The overall goals of this 4-year study are to improve the quality of preventive care for veterans by developing and implementing a new performance measure to assess completion of colorectal cancer (CRC) screening and to evaluate interventions and innovations to improve CRC screening within the VA.
The specific aims of the study are: (1) to use VA administrative data to construct VA-wide, veteran-level longitudinal data on CRC screening adherence among veterans at average risk for CRC, and (2) to examine the effects of VA structural characteristics on CRC screening adherence. To accomplish the first aim, the investigators will use VA Medical SAS (1997-2010), Fee Basis (1997-2010), and VA-Medicare (1997-2010) data to construct a longitudinal data set for the period 2001 to 2010. Using those data sources, the investigators will construct 2 study cohorts at average-risk for CRC - one for non-elderly veterans (50-64 years) and one for elderly veterans (65-75 years). Data collected for aim 1 also will be used to examine the effects of 3 VA structural characteristics on CRC screening adherence: improved access to primary care through community-based outpatient clinics; the implementation of the OncWatch intervention; and the dual-use system. Importance of the problem addressed. The study has the potential to improve the delivery of preventive health care, in this instance CRC screening, by developing a standardized measure to identify veterans who have not been screened or are overdue for screening. The measure can also be used to monitor trends in screening within and across VA settings and to evaluate the effects of interventions to increase screening adherence. Currently, performance is measured with small samples of medical records at each facility. The proposed measure would be available on all VA users and older dual-users, thus resulting in a more stable and presumably more accurate estimate of CRC screening. This proposal addresses the VA HSR&D priority areas of healthcare informatics and research methodology. Contribution to the VHA. The measure of screening adherence developed in this research study could be used throughout the VA to improve the delivery and surveillance of CRC screening. In addition to standardizing performance assessment, the use of existing data is likely to be more cost-effective than chart reviews. Adequacy of response to previous feedback. Not applicable. Methods. This is a well written proposal by a team of investigators with substantial expertise in health services research, particularly in the use of VA administrative data and large national data sets. The investigators provide detailed inclusion and exclusion criteria for selecting average-risk veterans for the cohorts. Likewise, the description of the variables and the analysis plan are very detailed and thorough. Using secondary data, they propose to develop a measure of CRC screening adherence that could be used system-wide to permit the tracking of CRC screening performance and comparison across sites. Another strength of the study is the plan to create a longitudinal data set to monitor trends in screening at the level of the individual patient, the VA clinic or hospital, and nationally. Such a system would increase the chances that veterans received timely and appropriate screening and follow-up. Such a measure also could reduce the problem of over-screening which could occur if medical records are difficult to access or incomplete, for example, due to dual-use of health care systems. A potential problem with the study is feasibility. The linked VA-Medicare data are currently unavailable and it is not clear if or when they will become available. The investigators plan to use that data set to study patterns of utilization for the older veteran cohort who are dual users of both VA and Medicare services. If those data do not become available, the investigators will revise this aspect of the study. A related issue is that the VA-Medicare data sets are currently available only for the period 1997 to 2004, and there is an approximately 2-year time lag for release of these data sets. This means that the data for 2010 will not be available, if at all, until 2012. So even if the data sets do become available during the time frame of the study, there may be a lot of down time waiting for the data from 2010 due to the 2-year time lag. Adequacy of data. The data are adequate to address the research questions proposed. Project organization and management. The roles and responsibilities of the research team are clearly described. As described above, the scope of work is uncertain due to questions about the availability of the VA-Medicare data and the time lag for receiving the data even if it becomes available. Plans to disseminate the intervention include through presentations and publications and through working with VA policymakers to implement the performance measure and extend it to other types of performance measures. Investigator qualifications. All of the investigators are well-qualified to conduct the proposed project. Dr. John Bain is PI of the project. He is a health services researcher with extensive experience in the analysis of large-scale secondary data sets. Currently, he is the recipient of a VISN 7 Career Development Award through which he has conducted pilot work to support aspects of the feasibility of the proposed project. Dr. Joseph Lipscomb is a health economist and is nationally and internationally recognized for his expertise in outcomes research. Dr. Deborah Fisher is a gastroenterologist and health services researcher at the Durham VA whose research has focused on CRC screening and surveillance, including quality assessment. Dr. Maria Ribeiro was involved in initiating and implementing the VISN 7 CRC screening OncWatch intervention that was designed to improve CRC screening adherence in the VA. Human subjects. No concerns. Inclusion of women and minorities. No concerns. Facilities and resources. The facilities and resources are adequate to support the project. Budget. Any concerns about the budget relate to the likelihood that the investigators will be able to conduct the project as proposed. Overall strengths. Strengths of the project include: (1) an innovative use of VA administrative data and other secondary data sets to develop a performance measure for CRC screening that can be used throughout VA health care settings and has the potential to improve the delivery of preventive health care for veterans as well as to monitor performance at a systems level; (2) preliminary data demonstrating feasibility of the proposed project; (3) a detailed description and justification of the design and methods; and (4) a strong team of investigators with relevant expertise to conduct the project. Overall weaknesses. A weakness is that the VA-Medicare data sets are currently available only for the period 1997 to 2004, and there is an approximately 2-year time lag for release of these data sets. Thus, even if the Medicare data become available there could be a delay waiting for the data from 2010 due to the 2-year time lag. Key issues (summary bullets): 1. A key issue is uncertainty about if or when the VA-Medicare data set will be available. 2. A second key issue is the time lag for acquiring the data even if it becomes available. This lag may unnecessarily lengthen the duration of the project.
Taking advantage of high-quality VA administrative data, this study will construct VA-wide, veteran-level, longitudinal data measuring CRC screening adherence that may be used as a potential alternative to the current chart review-based measurement. Moreover, using the newly constructed longitudinal data, this study will assess the effects of VA structural characteristics on performance of CRC screening adherence. The understanding gained from these empirical analyses may assist the VA in designing better interventions to improve CRC screening adherence by millions of veterans in the future. 1