Breast Cancer Surveillance Consortium: Progress in Understanding Screening Delivery and Early Detection

1
Breast Cancer Surveillance Consortium Progress
in Understanding Screening Delivery and Early
Detection
Rachel Ballard-Barbash, MD, MPH, Associate
Director, NCI/DCCPS/ARP National Cancer Institute
2
Establishing the Breast Cancer Surveillance
Consortium

• Origins and Purpose of the BCSC and SCC
• Complexities of creating the Consortium
• Resource for research
• Research Evidence
• Key factors for success
• Challenges and opportunities remain

3
Establishing the BCSC and SCC

• In the beginning, much was unknown
• No community measures of mammography quality and
  no source of national data
• Limited experience collecting data in the course
  of care required protection for providers as
  research subjects
• Many challenges to establishing the BCSC
• Shifted from independent RO1 to coordinated
  pooled data, mapping to CDE, new statistical
  methods for complex data
• Moved from paper to electronic data capture in
  early years
• Field of delivery research in practice was new
  and many of the Principal Investigators were new
  researchers

4
Scenes from the Beginning
5
Breast Cancer Legislation and Funding

• The BCSC began as mammography screening was
  increasing
• 1990, CDCs National Breast and Cervical Early
  Detection Program
• 1991, Department of Defense Breast Cancer
  Research Program
• 1991, NIH launches Women's Health Initiative
• Mammography Quality Standards Act of 1992 (MQSA)
  mandated NCI to develop a breast cancer screening
  surveillance system
• NCI Response
• Pilot studies in SEER registries supported
  development of 1993 RFA
• Expanded with 1994 RFA (new sites and Statistical
  Coordinating Center) to address racial/ethnic,
  geographic, and health system diversity in
  screening
• BCSC renewed in 2000 and 2005

6
BCSC Purpose

• Evaluate performance of mammography screening in
  practice
• Individual, health professional and system level
  factors
• Increase capacity to examine provider and system
  factors
• Define biologic characteristics of cancers that
  influence detection
• Quantify population effect of screening
• Longer term survival and mortality
• Track new technologies in screening
• Imaging, tissue, molecular markers, proteomics

7
(No Transcript)
8
BCSC Sites
9
BCSC Local Facilities (N164)
10
Demographics of Women

• 7,335,521 mammograms from 1994- 2008

Age BCSC - N () US population 2008 N ()
35-39 314,294 (4.3) 11,387,968 (16)
40-49 2,164,444 (29.) 21,515,659 (30)
50-59 2,208,148 (30.1) 15,938,332 (22)
60-69 1,465,980 (20) 10,802,003 (15)
70-79 961,472 (13) 9,134,000 (13)
80-84 221,183 (3) 3,110,470 (4)

Race/Ethnicity  BCSC - N () US population 2008 N ()
White, non-Hispanic 5,218,642 (71) 57,167,145 (71)
African American, non-Hispanic 424,840 (5.8) 9,460,539 (11.8)
Asian, non-Hispanic 453,569 (6.2) 3,637,776 (4.5)
Hispanic 636,119 (8.7) 8,716,664 (10.4)
Native Hawaiian or Pacific Islander 3,073 (lt 1) 114,817 (lt 1)
American Indian or Alaska Native 92,044 (1.3) 653,440 (1)
Mixed (Two or more) 54,974 (0.7) 766,436 (1)
Other 31,872 (0.4) n/a
Unknown 420,388 (5.7) n/a
11
Cumulative Number of Mammograms by Submission
Year
12
Cumulative Number of Cancer Cases by Submission
Year
13
Core Pooled BCSC Data Women Physician Level
Variables Outcomes
Self-administered questionnaire
Direct data entry or questionnaire
Annual linkage
14
BCSC as a Research Resource

• Since 1994, BCSC collected data on a cohort of
  over 2 million women
• 8,374,024 million mammograms (2,323,252 unique
  women)
• 86,700 breast cancers (65,313 invasive and 13,263
  In Situ)
• Screening data linked to Medicare data
• 107 radiology facilities and 1300 radiologists
• Collective insight of BCSC PIs about breast
  cancer risk factors, screening, and related
  outcomes
• Data complexity
• Statistical methods
• Research utilizing the core BCSC data focuses on
  delivery, performance and quality of care

15
Uses of Pooled BCSC Research Resources

• Research and modeling
• Data source for simulation models (CISNET)
• Investigators have collaboratively published 374
  papers
• Engaged new and junior investigators
• 36 publications by junior investigators (2005-8)
  most non-BCSC
• Three career development awards
• Enabled new grants
• Supported the generation of more than 65 research
  grants from many agencies many investigators
  from outside the BCSC
• New data linkages BCSC-Medicare linked data

16
Selected Ancillary Studies

• Assessing and Improving Mammography (AIM)
• Assesses accuracy of interpretation of mammograms
• Develops tools and guidance for training of
  radiologists
• Co-funded by ACS (Longaberger funds) and NCI
  (Breast Cancer Stamp )
• Factors Affecting Variability Of Radiologists
  (FAVOR)
• R01 utilizing BCSC data to study the variability
  in radiologists in community mammography settings
  (PI Joann Elmore)
• Comparative Effectiveness Research
• Comparative Effectiveness of Breast Imaging
  Strategies in Community Practice GO Grant (ARRA
  funds, PI Diana Miglioretti)
• Collaboration to evaluate digital vs. film-screen
  mammography BCSC-CISNET-EPC (ARRA funds, PI
  Diana Miglioretti)

17
Use of BCSC Research Evidence

• Delivery research generates questions for
  discovery and development research
• Within the BCSC, special research projects at
  individual sites used for discovery and
  development questions
• Address targeted translation issues
• Eg Develop quantitative, automated method for
  measurement of breast density
• Individuals sought as members of panels related
  to breast cancer on a diversity of topics (IOM,
  ACR)
• Contributed evidence to federal reports and
  policy
• IOM, GAO, WHO

18
Factors for Success

• Team Science approach, utilized variety of
  disciplines within each site
• A secure, centralized resource, shared by many
• Incorporation of collecting patient data for
  research purposes into clinical care practice
• Anticipate and understand the complexities of
  building a longitudinal dataset
• Creating new ways to provide feedback on
  performance

19
Challenges Remain

• Delivery, performance and quality of care is
  dynamic need ongoing data reflecting current
  clinical practice
• Requires prospective, longitudinal data
• Evaluate longer term outcomes beyond process
  measures
• Large, multiregional data to answer questions in
  specific groups
• Growth in investigator-initiated research
  utilizing the BCSC research resource indicates an
  enormous potential for addressing questions in
  delivery beyond the current scope
• Comparative effectiveness of digital and
  screen-film
• Innovative template for the future

20
(No Transcript)