, version 2 effective with cases diagnosed January 1, 2010 and fon/vard. To facilitate data collection for these new reportable diseases and using the new coding rules, the GBACR continued its practices to: 1) notify cancer registrars and all reporting sources of the new requirements; 2) provide training on the newly reportable sites and coding rules; 3) provide regular feedback to hospital registries regarding coding and abstracting errors; and 4) modify casefinding procedures to collect the newly reportable diseases. The GBACR Data Collection unit has worked with three groups of reporting facilities: 1) hospitals that submit their own data, 2) hospitals that contract with the GBACR for casefinding and abstraction, and 3) other facilities that diagnose or treat cancer patients. The following addresses the data collection activities of the GBACR during the contract period, by type of 5 SEER Final Report: August 1, 2003?July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-351 36 reporting source. The number and percentage of cases by type of reporting source are included as Appendix 21. a. Acute care facilities and outpatient departments (hospitals) Over the contract, the GBACR has collected more than 95% of all reportable cases each year from hospital acute care facilities and outpatient departments; this represents approximately 45,000 case reports abstracts), a number that has been stable in the past seven years (see Appendices 13-18). As of the end of the contract period, the GBACR catchment area contained 60 hospitals, which include 21 with an American College of Surgeons (ACOS)?approved program, and 39 without ACOS-approved programs (see Appendix 22). Under California?s mandatory reporting law, hospitals in our region report all newly diagnosed cases to the GBACR regardless of the residence of the patient. All reporting hospitals have signed letters of agreement with the GBACR to report cancers as required by the State mandate, and to identify their reporting method. These methods include reporting their own data (designated as Option A) or contracting with a third party, such as the GBACR, for abstracting services (designated as Option B). For both methods, data must be submitted in machine-readable format. The majority of non-VA reporting hospitals in California report electronically using the CCR-created software. The VA hospitals report data electronically using a VA?developed data management system called VISTA. Military and VA hospitals are exempt from reporting requirements under the State law, and submit cancer cases voluntarily. As of the end of the contract period, the GBACR catchment area included two VA hospitals and no military hospitals. Because of issues related to confidentiality, security and use of VA data for research, the VA suspended case reporting in 2005 but resumed reporting in July, 2009, after the CCR and its regional registries implemented the confidentiality/data security requirements that the VA required and is described in Section 8. Registry Information Management Systems, g. Production data b. Multi-specialty group practices Under the California state cancer reporting law, large multi-specialty physician group practices that diagnose more than 100 cancer patients each year and have a uniform medical record system are considered cancer reporting facilities and must bear the burden of reporting the cancer cases to the regional registry. Throughout the contract period, we have had three such practices in our area. c. Other reporting entities (non-hospitals) Over the contract period, approximately five percent of all GBACR cases were reported from non-hospital sources, such as pathology laboratories, physician offices, nursing homes, free- standing treatment centers, and death certificates. i. Patholoqv laboratories The GBACR Data Collection staff performed casefinding in all regional pathology laboratories that provide cancer diagnostic services. Casefinding methods included on-site review of pathology reports, with capture of pertinent information (such as the patient?s name, physician's name, primary site, and specimen data) by data entry into a laptop computer from paper pathology reports or by receipt of electronic pathology reports via AIM e-path software. 6 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-351 36 information from all pathology reports collected was used to link newly identified cases to the registry database to avoid duplicate case reporting. Pathology data were also used for rapid case ascertainment of subjects for studies requiring patient contact close to diagnosis. The GBACR also received pathology reports for Greater Bay Area residents whose physicians sent pathology specimens to be read at laboratories outside of our catchment area, including out of state. ii. thsicians? of?ces California requires that physicians report cancer cases diagnosed and/or treated only in any physician office. For physicians who do not or cannot send cases, the GBACR protocol during the contract period has been for staff to visit these offices or fax to them a Confidential Physician Reporting Form, a report used in California for reportable physician cases, to obtain case information. This has ensured that all cases diagnosed in a physician office are reported to the GBACR. During the last contract year, over 2,000 cases were abstracted from physicians? offices. Nursing homes Determination of the need to visit nursing homes and convalescent hospitals has been made on the results of death clearance. If a significant number of cases is reported solely from these facilities, they were added to our routine follow-back process. iv. Free-standing radiation treatment facilities By the end of the contract period, the GBACR was performing casefinding in seven free- standing radiation treatment facilities, two of which opened in 2009. v. Ambulatory care and hospice facilities Throughout the contract period, the GBACR received case listings routinely from several free? standing surgery centers, and we followed back on these cases to make sure that they were reported by a hospital or a physician. On the rare occasion that a death certificate identified a hospice as the reporting source, such cases were followed back to the hospice source to obtain additional information. vi. Death certificates Processing of information from death records was conducted by the CCR annually and On a yearly basis, CCR matches the statewide cancer case database against death records from the California Vital Statistics Department, which documents all deaths occurring in California in a given year. From this match, the CCR data management system (CCR DMS, also called Eureka) creates a source document with information from the death certificate for cases that positively or possibly match to a patient in the statewide database. Positive matches are processed, and the death certificate information is automatically applied to the patient-level record. The possible matches require manual review; if the death certificate is deemed to match to a patient, it is linked to the patient-level record, and the death information is applied. For those death records that do not link or for which the death record does not contain a cancer diagnosis, the death certificate source document is discarded. If the death record does contain a cancer diagnosis, the ?case? is sent to the CCR DMS casefinding system for follow-back to the 7 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 36 most likely reporting source hospital, coroner, nursing home, or physician) to determine if the case is reportable and, if so, to obtain the information needed to complete a cancer abstract. In addition, the CCR also received files of death records from the California Vital Statistics Department on a (incremental) basis. These files are linked to the CCR (statewide) database and the death information (date of death, cause of death, place of death, etc.) is applied to the patient-level record for patients with a positive match, as described above. The State death certificate files contain all deaths occurring in California, as well as deaths of some California residents that occurred outside the state. For the deceased Greater Bay Area residents whose cancers are diagnosed or treated outside of our catchment area, the California regional registry in the region where the patient was diagnosed or treated performs follow-back and enters the information in Eureka. The GBACR similarly performs follow-back and case abstraction for deceased residents of other regions but were diagnosed or treated in the Greater Bay Area. In the contract period, the follow-back process identified approximately 300 cases each year that were not previously identified and reported by cancer reporting facilities or from physician offices. For the SEER standard that cancers diagnosed only by a death certificate not exceed 1.5% of all cancers registered in a given year after follow-back, the GBACR met the SEER quality marker for the 2005 data submission the 2006 data submission the 2007 data submission and the 2008 data submission as indicated in the CPIC Data Quality Profiles for these submission years, included as Appendices 23-26. d. Resident cases diagnosed outside of our SEER region All resident cases diagnosed and/or treated outside of the GBACR catchment region were abstracted by the reporting facility and submitted to their local CCR regional registry for inclusion in the statewide database that is accessible to all CCR regional registries in California. Prior to the fall of 2004, case sharing of non-resident cases occurred directly between the GBACR and other California regional registries. However, this activity became obsolete with the migration in fall, 2004, of the GBACR database to the CCR DMS, which takes care of case- sharing centrally. For cases that were not abstracted, other regions in California sent copies of pathology reports to the GBACR. These were followed back and, if determined to be resident cases, added to the statewide database. Such cases represented a very small percentage of our total caseload. The CCR has reciprocal reporting agreements with more than 25 states with which cases are exchanged regularly. These cases represented an extremely small percentage of our total cases. e. Data Collection challenges Military and VA hospitals are exempt from reporting requirements under the California State law, and submit cancer cases voluntarily. Because of issues related to confidentiality, security, and use of VA data for research, the VA suspended case reporting to state-mandated, population- based cancer registries in 2005, including the two VA hospitals in the GBACR catchment area. In response to this nationwide issue, our Registry Compliance Program explored the statutory authority of the Secretary of the Department of Veterans Affairs to use discretion in authorizing hospital participation. We wrote two letters, one to the Secretary of Veterans Affairs, the other 8 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-351 36 the Secretary of Health and Human Services to address these reporting concerns. The response from the VA expressed a renewed commitment to provide VA cancer data in California, and ultimately the VA resumed reporting in July, 2009, after the OCR and its regional registries implemented the confidentiality/data security requirements that the VA required. In this process, the GBACR was the only registry in the country to adopt an approach of statutory interpretation together with an explanation of the benefits of cancer registration in approaching a nation-wide policy issue. 4. Quality Control Throughout the contract years, the GBACR has had, as a top goal, the accurate, complete, and timely collection and reporting of cancer data. To meet this goal, we employed several quality control and improvement activities, including training and education of regional cancer registrars. The components of our comprehensive quality control/quality improvement program are detailed below. a. Visual editing GBACR staff performs quality control by visual editing, which is the process of visually reviewing case information by comparing text information with coded data and resolving any edit errors. This quality control work involves the following components. i. Visual editing of data submitted by reporting facilities The GBACR staff visually edited every cancer case submitted until the fall of 2007. At that time, the OCR reduced the proportion of cases required to be visually edited from 100%, to accommodate cuts to the state budget. The first revision of the standard specified visual editing of all COR-required data items for 10% of the cases from high-quality abstractors (accuracy rate of 98% or higher) and for 100% of the cases from the remaining abstractors. In late February 2008, the OCR, in response to further budget cuts, introduced a random-sampling plan that automatically selected for visual editing of approximately forty percent of all cases but left visual editing at 100% of all cases with uncommon or unusual sites thymus, unknown primary site, other and ill-defined, etc.), class 3 cases, cases with edits, and cases that required review based on collaborative stage conversions or SEER treatment errata. Even as these standards changed, GBACR Quality Control staff visually edited all selected cases, looking for missing data items, inconsistencies between codes and text, and other problems. Cases that lacked supporting text and/or codes, or exhibited inconsistencies between text and code, were fixed by the visual editor or a query was made to the abstractor submitting the case when there was conflicting or suspected missing information. Changes made to coded fields were automatically recorded on the Eureka Visual Editing Discrepancy Report, which shows the original value, the corrected value and text?based comment areas for the visual editor and the abstractor. When a discrepancy was identified by visual editing, the editor provided constructive and educational feedback in the comments section. Errors or other problems, such as inadequate documentation, noted during the visual editing process were used to plan our abstracting and coding workshops. Accuracy rates were tracked for every regional facility and individual abstractor. When the rate of data accuracy consistently dropped below 97% for a given abstractor, we provided that abstractor with error?specific 9 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 training, including a one-on-one session and direction to resource documents to enhance the visual editor?s understanding of the abstracting and coding rules for speci?c data items. ii. Visual editing of data collected by internal staff On a basis, we assessed the quality of a random sample of data collected by our internal Oncology Data Specialists (ODS) who perform cancer reporting services at the six reporting facilities that contract with the GBACR for this purpose. The goal of this activity was to achieve a consistent and sustainable high level performance across our ODS staff, with uniformly accurate interpretation and understanding of SEER's coding guidelines. The concept and procedures for this quality control process were the same as for the visual editing process described above for data coming from abstractors at the reporting facilities. Visual editing of a sample cases previously edited (peer review) On a basis, quality control staff visually edited a random sample of cases recently visually edited by another quality control staff member. The purpose of this activity was to evaluate and ensure the consistent and uniform application of coding rules and procedures among our quality control staff. We provided feedback, with an opportunity to discuss and challenge any discrepancy, and additional training was provided when needed. b. Data quality improvement activities performed in the CCR DMS When edits or data inconsistencies arose that could not be resolved by automation, they were evaluated and resolved manually through several processes, including linkage resolution, merging of cases, consolidating of case information, processing of correction records, processing of hospital deletion records, and processing of potential duplicate cases. A summary of these activities is included in Appendix 39. c. Quality control activities with Surveillance Research During this contract period, the GBACR Surveillance Research and Registry Operations groups jointly developed a quality control program that identified areas for improvement in our data collection and reporting. The two staffs worked together to evaluate inter-record patterns of the database as a strategy for understanding completeness, timeliness and accuracy. d. Special quality control activities An important component of the GBACR's overall strategy to main high-quality cancer data has been full support and active participation in all SEER and CCR quality control activities, a number of which are detailed as follows. We actively participated in the various SEER quality control/quality improvement activities, such as SEER Ql video teleconferences, SEER workshops held in conjunction with the annual meeting of the National Cancer Registrars Association, all of the Reliability Studies CS, and Hematopoietic), and the beta testing of the electronic SEER Program Code Manual. GBACR staff participated in SEER audits, the last of which was the 2004 Casefinding Quality Improvement Audit finalized in December 2007. Several deficiencies in our casefinding procedures were discovered that resulted in 12 late cases and 76 missed cases. The 10 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-35136 deficiencies included: 1) not realizing a large university-based hospital had segregated the hospital (in- and out-patient) pathology reports from the non-hospital reports, which resulted in the non-review of the non-hospital reports; 2) a GBACR casefinder misinterpreting the reportability rules and failed to identify VIN cases; and 3) the late cases having an incorrect year of diagnosis that was not detected and corrected by GBACR staff. Corrective action included: 1) incorporating the non-hospital pathology reports review in the casefinding procedure at the university hospital and implementing a yearly survey of each pathology department to determine if practices have changed and recording the source of their specimens; 2) additional training on the reportability requirements; and 3) annual electronic matching of the year of diagnosis in our casefinding system with the year of diagnosis in the corresponding case in the statewide database, and making the appropriate corrections. i. SEER reliability studies During the contract period, GBACR quality control staff routinely participated in the SEER reliability studies and the reconciliation process through 2008. In 2008, all GBACR staff involved in data collection and quality control activities participated in the 2008 Multiple Primaries and Histology and the Collaborative Stage (CS) Reliability Studies. The results are summarized below: 0 2008 MPIH reliability study Results of the 2008 MPIH reliability study indicated that GBACR staff scored well for bladder head and neck and brain sites (89.81%) compared to the All SEER Registries and the All Participants scores. The lowest score was for Other Sites The GBACR average score for the all cancer sites category was 80.76%, which was lower than the average for All SEER Registries and All Participants. At the GBACR 2010 spring educational meeting, held during National Cancer Registrars Week and attended by the vast majority of registrars in our region, we conducted a comprehensive session on the rules including several case exercises the attendees were required to code. The cases were extensively reviewed with the group; the preferred answers were discussed along with the rationale for each answer. 0 2008 Collaborative Stage reliability study Results of the 2008 CS Reliability Study indicated that GBACR staff scored higher than All SEER Registries and All Participants for prostate cases our lowest score was on Lung cases The GBACR average score for All Cancer Sites was 77.95%, which is lower than, but close to, the average for All SEER Registries and all Participants. 0 2010 Hematopoietic and Neoplasm mini-reliability study This study was completed in March of 2010. Results have yet to be published. 11 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-351 36 ii. 2010 Collaborative Stage Data Collection System. version 2 Coding Practice Studies ROU staff participated in the breast, colon and lung coding practice studies conducted in May and July of 2010 and will participate in the prostate and melanoma coding practice study scheduled for October. Data Availability Assessment In May of 2010, GBACR Quality Control staff coordinated the submission of 180 case abstracts from ?ve Bay Area hospitals for this assessment study. All participating hospitals submitted data using the SEER Abs* software application. iv. CCR re-abstractinq audit for data completeness and accur? In May of 2008, CCR conducted re-abstracting audits in several hospitals, two of which were in the GBACR (referred to here as Hospital A and Hospital B). Ten cases were randomly selected from each, with 17 data items re-abstracted for each case. Hospital A had an overall data item accuracy rate of 90.6%, with most discrepancies in the race/ethnicity and date of diagnosis data items. Hospital had an overall data item accuracy rate of 92.9%, and discrepancies were found in the race, primary site and tumor size data items. The audit findings were provided both to the hospitals and the GBACR. GBACR quality control staff contacted the hospital registrars and reiterated our availability to answer abstracting and coding questions; we also provided them with a list of training activities, and identified coding and abstracting guidelines that we recommended they review during the abstracting and coding process. GBACR quality control staff also participated in four COR-initiated recoding audits during the period December 2007 to June 2009. These audits used the peer review method to audit specific data elements for specific anatomic sites. The audit team consisted of a member each from the GBACR and CCR quality control staffs. The two auditors independently audited cases from another reporting region in California, reconciled any differences, and provided the results to the region audited. Three of these audits involved cases that were not visually edited (bypassed) and assessed the accuracy of the data submitted by the facility registrar. The accuracy rates for these three audits were: 94.8% for the Colon and Lung Recoding Audit; 95.9% for the Recoding Audit; and 95.1% for the Bladder Recoding Audit. The other two audits were conducted on cases that had been visually edited and assessed the accuracy of the visual editors that initially visually edited the cases. The accuracy rate for GBACR visual editors was 99.1% for the Ovarian Recoding Audit and 99.5% for the Benign Brain Tumors Recoding Audit. The results of the re-coding audits focusing on the quality of the visual editing of GBACR cases were extremely favorable for our quality control staff, with an accuracy rate of little over 99%. The audits focusing on the quality of the data in GBACR cases that bypassed visually editing ranged from 94.8% to 95.9%, below the standard of 97%. The CCR initiated the audits on cases that bypassed visual editing process to assess how the cuts to the OCR budged had impacted the quality of the data. To date, this assessment is still in process. In June of 2009, GBACR staff participated in a breast re-coding audit of cases that were not visually edited (bypassed) and assessed the quality of the abstracting. The overall average accuracy rate for this audit is 93.7%. The accuracy rate among the regions ranged from 88.8% 12 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 to 97.7%, with the GBACR at 92.4%. The results of this audit clearly demonstrated the need to develop training modules on breast cancer cases for abstractors and visual editors targeting the following data items: CS site specific factor CS nodes; coding grade, CS tumor size and histology. CCR undertook this activity. e. Quality Control challenges i. SEER extract In 2005, the GBACR undertook a major quality control project to identify the source of data quality issues in our 2004 data submission that SEER had brought to our attention. We identified the source of the problem, which was within our SEER data extract program as a result of conversion of code from Speedware to SQL during a change in platforms for our Merlin database. We then assembled a team to examine the extract code with a Iine-by-Iine review; to do this, we researched SEER documentation and past extract programs to help identify errors. As errors were identified, we corrected them and, further, produced a completely new program created in SAS using old code and verifying coding changes with SEER documentation. Our commitment to keeping a high level of data quality then led us to develop a part-time Quality Control Analyst position. Julia Song was hired in September 2006 and tasked to assist the quality control team to identify and resolve data quality issues on an ongoing basis, and to refine and document the SEER extract program to ensure that the data collected, processed and reported are of the highest quality. Subsequently, changes also occurred in registry leadership. As a result of this robust quality improvement program, we received recognition from SEER for overall data quality (second place) for our 2007, 2008 and 2009 data quality profiles. ii. Reduced visual editinq In September of 2007, the OCR lowered the proportion of cases required to be visually edited from 100%, due to cuts to the state budget. The first revision of the standard specified visually editing all OCR-required data items for 10% of the cases from high-quality abstractors (accuracy rate of 98% or higher) and for 100% of the cases from the remaining abstractors. In late February 2008, the OCR, in response to further budget cuts, introduced a random-sampling plan that automatically selected for visual editing approximately forty percent of all cases but left visual editing at 100% of all cases with uncommon or unusual sites thymus, unknown primary site, other and ill-defined, etc.), class 3 cases, cases that have edits, and cases that require review based on collaborative stage conversions or SEER treatment errata. By the end of September 2009, additional budget cuts are expected to lead to the replacement of the 40% sampling plan with a new plan specifying the visually editing of 100% of the cases from abstractors with an accuracy rate of less than 97%. Even after the visual editing sampling plan was put in place, the GBACR continued to identify and visually edit cases from our new and inexperienced cancer registrars, ensuring that their work was reviewed and that they received the feedback essential for improving their abstracting and coding skills. 13 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-351 36 2010 data chanqes The extensive 2010 data set changes based on the new version of the Collaborative Stage Data Collection System the 2010 Hematopoietic and Neoplasm reportability and coding rules and the new and/or revised data items contained in the NAACCR version 12 data set have represented a major challenge and concern for the timeliness and quality of cases diagnosed January 1, 2010 and forward. Our approach and progress to prepare the GBACR staff and community registrars are described in the section below. 5. Education and Training Education and training are integral to complete, timely and quality cancer data. a. Ongoing education and training opportunities - We conducted quarterly registrar meetings that included lectures by GBACR Surveillance Research scientists, who presented registry data research results, physicians and other staff, who addressed reporting issues, data set changes, training, etc. Also, we conducted coding exercises on new changes or difficult areas to ensure consistency and accuracy. The meeting agendas for the current contract period are included as Appendices 27-30. 0 We hosted the 2007?2008, 2008-2009, and 2009-2010 NAACCR Hospital Registry Webinar series, which were very well attended. We took advantage of the recorded webinars and hosted them at times convenient for our community registrars. Plans are underway to schedule the webinars in the 2010-2011 series. We also hosted the annual education meeting of the Northern California Cancer Registrars Association. 0 We provided comprehensive training to new GBACR oncology data specialists and quality control specialists. We provided remedial training to both internal and external cancer registrars when a pattern or opportunity for improvement is identified. 0 We assisted reporting facilities in identifying training needs for their personnel through the provision of visual editing discrepancy reports. a We provided technical assistance and expertise to staff in our cancer reporting facilities by a) responding to abstracting and coding questions via telephone, email and the RegWeb; b) researching difficult or unusual cancer case information; and 0) providing timely notification of changes in abstracting and/or coding rules and procedures, final SEER SINQ and CCR Inquiry and Response information, and training activities offered by other organizations. To address the 2010 data changes, GBACR quality control staff attended the Train the Trainer session in July 2009, all the ?dry run? sessions (thirteen videoconferences) in preparation for the release of the 2010 Hematopoietic and Neoplasm (HLN) Reportability and Coding Rules and HLN database, and the 2010 SEER Advanced Topics for Registry Professionals held prior to the National Cancer Registrars Association?s annual meeting in Palm Springs in May 2010. With this education and training, GBACR quality control staff conducted two workshops that covered colon, lung, prostate and breast tumors. The program was presented twice, on March 10, 2010 and on March 31, 2010, to ensure attendance by all internal and external cancer registrars. 14 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 Our most recent quarterly registrars? meeting (July 21, 2010) primarily focused on the Hematopoietic Neoplasm Reportability and Coding (hematopoietic manual and data base, and diagnostic con?rmation). Over 60 cancer registrars attended and gave glowing reviews of the content, which was presented by Shawky Matta, GBACR Quality Control Manager. b. Registry Compliance Program In the last contract year, the Program has initiated a series of ?Topics in Compliance? presentations for the Cancer Registrars who attend quarterly meetings and who work throughout the Greater Bay Area in order to establish a community of well informed experts in cancer data and the uses and applications thereof in surveillance and research. Upcoming ?Topics in Compliance? presentations will address the following: New Guidance from OHRP on Engagement in Human Subjects Research New Guidance on GINA Genetic Information Nondiscrimination Act) IRB Approvals for Research Uses of California Cancer Registry Data Look What the Stimulus Package Dragged In: New HIPAA Regulations on Breach Notification (and more)! - Cancer Registration and the Citizen?s Right to Privacy 6. Monitoring Completeness and Timeliness of Reporting a. Completeness of reporting As described above, the GBACR employs many sources of casefinding to identify potentially reportable cases. All casefinding documents are in an electronic format through the creation of an electronic record from a paper source or e-path. The electronic casefinding documents are first linked to the cases previously reported by GBACR reporting facilities. Electronic lists of patients who do not have a corresponding database case are then compiled and returned for further investigation to the reporting facility via the RegWeb. These potentially missed cases are tracked until receipt of an abstract or notification that the case is not reportable. Completeness of reporting is monitored using routine reports from the CCR DMS system. On a basis, we post a report for each reporting facility to the RegWeb, which automatically sends an email notification of the post to the facility cancer registrar. On a quarterly basis, we evaluate completeness based on the expected caseload for the facility against the actual number of cases received and then send a letter along with a case count report to the cancer registrar and the hospital administrator detailing their completion status. Facilities that fail to submit their expected caseload each quarter (Le, 25% in the first quarter, 50% in the second quarter, 75% in the third quarter and 100% in the fourth) and do not have a valid reason are then considered delinquent in their reporting. The letter requests the submission of the delinquent cases within a specified time period or, if this timeline cannot be met, submission of a plan of action detailing the steps they will take to become compliant with the reporting standards. The completeness estimates from the SEER Data Quality Report for cases diagnosed in 2008 and 2009 are as follows on the next page: 15 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 2008 2009 SFO SJM SFO SJM Total 100.09 97.82 87.95 84.9 Invasive 99.58 97.75 87.65 85.56 In situ 105.84 96.2 91.01 76. 74 b. Timeliness of reporting The CCR reporting standard requires that 97% of cases to be reported within the mandated six months of the time the case comes under the care of, or is admitted to, a cancer reporting facility, as described above. Physicians caring for cancer patients who are not referred to a cancer reporting facility are to report the case within 30 days of first seeing the patient for the cancer. Based on this standard, timeliness of reporting from the GBACR reporting facilities has been monitored during the contract period using reports routinely obtainable from the OCR DMS system. As with the Hospital Case Count Reports, a Hospital Submission Timeliness Report specific to each reporting facility is posted to the RegWeb for review by that facility?s personnel. Twice a year, we generate and send a Semi-Annual Timeliness Report to the reporting facility administrators with a letter noting the extent of their adherence to the timeliness standard. Facilities that do not meet the standard are requested to submit a plan detailing the steps that will be taken to bring their facility into compliance. These action plans are monitored and in time, the majority of the facilities become timelier in their reporting. During the last contract year, one delinquent facility was unable to hire additional cancer registrars to work on their backlog and GBA CR staff was dispatched and abstracted 380 cases over a three week period. Appendix 31, a Regional Hospital Timeliness Report for the first six months of 2010, shows that the timeliness of reporting of all GBACR hospitals is 75%, which is a 10% improvement over the last time period of July to December 2009 as depicted in Appendix 32. c. Completeness and timeliness challenges Unfortunately, our recent progress made in the timeliness of reporting will be negated because the OCR-generated software (CNEXT) used by GBACR hospitals to enter case information for 2010 diagnoses is not yet fully available. In June, CINET Solutions released CNEXT version 1.10, which does allow entry of 2010 cases. However, several components of the software are incomplete, such as the complete edit set, and the online help that links the user to reference documents. In addition, the software does not meet all of the specifications for 2010 case data, which are still being finalized by CCR staff. The final version of CNEXT is due to be released in August of 2010. Approximately one third of our reporting facilities chose to install version 1.10, and the remaining facilities are using the current version and collecting the 2010 data using the SEER abstract addendum generator or other manual methods. The CCR anticipates that the Eureka DMS (version 9.0) will be ready to receive 2010 data in late September 2010. CCR has suspended the monitoring of timeliness of reporting until September of 2010 and will re-evaluate at that time. Timeliness is calculated using the date of admission and the date the case is received at the registry. Reporting facilities have been instructed to abstract 2010 cases but to not transmit them to their regional registry. 16 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract NO1-PC-35136 Once the Eureka DMS is ready to receive 2010 data, it will be deluged with transmissions. Subsequently, there will be a backlog of cases to be processed (visually edited, linked, consolidated, etc.). 7. Follow-up Methods (Passive and Active) Throughout the contract years, the GBACR employed both passive and active follow-up methods to obtain the vital status for all cancer patients, regardless of date of diagnosis. Passive follow-up involves linking GBACR records with various electronic files from state and federal agencies, and also applying follow-up records received from reporting hOSpitals with ACOS-accredited cancer programs. Active follow-up includes obtaining the vital status of patients through subject searching services, physician contacts, and research involvement. For the San Francisco-Oakland region, both active and passive follow-up methods have been used for all cases since 1973. For the San Jose-Monterey region, follow-up from 1988 through 1991 was conducted only with information from ACOS-accredited hospitals, since at the time it was not required by the CCR. However, once the GBACR received the SEER contract for this catchment area (1992), passive follow-up methods were applied to all cases, and active follow- up methods have been used for cases from this four-county catchment area since 1992. The CCR routinely links all cases in the statewide database diagnosed since 1988 with files from state and federal agencies, and GBACR data are included in these linkages. Until 2009, SEER ?legacy? cases those diagnosed in the period 1973 through 1987) were submitted independently by the GBACR for linkage with the National Death Index (NDI), Social Security Administration and the Center for Medicare and Medicaid Services (CMS). Since then, the CCR has included the 1973-1987 GBACR legacy data in its linkage files. Results from CCR and GBACR linkages are loaded into Eureka using the passive follow-up record layout. a. Passive follow-up For passive follow-up, GBACR data are electronically linked with databases from the following entities: i. Acute care and outpatient facilities (hospitals) Hospitals that do their own follow-up submit the follow-up data directly to the GBACR using and GBACR staff uploads the files to the CCR DMS. The hospitals obtain follow-up from a variety of sources, but primary from the patient?s readmission to the hospital or from the patient's physician. ii. California Department of Public Health, Office of Vital Records The CCR uses probabilistic linkage algorithms to match passive follow-up records to cases in the statewide Office of Vital Records database. CCR performs this linkage from incremental death files and the yearly Statistical Master File of Deaths ("Stat Master") ?le to obtain vital status information. Results from the yearly linkage to the Stat Master file are matches, possible matches or non-matches. The majority of the cases are linked automatically, and the incremental files are automatically applied to the statewide database cases when the linkage algorithm finds an exact match, including cause of death; the case is then 17 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract NO1-PC-35136 consolidated with the existing case data. Possible matches are manually reviewed by GBACR staff, and matches found to be positive are applied to the statewide database. When the record is a non-match but found to have a cancer cause of death, follow-back is performed to check for a possible missed case from a facility or physician. Those found to have been missed are then abstracted. Any cases that cannot be identified through our follow-back procedures are then considered ?death certificate-only? cases. All other non-matches are discarded. Social Security Death Master File Another source of vital statistic updates for cancer cases utilizes the Social Security Death Index database provided through an agreement with the National Program of Cancer Registries (NPCR). The CCR performs this linkage on an annual basis and uploads updated vital status into the OCR DMS. iv. CMS Medicare and Medicaid Each July, the OCR submits a data file to CMS for linkage and CMS provides the date of death for the expired cases, and the latest transaction date for cases that are alive. Follow-up information is updated in the CCR DMS. v. Kaiser Permanente Northern California Twice a year, the Kaiser Permanente Northern California HMO, which transmits roughly twenty- five percent of our yearly admissions, undertakes a focused effort to update patient name, Social Security number, date of birth, date of last contact, and vital status in its system. This is accomplished by linking the Kaiser cancer registry database to various internal Kaiser systems. For the date of last contact and vital status information, Kaiser provides us with file containing several thousand records, which we link to the CCR DMS data. If we prove not to have a current follow-up date or vital status, we create follow-up records and apply the data. The updated information change in name, Social Security number, etc.) is submitted to us in the form of update (correction) record that is then applied to the OCR DMS. vi. Voter registration During this contract period, the OCR annually obtained voter registration records for the 58 California counties. This database included the date of registration and identifies the persons who voted in the most recent elections. For linked cases, the voter registration or election date was considered as the date these patients were alive, and their last follow-up was updated with this date. vii. California Department of Motor Vehicles (DMV) Twice a year, the OCR links all living cases in the statewide database to the DMV files. This linkage identifies any cancer patients with any activity related to their driver?s licenses, such as renewals or traffic tickets. The date of this DMV activity is considered a date that the patient was alive. 18 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract NO1-PC-35136 California Birth Statistical Master Files On an annual basis during the latter half of the contract period, the OCR has linked all living female patients under the age of 65 in the OCR database to a state file of the birth certificates for a given year. Any woman listed as a mother on a birth certificate is assumed to have been alive as of the date of the birth. This linkage has proven to be helpful for follow-up of female patients diagnosed with cancer prior to or during their child-bearing years. ix. National Change of Address Once a year, the CCR links cases to the National Change of Address file. This gives the registry new addresses for many cases, which then aid in any passive and active linkages for follow-up. This information is held in afield separate from address at diagnosis. x. Office of Statewide Health Planning and Development (OSHPD) The California OSHPD provides to the public datasets of inpatient information collected from licensed hospitals in California. For each inpatient discharged from a California-licensed hospital, the record includes general acute care; acute chemical dependency recovery; and health facilities. Linkage of these files with the statewide database is conducted annually by CCR. xi. Social Security Administration (SSA) Once a year, CCR submits a data file to SSA to obtain Epidemiology Vital Status data, and the statewide database is updated with current information. xii. National Death Index (NDI) Yearly, the CCR matches patients lost to follow-up and those expired with an unknown place of death with the NDI files. Positive matches provide coded cause of death and place of death, which is then applied to the statewide database. GBACR staff creates files for our 1973-1987 patient set for inclusion with the files submitted by CCR. This linkage is performed under an interagency agreement. Funded Research studies Throughout the first six contract years, our Surveillance Research group transmitted follow-up information on all GBACR patients eligible for ongoing research studies twice yearly. For these patients, study investigators were asked to submit updated vital status and, for living patients, recent contact information. ROU linked the research data follow-up file with the CCR DMS data and created passive follow-up records for patients who have a more current date of last contact and/or vital status from the research contact. In the last contract year, we contacted investigators about increasing the frequency of this process to three times annually so that we can further capitalize on this important and cost-effective source of vital status and updated contact information. 19 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-351 36 b. Active follow-up Throughout the contract years, the GBACR has also conducted active follow-up activities. One component has involved obtaining information from physicians. Whenever the GBACR writes to a physician for other purposes, we inquire about the vital status of the patient. For patients reported to the GBACR only by a physician, lists of those lost to follow-up are generated, and follow-up information is obtained when GBACR staff visits the physician offices to abstract physician?only cases. For follow-up for pediatric cancer patients, who traditionally have been difficult to locate, our practice has been to complete all of the linkages mentioned above, process all of the follow-up records from the reporting facilities, and then run ?lost to follow-up" listings for the three hospitals that diagnosis and treat the vast majority of patients in these age groups. We request that these hospitals access their various internal systems to see if the patient has been seen as an in- or out-patient. If the hospital is unable to perform this activity, GBACR staff visits the reporting facility and obtains the needed information. In our region, follow-up of pediatric cancer patients is particularly challenging because this group includes patients diagnosed as far back as 1973, now 37 years ago. Therefore, in 2008, we enhanced our active follow-up for this group by applying subject-searching methods we have developed for research studies. We used a combination of Acxiom (Insight America), MasterFiles, and Haines CrissCross internet resources to search for patients and/or patient families, guided by specified standards for considering information we identify to be new. Acxiom links to the Social Security Death Index, both Acxiom and MasterFiles link to property title records, and MasterFiles links to ?real-time? telephone listings. Using these databases to search for updated information on 1,110 non-Kaiser patients diagnosed with cancer as children or young adults and considered lost to follow-up, we found some new information for approximately half the patients. This included an updated date of last contact and vital status on 343 and a correct Social Security number on 38 of these patients; this latter information was very useful because it greatly enhances the success of finding matches for the other linkages described above. We continue to use this approach for our pediatric cancer patients and other difficult to follow patients. According to the SEER Data Quality Reports for the GBACR for the years 2006-2009, summarized below, the follow-up rates consistently met the contractual standard. Goal 2006 2007 2008 2009 Follow-up age <20 390*; 380? 92.42% 92.05% 92.52% 94.91% Follow-up age 20-64 390*; 380** 96.94% 96.80% 96.38% 96.55% Follow-up age 65+ 395*; 98.83% 98.80% 98.82% 98.64% Follow-up all in situ diagnosis 390*; 380** 98.09% 95.79% 94.96% 95.12% *Contractual standard; Minimum acceptable For the last contact year, many of the activities mentioned above have been completed or are currently underway as we prepare for our November 1, 2010 data submission. 20 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 c. Follow-up challenges In June of 2010, SEER notified us that the 2008 National Death Index (NDI) file would not be available until August/September. The data we receive from NDI linkage impacts our follow?up rates and the unknown cause of death quality indicator. The CCR routinely re-Iinks previous California vital statistics files so we are assured that information for all deaths occurring in California is loaded into the statewide data management system. If we are unable to link with the 2008 NDI file, we may not be able to meet the data quality indicator for unknown cause of death. 8. Registry Information Management Systems At the outset of the current contract period, GBACR data management occurred through use of two in-house systems. 1) The first system, called "Merlin", was written in Speedware using Image (Hewlett Packard?s database) and was used for processing, managing, and reporting cancer data. The system allowed GBACR staff to upload new cases, follow-up and correction records; edit; update patient and tumor-specific information; and prepare SEER data submissions. 2) The second system, a casefinding system, also was developed using Speedware software to be the recipient database for all pathology-based casefinding documents collected using our casefinding data entry software application and for the electronic pathology reports received via AIM e-path software. The casefinding system works in tandem with the Merlin data management system to match the casefinding documents with cases reported by our cancer reporting facilities. The casefinding system also generates lists and letters used in the follow-back for missed hospital cases and for data collection of non- hospital cases. In 2003, the CCR mandated that reporting of cancer cases in California occur via Eureka. Thus, our SEER San Francisco-Oakland and San Jose-Monterey registries, which represent two of the regional registries of the CCR, were required to migrate our data to Eureka. At that time, SEER leaders indicated that the GBACR would not be required to go onto SEER DMS. Therefore, starting in the summer of 2003, the GBACR migrated its database to Eureka by loading all data from Merlin to Eureka, which became the central cancer data storage and process system for GBACR data loading, visual editing, follow-up, and reporting. However, because the Eureka casefinding system was not completed during the contract period, the GBACR has continued to maintain its own casefinding software. During the current contract period, the GBACR registry information systems have gone through a significant evolution and integration process driven by: 1) the evolution in computer technology, 2) the desire to integrate many small and isolated applications into one automated system, and 3) the successful development and deployment of Eureka as a centralized data repository and process engine for all SEER regions in California. As a result of this integration and evolution, GBACR production and operations became supported by a federated information service infrastructure consisting of the OCR DMS and the GBACR production system (see figure on next page). Each entity remains autonomous, but information is interconnected, shared, and interoperable based on SEER and NAACCR standards. Security is the individual responsibility of CCR DMS and GBACR production, but both entities follow National Institute of Standards and Technology (NIST) and the Federal Information Systems (FIPS) standards. 21 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 GBACR . ?#111 RegWeb . .. w? Eureka at Replica DMS . Doctor DB Case Me?'n Finding Ill 1* littmi i Figure 1. Federated Information Services Structure between OCR and GBACR in 2010 During the contract period, the ultimate combination of DMS with GBACR local applications has been advantageous for the efficiency and effectiveness of registry operations. Centralization has been helpful, as it allows us to use DMS as a central cancer data repository for easy sharing and of processes with other California regions. However, having local GBACR systems were useful for regional communication, processing pathology data, and performing both casefinding and facility follow-back. As part of these systems, he GBACR staff developed two communications options. One is RegWeb, a secure web portal that creates a popular and versatile communication window between GBACR and all regional hospitals, surveillance study groups, and other California SEER regions for daily data exchange, data sharing, reporting, viewing, and discussions. Another is the SEER ROCS report system, developed in 2007, to track registry work load and time allocation according to the SEER-required functions. The GBACR shared this system with OCR and hosted ROCS services for four of the OCR regional registries. a. Data entry system for case?nding During the contract years, the GBACR has used a PC-based, stand-alone data entry application developed in-house to collect (abstract and code) pathology information from hospital pathology departments and free-standing pathology laboratories and to identify cases for our early case ascertainment (ECA) studies. Over the last several years, this data entry system 22 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser. Principal Investigator Contract N01 -PC-351 36 was redesigned in the .net platform that streamlined workflow and allowed for the efficient processing of the casefinding documents through the following features: it became web-based and includes additional fields to collect more patient, tumor, and physician information that facilitate linkage of these reports to cases submitted by reporting facilities and to follow back to the appropriate non-hospital reporting entity. In addition, it automated several aspects of this process, including the automatic uploading of pathology records from the laptop to the Pathology Data Management system (PDM), which is the repository for all of the pathology data collected or received by the and the receipt of new and updated information from the physician database (name and license number of new physicians, updates to addresses, specialty, etc). During the last contract year, this application was modi?ed to allow easy identification of pathology reports for other California regional cancer registries. Also, the application was migrated from Visual Studio 2005 (ASP .Net 2.0) to Visual Studio 2008 (ASP .Net 3.5), which allows for the use of new features for data handling and user interface development. b. Pathology data management system (PDM) During the contract period, the PDM, which was initially designed for the receipt and classification of electronic pathology data, also undenrvent extensive redesign. Phase I of the PDM redesign, completed in 2008, consisted of re-programming the application to share the Microsoft Office .Net framework with the data entry software and the GBACR web portal. Using the .Net framework allowed for a common infrastructure and provides the ability to achieve a data model that will accommodate the three record formats existing in the PDM: the old record format, the new record format, and the e-path record format. Phase II, completed in during the last contract year, consisted of integrating into the PDM a stand-alone, currently manual process that consolidates multiple pathology reports for the same tumor into one report that is subsequently submitted to the our local case?nding system for linkage with all of the GBACR cases. The report function in PDM was enhanced to include a productivity report for the user that detail the number of reports submitted for a specific facility and the number of electronic reports reviewed. A report based on cancer site was also added to aid in the identi?cation of particular cancer sites for use to respond to inquiries by researchers. These redesigns significantly increased the efficiency of these systems and eliminated several time-consuming manual operations associated with the collection and processing of case?nding source documents. c. Electronic pathology reporting Registry Information Systems (RIS) staff played an integral role in the GBACR effort to implement electronic pathology (e-path) reporting from regional hospitals. Our e-path process involves hosting the e-path reporting application called TransMed, developed by AIM. RIS staff has worked with the AIM e-path team and assisted in the planning, testing, and implementation of new e-path installations. RIS staff also regularly monitored the TransMed software, ensuring timely data integration into the PDM for processing by data collection staff. During the last contract year, improvements were made to several processes including automating the various steps to load incoming messages and generating a report of received and duplicate messages by facility. Improvements also were made to the programming logic for the de-identified and identi?ed Kaiser messages, and the process to generate a request to Kaiser for protected health information was automated. 23 SEER Final Report: August 1, 2003?July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 d. Web portal (RegWeb) The web-based portal application, RegWeb, was implemented in March 2007. This secure web portal allows internal and hospital registrars to upload files to the internet, respond electronically to visual editing discrepancies, and download weekly case load list reports as well as timeliness, accuracy, and case count (completeness) reports. We expanded the use of this application to include information exchange associated with death clearance follow-back; lists of patients and their associated death certificate images are posted for the reporting facilities, and completed follow-back is returned by the facilities via RegWeb. RegWeb also has been used by CPIC researchers and the OCR to securely send and receive confidential data. RegWeb continues to prove to be an efficient, timely, and cost-effective means for communicating with our constituents. During the last contract year, new administrative reports were added to track users added or deleted and to manage the files uploaded and downloaded. The RegWeb application was migrated from Visual Studio 2005 (ASP .Net 2.0) to Visual Studio 2008 (ASP .Net 3.5), which allows for the use of new features provided by Microsoft for data handling and user interface development. e. GBACR doctor database Throughout the contract period, the GBACR maintained a database of physician names, addresses and California license numbers for use in data collection, casefinding and follow-back activities. This application also was used by the CPIC researchers for physician contact associated with research studies. Because the doctor database application was initially developed using Access, a Microsoft application, it had had many performance issues, such as slow pop-up windows, slow response time, difficulty navigating the various data fields, and the need for some manual processes. During the last contract year, the RIS team rewrote the application to use the .Net technology. The user interface was redesigned, including the login, adding a new doctor and search features. Performance was greatly improved for searching, listing and saving new entries and updates. The program between the doctor database and the case?nding system was rewritten and now automatically loads new and updated information into the casefinding system. These changes significantly improved the workflow associated with the doctor database. f. SEER Registry Operations Cost Study (ROCS) reporting tool To facilitate the ROCS-mandated documentation of time expended for each SEER registry function, we developed a web-based reporting application. We implemented this tool in November 2007 but continued to enhance it, providing a friendlier user interface and reports to assist management in monitoring and entering information in the SEER ROCS CTS system. We shared this application and associated expertise with the Los Angeles SEER program and with some of the OCR regional registries that are part of the Greater California SEER program. 9. Production data To comply with the new security requirements associated with the receipt and storage of Veterans Administration (VA) data, the RIS team collaborated with CCR information system staff to identify, implement, and test an acceptable solution. During the last contract year, we 1) migrated to SQL Server 2008 Enterprise edition to utilize total database 24 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-35136 2) designed an key infrastructure in?uencing backup requirements and security protocols for storage and management of the keys; and 3) databases in the SQL Server 2008 infrastructure that contain con?dential patient data. Migration and of the remaining SQL Server 2000 and 2005 databases are scheduled to be completed by October 2010. h. Casefinding system The GBACR casefinding system has been the data warehouse for all casefinding documents. It uses algorithms to match incoming casefinding documents with documents existing in the casefinding database, to match incoming casefinding documents to cases already reported, and to produce follow-back letters, listings and reports. Throughout the contract period, over 100,000 casefinding documents annually have gone through this match that we call "resolution." Typically we have sent over 750 cases back to the facilities to either abstract and submit a case, or inform us why the case is not reportable or that the case is not their reporting responsibility. This matching process has been our primary method for identifying physician-only cases. i. Eureka and GBACR casefinding systems The Eureka casefinding system was initially developed to receive and process electronic pathology reports from the AIM e-path installations in the Greater California SEER region, and to process death certificate information and perform the death clearance follow-back process. While the Eureka death clearance follow-back function is used by GBACR staff, the Eureka casefinding system during the contract period othenNise did not meet the casefinding processing needs of the GBACR, which generates electronic reports for its 100% casefinding efforts (over 100,000 reports each year). Because the efficiency and completeness of GBACR casefinding activities required an adequate casefinding system to manage these processes, and because the time frame for completion of the Eureka casefinding system remains uncertain, the GBACR had planned to continue to use our own casefinding system in the upcoming contract period until such time as the Eureka casefinding system was completed to our specifications. Since we cannot anticipate when that will occur, we planned to upgrade our legacy casefinding system in the first year of the upcoming contract period so that it can continue to function until the Eureka casefinding system is completed. However, during the negotiation phase of the CPIC proposal for the contract period beginning August 1, 2010, SEER expressed concern about our ?hybrid? system and declined to support any additional upgrades to the registry legacy systems. Consequently, the above mentioned work effort for our own casefinding system along with the associated staff resources were eliminated from the statement of work and budget for the next contract year. SURVEILLANCE RESEARCH During the last contract period, the GBACR was very active in data dissemination and research facilitation, and in original registry data-based research. We made our SEER data widely available for conducting regional and national cancer epidemiology studies, as evidenced by the large number of published scientific manuscripts using GBACR data (Appendices 33-34), and research studies that accessed confidential data through Letters of Agreement. Our data have contributed routinely to linked databases, Such as the SEER-Medicare and the SEER-NLMS databases, and to special data linkages such as the National Registry matches and 25 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 ongoing National Transplant Registry matches coordinated by the NCI DCEG Infections and Immunoepidemiology Branch. GBACR data continued to be included in the International Association of Research on Cancer?s publications ?Cancer in Five Continents.? Over the contract period, GBACR Surveillance Research staff also used the data extensively and creatively for the development and testing of etiologic hypotheses and cancer control efforts, supported by either our core SEER contract or by SEER RRSS, publishing a total of 96 papers in peer-reviewed scientific journals. Our Surveillance Research investigators continued to serve as national leaders in identifying and studying issues of quality and completeness of specific registry variables and of the feasibility of collecting additional variables (race/ethnicity, socioeconomic status, language, treatment details, malignancy subtypes, and body size). This work has led to broader and more coordinated efforts to improve these data in the SEER and the national cancer registry community. Over the contract period, the GBACR participated in 29 SEER Rapid Response Surveillance Studies (RRSS) (described in Appendix 35, addressing questions of etiology, treatment, screening outcome, and tissue availability, among other topics; many of these studies involved collaboration with other SEER registries. We also coordinated other special studies including the annual Patterns of Care data collection and the Adolescent and Young Adult Health Outcomes and Patient Experience Study. During the last contract year, Surveillance Research staff continued with these activities. GBACR Surveillance Research staff responded to many requests for registry information, facilitated confidential data release for use in 24 studies (Appendix 36), and continued to provide an active Early Case Ascertainment system for investigators. From January 2009 to the end of the contract, at least 260 manuscripts, correspondence, and book chapters utilizing data collected by the GBACR, including 27 manuscripts funded by our SEER core and RRSS contracts, were published in the medical literature, predominantly peer-reviewed scientific journals. We participated in four SEER RRSS, and several multi?registry research and strategic planning efforts and offered six new ideas to the RRSS enterprise in the last contract year. The Surveillance Research staff also made numerous scientific and community presentations involving registry data. 1. Dissemination of GBACR and Related Data We consider one of our major successes during the contract period to involve the substantial use and dissemination of our local and national SEER data, as follows. a. Production and dissemination of annual reports Prior to 2008, the GBACR annually produced a 400-page incidence and mortality report describing the distribution of cancer incidence and mortality for 32 anatomic sites (including 27 specific sites, four leukemia subtypes, and all sites combined) by age, sex, race/ethnicity, year of diagnosis and county. In addition to distributing hard copies of this report, we advertised its availability via distribution of postcards, media press releases, and website postings. In 2008, we initiated a companion report, ?Cancer Risk Factors and Screening Behavior in the Greater Bay Area: Data from the California Health Interview Survey,? also available on our website, that described the regional and county-level distribution by sex and race/ethnicity of selected cancer risk factors, including physical activity, ovenrveightlobesity, hormone replacement therapy use, smoking, alcohol consumption, and cancer screening practices including mammography, Pap screening tests, and colorectal cancer screening. As users have been able to access county-, 26 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-351 36 sex-, race/ethnic- and age-specific cancer incidence and mortality rates directly from the OCR website, we ceased the production of the annual incidence and mortality report in 2008. Instead, we have been producing a yearly, more lay-friendly executive summary of yearly changes in incidence and mortality and distributing it through our CPIC e-newsletters, web and social media outlets Facebook, Twitter), press releases, and various other channels facilitated by our organizational Communications Department. In the last contract year, we produced several press releases summarizing highlights from our annual reports and disseminated them to the local media and our social media outlets. Working with the science writer hired by CPIC, we also produced more engaging communications detailing findings from our surveillance studies for similar distribution. b. Fact sheets Throughout the last contract period, the GBACR annually prepared Fact Sheets on the cancer sites being studied in a given year as part of the SEER Pattern of Care studies. These reports, mailed to approximately 1,200 Greater Bay Area physicians participating in the studies or with the same specialty and having cared for patients in our registry, described site-specific regional trends in incidence and mortality. In the past year, we produced and disseminated two fact sheets. c. Data access/communication In the past contract period, we made advances in directing many telephone callers and internet browsers to the California statewide interactive tool for obtaining the most recent county-level incidence and mortality data for our nine-county region. In the last year, we further updated our data release policies and procedures and sought to provide more clear and updated documentation of all GBA CR and CCR policies and procedures. In addition to improving access to online cancer statistics, we initiated new collaborations to develop more effective means of communicating cancer statistics to our various constituencies. Little attention had been paid to maximizing the impact of cancer statistics through communication science, but strategic visual display of surveillance data to the public, policy makers and news media may increase awareness and social action for cancer control. In the last year, Dr. Clarke has worked closely with Dr. Matt Kreuter, Founder and Director of the Health Communication Research Lab at Washington University in St. Louis to develop effective interactive internet displays that reveal cancer registry data with ef?ciency and clarity, at multiple levels of detail, capturing multivariate complexity, and showing how data change over space and time. As part of Dr. Kreuter?s research on cancer risk communication, we are conducting randomized experiments to evaluate the effects of at least four variants of visual display of cancer surveillance data on six types of reactions among the US general population. Our goal is to use the results of this research to advise cancer registries, including ours, on how to improve the visual impact of local cancer statistics. 2. Surveillance Research Conducted by Surveillance Research Staff Over the past contract period, GBACR surveillance researchers have addressed a broad range of regionally and nationally relevant cancer issues, employing state-of?the-art analytic methods. 27 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 a. Improving the utility of SEER data through research One major research focus has been on improving the utility of SEER data. Through using SEER data in analyses to address issues relevant to cancer prevention and control, we have identified questions about the quality and/or availability of relevant registry data items. These practical questions have led to the generation of research studies, many funded by SEER RRSS. These database quality/enhancement studies have focused on data item accuracy and completeness, introduction of new variables, completeness of case ascertainment, denominator development, and linkages, as summarized below. i. Accuracy/completeness We conducted studies to evaluate the quality of certain registry data items including race, ethnicity, as well as to determine the variability in hospital policies and staff practices in the collection of patient sociodemographic information. We also developed and tested methods to determine patient nativity and timing of immigration and develop associated population estimate data. ii. New variables We examined the feasibility of collecting proposed new data items relevant to cancer control, including patients? body size, primary language, and socioeconomic status. Completeness of case ascertainment With support from a SEER RRSS, we documented the extent of melanoma underreporting and developed strategies to remediate it, as described further below. As part of separate work, we evaluated how better to ascertain hepatocellular carcinoma cases reported to the registry with delayed, incomplete, or poor-quality data. iv. Denominator development We undertook significant work in this area. In collaboration with a demographer, we developed population estimates needed to compute cancer rates among the oldest-old, which we then utilized as part of a SEER RRSS. In other work, we developed annual population estimates needed to compute cancer trends among Asian and Hispanic immigrants. v. Database enhancement throuqh linkaqe Under Dr. Gomez? lead, we assembled a database of contextual (neighborhood-level) data on the social, built, and immigration environments, which has been well-used to answer numerous research questions. We also linked SEER data to detailed treatment and comorbidity data from Clinical Research Networks (CRN), and to detailed treatment and comorbidity data from electronic medical records, specifically the OncoShare project, involving data from Stanford Medical Center and Palo Alto Medical Foundation. vi. Improving the completeness of melanoma case counts We facilitated complete and accurate melanoma reporting to enable ongoing characterization of this epidemic. Then, we disseminated information regarding reasons for and solutions to 28 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract underreporting, including a co-authored editorial on melanoma under-reporting in the Journal of the American Academy of Dermatology and the development of a website to provide dermatologists with helpful information and links regarding their cancer registry reporting obligations and mechanisms for reporting. b. Other work initiated by Surveillance Research investigators Over the contract period, we also focused on a number of specific research questions, many using the data resources described above. As the affluent populations within the Greater Bay Area have higher than average rates of breast cancer, prostate cancer, melanoma, and malignancies, we engaged in detailed surveillance studies of these cancers. Further, we focused on cancer incidence and outcomes in Asian populations. Finally, we also studied individua ?, institutional-, and contextual-level effects of immigration-related factors on cancer incidence and outcomes. These accomplishments are described in detail below and include information about specific milestones achieved during the last year of the contract. i. Breast cancer incidence patterns in Marin Countv and beyond Drs. Clarke, Glaser, Chang, and Keegan continued to monitor and characterize changes in breast cancer occurrence in both the local (Marin County, Bay Area as a whole) and national populations. Our work with this public health cancer crisis has allowed the GBACR to fulfill a strong public health function, to conduct significant surveillance research studies that informed breast cancer etiology, and to position our registry to mount other community-engaged and scientifically informative responses to unusual regional cancer patterns. Drs. Clarke and Keegan worked closely with the Marin C0unty Health Department to monitor closely breast cancer occurrence and risk factor patterns in the non-Hispanic white and Hispanic populations of Marin County. The results of these analyses were published in several peer-reviewed manuscripts and official reports. We also spent substantial effort responding to requests for information or data from the scientific, legislative, advocacy, and lay communities concerned about the Marin County excess of breast cancer. Some of our efforts to understand changes in risk factor distributions resulted in our submission of a SEER RRSS (2004 cycle) to aggregate and analyze risk factor data from epidemiologic studies and Kaiser Permanente. Using data from this RRSS, we were the first to detect and describe the rapid decline in breast cancer incidence observed immediately after the release of the Women's Health Initiative clinical trial results in 2002 and to relate them to the recent mass cessation of hormone therapy. We followed this report with an invited commentary on this topic in the journal Cancer Causes and Control, and an ecologic analysis showing a strong correlation between the decline in county- Ievel prevalence of estrogen-progestin hormone therapy use and the decline in county-level breast cancer incidence rates. We then reported on the extent of the decline of breast cancer in population subgroups defined by race/ethnicity, socioeconomic status, and rural/urban residence, using SEER, OCR, and NAACCR data. In 2010, we senior-authored a manuscript describing declines in hormone therapy use and breast cancer incidence rates in Marin County for the period 2000-2007 based on data from the independently funded Marin Women?s Study, On which Dr. Clarke serves on the steering committee. ii. Cancers in Asians Attentive to the high concentration and sociodemographic diversity of Asians in our SEER catchment area and throughout California, we have conducted surveillance research on cancer 29 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 incidence, diagnosis, treatment, survival, and mortality in specific Asian populations. During this contract, the GBACR Surveillance Research portfolio on cancer in Asians grew to cover a broad range of topics and approaches, in large part through our SEER RRSS work to develop denominator data and other methods for examining cancer trends in specific subgroups (SEER RRSS 2006 cycle). This portfolio included a detailed report and several scientific papers describing incidence trends for Asian populations (Chinese, Japanese, Filipinos, Vietnamese, Koreans, South Asians), as well as a book chapter by Drs. Gomez and West (Chapter 1 in Praeger Handbook of Asian American Health, 2009) reviewing variations in cancer rates across six Asian groups in California. Dr. Gomez and GBACR staff have been leading the current SEER effort to examine national trends in cancer incidence among several Asian/Pacific Islander groups. For this topic area, our registry?s public health effort has been mounted in response to the high rates of liver cancer among Asians, as Dr. Chang has been working in conjunction with the Asian Liver Center at Stanford University and local community-based Asian organizations to disseminate information about liver cancer prevention, screening, and control. We also have worked with the linked SEER-Medicare database to examine racial/ethnic disparities in colorectal, breast, and prostate cancers, with a focus on survival patterns among Asian populations. Dr. Gomez participated with Dr. Clegg and team on the SEER-NLMS papers. In the past year, our RRSS-funded work led to two published papers in special issues of the American Journal of Public Health on cancer incidence and survival among Asian immigrants. These papers have led to broad media coverage, including a presentation by Dr. Gomez at a National Press Club event. Over the last contract year, we have used the data to publish seven articles, and have several more that have been submitted or are in progress. Cancers in immigrant populations Because both cancer etiology and cancer control are advanced by a clear understanding how and why cancer rates vary between foreign-born and US-born Asians and Hispanics (the predominant immigrant populations in California), we have focused on this area in our SEER Surveillance Research work. Working with a demographer, Dr. Gomez led an effort to devise methods to improve numerator data for relevant variables and to develop denominator data needed to calculate cancer rates and trends by and immigration history, as she presented in a plenary session at the 2009 NAACCR conference. In addition to developing the methodologic tools needed to assess nativity and immigration among individual cancer patients, the GBACR Surveillance Research team worked to improve our capability to examine multilevel impacts of neighborhood, institutional, and individual-level immigration and racial/ethnic factors on cancer incidence and outcomes. With SEER RRSS funding (2006 cycle), we developed the data used to characterize contextual and institutional race/ethnicity and immigration factors by using US census data to derive neighborhood (block group) measures of Asian and Hispanic ethnic enclaves, compiling census-based measures of residential segregation, and developing measures of patients? racial/ethnic and socioeconomic composition at the hospital level. In the last contract year, we have expanded our research portfolio on studies of cancer in Hispanic migrants and conducting descriptive migrant studies among Hispanics, as we have already done for Asians. This work has led to five publications, and at least eight more currently in preparation. iv. Role of contextual environments on cancer In this contract period, Drs. Gomez and Keegan, together with a range of regional epidemiologists, geographers, and geospatial statisticians, have built a multi-disciplinary 30 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser. Principal Investigator Contract N01 -PC-351 36 program in cancer surveillance studies of the neighborhood social, built, and immigration environment. Through support from a SEER RRSS study (2003 cycle), Dr. Gomez and her team compiled existing geospatial- data characterizing neighborhood social, built, and immigration environments and are currently writing a manuscript describing the specific measures in the data resource and the value of the data for enhancing cancer surveillance. As above, we have leveraged the support of SEER RRSS funds for building this resource to yield very productive subsequent research on the effects of the social and built environment, segregation, discrimination, immigration, and other individual- and contextual-level factors on breast, prostate, and colorectal cancer risk, stage, treatment, survival, and quality of life. The availability of this SEER-supported data resource also has generated considerable interest among other researchers, such that the resource has become a platform for six currently funded studies multilevel studies of cancer occurrence and outcomes, and several more being reviewed. In the last contract period, we have initiated four newly-funded studies that incorporate the contextual data to examine the association of neighborhood factors with breast and prostate cancer incidence and outcomes. Three more grants are currently pending. Three manuscripts are currently in preparation. v. Surveillance of maliqnancv subtypes The GBACR surveillance research group has had a long and productive involvement in the descriptive epidemiology of malignancies, including Hodgkin and non- Hodgkin subtypes. In the last contract period, Drs. Glaser, Clarke, Chang, and Keegan have evaluated outcomes related to EBV tumor status (based on previous SEER RRSS funding), and identified the appropriate classification and categorization of individual subtypes by ICD-O-3 type research so as to address a major consideration for the meaningful study of malignancy subtypes. For the latter work, Drs. Clarke and Keegan completed two RRSS-funded efforts (cycles 2004 and 2007) to better understand the reasons for poor reliability of ICD-O-3 histologic coding for histologic subtypes. We found that complex referral patterns may complicate the abstraction of the final histologic code from consult reviews, and that electronic pathology reporting will greatly aid the reconciliation of histology codes in the setting of multiple consult reviews. Through such efforts, we have continued our leadership in the refinement of surveillance methods for monitoring malignancy subtypes. In addition, we have worked actively to bring an appreciation of the utility of SEER data to, and to integrate our research efforts with, the NCI-sponsored consortium, an international working group of epidemiologists. In the last year, we have completed a preliminary manuscript describing the results of our recent RRSS, presented these findings to the regional cancer registrars, and cooperated with our pathologist collaborators as part of the consortium to update the recommended groupings for use in epidemiologic studies of malignancies. OTHER ACTIVITIES OF NOTE 1. Registry Leadership During the last contract period, GBACR leaders were active participants in the regional and national registry community. Ms. Davidson-Allen was a member of the CCR Strategic Planning Group, the NAACCR Pathology Data Workgroup, and the Advanced E-Cancer Reporting and Registry Operations (AERRO) Group. She also was active in the leadership of the California Cancer Registrars Association, for which she served as Treasurer for four years. 31 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract In addition, she has been an Adjunct Instructor in Cancer Information Management classes at Santa Barbara City College, contributing to the training of cancer registrars. Dr. Sally Glaser was involved in initiating and planning the recent SEER strategic Visioning activities, including hosting the Visioning meeting in San Francisco in April 2008 and, with Dr. Clarke, co-chairing the SEER Visioning Marketing/Data Relevance/Revenue Generation working group. She participated with Dr. Brenda Edwards and Dr. Dennis Deapen of the Cancer Surveillance Program of Los Angeles in presenting a SEER update to the NCI Board of Scientific Advisors in 2008. Dr. Gomez has been a participant in NAACCR planning activities involving GIS technologies and the evaluation of the NHIA and NAPIIA algorithms for improving racial/ethnic classification, as well as efforts involving surveillance of cancer trends among Asian populations. Dr. Clarke co-chairs a NAACCR working group to improve melanoma reporting. Our staff were regular attendees and presenters at SEER, NAACCR, and CCR scientific and strategic meetings. In the last contract year, Drs. Glaser, Clarke and West participated in the 2009 NCI surveillance strategic planning meeting. Dr. Clarke served on the board of a local ACS unit. Dr. West continued as a member of the NAACCR Data Use Committee, a member of the California Evaluation and Outcome Committee, and as Chair of the National Coordinating Council for Cancer Surveillance The GBACR Surveillance Research group continued to take advantage of our location in a community with several public health degree programs to introduce master and doctoral degree students to SEER data and to the practice of cancer surveillance research. In this way, we have served as an informal training site for 18 student interns during the contract period. Our work with interns continued to be productive with respect to scientific publications and to training of students to understand and use SEER and other surveillance data and methods. In the last year, we mentored two master?s degree epidemiology interns. 2. Legal/Regulatory Issues As the legal and regulatory environment for cancer registration, surveillance, data release, and related research activities have grown more complex, the GBACR developed a Registry Compliance Program under Dr. Robert McLaughlin, J.D., Legal and Regulatory Affairs officer and IRB administrator, to help ensure complete and consistent cancer reporting, access for Rapid Case Ascertainment, records review for audit and quality control purposes, access to medical records for research studies, and release of patient information for epidemiologic and cancer control studies. Through development and clarification of policies and procedures, this program helped advance protocols for large data linkage studies using CCR data, inclusion of CCR data in international cancer consortium activities, ef?ciencies in IRB review, or enhancements of data security standards. it also conducted independent research, as evidenced by its publication in the last contract year of a comparative legal analysis of state statutes that mandate cancer reporting. The Registry Compliance Program has provided strong support to all three of California?s SEER registries by facilitating cancer registration and surveillance research in a state regulatory environment marked by continued change. The GBACR Surveillance Research staff provides access to cancer registry data as needed by health department and scientific investigators for health planning and research studies and with attention to all appropriate data-release regulations. In this capacity, we act as an administrative and scientific liaison to maintain the cooperation of physicians and hospitals in special studies, including, but not limited to, studies pertinent to the NCI. Over the last contract period, however, this service has become 32 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 increasingly subject to disruption as the disclosure of confidential data (and protected health information) to researchers has changed with enhanced institutional scrutiny and the evolution of HIPAA definitions, policies, and procedures. With the support of Dr. McLaughlin, data release staff have worked actively to educate reporting facilities, and collaborated with the OCR and other California state entities to help with the interpretation of relevant state laws and the setting of registry policies and procedures for the release of patient data. As a result, our ongoing data release efforts reflect a high level of active attention and responsiveness to applicable state and federal laws and regulations. The Registry Compliance Program also worked to support the implementation of new California statewide data security standards mandated through State Bill 13 while accommodating the efficiency and efficacy of population-based cancer research. With input from this Program and IRB, the latter became the first in California on which the state IRB would rely for a class of research studies pursuant to a cooperative research agreement. Development of this agreement was a substantial benefit, as it enables a reduction of state committee-level review and delays in advancing population-based cancer research while preserving local review and control for data security. Also to better facilitate SEER-supported research, the Registry Compliance Program sought clarification of the data destruction rules of the OCR, and of ambiguities in state access and disclosure rules including those in definitions of "researcher" and ?research institution." 3. Biorepository In 2003, SEER began a Residual Tissue Repository (RTR) program to determine the feasibility of obtaining stored tumor tissue, especially tissue scheduled for discard, and of establishing biorepositories in SEER areas. Dr. Glaser received SEER RRSS funds to survey pathology facilities in the GBACR catchment area to identify discard policies and the availability of paraffin stored tissue for population-based research. Around that time, a regional hospital pathologist contacted the GBACR regarding our interest in tissues they were ready to discard. We accepted these materials, thus initiating the GBACR Discard Repository. After receiving these materials, we developed methods to distinguish cancer from non-cancer cases and to store cancer specimens from over 6,000 cancer patients (24,000 blocks). The blocks have been tested for quality and found suitable for high?throughput molecular analyses in two different laboratories. To operationalize the Discard Biorepository as the Biorepository Resource?, Dr. West and the GBA CR Registry Compliance Program developed a technical approach that separates data and biospecimen requests and that relies on the coding of specimens so that the GBACR may link specimens to registry data for researchers on a study-specific basis. The GBACR is the sole custodian of the key linking tumor specimens to cancer case records. This approach eliminates the need for the Discard Repository to become an independent custodian of identifiable patient data, including all cancer data collected and managed by the GBACR, and limits the release of identifiable con?dential data to researchers, thereby protecting patients while facilitating research. Under the protocol, the CPIC Of?ce of Legal and Regulatory Affairs provides a variety of administrative and record-keeping functions, including oversight of Material Transfer Agreements that set the terms for the use and return of tissue materials. Electronic and hard- copy material transfer agreements are included with the IRB ?les for each research project to 33 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-351 36 inform and complement the continuing reviews, principally with regard to safeguarding the con?dentiality of identi?able patient information. Investigators responsible for biospecimen- based resource protocols are required to document the recipient(s) of these resources so that the CPIC protocols contain ongoing, comprehensive information about the exact dissemination, use, and findings associated with CPIC biospecimen resources. a. California Biorepository Research Network The GBACR Discard Repository became a partner in the California Biorepository Research Network (CBRN). In 2008, Dr. Moon Chen, Associate Director for Population Sciences at the University of California, Davis, Cancer Center and the Principal Investigator for the NCI-funded Asian-American Network for Cancer Awareness and Training (AANCART), assessed the feasibility of establishing an infrastructure for collaboration among nine NCI-designated Cancer Centers in California to identify all cancer cases at these institutions with tumor tissue that might be used to enhance translational research on health disparities. A subsequent survey was conducted among these Cancer Centers regarding types of specimens collected, patient consent practices, and procedures for processing and storing specimens, and processes to distribute materials. The survey was followed by a meeting with representatives from all of these facilities, including Dr. West, to discuss the creation of the CBRN. The intention was for the CBRN to be built on the platform of NCl-funded Region 6 (Western US and Hawaii) Biorepository Management of Programs and to initially create a ?virtual repository? that includes these nine facilities and the OCR. The CCR was intending to create the database and link the cancer cases to the facilities where the tissue resides. However, as of the end of the contract period, the CBRN had not been funded. b. Change to California state law GBACR researchers have long been interested in modifying State legislation to facilitate access to tumor biospecimens so that they might be tied to existing cancer registry data and thus facilitate and enhance registry-based research at the molecular and genetic level. Previously the OCR had put such a request fonivard to the legislature for consideration, without success. During the contract period, the GBACR Registry Compliance Program evaluated whether and how the existing infrastructure of state?mandated cancer registries might be used to advance direct registry access to biospecimen resources that may be archived at the reporting entities from which pathology data is routinely gathered and assembled as registry data. Based on a 2007 survey of state registry statutes in SEER regions and their potential support for registry access to biospecimen resources, the Registry Compliance Program began to conceptualize biospecimen resources as texts and as extensions of the data present in medical records. The Program also began to assemble articulations of proprietary interests in biomedical research specimens as well as various ethical issues that impact the disposition of biospecimens, including the relationships between a doctor and a patient, a researcher and a human subject/research participant, and a researcher and a research institution. In 2008, the GBACR begin to work with the California Division of the American Cancer Society (AC8) to gauge interest in supporting a measure to change the 1986 statewide cancer reporting mandate to include access to biospecimens. The ACS endorsed this idea and worked with Drs. West, Glaser, Clarke, McLaughlin and others to explore the justification for this change and to 34 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-351 36 identify a legislator to champion this effort. As this concept moved forward, the ACS determined that the timing was not right, given the current budget situation in California. 4. Process Management/Continuous Process Improvement a. Overview of methodology CPIC has been adopting and implementing Process Management methodology since 2005. We are committed to a strongly integrated Process Management culture, and the GBACR has been the focus of some of our earliest Process Management efforts. Our goal in this work has been to have our Continuous Process Improvement and embedded Process Management methodologies work effectively throughout the organization, with a focus on ways to constantly improve and make more efficient all of business practices. In this work, we use a hybrid of Process Management methodologies based on combining the basic steps and concepts of Six-Sigma, Lean, Lean-Six?Sigma, and TQM, and including the following functions: 0 Discovery/Identi?cation: This initial phase focuses on the discovery, identification, and documentation of the current processes. 0 Evaluation/Review: This phase creates an initial evaluation and review of the current processes ?as is?, defines and measures the level and status of requirements met, along with the measurement of key quality indicators met, and the specific methods for identification of root-causes of errors, issues, or inefficiencies. - Assessment: This phase fully assesses the evaluation and review (measurements and status) of the requirements and the key indicators from the previous phase, creating a full set of updated and documented requirements (as needed) for each of the organizational processes and procedures. 0 Recommendation: This phase defines, plans, and designs the recommended improvements, including methods of audit, report, and measurement for post implementation, and evaluation and review during the neXt round of Continuous Process Improvement. 0 OptimizeHon/Implementation: This phase implements the newly defined, designed, documented processes. 0 Audit/Measurement: This phase is a verification of successful implementation. It is how success is measured and reported for all process improvements and optimizations gained from the general Continuous Process Improvement. This hybrid methodology combines Process Management and Continuous Process Improvement components into a form of continuous process management that is outlined in the diagram on the next page. The diagram on the next page illustrates how the methodology built and used in a continuous circular method have been providing a key foundation for CPIC and GBACR operations. Each of the phases of this methodology, which are simple to follow and execute, dovetails directly into the next phase. Additionally, within each of these phases, the same methodology is deployed. For example, within the ?Assessment Phase? exists the entire set of five phases Evaluation/Review, Assessment, Recommendation, Optimization/ Implementation, Audit/Measure), embedding this Continuous Process Management and making the entire methodology a self-reinforced tool for all parts of the organization. 35 SEER Final Report: August 1, 2003?July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 Continuous Process Management Discovery/ Identification Each of the top level phases of the cycle have sub-process methodologies but Evaluation Review leverage the same top level 5 to reinforce the general PM methodology Assessment kl Optimization Recommendation Implementation Figure 2. Process Management/Continuous Process Improvement b. Process Management/Continuous Process Improvement activities in the GBACR Using this approach within the GBACR during the contract period, we have begun to continuously assess and improve registry operations using process evaluation methodology with the following components: 1) documentation of current processes; 2) definition and measurement of requirements and key quality indicators; 3) specification of methods for identifying the root-causes of errors or issues, and 4) a system for continuous improvement. To maximize the effectiveness and ef?ciency of GBACR processes, we identi?ed the registry's essential functions with the goal of gradually applying Continuous Process Improvement to all of them. These functions are data collection, data processing management, data submission, and data sharing. Because some of the key challenges the GBACR faces have been in the areas of data processing and management, our first application of Process Management was to examine the manual and automated processes and the associated workflow starting with the point of entry of a casefinding document to the end point, which is the creation of a physician case or the receipt of a missed hospital case. These are critical activities, as they encompass the numerous steps on the long path of data reporting to SEER and thus create the foundation for all data management actions. We completed the discovery steps (charting the processes flows and creating the documentation narratives), the preliminary evaluations, and the documented assessment of these processes and procedures, along with a clear and understood set of preliminary recommendations. This allowed the GBACR to start with a basic list of processes to address, within the scope of Process Management. The GBACR RIS staff successfully executed some of these recommendations, which have helped to optimize some of the day?to-day execution of tasks and operations. These include: 36 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract NO1-PC-35136 0 Additional automated business rules built into the applications, enabling faster data entry processes; 0 Interactive data checking during the input processes, eliminating errors to be ?xed later; 0 Automating duplicate checking and processing, eliminating the need for personnel to go back and manually check for possible or known duplicates of information; Adding a user productivity report in the PDM system that eliminated manual tallying; Upgrading the Doctor database to .net technology with a more user-friendly interface and automated functions; 0 Automating the several manual steps to load e-path messages; 0 Adding reports for the messages received by facility for monitoring purposes; 0 Modifying the visual editing discrepancy feedback process to more easily search for a particular case. c. Other Process Improvement activities i. E-Path Process Improvement Project Along with the other two California SEER regions, the GBACR was invited by SEER in the last contract year to participate in a process improvement activity led by Lilia O?Connor and Chuck Sweet. The mission of this project was to improve the current process to recruit, enroll, and begin electronic exchange of cancer data between a facility and the OCR. A significant amount of time was spent developing a project initiation document and mapping out deliverables. ii. ROU Planning Team In 2009, hired a consultant for an institution-wide process to enhance ef?ciency. After information gathering, including an institution-wide survey, an ROU Planning Team was organized as one of several theme-oriented Planning Teams. The ROU team, consisting of eight volunteers, was charged with working on new procedures to enhance efficiency, including improving the orientation process and on-boarding experience for new hires, improving training for ROU staff on core applications, and clarifying requests for information from doctor?s offices so data could be gathered more quickly and accurately. In addition to this task?oriented work, the organization-wide review yielded other input that indicated broader needs within ROU to improve operational efficiency and effectiveness and to build employee morale at a time of increasing workload and evolution to more electronic processes. Through the end of the last contract year, the organization consultant provided additional in-depth assessment to more accurately identify areas for improvement. Then, from these ideas, she worked with Dr. Glaser and Ms. Davidson-Allen to develop detailed concrete plans for improving day-to-day operations and staff morale, given pending changes to our case-?nding protocol and increasing reliance on technology and electronic communication. The consultant also provided input on succession plans for senior managers, including options that involve internal promotion opportunities for existing staff. (2) LIST OF STUDIES A list of the peer-reviewed scientific publications for the contract period authored by GBACR staff and supported by either our core SEER contract or by SEER RRSS is included in Appendix 37. In this contract period, the GBACR Surveillance Research group produced 96 peer- 37 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-351 36 reviewed publications in a variety of scientific journals including high-impact titles like the Journal of the National Cancer Institute and the Journal of Clinical Oncology. In the last contract year (2009-2010), we published or have in press 27 manuscripts, with several more still under review (see ?submitted? section of Appendix 37). (3) CURRENT STATUS AND FUTURE PLANS 1. Organization Chart and Current Staffing a. Organization Chart The organizational charts for the Registry Operations Unit and Surveillance are included in Appendix 38. b. Staffing i. Registry Operation Unit changes The ROU experienced a number of changes in staff over the contract period, as have been well documented in our annual progress reports. The following personnel changes took place in ROU during the last contract year (August 2009-July 2010): Personnel leaving the organization: Lois Inferrera, Quality Control Specialist; Susie Huey-Lee, Administrative Assistant; Pui Pui Cheng, Quality Control Specialist/Oncology Data Specialist (split position) and Nadine Dyke, Quality Control Specialist. Personnel hired: Eileen Leong, Oncology Data Specialist, and Angela Johnson, Administrative Assistant. Staffing ROU with experienced and credentialed cancer registrars continues to be a challenge, given the relative scarcity of a workforce with this highly specialized training. We have been proactive in providing an attractive work environment, such as flexible work schedules and working remotely from home. Throughout the contract period, we also provided a structured approach to supporting staff to become Certified Tumor Registrars. ii. Surveillance Research (SR) changes The following personnel changes took place in SR during the last contract year (August 2009- July 2010): Personnel leaving the organization: none. Personnel hired: Juan Yang, Surveillance Research Intern; Melissa Ongpin, Surveillance Research Intern 38 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 36 During the contract period, we did not experience major challenges staffing the Surveillance Research group and had a healthy response to our request for internship applications. 2. Rules for Reporting to the Registry in 1985, the California legislature enacted a state law that made cancer a reportable disease and established the CCR as the repository for the reports. Under California Health and Safety Code, Section 103885, resident cases diagnosed or treated for cancer beginning January 1, 1988 are to be reported by hospitals, physicians, and certain other health-care providers to the GBACR within six months of the time the case comes under the care of or is admitted to the facility, and then to the CCR. Cases are reported to the CCR and to SEER according to procedures described in this section. 3. Methods Used to Obtain Data and Percent of Records from Each Method See above Section 3. Data Collection and Appendix 21: Reporting Source by Year of Diagnosis. 4. Follow-up Methods (including percent passive follow-up by age group) See above Section 7. Follow-Up Methods (Passive and Active). 5. Key Relationships with Other Registries and Organizations a. OCR and its regional registries As described in the sections above, our staff collaborate closely with the other California regional registries in all aspects of registry operations and surveillance research. b. Stanford Cancer Center With the establishment in 2005 of the Stanford Cancer Center (of which CPIC is an affiliate), GBACR scientists broadened our data use efforts to include clinical cancer surveillance, which we define as identifying research questions of immediate importance to clinical practice and answering these questions expediently using existing large, population-based cancer surveillance resources, predominantly SEER. The collaboration of active, research-oriented clinical oncologists with experienced, public-health-minded GBACR cancer surveillance researchers has provided a synergistically effective opportunity for maximizing the quality and impact of our SEER data use and resulted in numerous surveillance publications with clinical centers, predominantly Stanford. c. UCSF We have a history of scientific collaboration with a variety of investigators at UCSF. In 2009, we began working with UCSF breast surgeon Dr. Laura Esserman to map variables from her ATHENA breast health network electronic health record system to NAACCR standardized fields. Dr. Clarke is an active member of the Athena Breast Health Network led by UCSF but including representatives of all five NCl-designated UC Cancer Centers. 39 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 36 d. American Cancer Society We have improved our collaborative relationships with our state and local ACS offices. Dr. West serves on the statewide California Evaluation and Outcome Committee. In the last contract year, Dr. Clarke has served actively on the board of the Santa Clara unit and has presented a yearly summary of cancer incidence trends in the region to both this and the San Mateo county leadership. As described above, Drs. Glaser, McLaughlin, Clarke and West worked with the state ACS legislative action committee to change state cancer reporting laws to include biospecimens. e. NAACCR and SEER We have been strong collaborators within the greater cancer registry community throughout the contract period. Dr. Glaser was involved in initiating and planning the recent SEER strategic Visioning activities, including hosting the Visioning meeting in San Francisco in April 2008 and, with Dr. Clarke, co?chairing the SEER Visioning Marketing/Data Relevance/Revenue Generation working group. She participated with Drs. Edwards and Deapen in presenting a SEER update to the NCI Board of Scientific Advisors in 2008. Drs. Glaser, Clarke and West participated in the 2009 NCI surveillance strategic planning meeting. Dr. Gomez has been a participant in NAACCR planning activities involving GIS technologies and the evaluation of the NHIA and NAPIIA algorithms for improving racial/ethnic classification, as well as efforts involving surveillance of cancer trends among Asian populations. Dr. Clarke co-chairs a NAACCR working group to improve melanoma reporting. Dr. West continues as a member of the Data Use Committee and as Chair of the Similarly, all GBACR Surveillance Research staff participate in CCR statewide research and coordinating efforts. 6. Current Process Management Methodology and Plans for the Future See above section 4. Process Management/Continuous Process Improvement and below Section 7. Future Plans, Challenges, Opportunities. 7. Future Plans, Challenges, Opportunities a. Registry Information Systems In the upcoming contract year, the GBACR will actively investigate external casefinding systems, select a system, and begin the process of transitioning casefinding activities to the new system. Until the transition is complete, we will continue to use our current casefinding systems. b. Physician self-reporting The CCR has developed a new feature in the DMS for physicians to directly report cancer cases to Eureka. This Eureka Physician Office Access feature will be deployed for dermatology offices in the fall of 2010. It will provide a select set of data items for the physician or the of?ce staff to complete using Internet?based access to this reporting module. The GBACR plans to implement this option for physician reporting, starting in selected dermatology practices in our area. 40 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01-PC-35136 c. Process Management/Continuous Process Improvement In the next contract period, we will continue working with Process Improvement. In ROU, we will address registry data collection activities to reduce process that are manual, labor intensive, and created over years of experience. An early target will be physician case reporting. We also will continue with the SEER-supported California process improvement activity led by Lilia O?Connor and Chuck Sweet to improve the current process to recruit, enroll, and begin electronic exchange of cancer data between a facility and the CCR. We also will apply Continuous Process Improvement and Process Management to activities such as the processing of external and internal data sharing requests. The GBACR Surveillance Research group processes numerous data requests that are generated both internally and externally. The many steps in responding to these requests can be reviewed, evaluated, and improved using the Process Management methodology. d. Electronic health records and cloud computing i. E-path At the beginning of the new contract period, we will hire an e-Path manager. This person will lead, plan and manage a concerted work effort to implement electronic reporting software in the remaining GBACR reporting facilities that have high volume pathology laboratories. This work effort, which will span two years, will allow us to gain greater efficiencies in casefinding and realize greater cost-savings. It also will allow us to accelerate the pace at which the GBACR implements the capabilities of electronic health reporting. ii. Other aspects of electronic reporting As part of the ATHENA project, we are participating in the design of a comprehensive system of electronic health records to be deployed at all five University of California breast health clinics. Our preliminary efforts have involved mapping of most of the clinical staging and pathology variables to NAACCR standard variables as part of the UCSF system If appropriate, we are suggesting that the NAACCR standards be incorporated directly into the data entry screens of the system. We also are building the open-source software needed to port data from the system directly into the UCSF hospital registrars? databases. The possibility of autopopulating cancer registry fields directly from electronic health records lays the groundwork for same-day reporting of cancer cases or updating vital status to a central cancer registry. In the spring of 2010, the GBACR collaborated with Drs. Lewis Frey and Nan Stroup of the Utah Cancer Registry on an application to build dedicated software for the expanded reporting of surveillance data from electronic health records. Dr. McLaughlin has initiated consideration of the data security and confidentiality issues inherent in electronic health records and cloud computing for data analysis. We will continue to build out our multidisciplinary efforts in incorporation of electronic health information into cancer surveillance and research. 41 SEER Final Report: August 1, 2003-July 31, 2010 Sally L. Glaser, Principal Investigator Contract N01 -PC-351 36 e. Education and training We have numerous future plans for education and training. We will provide targeted training for the 2010 data set changes to include training for and the new Hematopoietic and Neoplasm Rules. On a basis, we will host NAACCR webinars for all regional registrars using the NAACCR Webinar Series. On a quarterly basis, we will conduct regional registrars meetings at the GBACR to share information and conduct training activities. Annually, we will send our GBACR staff to educational conferences (SEER, NCRA, CCRA, NAACCR) and share the information we learn with internal and external cancer registrars. On an ongoing basis in our quality control work, we will identify community registrars in need of additional training in abstracting and coding and identify and/or provide targeted training. We also will disseminate cancer reporting-related information in a timely and efficient manner, including essential materials for the 2010 data set changes, quality control memos from the OCR, and finalized SEER SINQ and CCR questions. f. Improved internet-based dissemination of cancer statistics We are continuing website enhancements and, in the next contract period, our master?s level epidemiologist, who also has training in health education, will include a mechanism for conducting satisfaction surveys among GBACR website users so that we can understand how well our internet resources are meeting their needs. We will incorporate findings from our collaborations with Dr. Matt Kreuter and the Health Research Communication Laboratory at Washington University regarding the best practices for communicating and visually displaying cancer statistics to lay audiences. g. Biorepositories We will continue to maintain our Discard Repository, and if SEER funds become available, we will enhance and expand this effort as described in our RFP response. If the CBRN (described above) is funded, it may be very helpful for the proposed biorepository, in that it could help regional pathologists understand the importance of tumor tissue-based research and develop common specimen preparation, storage, and distribution policies. In addition, we will continue to move fonNard a possible change in the State cancer mandate that would provide access to tumor specimens related to the registry database, since obtaining access to tissues as part of the cancer reporting regulations would greatly facilitate our development of the Discard Biorepository. 42