Original Article Introduction to the National Program of Cancer Registries Modeling Electronic Reporting Project (NPCR-MERP) Sandra F. Thames a ; Kenneth Gerlach, MPH, CTR a ; Howard J. Martin, PhD, MS, MA b ; Timothy Carney, MPH, MBA c ; Lynne T. Penberthy, MD, MPH d ; Michael Lanzilotta e ; Steven Peace, BS, CTR f Abstract: Introduction: In 2004, the Centers for Disease Control and Prevention s National Program of Cancer Registries (NPCR) launched the Modeling Electronic Reporting Project (MERP) in collaboration with the Virginia Commonwealth University Health System (VCU/HS), the Virginia Cancer Registry (VCR), and the National Cancer Institute s Surveillance, Epidemiology, and End Results (SEER) program. The purpose of this project is to develop a comprehensive best practice model to promote electronic reporting for cancer surveillance that takes advantage of the electronic medical record. Methods: The NPCR-MERP has developed unified modeling language as-is and to-be model diagrams based on the VCU/HS and VCR infrastructure. The NPCR-MERP will host consensus-building sessions with cancer surveillance community representatives from across the nation to develop a best practice model. Discussion: The refined, and more detailed, model will serve as a best practice guide for the development of interoperable, standardsbased cancer registry software. Key words: best practice model, cancer surveillance, electronic medical record; electronic reporting, unified modeling language Introduction Methods for cancer surveillance have existed for many years, although this continues to be among the most complex areas of disease surveillance. The complexities derive from a number of factors unique to cancer, including the breadth and complexity of data that are captured, the multiple and varying sources from which these data are obtained, the many providers from which cancer patients receive their care, the longitudinal nature of the data, and the numerous levels of users of cancer surveillance information. 1 The latter includes hospitals, central registries, national organizations, and researchers at the local, state, and national levels. The development and dissemination of new clinical information about biomarkers and genetic factors will affect the diagnosis and treatment of cancer and will require cancer surveillance programs to enhance the range and depth of the data collected and stored, further increasing the complexity of cancer surveillance. Furthermore, the cancer surveillance community and other data sharing partners demand timelier and more relevant cancer information. The authors of a paper on a national framework for cancer surveillance recommend integrating information technology into current systems to improve the completeness, timeliness, and quality of reporting and to facilitate the transition from paper to electronic medical records, including electronic messaging, standardized vocabularies, and interoperable systems. 2 New methods of data collection, validation, storage, and transmission are needed to fill these needs. A possible way to enhance completeness, timeliness, and quality in cancer surveillance would be to use the emerging electronic medical record (EMR). An EMR is a compilation of medical information about an individual s health and health care services received within a specific hospital or facility. 3 Using the EMR to automate data capture could improve completeness, timeliness, and quality of cancer incidence data and offers the opportunity to expand data collection to include automated electronic capture of diagnostic test results, treatment, comorbidity, and followup. In addition, use of the EMR and automating more of the hospital cancer registry operations could streamline the duties of the cancer registrar. Automation could eliminate some of the clerical tasks such as manually searching paper medical records and multiple data systems that do not communicate with one another, allowing more time for quality control and analysis of the data. There are certain terms and concepts that must be understood to appreciate the approach the collaborating partners have taken in this project. Several information technology terms are used throughout the paper that may be new to some readers. Therefore, a glossary of terms has been provided as Appendix 1. A few of the key terms and concepts include: interoperability, which is the ability of systems (software and hardware) to communicate and share data with other systems; electronic messaging protocol are standards, guidelines, or specifications required to securely transmit data electronically; a robust and transportable model provides tools and processes that allow easy implementation in any environment; and granular refers to the level of detail that is described in the model. Please refer to Appendix 1 for additional definitions of key terms and concepts. Introduction to the National Program of Cancer Registries Modeling Electronic Reporting Project (NPCR-MERP) a National Program of Cancer Registries, Centers for Disease Control and Prevention, Atlanta, GA. b Virginia Cancer Registry, Virginia Department of Health, Richmond, VA. c Northrop Grumman IT, Contractor to Centers for Disease Control and Prevention, Atlanta, GA. d Massey Cancer Center, Virginia Commonwealth University, Richmond, VA. e Coordinating Office for Global Health, Centers for Disease Control and Prevention, Atlanta, GA. f Surveillance, Epidemiology, and End-Results (SEER) Program, National Cancer Institute, Washington, DC. Address correspondence to Sandra Thames; Centers for Disease Control and Prevention, 4770 Buford Hwy, NE, MS-K53, Atlanta, GA 30341. Telephone: (770) 488-5689; e-mail: sthames@cdc.gov. Journal of Registry Management 2006 Volume 33 Number 3 97
Purpose The Modeling Electronic Reporting Project (MERP) is an effort led by the Centers for Disease Control and Prevention s National Program of Cancer Registries (NPCR) in partnership with the National Cancer Institute s Surveillance, Epidemiology, and End Results (SEER) Program, the Virginia Cancer Registry (VCR) in the Virginia Department of Health, and the Virginia Commonwealth University Hospital System (VCU/HS). The purpose of the NPCR-MERP is to develop a comprehensive, best practice model to promote electronic reporting for cancer surveillance that takes advantage of the emerging technology that drives the EMR. Modeling is a tool used to describe or visualize an operation such as cancer registration. This tool is widely used in the health care information technology community to describe complex operations and the associated information systems behind them. Modeling can serve to bring experts in a particular operation such as cancer registration together with information technology analysts who design and develop the associated software. The process of modeling involves a description of the operations, the visual representation of the operations, and the refinement and enhancement of the visual representations to obtain consensus or best practice. The modeling process has the ability to capture existing knowledge, standards, and practices in a simplified visual format ( blueprint ) that facilitates communication among stakeholders, promotes accurate analysis of data and processes, and enables improvement of the essential aspects of cancer registration. 4 The aim of this project is to create a model of an ideal cancer surveillance infrastructure that could exist within 3 distinct organizational environments: (1) a hospital environment (HOSP); (2) the state-wide, population-based central cancer registry (CCR); and (3) a national cancer program (NCP). To achieve this end, the model must incorporate the organizational goals, structures, processes, rules, and events from each of the organizations that participate in cancer registry reporting. In addition, it aims to incorporate the many interfaces between these 3 types of organizations and define nationally-accepted standards and practices. Modeling provides the required analysis and system tools to achieve these objectives. The NPCR-MERP seeks to create a model that has several qualities essential for cancer surveillance. The model must be robust, scalable, and transportable, so that it can be applied in different reporting facilities and be capable of handling some or all of the EMR sources available at a facility. Different reporting facilities will be able to implement the model or its components, because it will not be tied to a particular operating system or application. Unified modeling language (UML) is a commonly-accepted notation standard. 5 The model based on the electronic capture of data from the EMR must be in a form that can be transformed into software applications for use in the many varied operations that collection of cancer data involves. In addition to developing the model, its feasibility will be tested through a pilot implementation at 2 of the partner sites: VCU/HS and VCR. The successful implementation of the model will enhance the capabilities of cancer registries to produce complete, timely, and high-quality information, as measured against existing registry standards. The NPCR-MERP model will include the electronic transfer of information within and between hospital data sources and will possess key characteristics to ensure that other facilities can adopt the products with minimal effort. Thus, NPCR-MERP products are designed to use national standards for coding and transmitting data that are well understood and widely available. For example, NPCR- MERP has adopted Health Level Seven (HL7 ) 6 as the standard electronic messaging protocol for transmitting cancer data. The project incorporates the standard vocabularies and code systems native to the various EMR sources available to most facilities (LOINC, SNOMED CT, etc.). Messages transmitted between hospital and central registries will comply with Public Health Information Network (PHIN) 7 standards, based on national industry standards (for more details, refer to Table 1.) Methods The NPCR-MERP is a complex project that requires careful coordination, cooperation, and communication among the participating agencies and organizations. The roles that each partner plays are complementary and focus on distinct areas of expertise. VCR and VCU/HS function at the CCR and HOSP levels, respectively. At the CCR, and especially the HOSP level, the focus is more on implementing the developed model and is thus more granular and practical. At the NCP level, through NPCR and SEER, the focus is on guiding the process to develop a high-level conceptual model that describes the NPCR-MERP structure and operation. The final product will be a best practice, consensus model that is nationally applicable, with sufficient detail for the development and implementation of software applications. In the first phase of the modeling process, the NPCR- MERP team is using iterative techniques to develop and refine the model. As the team develops a model of automated and electronic processes for casefinding, follow-up, and quality assurance/editing, the processes are being piloted for implementation at the hospital and CCR level by VCU/HS and the VCR. To date, the pilot implementation includes the automated capture of cases from a number of electronic data sources including not only electronic pathology reports, but also claims data, discharge summaries and radiology reports. The pilot implementation will inform subsequent iterations of the model and their implementation. In the second phase of the project, NPCR-MERP partners plan to engage the broader national cancer surveillance community (refer to Figure 1) to corroborate, refine, and/or modify the proposed model developed in Phase 1, and to obtain consensus on a best practice model for using health information technology in the cancer surveillance community. Achieving this task calls for work with interested members of the national cancer surveillance community to revise and extend the NPCR-MERP Phase 1 model and to develop a plan or strategy to move forward with implementation of the model. Approach The highest level of perspective and the most comprehensive look at the NPCR-MERP modeling effort is represented in the NPCR-MERP Context Diagram (Figure 1). Figure 1 describes the cancer registration domain for NPCR-MERP. This description is a visual representation in the form of a context diagram. This cancer registration context diagram shows the major business entities (NPCR, SEER, VCR, VCU/HS, CoC, NAACCR), their relationships, and their responsibilities. This diagram provides the foundation for other modeling diagrams. The NPCR-MERP has 98 Journal of Registry Management 2006 Volume 33 Number 3
developed many diagrams in addition to Figure 1 and these can be viewed at http://www.cdc.gov/cancer/npcr/merp. These diagrams model hospital cancer surveillance processes including casefinding, treatment, follow-up, and quality assurance/editing with the primary focus on automated electronic reporting. The model provides a view of the landscape by domain, activity, level of perspective, or any combination of these categories. Discussion A primary reason for cancer registration complexity is that data are compiled and consolidated from many different systems and in many formats. Currently, methods to access and capture registry data are predominantly manual. The surveillance data that the national programs (NPCR, SEER) receive are of high quality, 8 however, the resources expended to find, access, collect, and collate the required data are enormous. Using the EMR and other sources of electronic information could potentially improve the completeness, timeliness, and quality of the data reported, and over time reduce costs significantly. Historically, it takes approximately 4 6 months to obtain from a medical record the basic information necessary to abstract a specific case. The published national and state cancer statistics are not available until 3 years after diagnosis. The Commission on Cancer Standard 3.3 notes, For each year between survey, 90% of cases are abstracted within 6 months of the date of first contact. 9 In some circumstances it takes longer, depending on the series of treatments the patient receives. The use of the EMR and other electronic data sources in cancer surveillance programs will provide an opportunity for faster case identification and reporting. Instead of waiting for the report, notes, and procedures to be dictated, printed, and compiled into a paper medical record and filed, electronic versions can be immediately available when signed by the responsible physician. This should obviate the chore of searching for medical records in the medical records department. The technology to access the EMR and other electronic sources of information will make data acquisition much easier and faster and will enhance data capture speed and reduce transcription errors through automated upload of the information. Applying the NPCR-MERP model will involve two levels of activity. The first will be using the model to support standardization and applications development throughout NPCR-MERP program areas. NPCR-MERP will be consistent with national standards adopted by the Department of Health and Human Services-funded Regional Health Information Organizations (RHIOs) to ensure interoperability of the model with other systems. 10 The model will provide input into the development of design-level applications. The goal is to ensure that with each iteration, the model more closely approximates national representation. NPCR programs can use it for strategic planning, systems development, and operational and process improvement. The second level of activity will be directed at using the model as a tool to aid in long-term cancer surveillance infrastructure improvement. Currently, there is consistent movement toward defining what a cancer informatics infrastructure should be. The Centers for Disease Control and Prevention s PHIN Messaging System 11 and the National Cancer Institute s Cancer Biomedical Information Grid (cabig ) 12 are two national efforts focused on the development of standards-based, interoperable infrastructures to meet data exchange needs for their respective agencies and partners. The NPCR-MERP model is one of several components of a national effort to define best practice, maximize the use of technology, and create an environment that encourages continued innovation and improvement. The time is right for the cancer surveillance community to identify methods to capture and report data electronically, in real time. The cancer community has a unique opportunity to work with national standard setters to develop an interface between the EMR and cancer registries that will meet both clinical and public health needs. The NPCR-MERP initial diagrams demonstrate the current and proposed business processes within the hospital, using VCU/HS and VCR as a pilot. After the model has been distributed to the broader cancer community and input has been compiled, it will be modified to more universally represent hospital and central cancer registry requirements. The refined and more detailed model will serve as a guide for the development of interoperable, standards-based cancer registry software. Acknowledgements This project was developed in conjunction with work performed with the Centers for Disease Control and Prevention s National Program of Cancer Registries Modeling Electronic Reporting Project (NPCR-MERP). Software developed at Virginia Commonwealth University was partially funded through an agreement with the Virginia Cancer Registry which received funds for this project from the Centers for Disease Control and Prevention s National Program of Cancer Registries, cooperative agreement number U55/CCU321957. References 1. Silva JS, Ball MJ, Chute CG, et al. Essential technologies for clinical trials: health informatics. Cancer Inform. 2002;XXVI:377, p.62 illus. 2. Wingo PA, Howe HL, Thun MJ, et al. A national framework for cancer surveillance in the United States. Cancer Causes Control. 2005;16:151-170. 3. Waegemann CP. EHR vs. CPR vs. EMR. Healthc Inform. May 2003. Available at: http://www.healthcare-informatics.com/issues/2003/05_03/cover_ehr. htm. 4. Williams W, Lyalin D, Wingo PA. Systems thinking: what business modeling can do for public health. J Public Health Manag Pract. 2005;11(6):550-553. 5. Bell D. UML Basics: An Introduction to the Unified Modeling Language. IBM. June 15, 2003. Available at: http://www-128.ibm.com/developerworks/ rational/library/769.html. 6. Health Level Seven, Inc. Health Level Seven (HL7). 2005. Available at: http:// www.hl7.org. 7. Centers for Disease Control and Prevention. Public Health Information Network. Available at: http://www.cdc.gov/phin. 8. US Cancer Statistics Working Group. United States Cancer Statistics: 2001 Incidence and Mortality. Atlanta, GA: Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute; 2004. Available at: http://www.cdc.gov/cancer/npcr/uscs/ pdf/2001_uscs.pdf. 9. American College of Surgeons, Commission on Cancer. Cancer Program Standards 2004: Standard 3.3 (p 28); 2003. Available at: http://www.facs. org/cancer/coc/cocprogramstandards.pdf. 10. Department of Health and Human Services. Summary of Nationwide Health Information Network (NHIN) Request for Information (RFI) Responses. June 2005. Available at: http://www.hhs.gov/healthit/rfisummaryreport.pdf. 11. Centers for Disease Control and Prevention. Public Health Information Network Messaging System. Available at: http://www.cdc.gov/phin/software-solutions/phinms/index.html. 12. National Cancer Institute. Cancer Biomedical Informatics Grid (cabig). Available at: http://cabig.nci.nih.gov/overview/interactive_overview/document_view. Journal of Registry Management 2006 Volume 33 Number 3 99
Table 1. Project Objectives for NPCR-MERP 1. 2. 3. 4. 5. 6. Analyze the current and potential status of EMR sources and electronic reporting systems in hospitals to identify opportunities to automate processes in cancer registries. Create a robust and transportable model for electronic case ascertainment and the transmission of cancer abstract data from the EMR inter-field, and other electronic and inter-record). sources of information to the hospital and central cancer registries. Use business modeling techniques and UML for the model notation. Conduct a pilot project to implement the model within a hospital setting. Assess the appropriateness of the model, adjust it accordingly, and ensure that messages comply with HL7 and PHIN standards. Model the technical architecture for acquiring and transmitting cancer registry surveillance data from the EMR and other to move implementation of the model forward. electronic sources of information to the hospital, CCR, and any other direct interfaces. Create a template for PHIN-compliant electronic data exchange messages (HL7). Use standard, national vocabularies to transmit cancer abstract data electronically from the hospital registry to the central registry. Analyze current methods of transmitting cancer abstract data from the hospital registry to the central registry. Develop a model for priorities more frequent is appropriate or staggered submissions. and feasible. Maintain the capacity to perform electronic data edits (single-field, inter-field, and inter-record). 7. Develop a plan or strategy that will define priorities and leverage limited resources 8. Explore whether expanding modeling activities to other health care domains and 7. Develop a plan or strategy that will define priorities and leverage limited resources to move implementation of the model forward. Figure1. Cancer registration: context diagram for the NPCR-MERP project 8. Explore whether expanding modeling activities to other health care domains and priorities is appropriate and feasible. Figure1. Cancer registration: context diagram for the NPCR-MERP project Revision Date: 02-07-06 Hospital level State / Regional level National level Reporting: from Region to State; from State to State NAACCR collects data from Reporting: from Hospital to Hospital Hospital Cancer Registry 0..1 a part of 1 [for Hospital w/o Registry] Hospital a part of In-Hospital Source of Cancer Data data sharing Central Cancer Registry non-hospital Source of Cancer Data collects data from National Population-based Cancer Program NPCR @ CDC SEER @ NCI National Hospital-based Cancer Program Pathology Laboratory Claims Department Pathology Laboratory Healthcare Provider NCDB @ CoC Patient Abbreviations: CDC: Centers for Disease Control and Prevention CoC: American College of Surgeons Commission on Cancer NAACCR: North American Association of Central Cancer Registries NCDB: National Cancer Data Base NCI: National Cancer Institute NPCR: National Program of Cancer Registries SEER: Surveillance, Epidemiology, and End Results Program How to read and interpret the context diagram: Relationships between entities are visualized by connecting lines. Names associated with these lines describe the type of relationship between entities. Example: A relationship between Hospital Cancer Registry and Central Cancer Registry is shown as a connecting line with the name. Such a relationship should be read as Hospital Cancer Registry Central Cancer Registry. 100 Journal of Registry Management 2006 Volume 33 Number 3 12
Appendix 1: Glossary of Terms Cancer Biomedical Informatics Grid (cabig ) The Cancer Biomedical Informatics Grid (cabig) is a voluntary network or grid connecting individuals and institutions to enable the sharing of data and tools, creating a World Wide Web of cancer research. The goal is to speed the delivery of innovative approaches for the prevention and treatment of cancer. The infrastructure and tools created by cabig also have broad utility outside the cancer community. cabig is being developed under the leadership of the National Cancer Institute s Center for Bioinformatics. Electronic Messaging Protocols Electronic messaging protocols are the standards, guidelines, or specifications required for the secure transmission of data electronically. Granular Granular refers to the level of detail that is described in the model. Health Level Seven (HL7 ) Organization Health Level Seven is one of several American National Standards Institute (ANSI)-accredited Standards Developing Organizations (SDOs) operating in the health care arena. Most SDOs produce standards (sometimes called specifications or protocols) for a particular health care domain such as pharmacy, medical devices, imaging, or insurance (claims processing) transactions. Health HL7 s domain is clinical and administrative data. Our mission is: To provide standards for the exchange, management and integration of data that support clinical patient care and the management, delivery and evaluation of health care services. Specifically, to create flexible, cost effective approaches, standards, guidelines, methodologies, and related services for interoperability between health care information systems. Health Level 7 (HL7) Standard HL7 is a formatting standard for structuring, storing, and messaging clinical data. The standard also supplies a basic set of vocabularies to be used for the attributes in the HL7 Reference Model. HL7 v.3.0 specifications describe 6 basic components: 1) The sets of fields or attributes that make up a message 2) The vocabularies that are needed to enforce consistent data entries in the fields 3) The logical database structure for storing the records 4) The messaging or transport method that the records are shared by 5) The structure of the message to be shared, XML 6) The relationships of the various components in an HL7 message that follow a hierarchy. Interoperability Interoperability is the ability of systems (software and hardware) to communicate and share data with other systems. Logical Observation Identifiers, Names, Codes (LOINC ) LOINC is a clinical terminology important for laboratory test orders and results, produced by the Regenstrief Institute. LOINC is one of a suite of designated standards for use in US Federal Government systems for the electronic exchange of clinical health information. Public Health Information Network (PHIN) The PHIN is CDC s vision for advancing fully capable and interoperable information systems in the many organizations that participate in public health. PHIN is a national initiative to implement a multi-organizational business and technical architecture for public health information systems. With the acceptance of information technology as a core element of public health, public health professionals are actively seeking essential tools capable of addressing and meeting the needs of the community. Regional Health Information Organization (RHIO) A RHIO is a multi-stakeholder organization that enables the exchange and use of health information, in a secure manner, for the purpose of promoting the improvement of health quality, safety, and efficiency. Officials from the US Department of Health and Human Services see RHIOs as the building blocks for the National Health Information Network (NHIN). Robust and Transportable Model A robust and transportable model provides tools and processes that allow easy implementation in any environment. Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT ) A dynamic, scientifically-validated, clinical health care terminology and infrastructure that makes health care knowledge more usable and accessible. The SNOMED CT core terminology provides a common language that enables a consistent way of capturing, sharing, and aggregating health data across specialties and sites of care. Journal of Registry Management 2006 Volume 33 Number 3 101