Health Information Technology Call for Papers Clinical Health IT in Hopsitals & Ambulatory Settings (Including E-Prescribing) Chair: Christine Elnitsky, Department of Veterans Affairs Monday, June 26 • 10:30 am – 12:00 pm ●The Promise of Clinical IT Systems? Implications for Practice from Qualitative Research Nance Goldstein, Ph.D., MSc, MPhil, BA Presented By: Nance Goldstein, Ph.D., MSc, MPhil, BA, Professor / Resident Scholar, Economics Department / WSRC, University of Southern Maine/ Brandeis University, 14 ½ Fayette St, Cambridge, MA 02139; Email: nance@brandeis.edu Research Objective: To capture the experience of clinicians in adapting to new IT systems that aimed to improve the quality of patient data, the consistency of care practices and the use of scarce resources. The study begins to identify possible effects of IT-mediated documentation on patient care and workforce attitudes. Study Design: Observational case study in cardiac/pulmonary department of urban rehabilitation hospital before and after implementation of information technology clinical documentation system in 2002 and 2003. Includes pre- and post-implementation interviews with 25 clinicians/users. Also senior hospital and corporate healthcare network executives, IT managers and planners and professional practice managers, using a semi-structured protocol. Population Studied: Observations and interviews of patient care clinicians and IT system users included nurses, case managers, physical/occupational/respiratory/ speech therapists, nurses’ aides and multiple levels of professionals involved in IT planning and implementation. Principal Findings: - Clinical users felt they suffered from design errors, IT shortcuts and the additional required tasks without acknowledgement or compensation. Clinicians individually faced the dilemma – sacrifice time with the patient or stay late to complete IT requirements. - Users noticed the benefits gained by higher status clinicians (e.g., the opportunity to work remotely in quiet offices). Also the implementation excluded key classes of clinicians (for differing reasons) which increased the workload for system users. - Resource decisions for IT implementation may ignore or underestimate the need for supporting time for clinical users both to enable their contributions to design and implementation planning and to compensate the extra time required for successful use of IT as part of patient care work practices. - While improving the quality and access of up-todate patient data, IT systems also associate data capture with individual clinicians, threatening to isolate individuals as the source of medical errors. - One month after the IT implementation, one-fifth of clinician respondents reported a decrease in their satisfaction with their job and their performance. Thirty-five percent disagreed with the statement that the system’s implementation had increased the quality of care. The IT department had a backlog of 12 website pages of changes requested from clinical users. Conclusions: Clinical IT systems dramatically disrupt patient care. The new systems represent complex adaptive changes and important two-way flows as practitioners alter the system and the IT tools influence work practices, workplace dynamics and organizational culture. This suggests the need for detailed studies specific to healthcare contexts – their work practices, interdependencies and cultural norms. Implications for Policy, Delivery, or Practice: The study suggests that researchers, IT professionals and healthcare administrators consider in much more detail the context for healthcare IT systems. A wider socio-technical framework for planning and implementation may optimize the intended strategic organizational and patient care benefits. Primary Funding Source: No Funding ●Physicians’ Use of E-prescribing Systems in Today’s Market Joy Grossman, Ph.D., Anneliese Gerland, Cheryl Fahlman, Marie Reed Presented By: Joy Grossman, Ph.D., Senior Health Researcher, , Center for Studying Health System Change, 600 Maryland Ave, SW, #550, Washington, DC 20024; Tel: (202) 484-3298; Fax: (202) 484-9258; Email: jgrossman@hschange.org Research Objective: In what ways are physician practices using electronic prescribing? What are the major barriers and facilitators to electronic prescribing adoption and use? Study Design: Qualitative analysis of discussions with 12 physician practices with e-prescribing and vantage organizations such as health plans conducted between November 2005 and January 2006. The practices were selected from among the 12 Community Tracking Study (CTS) sites to provide a mix of practice size and ownership; physician specialties and geographic location. Respondents included administrators, physicians and other clinical and administrative staff who provided perspectives on how the system is used in the practice; the barriers and facilitators to adoption and use; and the impact on practice operations, physician prescribing patterns and patient satisfaction. Population Studied: Physician practices using e-prescribing located in the 12 CTS sites. Principal Findings: Respondents see significant benefits to using e-prescribing and few would go back to paper. However, they also identified substantial challenges to implementing and using these systems. Respondents value the basic documentation features of e-prescribing systems most highly. They are more mixed in their assessment of the value of safety alerts, with many physicians finding them disruptive and having little impact on prescribing choices. Pharmacy callbacks for clarifications due to illegibility have declined dramatically. However, most practices still get formulary callbacks, because physicians either do not have access to formulary data or find that it is not complete or accurate. Similarly, most practices have faced hurdles with local pharmacies and mail order companies in implementing electronic faxing. Few have true “end-to-end” electronic transmissions with pharmacies because of a variety of factors—the feature is not offered, state regulatory constraints and/or the lack of readiness of local pharmacies. Respondents also identified organizational challenges to implementation; the most substantial being overcoming physician resistance and changing practice workflow. Conclusions: There is a substantial gap between the vision of e-prescribing based on a completely electronic system with advanced clinical support tools and the reality of how it is used in physician practices today. While much attention has focused on physicians’ reluctance to adopt e-prescribing, our findings suggest that there are substantial challenges to successful implementation that are due to external factors outside of the physician practice. Implications for Policy, Delivery, or Practice: Our findings suggest policy makers should continue to support efforts focused on improving the value proposition to physicians of adopting commercial e-prescribing products. For example, certification efforts could help ensure the availability of products that have the most important features to improve efficiency and quality. Promoting the availability of complete formulary information would be very valuable, especially with the introduction of so many Medicare Part D plans. Similarly, efforts to resolve state regulatory issues and assist pharmacies in becoming adept at handling electronic transmissions are critical. Primary Funding Source: AHRQ ●Case Statement for Implementation of CPOE in all Massachusetts Hospitals Doug Johnston, MA, Erica Drazen, Sc.D., Mitchell Adams, M.B.A., Wendy Everett, Sc.D., Keith MacDonald Presented By: Doug Johnston, MA, Director of Research, Research, New England Healthcare Institute, One Broadway, Twelfth Floor, Cambridge, MA 02142; Tel: (617) 225-0857; Fax: (617) 225-9025; Email: djohnston@nehi.net Research Objective: Computerized physician ordered entry (CPOE) systems have been shown to reduce medical errors, improve patient safety, and reduce costs in inpatient settings. However, CPOE adoption rates in Massachusetts hospitals remain low; only 10 percent of the state’s acute care hospitals are currently equipped with CPOE, and only 20 percent have begun installing these systems. One barrier to CPOE adoption is a compelling analysis of financial value to payers, providers, and patients. The objective of this study was to analyze the costs and savings associated with implementing CPOE in Massachusetts hospitals that do not already have these systems. Study Design: A team of researchers from the New England Healthcare Institute (NEHI), the Massachusetts Technology Collaborative and First Consulting Group developed a costbenefit model that projected the net savings from inpatient CPOE adoption in Massachusetts. Cost estimates were derived using data from hospitals that had previously implemented vendor-supplied CPOE systems. NEHI projected savings using data from peer-reviewed literature on reduced inpatient adverse drug events (ADEs) and improved utilization of inpatient and emergency department (ED) resources from CPOE. This model predicted to whom and in what proportions any potential savings would accrue. Based on discussions with managed care and hospital reimbursement experts, the team defined the predominant payment mechanisms in Massachusetts and calculated the inpatient and ED benefits following CPOE implementation under each mechanism. NEHI identified which stakeholder—hospitals, payers/employers, and patients—would most likely accrue the benefit in each category. These proportions were then applied to the calculated total to yield the potential savings accrued to each stakeholder. Population Studied: NEHI limited the analysis to benefits that could be generalized across all Massachusetts hospitals, payers, and the entire inpatient population. Principal Findings: NEHI found that the projected costs for CPOE installation in Massachusetts hospitals amounted to approximately $210 million over three years, with ongoing operating costs of $25 million annually. The total estimated annual benefit for inpatient CPOE in the entire Massachusetts health care system was $1.48 billion. Reflecting only those hospitals that have not yet begun CPOE implementation, the total projected savings was $803.4 million. Adjusting for fixed hospital costs, and assuming that only 75 percent of the $803.4 million would be realized, NEHI predicted that CPOE would save the Massachusetts health care system a total of $299.4 million. Of this total, $202.1 million, $89.6 million, and $7.8 million would accrue to hospitals, payers/employers, and patients respectively. Spreading one-time implementation costs evenly over three years, annual net savings from these hospitals would be approximately $275 million. Conclusions: In addition to patient safety and quality improvements, NEHI concluded there is a compelling financial case for the installation of CPOE in all Massachusetts hospitals. Implications for Policy, Delivery, or Practice: Achieving projected cost savings from inpatient CPOE in Massachusetts requires a coordinated, collaborative initiative between providers, payers, and vendors of these systems. Such an initiative should specify system performance and functional standards, negotiate with key vendors, develop a funding and incentive program, and establish time frames for achieving universal inpatient CPOE adoption in Massachusetts. Primary Funding Source: New England Healthcare Institute and Massachusetts Technology Collaborative ●Physician Organizations’ Use of Clinical Decision Support for Order Entry Jodi Simon, MS, Thomas G. Rundall, Ph.D. Presented By: Jodi Simon, MS, School of Public Health, University of California at Berkeley, 9 Longview Court, San Francisco, CA 94131; Tel: 415-285-2810; Fax: 415-648-5524; Email: jodi_sacks@yahoo.com Research Objective: to understand what organizational structure and market characteristics of physician organizations are associated with the use of order entry with decision support for the care of patients with diabetes, asthma and congestive heart failure. Study Design: a quantitative nationwide survey of all physician organizations in the United States with 20 or more physicians. The data were collected by trained interviewers at the National Opinion Research Center at the University of Chicago in 60-minute structured interviews with the CEOs, presidents, or medical directors of the physician organizations. Population Studied: data were collected on 1104 physician organizations, representing a 70% response rate. Respondents and non-respondents did not differ by size or state where they were located. Of the 1104 organizations responding, 738 were medical groups and 366 were independent practice associations. Principal Findings: the use of order entry with decision support for chronic disease care is not common in physician organizations; 73% of the organizations did not use order entry with decision support for any of the three chronic conditions. The presence of external incentives for quality was significantly associated with the use of these tools. We found no statistically significant net relationship between adoption of these tools and organization size, organization age, hospital or HMO ownership, level of HMO penetration in the county, or urban versus rural setting. Conclusions: the use of order entry with decision support is not being greatly utilized in physician organizations to help manage patients with chronic disease. Those organizations that are leveraging these tools are more likely to be those experiencing compelling external incentives to improve quality. Implications for Policy, Delivery, or Practice: the strong relationship between external incentives for quality and the use of order entry with decision support suggests that environmental factors play an important role in physician organizations’ decisions to adopt this type of technology. Policies requiring reporting of chronic care measurements and rewarding improvement as well are financial incentives for the use of specific clinical information technology tools are likely to accelerate adoption in physician organizations. Primary Funding Source: RWJF ●Adoption of Electronic Prescribing in Community-based Medical Practices Christine Vogeli, Ph.D., Michael A. Fischer, M.D., MS, Rainu Kaushal, M.D., M.P.H., Timothy G. Ferris, M.D., M.P.H., Daniel Z. Sands, M.D., M.P.H., Joel Weissman, Ph.D. Presented By: Christine Vogeli, Ph.D., Instructor of Medicine, Institute for Health Policy, Massachusetts General Hospital, 50 Staniford Street, 9th floor, Boston, MA 02114; Tel: 617 7240984; Fax: 617 724 4738; Email: cvogeli@partners.org Research Objective: Emerging data suggests that electronic prescribing is an important tool for improving both the safety and efficiency of prescribing. Despite this evidence, adoption of electronic prescribing systems remains limited, particularly in the ambulatory setting. This study sought to understand how quickly physicians adopt electronic prescribing systems when available, and the patterns of use after adoption. Study Design: In October 2003, two major health insurers, their PBMs, and a large e-prescribing company collaboratively launched a voluntary e-prescribing program in Massachusetts in which prescribers were provided with eRx software and free handheld PDAs. The e-prescribing system allows prescriptions to be written either with a handheld PDA device or on a personal computer linked to the internet. Prescriptions are immediately transmitted to pharmacies electronically or by fax. In order to study adoption rates, we identified physicians who had been enrolled as users of the electronic prescribing system and tabulated the frequency with which they wrote electronic prescriptions between April 2004 and March 2005, the first year the system was in use. Population Studied: Physicians practicing in a large sample of community practices in Massachusetts in 2004-2005. Principal Findings: The number of physicians using the electronic prescribing system increased steadily, from 131 prescribers in April 2004 to 1,024 in March 2005. The number of prescriptions per month increased rapidly, from 4,000 in April 2004 to over 55,000 in March 2005. Over time, electronic prescribers became more active users of the system, growing from 30 prescriptions per month to 54 prescriptions per month during the study period, an 80% increase. In the first six months of the study period, only 6% of enrolled physicians were writing more than 100 electronic prescriptions per month. By the final month studied this proportion had more than doubled, with over 15% of enrolled physicians writing more than 100 electronic prescriptions per month. Conclusions: Physician use of electronic prescribing increased almost ten fold with the provision of an insurance company subsidized electronic prescribing system. Among adopters, actual use of electronic prescribing systems by individual physicians in community settings increased rapidly An increasing proportion of physicians in our sample became high-frequency electronic prescribers. Once electronic prescribing is implemented, actual use can accelerate rapidly. Implications for Policy, Delivery, or Practice: Electronic prescribing has been recognized as an important tool for reducing medication errors and has the potential to allow for more cost-effective prescribing in inpatient settings. Adoption of e-prescribing tools in the community setting has the potential to extend these benefits. Primary Funding Source: AHRQ Call for Papers Making Electronic Health Exchange Real for Providers and Consumers: Internet, Portal, EMR & Business Case Chair: Patricia MacTaggart, EDS Tuesday, June 27 • 8:45 am – 10:15 am ●Developing a Business Case Model for Integrated Child Health Information Systems Tim Dall, Kristin Saarlas, MPH, Alan Hinman, MD, MPH Presented By: Tim Dall, Vice President, The Lewin Group, 3130 Fairview Park Dr., Suite 800, Falls Church, VA 22042; Tel: 703-269-5743; Fax: 703-269-5503; Email: tim.dall@lewin.com Research Objective: To improve state public health agencies’ capacity to assess the costs and benefits of integrating various child health information systems (e.g. newborn screening, early hearing detection and intervention, immunization registries, WIC, lead screening and birth defects) by developing a business case modeling tool. Study Design: The business case model for integrated child health information systems (ICHIS) was developed with expert input by a workgroup representing state public health agencies, physicians, health plans and family advocates. Assumptions on the marginal costs and benefits of integration were developed using published literature and unpublished estimates by the experts in the field. The model, and its underlying assumptions on effectiveness and efficiencies of integration, was pilot tested in five states in January 2006. Population Studied: Costs and benefits associated with state public health programs related to child health. Principal Findings: The conceptual framework and underlying assumptions for the model have been agreed upon by the workgroup members and represent best available data. Findings from five states that pilot tested the business case model in Jan 2006 will be presented, including the feasibility of using the modeling tool, the validity of the data, and preliminary cost/benefit results based on different state information system models. Several areas of the model, especially those associated with costs of integration, are lacking and require additional efforts. Conclusions: Public health agencies need tools to assess the benefits and costs (i.e., return on investment) of integrating separate health information systems in order to justify the additional costs of integration to various stakeholders. The business case model for ICHIS allows for states to develop a customized business case to fit their current integration model as well as estimate a future scenario. Flexible models such as this tool can provide states with information to decide which systems should be integrated and provide results from a societal perspective as well as specific to government, providers and families/payers. Additional support to states to collect data and interpret the results of their business cases are required to maximize the benefit of this tool. Implications for Policy, Delivery, or Practice: The business case model for ICHIS can be used to inform policy makers, public health leaders and providers on the costs and benefits of health information systems and the impact of such systems on the delivery and coordination care for children. Primary Funding Source: HRSA, RWJF ●Time to Reap: Improving Quality by Harvesting Data from the EMR Marie Eidem, B.S., Aaron Kurtzhals, B.S., James Naessens, M.P.H. Presented By: Marie Eidem, B.S., Lead Analyst/Programmer, Health Care Policy & Research, Mayo Clinic, 200 First Street SW, Rochester, MN 55905; Tel: (507)284-5723; Fax: (507)2841731; Email: eidem.marie@mayo.edu Research Objective: To assess the accuracy and reliability of immunization data extracted from an electronic medical record for publicly-reported quality measures. Study Design: Pneumococcal vaccination data was queried from our electronic medical record. A computer algorithm was applied to determine whether this vaccination was received according to reporting guidelines for ORYX Core Measures. The algorithm was verified against a small sample before being fully assessed. These results were compared, at the patient level, with responses previously collected based on manual record review by registered nurses. Discrepancies between the computer-calculated result (EMR) and nurseabstracted result (RN) were reviewed by another nurse, who determined the “true” (T) immunization status. Population Studied: Primary and secondary pneumococcal vaccination data from electronic medical records of 523 patients hospitalized with community acquired pneumonia (CAP) from 7/1/2004-3/31/2005. Because 289 patients did not meet pneumococcal vaccination age requirements under ORYX Core Measures, RN results were not available for them and they were excluded from analysis. Our current EMR does not differentiate between refusals and contraindications; therefore, an additional 7 records without EMR results were excluded. Also excluded were 3 patients not granting research authorization, leaving 224 patient records studied. Principal Findings: Initial discrepancies between EMR and RN results pointed out minor logic errors in the computer algorithm. After this was corrected, EMR and RN assessments matched in 164(73.2%) records: 23 agreed that vaccination was given during this hospitalization; 119 agreed it was given prior to hospitalization; and 22 agreed that no documentation was found. Of the 60 records with discrepancies, 51 had an RN but no EMR result. Four of those records plus the remaining 9 with discrepancies were reviewed by another RN. The EMR result was correct in 12(92.3%) of these cases. Overall, a positive indication from the EMR was correct in 151 of 151 records(100% positive predictive value). A negative result from the EMR was correct in 25 of 73 records(34.2%), due to vaccination documentation occurring outside of the EMR immunization module. Evaluation of the discrepancies also brought to light two potential areas of clinical practice improvement: vaccinations given multiple times(6) and vaccinations not given until after dismissal from hospital(3). Conclusions: The ability to compare manually abstracted and electronically captured data at the patient level was key in identifying errors in the computer algorithm. Moving to a more automated method, where RNs focus abstraction efforts on cases with no immunization data found in the EMR will improve accuracy and save valuable staff time. Review and categorization of discrepancies may identify areas for both documentation and clinical practice improvement. Implications for Policy, Delivery, or Practice: Manual abstraction of quality care measures is a resource-intensive process that is also subject to human error and bias. With pay-for-performance initiatives on the horizon, the magnitude of this problem will certainly grow. Reporting of quality care measures can no longer rely on manual abstraction by even the best trained clinical personnel, but must take advantage of the increasing capabilities of the electronic medical record. Primary Funding Source: No Funding ●How Older Adults use the Internet to Look for Health Information Kathryn Flynn, Ph.D., Maureen A. Smith, M.D., Ph.D., M.P.H., Jeremy Freese, Ph.D. Presented By: Kathryn Flynn, Ph.D., Postdoctoral Fellow, Center for Clinical and Genetic Economics, Duke University, PO Box 17969, Durham, NC 27715; Tel: (919) 668-4617; Fax: (919) 668-7124; Email: kathryn.flynn@duke.edu Research Objective: Many patients report that they want more information than their physicians have time to provide. Health care decision aids may be more successful if we better understand how patients use non-physician sources of health information, such as the Internet, to seek health information. Preparing for an upcoming doctor visit by seeking information online before a visit may help patients participate in decision making during a visit. Conversely, seeking information after a visit may suggest that patients need more information or support than they received during the visit. Seeking information online instead of visiting a doctor could be problematic given the variable quality of information available. Our objective was to determine how older adults use the Internet to seek health information, especially with regard to timing in relation to doctor visits. Study Design: The Wisconsin Longitudinal Study Graduate Survey follows a one-third random sample of graduates from Wisconsin high schools in 1957 using telephone and mail surveys. Two dependent variables were use of the Internet to search for health information and timing of use. Explanatory variables included self-reported effort to stay healthy, preferences for information and decision making during health care visits, personality traits, and length of relationship with a usual provider of care. Multinomial logistic regression adjusted for gender, marital status, number of children, rural/farm origin, cognitive ability in high school, educational attainment, health insurance, SF-12 physical and mental component summary scores, number of common conditions, and number of regularly taken prescription medications. Population Studied: All 5919 respondents, most aged 63 to 66 years, who completed the 2004 telephone and mail surveys. Principal Findings: One-third of respondents had searched the Internet for information about their own health or health care. Half of these had searched for health information unrelated to their last doctor visit, while one-third searched after a visit and one-sixth searched before a visit. Educational attainment, cognitive ability in high school, and greater openness-to-experience were positively associated with searching the Internet for health information irrespective of timing. Compared to those who had never sought health information online, sicker individuals (especially those with cancer) were more likely to seek information online either after or unrelated to a doctor visit. A preference for being given many treatment choices rather than letting a doctor make decisions, greater neuroticism, and greater efforts to stay healthy were related to seeking health information online either before or unrelated to a visit. Conclusions: A significantly smaller percentage of older adults seek health information on the Internet than some previous estimates have suggested. Although the majority of respondents seek health information online unrelated to a doctor visit, there are important differences in timing. Our findings provide new insight into how sociodemographic characteristics, health, cognitive factors, and personality relate to seeking health information outside of doctor visits. Implications for Policy, Delivery, or Practice: Understanding how patients use alternative sources of health information is important to facilitating shared decision making. An important next step will be to examine outcomes associated with seeking health information online and whether outcomes vary by the timing of or motivation for search behaviors. Primary Funding Source: AHRQ ●Hospital-Physician Portals: A First Step In Sharing Patient Data Across Care Settings Joy Grossman, Ph.D., Thomas S. Bodenheimer, M.D., Kelly McKenzie Presented By: Joy Grossman, Ph.D., Senior Health Researcher, , Center for Studying Health System Change, 600 Maryland Ave., SW, #550, Washington, DC 20024; Tel: 202484-3298; Fax: 202-484-9258; Email: jgrossman@hschange.org Research Objective: To assess the extent of the adoption of information technology (IT) to share clinical data among providers in local communities. Study Design: Qualitative analysis of site visit interviews with health care executives in 12 Community Tracking Study (CTS) markets selected to be nationally representative of communities with populations over 200,000. CTS Round 5 site visits were conducted between January and June 2005. Over 256 respondents were asked how IT was used to share clinical data across patient care settings within and across organizations and the extent to which stakeholder organizations were engaged in activities to promote community-wide clinical data sharing activities. Population Studied: Largest hospitals/hospital systems, physician groups, safety net providers, and health plans and local health IT organizations and experts in the 12 CTS communities. Principal Findings: Most large hospitals have or are developing physician portals to provide admitting physicians with remote access to critical patient records. Portals may also give physicians the ability to perform transactions such as ordering tests and signing charts. We found little data sharing among unaffiliated organizations beyond the evolving community-wide efforts in Indianapolis and Boston and initial collaborative discussions in several other sites. Competition among hospitals for physicians is a key factor driving adoption of proprietary physician portals. In contrast, provider and health plan competition and adversarial relationships between providers and plans are viewed as significant barriers to community-wide clinical data sharing. Conclusions: Physicians portals have more limited data sharing than that envisioned under community-wide health information exchanges. Access is restricted to physicians affiliated with the hospital. Moreover, portals are typically confined to hospital records and in only a few cases provide access to ambulatory care data. Nonetheless, given the long time-horizon to developing a national health information network, physician portals are a small but significant step towards more complete clinical data exchange. Whether these proprietary systems are an intermediary step in the evolution of a nationwide system of clinical data exchange remains to be seen. Implications for Policy, Delivery, or Practice: Policy makers need to consider the important role market competition plays in spurring or hindering IT adoption. Proposed changes to the Stark and anti-kickback rules may make it easier for hospitals to support IT adoption among physicians. They are more likely to do so if they perceive a competitive advantage to tie physicians more closely to their institutions. In contrast, market competition is a substantial barrier to developing collaborative community-wide efforts, particularly in larger, more competitive markets where health care organizations are already making substantial proprietary investments in IT. Primary Funding Source: RWJF ●New York State Physicians and the Adoption of Innovation: Use of Internet and E-Mail in Practice, 19992004 Sandra McGinnis, Ph.D., Jean Moore, BSN, MSN, Chris Morrett, Ed Salsberg, M.P.H. Presented By: Sandra McGinnis, Ph.D., Research Associate, Center for Health Workforce Studies, School of Public Health, University at Albany, 7 University Place, Rm-B334, Rensselaer, NY 12144; Tel: 518-402-0250; Fax: 518-402-0252; Email: slm12@health.state.ny.us Research Objective: To examine the predictors of physician use of Internet/e-mail for five practice-related functions, and to examine changes in the profiles of users over time as the prevalence of use increases. Study Design: Data are taken from the New York State Physician Reregistration Survey, which surveys physicians on their characteristics and practice patterns as they renew their license on a three-year cycle. Since 1999, the survey has included a question about use of the Internet/e-mail for five functions: obtain lab results, x-rays or hospital records; obtain information about treatment alternatives; communicate with/answer questions from patients; obtain Continuing Medical Education credits; and transmit prescriptions to pharmacies. Ordinary least squares regression is used to analyze the predictors of technology use during the three survey cycles. Predictors are compared for the same types of use during different periods, to provide a profile of how physicians move through the five stages of adoption of innovation introduced by Rogers (1995). Population Studied: Physicians active in practice in New York State, 1999-2004. Principal Findings: Use of Internet/e-mail by physicians has increased for all five of the functions analyzed between 19992004. Use rates in the 2003-2004 do not exceed 50% for any function, however, indicating that physician use of technology remains in either the early adopter or early majority stage for all five functions. Predictors of technology use vary between the five functions. The profile of users for each function also changes over time, with age and gender differences in particular becoming less salient between users and non-users between the innovator, early adopter, and early majority stage. Conclusions: Use of Internet/e-mail by physicians in New York State has increased markedly, but has not yet entered the late majority stage. Use of Internet and e-mail is being adopted more quickly for some functions (e.g. obtaining CME) than for others (e.g. transmitting prescriptions). The profile of physicians who use Internet/e-mail for functions that are less common differs in some important ways from the profile of physicians who use the technology for functions that have become more common. In particular, demographic characteristics of physicians are more associated with use in early stages of adoption, while practice patterns such as specialty and setting are more associated with use in later stages of adoption. Results also imply, however, that the early users of technology for one function are not necessarily the early users of technology for other functions, pointing to a need for further research focused on the adoption of specific types of technology. Implications for Policy, Delivery, or Practice: Given the tremendous attention focused on implementing better health information technology systems, the results may prove a valuable tool to identify physicians who are the first to use Internet/e-mail technology for various functions and those who still do not use Internet/e-mail for these functions. This has important implications for marketing new technologies towards those who are more likely to be innovators, while targeting policies designed to encourage technology use towards physicians who may be most resistant to use of technologies. This may also inform the design of new products and services to complement and capitalize on existing patterns of use, leading to more efficient transitions to new technologies. Primary Funding Source: Other Related Posters Health Information Technology Poster Session A Sunday, June 25 • 2:00 pm – 3:30 pm ●The Influence of Environmental Factors on EHR Adoption by Physicians Maziar Abdolrasulnia, M.B.A., M.P.H., Richard M. Shewchuk, Ph.D., Nir Menachemi, Ph.D., W. Jack Duncan, Ph.D., Douglas J. Ayers, Ph.D., Robert G. Brooks, M.D. Presented By: Maziar Abdolrasulnia, M.B.A., M.P.H., Doctoral Student, Health Services Administration, University of Alabama at Birmingham, 2010 5th Street South, Birmingham, AL 35205; Email: maziabdo@uab.edu Research Objective: Despite the benefits associated with electronic health record (EHR) systems, physicians have been slow to adopt this technology. Existing studies that have examined EHR adoption have largely ignored environmental factors that may influence the decision to adopt. We examined the relationship between county level characteristics and EHR adoption in small to medium-sized physician practices. Study Design: This project combined both primary and secondary data. The two main sources of the data include (1) A physician IT survey conducted in Florida during the spring of 2005; and (2) 2004 area resource file (ARF) containing county characteristics. Based on the associated healthcare literature, we operationalized environmental factors per county as physicians per capita, Medicare managed care penetration, and poverty level. We used hierarchical logistic regression analysis to examine the relationship between these environmental factors and EHR adoption by physicians. Population Studied: A total of 2961 physician practices were analyzed. The majority (94%) of physician practices were in urban area, 38% were solo practices, 48% were in groups of two to five, and 14% were in groups larger than 6. Mean (std. dev.) years since medical school of the sample was 21.1 (±10.0) years. Principal Findings: Overall 17.4% of physicians self-reported adoption of EHRs. After controlling for practice and county level characteristics, physicians per capita was positively and significantly associated with EHR adoption (OR = 1.15, p = .015) and percent Medicare managed care penetration was negatively and significantly associated with EHR adoption (OR = .981, p <.001). Physicians practicing in an HMO setting were more likely than single specialty practices to adopt EHRs (OR = 2.46, p = .039). Additionally, physicians in groups of six or more were more likely to adopt EHRs than those practicing solo (OR = 1.73, p = .001). Moreover, physicians adopting other type of information technologies were more likely to adopt EHRs than those not adopting or adopting few information technologies (OR = 1.46, p < .001). Lastly, years since medical school and county poverty rate were not related to EHR adoption by physicians. Conclusions: Competitive, regulatory, and practice characteristics are associated with the decision to adopt EHRs by physicians. Adoption of EHRs may be even more advantageous in certain environments. Implications for Policy, Delivery, or Practice: Our findings suggest that environmental factors influence the adoption of EHR systems by physicians. These findings may assist both policy makers and physicians who are interested in promoting EHR adoption and/or better understanding the strategic environment and it’s affect on the adoption of this technology. Primary Funding Source: CMS ●The Foundation for Medicaid’s Role in the Adoption of Health Information Technology Shaun Alfreds, M.B.A., Michael Tutty, MHA, Jay Himmelstein, M.D., M.P.H., Mark Frisse, M.D., Patricia MacTaggart, M.B.A., Bruce Greenstein Presented By: Shaun Alfreds, M.B.A., Project Director, Center for Health Policy and Research, University of Massachusetts Medical School, 222 Maple Avenue, Shrewsbury, MA 01545; Tel: (508) 856-8634; Email: shaun.alfreds@umassmed.edu Research Objective: This paper defines the challenges and opportunities that state and federal agencies might play in leveraging HIT developments to improve the quality of healthcare for Medicaid beneficiaries and the efficiency of Medicaid operations. In addition, it provides actionable policy recommendations from national experts for federal and state agencies. Study Design: An advisory committee of thought leaders from federal and state governments, policy organizations, HIT vendors, and healthcare provider groups was convened to identify key issues and research necessary to inform an expert panel meeting held in March of 2006. Recommendations of the advisory committee and the expert panel are summarized and presented in this paper. Population Studied: Medicaid agencies, Medicaid beneficiaries, and providers serving Medicaid covered individuals throughout the country. Principal Findings: The value of HIT as a tool to provide higher quality of care and more efficient operations has been recognized. Interoperable health information systems are evolving into regional health information networks, which may further mature into a National Health Information Infrastructure. Medicaid agencies, the largest payer of healthcare for low income and vulnerable populations, must participate in these information sharing networks not only to realize the benefits for their beneficiaries and internal operations, but also in order to address the needs of Medicare beneficiaries, the commercially insured, and the uninsured. To date however, Medicaid agencies have had limited involvement with advanced HIT due to the complexity of the populations served, the providers serving them, and perceived legal, fiscal and regulatory barriers. Overall recommendations for Medicaid involvement in the HIT arena are provided with specific emphasis on five areas. These are: 1) Using HIT to Improve Healthcare Quality for Medicaid Beneficiaries, 2) Leveraging HIT to Improve Medicaid Agency Efficiencies, 3) Financial and Non-Financial Incentives to Illicit Medicaid Provider Adoption of HIT, 4) Legal and Regulatory Issues in a Medicaid Context, and 5) Case Studies of Medicaid and State Government Involvement in HIT Initiatives. Implications for Policy, Delivery, or Practice: This paper captures key recommendations made by thought leaders and policy experts on necessary policy decisions defining the roles, needs, and priorities of Medicaid programs in facilitating the use of HIT to achieve safe, high quality, accessible, and efficient healthcare for Medicaid beneficiaries as well as the citizens of the states they serve. Primary Funding Source: AHRQ ●Electronic Health Records in Four Community Physician Practices: Impact on Quality and Cost of Care Dawn Bazarko, RN, M.P.H., W. Pete Welch, Ph.D, Kimberly Ritten, Yo Burgess, , Robert Harmon, M.D., M.P.H., Lewis G. Sandy, M.D., M.B.A. Presented By: Dawn Bazarko, RN, M.P.H., Senior Vice President, Care Improvement Strategies, United Clinical Advancement, UnitedHealthcare, 5901 Lincoln Drive, MN012S117, Edina, MN 55436; Tel: 952-992-4283; Fax: 952-992-5028; Email: dawn_m_bazarko@uhc.com Research Objective: To assess the impact of the electronic health record (EHR) on cost (to the payer) and quality of care in physician practices in the community. Most other such studies to date have been limited to a single practice with inhouse research capabilities. Study Design: Retrospective before-after-study-control. Studying multiple community practices requires two types of information: a large claims database and local knowledge of which practices had EHRs. With the help of the medical directors of a large managed care organization (MCO), we identified four community physician practices (in three metropolitan areas) that had implemented EHR in the last five years. These practices were interviewed both before and after our data analysis. About fifty practices without EHR were identified in the same county as one of the study practices. As we defined them, periods before and after implementation varied across study practices but were the same for a study practice and its controls. We used two commercially-available software products to analyze episodes involving one of four chronic conditions. Episode Treatment GroupsTM (ETGs) were used to measure cost of care and adjust it for casemix. Allowed charges on claims were “re-priced” to remove any impact of price differences. EBM ConnectTM was used to analyze the rate of adherence to clinical guidelines as a measure of quality. Several guidelines were used for each of the four chronic care conditions. Tests for significance were applied to the difference in the change (in cost or quality) between study and control practices. Population Studied: Patients enrolled in a large MCO who had at least one of four chronic conditions: hypertension, hyperlipidemia, diabetes, and selected heart conditions. About 90,000 episodes of care were analyzed. Principal Findings: The implementation of the EHR did not significantly affect the cost per episode (minimum detectable difference of 3.5 percent of the mean). It had a significant, modest positive impact on the quality measure of guideline adherence for hypertension (2 percentage points) and hyperlipidemia (5 percentage point), but no significant impact for diabetes and coronary artery disease. Interviews with the study practices revealed that the timing and comprehensiveness of EHR implementation varied across practices, creating an intervention variable that was heterogeneous. Conclusions: Guideline adherence is increasing across practices without EHRs and slightly faster in practices with EHRs. The increase in non-EHR practices is consistent with reports that many of the functionalities of EHRs can be obtained in other ways. Measuring the impact of EHRs on cost per episode is challenging, because of the difficulty of capturing the later costs of a chronic condition. Implications for Policy, Delivery, or Practice: Our findings tentatively suggest that the advantages of current EHRs may be overstated. Primary Funding Source: AHRQ ●HIT and QI: A State-Wide Educational Quality Improvement RCT among Texas Rural Hospitals Giovanni Filardo, M.P.H., Ph.D., David Nicewander, MS, Percy Galimbretti, M.D., Ph.D., Susan McBride, Ph.D., RN, Josie Williams, M.D., MMM, David Ballard, M.D., MSPH, Ph.D. Presented By: Giovanni Filardo, M.P.H., Ph.D., epidemiologist, Institute for Health Care Research and Improvement, Baylor Health Care System, 8080 North Central Expressway, Suite 500, Dallas, TX 75206; Tel: 214-265-3633; Fax: 214-265-3640; Email: giovanfi@baylorhealth.edu Research Objective: Hospitals are under increasing pressure to measure and report their quality of care and to improve quality. Particularly with the growing popularity of “pay for performance” programs, hospitals’ continued viability is dependant on their ability to measure quality indicators and to improve performance relative to these indicators. Although a number of formal quality improvement education programs are offered nationwide, we are conducting the first to target rural hospitals. Study Design: A hospital-randomized controlled trial, following implementation of a web-based hospital quality of care benchmarking and case review tool in rural and small community hospitals in Texas. Sixty hospitals or more that demonstrate the successful implementation of the web-based tool will be randomly assigned to receive a rapid cycle educational program on methods and techniques for implementing and evaluating quality improvement initiatives or not (control group). The seven day (three 2 day methods sessions interspersed over 2 months with project coaching between sessions and during follow-up, coupled with a daylong project reporting conclave) educational program will enroll for each hospital a physician, nurse and administrator. The curriculum is designed to facilitate the development of skills and competencies needed to implement quality improvement efforts. The incremental benefit of this educational program – over and above the effect of the benchmarking and case review tool – will be evaluated by comparing quality of care measures (Centers for Medicare and Medicaid Services core inpatient indicators for eligible patients admitted with congestive heart failure [CHF] or community acquired pneumonia [CAP]), patient safety (selected Agency for Healthcare Research and Quality [AHRQ] indicators), and all-cause inpatient mortality between the treatment and control groups for 2 years following the educational intervention. Population Studied: Rural and small community hospitals in Texas, defined as hospitals located in a county in Texas with <150,000 inhabitants according to data from the U.S. Bureau of the Census 2000, or listed in the American Hospital Association 2003 dataset and designated as a Critical Access Hospital. Principal Findings: The minimum detectable differences between study groups, estimated for the selected patient safety and quality indicators using methods appropriate for cluster-randomized studies with a binary outcome have been calculated on a sample of 66 small community rural or small community hospitals in Texas. Given the estimated baseline rates for composite quality scores for CHF (62%) and CAP (56%), the size of the hospital cohort enables detection of 13% and 10% differences (a=.05, power=.8) respectively, between study groups. Conclusions: N/A Implications for Policy, Delivery, or Practice: This project begins the work of assessing the current information technology capabilities of rural and small community hospitals in the United States, and determining how these capabilities can best be enhanced to improve the quality in such hospitals. It also addresses the issue of whether extant quality of care measures are applicable to rural and small community hospitals, or whether their low patient volumes and other unique characteristics require the development of tailored measures. This project is supported by the AHRQ (RFA-HS-04-011). Primary Funding Source: AHRQ ●Evaluation of Ambulatory EHR Deployment: Quality and Fiscal Effects Neil Fleming, Ph.D., CQE, Donald Kennerly, M.D., Ph.D., Edmund Becker, Ph.D., Robert Mayberry, Ph.D., M.P.H., Steven Culler, Ph.D., David Ballard, M.D., MSPH, Ph.D. Presented By: Neil Fleming, Ph.D., CQE, Vice President, Health Care Research, Institute for Health Care Research and Improvement, Baylor Health Care System, 8080 North Central Expressway, Suite 500, Dallas, TX 75206; Tel: 214-265-3601; Fax: 214-265-3640; Email: neilfl@baylorhealth.edu Research Objective: The objective is to quantify the effects of an ambulatory electronic healthcare record (AEHR) on the 6 Institute of Medicine domains of quality and financial performance in primary care practices. Study Design: This is an observational cohort study comparing quality and financial measures pre-and post implementation of an AEHR across HealthTexas Provider Network (HTPN), the ambulatory care component of Baylor Health Care System (BHCS) in Dallas, TX. Comparisons will be based on patient-, physician-, and practice-level data, for measures of Safe, Timely, Effective, Efficient, Equitable, and Patient-centered care. Measures of practice financial performance (monthly practice margin [overall patient revenues minus practice expenses] per physician) and practice financial efficiency (practice expense per RVU and staffing expenses) pre- and post-AEHR implementation will be compared. AEHR implementation will be staggered across practices over 3 years, allowing both cross-sectional (between AEHR practices and non-AEHR at fixed time points) and longitudinal (between measures collected within the same practices before and after AEHR implementation) comparisons. Net financial impact of the AEHR, including actual costs of implementation and maintenance, will be determined. Patient safety will be measured according to the rate of adverse events (including drug-related events) detected in the elderly Medicare population using the Institute for Healthcare Improvement’s ambulatory care trigger tool preand post-AEHR implementation. Effective care will be operationalized using a composite measure for delivery of 11 adult clinical preventive services (CPS). Patient satisfaction survey data incorporate patient-centeredness (overall satisfaction and willingness to recommend) and timeliness indicators (time waited for appointment and office waiting time). Equity will be evaluated by comparing performance on safety, effectiveness, timeliness, and patient-centeredness measures across age, gender, race/ethnicity, and socioeconomic status. Efficiency will be measured using physician RVUs, based on CPT-4 codes from patient visits, for services related to laboratory/pathology, and radiology. Population Studied: HTPN includes 66 practices throughout Dallas-Fort Worth. Since processes of care and associated costs differ by specialty, this analysis will focus on primary care physicians and practices (32 practices and 183 physicians). Principal Findings: Baseline data show 14% of patients wait >30 days between making and keeping an appointment, and 15% have >30 minute in-office wait times; the composite score for the 11 CPS indicators (proportion of opportunities per patient achieved) is 0.87; 92% of patients report excellent or very good satisfaction with their physicians, and 97% report they would recommend their physician to other people. Baseline equity data show patients with higher SES are more likely to receive CPS that those with lower SES (0.87 mean composite score vs 0.83) and more likely to report “excellent” or “very good” satisfaction (93% vs 90%). Minimum detectable effects (? = 0.05, power =0.80) are 4.5% for adverse drug events; 0.11 standard deviations for time waited for an appointment and in-office waiting time; 2.2% for delivery of CPS; and 0.05 standard deviations for overall satisfaction with physician and willingness to recommend. Conclusions: n/a Implications for Policy, Delivery, or Practice: This study will inform decisions to invest in AEHR from both quality and fiscal perspectives. Primary Funding Source: No Funding ●Multi-Hospital System Membership and Diffusion of Health Information Technology Michael Furukawa, Ph.D. Presented By: Michael Furukawa, Ph.D., Assistant Professor, Health Management and Policy, Arizona State University, P.O. Box 874506, Tempe, AZ 85287-4506; Tel: (480) 965-2363; Fax: (480) 965-6654; Email: Michael.Furukawa@asu.edu Research Objective: Health information technology (HIT) is widely-regarded as a key strategic resource to increase efficiency and improve quality of care. Despite growing interest in HIT by managers and policymakers, few studies have examined the determinants of HIT adoption. In particular, little is known about the role of multi-hospital system membership in the diffusion of HIT across hospitals and affiliated organizations. This study seeks to address this gap in the literature by examining the effect of system membership on the adoption of HIT by hospitals, sub-acute, and ambulatory facilities. Study Design: The primary data source is the 2004 HIMSS Analytics (HA) Database, which contains detailed information on the adoption of HIT within integrated health delivery systems. The HA database is linked to the 2002 AHA Annual Survey Database, which provides information on hospital characteristics and system membership, and to the 2004 Area Resource File, which provides information on area characteristics. I test the overall effect of system membership and of system type (centralized, centralized physician/insurance, moderately centralized, decentralized, and independent) on HIT adoption, controlling for facility and area characteristics. Dependent variables include specific HIT applications (e.g., computerized patient records, computerized physician order entry), and the type of HIT defined by 4 categories: financials and business office; medical records and administrative; management and human resources; and clinical and ancillary departments. I estimate probit regressions of the probability of HIT adoption and negative binomial regressions of the count of HIT within each category. I address the potential endogeneity of system membership using propensity score weighting in a multinomial selection model. Population Studied: U.S. short-term, non-federal hospitals and affiliated organizations in 1,453 integrated health care delivery systems. The sample includes 3,989 hospitals, 3,007 sub-acute, and 18,008 ambulatory care facilities. Principal Findings: I find that multi-hospital system membership is a strong determinant of HIT adoption. Overall, system membership is associated with greater adoption across all categories of HIT, but most significantly for management and financial applications. Facilities in centralized physician/insurance health systems are much more likely to adopt clinical IT relative to independent hospital systems. Conclusions: Members of multi-hospital systems have greater levels of HIT adoption than facilities in independent hospital systems. However, system membership appears to have a differential effect on the type of HIT, with only marginally significant effects on the adoption of clinical IT. HIT adoption also varies by system type, suggesting that centralization, integration, and insurance offerings are key system characteristics that are associated with the adoption of clinical IT. Implications for Policy, Delivery, or Practice: The consolidation and integration of hospitals, physicians, and affiliated organizations into multi-hospital systems has important implications for the adoption and diffusion of HIT. System membership may confer network effects because HIT adoption confers benefits to the adopter as well as affiliated facilities. The findings of this study suggest that governance and organizational structure play a key role in the diffusion of HIT across facilities within a system. Primary Funding Source: No Funding ●Evaluation of an Integrated, Web-based Computer System for Patients Living with HIV/AIDS Daniel Gentry, Ph.D., M.H.A., Stacie Metz, M.P.H., M.S.W., MA, Kathye Gorosh, M.B.A. Presented By: Daniel Gentry, Ph.D., MHA, Associate Professor, Dept. of Health Management and Policy, Saint Louis University School of Public Health, 3545 Lafayette, Salus Center 300, St. Louis, MO 63104; Tel: 314-977-8152; Fax: 314977-1441; Email: dgentry@slu.edu Research Objective: To conduct a process and outcome evaluation of a patient-centered, web-based computer system for people living with HIV/AIDS; 7 process evaluation questions and 4 outcome evaluation questions were specified. Study Design: This was a prospective study that followed and compared four different cohorts who utilized the Community and Minority Education and Training Initiative for HIV/AIDS (COMET). The vast majority of data were systematically and routinely collected over a period of 18 months (November 2003-May 2005); the data collection was integrated into the web-based system. Additional qualitative data were collected through client focus groups. Population Studied: Four populations were studied: 79 English-speaking CORE Center patients who were returning users; 137 new English speaking patients; 108 Spanishspeaking users; and 86 users from a separate communitybased organization. All patients were living with HIV/AIDS. Total n=410. Principal Findings: Five types of principal findings are reported: client demographics, client COMET usage, provider COMET usage, COMET client focus group feedback, and service referrals for COMET clients. Selected findings include: 64% of COMET users on HIV/AIDS medications reported being non-adherent at some point; the most frequently accessed COMET activity was the "Knowledge" function (provides information on the immune system, medication, wellness, etc.). In addition to findings, limitations are also discussed. Conclusions: Stengths of the COMET program included: the opportunity to learn and use computers, and increased access to the patient's own medical information. Weaknesses of COMET included: delays in posting of lab results, limited computer and COMET training for clients, and very limited participation by physicians and other providers. Implications for Policy, Delivery, or Practice: Patient empowerment, self management and medication adherence are areas of growing importance in terms of improving patient access and quality and for cost-effectiveness. The COMET program has received additional funding not only for continuation but also for replication by other service providers. The results of this evaluation will be used to improve the program, designate additional outcome measures, and develop tools for program monitoring. Both the full evaluation report(34 pages) and an executive summary (3 pages) will be made available to session participants. Primary Funding Source: U.S Office of Minority Health ●Record Linkage Research and Informed Consent: Who Consents? Nicole Huang, Ph.D., Shu-Fang Shih, MS, Hsing-Yi Chang, Ph.D., Yiing-Jenq Chou, M.D., Ph.D. Presented By: Nicole Huang, Ph.D., Assistant Professor, Institute of Public Health, National Yang Ming University, 155, Li-Nong Street, Section 2, Taipei, 112; Tel: 886-2-28201458; Fax: 886-2-28261002; Email: syhuang@ym.edu.tw Research Objective: Recently, linking survey data on individuals with administrative records has become a popular data source in medical and public health research. Government requirement of written informed consent before accessing health records may introduce a selection bias if consenting patients differ from those who do not give consent. This study aimed to compare interviewees of a national health survey who consent and refuse for access to their national health insurance records in Taiwan. Study Design: A cross-sectional study Population Studied: A national representative sample (n=14,611 adults) of the general adult population aged 20 years or older were interviewed by the staff of the Taiwan National Health Interview Survey (NHIS)and asked for a written permission for their National Health Insurance records to be reviewed. Survey responses were used to compare interviewees by consent status. Principal Findings: Of the 14 611 NHIS participants, 12 911 (88%) gave consent, and 1 700 (12%) denied consent. The multivariable analysis indicates that the elderly aged 65 or above, the illiterates, people with monthly household income less than US$ 950, and the rural area residents were more likely to decline consent than their counterparts. On the other hand, health status did not seem to have a strong influence on consent -giving or -withholding. Conclusions: Consenters differed from non-consenters significantly in important aspects such as age, educational background, socioeconomic status and residential location. The non-uniform distribution of these variables may be significantly related to utilization and health outcomes. Implications for Policy, Delivery, or Practice: Therefore, record linkage research restricted to consenters only, runs the risk of miss-characterizing the utilization and health outcomes of the general population and lead to selection bias. Higher refusal rates observed among sub-populations of the elderly and the illiterate suggest that more effective communication strategies may help to reduce refusal rates among these groups. This may minimize possible selection bias in large scale record linkage studies in Taiwan. Primary Funding Source: Taiwan's National Science Council ●Facilitating the Implementation of Health Information Technology in the State of Louisiana Through Medicaid Michelle Lim Warner, M.P.H.c identifying ways to achieve widespread implementation at the national level. Primary Funding Source: No Funding Presented By: Michelle Lim Warner, M.P.H.c, Master in Public Health Candidate, School of Public Health and Health Services, George Washington University, 4615 North Park Avenue #909, Chevy Chase, MD 20815; Tel: (202) 841-9639; Email: milim@gwu.edu Research Objective: Louisiana has a unique opportunity to re-design its health care system by focusing its re-building efforts on creating an interoperable health information technology (HIT) infrastructure. Medicaid, as the single largest purchaser of health care in a state, is in a unique position to further this effort. This paper will explore and identify optimal strategies for Medicaid to employ in order to facilitate the implementation of HIT in the State of Louisiana. Study Design: Involves an extensive review of the literature related to HIT and discussions with thought leaders from Louisiana state government, the federal government and experts in the field of health care financing, health policy, Medicaid administration, and electronic health information systems. Population Studied: Louisiana Medicaid Agency, Louisiana Medicaid recipients and providers serving Medicaid covered individuals Principal Findings: HIT has the potential to improve health care quality and the efficiency of a health care delivery system. Yet widespread adoption and use of such technology has been occurring at a far slower pace than is desirable. Many barriers, such as cost and legal and regulatory impediments, have limited the rapid adoption of HIT. However, the ramification of Hurricane Katrina to the Louisiana health care system illustrated the critical need for HIT to preserve and protect patient health records and maintain a working health care delivery system. As any state’s largest single purchaser of health care, Medicaid is in a position to facilitate the implementation of HIT. Medicaid has many levers from which to further this effort – one of which is purchasing for quality. Louisiana is currently ranked 49th in the nation in terms of health care quality, and can benefit from having an HIT infrastructure supporting its health care system. Specifically, Medicaid can employ payment strategies such as pay-for-performance to incentivize its providers to utilize HIT. Medicaid can also employ regulatory and other providerrelated approaches, as well as partnering with other funders to facilitate and achieve HIT implementation. Conclusions: Medicaid has a significant impact on furthering the HIT movement and can be leveraged to facilitate the implementation of HIT in states. Implications for Policy, Delivery, or Practice: This paper provides a new lens from which to view the evolution of HIT implementation in two ways: (1) by focusing on HIT implementation at the state-level and (2) by identifying strategies for Medicaid to leverage in facilitating the effort and furthering the HIT movement. Much of the visible activity around HIT adoption and implementation are occurring at the national level. Yet the problems and issues related to health care delivery are addressed at the local level. Furthermore, the potential of HIT to improve health care is first realized at the local level. As such, taking a look at how the largest purchaser for a state can move the HIT agenda forward is critical to ●Improving Cancer Symptom Management through Information Technology Tami Mark, Ph.D. M.B.A., Barry Fortner, Ph.D., Gina Johnson, APRN, MSN, Xue Song, Ph.D. Presented By: Tami Mark, Ph.D. M.B.A., Associate Director, Outcomes Research, Thomson/Medstat, 4300 Connecticut Avenue, NW Suite 330, Washington, DC 20008; Tel: (202) 719 -7832; Fax: (202) 719-7801; Email: Tami.Mark@thomson.com Research Objective: To conduct a randomized trial to assess the ability of a computerized symptom screening and educational program to improve cancer symptom management following chemotherapy (specifically, depression, pain, and fatigue). Study Design: The PACE system was installed in two community cancer clinics naive to the intervention. Patients were randomized to receive the PACE system or standard care. Assessments were distributed monthly or bi-monthy to determine whether physicians were discussing symptoms, whether symptoms were being treated, the level of symptoms experienced, the level of satisfaction with information provided about cancer treatment, and attitudes towards pain management. Subsequent surveys were conducted with providers to determine their level of satisfaction with the PACE System™ (survey is presently in the field).Between June, 2004, and December, 2004, 350 patients were screened at two community oncology clinics. Of those patients, 78 met the inclusion criteria and were invited to participate in the study. Forty-eight (48) patients agreed to participate. The most frequent reason for exclusion was that patients were not newly initiating chemotherapy. Population Studied: Patients treated at two community oncology clinics who were starting chemotherapy, were 18 and older, and literate. Principal Findings: The monthly patient reported assessment showed a trend toward more provider discussion of pain symptoms among patients with pain (p 0.12) and less discussion of fatigue among patients without fatigue (p = 0.004). There was no difference in discussion of depression although patients with depression were less likely to report that the clinic could have done more to assess their symptoms (p = 0.09). Patients in the intervention group had lower scores in the memorial symptom assessment scale in follow-up months than patients in the intervention group (p = 0.03) indicating lower levels of symptoms. There were no differences between the control and intervention groups in the scores on the HADS-anxiety scale, the HADS-depression scale, and the Spitzer QoL life instrument. Patients in the intervention group were more satisfied with the information that they received on chemotherapy treatment and had less fatalistic attitudes towards pain management. Conclusions: There is evidence that The PACE System™ is having a positive effect on cancer symptom identification, targeting, and management, as well as patient information about and attitudes towards cancer symptom management. Implications for Policy, Delivery, or Practice: The NIH recently concluded that cancer symptom management has lagged behind cancer treatment and that frequently symptoms go undetected and untreated. Information technology can be a viable method for improving symptom assessment and management as well as educating patients about cancer treatment. Primary Funding Source: RWJF ●The Effect of IT Adoption on JCAHO Performance Scores in Florida Hospitals Nir Menachemi, Ph.D., M.P.H., Anol Bhattacherjee, Ph.D., Neset Hikmet, Ph.D., Onur Kayhan, B.S., Robert G. Brooks, M.D. ●Barriers to Ambulatory EHR: Who are “Imminent Adopters” and How do They Differ From Other Physicians? Nir Menachemi, Ph.D., M.P.H. Presented By: Nir Menachemi, Ph.D., M.P.H., Assistant Professor, Division of Health Affairs, Florida State University College of Medicine, 1115 West Call Street, Tallahassee, FL 32306; Tel: (850) 644-2362; Fax: (850) 645-2859; Email: nir.menachemi@med.fsu.edu Research Objective: This empirical study examines the relationship between healthcare information technology (IT) investments and operational performance in Florida hospitals. Unlike most prior studies in this area that have measured IT investment at the macro or facility level, we measure IT investment at the micro or technology level, which allowed us to explore technological differences among various types of IT investments. Study Design: This unique project combines both primary and secondary data. The primary data comes from a hospital IT survey conducted in Florida between May and October 2003. Secondary data, used to construct the dependent variable (operational performance), was obtained from the Joint Commission of Accreditation of Healthcare Organizations (JCAHO). JCAHO is an independent national agency responsible for auditing and certifying performance and quality ratings of US hospitals. Using the IT data collected, and previously devised methods, we operationalized IT capabilities in three main hospital functional areas; administrative, clinical, and strategic. Regression analyses were used to examine the relationship between these IT capabilities and overall JCAHO performance score. Each model specified bed-size and hospital case-mix as covariates. Population Studied: A total of 96 Florida hospitals were included. They were representative of hospitals in the state in terms of size, system-affiliation, and tax-status. Principal Findings: Our measure of overall IT utilization (p=0.04) was positively associated with operational performance as measured by the JCAHO performance scores (R2=0.115). However, clinical IT adoption (p=0.006), and not administrative (p=.232) or strategic IT (p=0.13) adoption, was responsible for this strong and significant relationship. Conclusions: Hospitals in our dataset that have adopted more clinical IT applications performed better on JCAHO performance evaluations. This occurred most probably because clinical technologies such as computerized physician order entry systems, chart tracking systems, and laboratory information systems, are directly targeted at improving and transforming the management and delivery of healthcare. Implications for Policy, Delivery, or Practice: This study provides evidence that the adoption of IT, particularly clinical applications, is associated with improved operational performance. Additionally, our findings provide preliminary evidence that not all categories of HIT have equivalent effects on facility performance. Therefore, unlike previous studies, different classes of health IT should be examined separately to achieve a better estimate of the performance effects of health IT. Primary Funding Source: No Funding Presented By: Nir Menachemi, Ph.D., M.P.H., Assistant Professor, Family Medicine and Rural Health, Florida State University College of Medicine, 1115 West Call Street, Tallahassee, FL 32306; Tel: (850) 644-2362; Fax: (850) 6452859; Email: nir.menachemi@med.fsu.edu Research Objective: Despite existing knowledge regarding EHR barriers in the ambulatory setting, little is known, specifically, about physicians who are likely to adopt EHR imminently. The current study identifies these imminent adopters and compares their barriers to other physicians. Study Design: Mailed survey of Florida physicians (n=14,921) about barriers to EHR and adoption intentions. Physicians were categorized as, current EHR users, “imminent adopters” (those planning to adopt EHR within 1 year), “interested adopters” (those planning to adopt EHR but not within 1 year), and those not considering EHR. Chi square analysis and binary logistic regression models were used to identify trends among imminent adopters and to compare barriers among physicians in each of the adoption categories above. Population Studied: Florida physicians practicing in the ambulatroy setting. Principal Findings: : Imminent adopters were significantly less likely to be in solo practice (19.6 vs. 40.0%, P<.001) and more likely to be in an urban area (P=.044) or in a multispecialty practice (P=.023). Imminent adopters were also more likely to be practicing family medicine (.014), or obstetrics/gynecology (P=.038). When compared to their physician colleagues, imminent adopters perceived EHR barriers very differently. For example, imminent adopters were significantly less likely to consider upfront cost of hardware/software [OR=0.35(0.30, 0.45)] or that an inadequate return on investment [OR=0.25(0.19, 0.34)] was as a major barrier to EHR. Moreover, imminent adopters differed from their colleagues with respect to numerous other productivity and technical barriers. Conclusions: Policy and decision makers interested in promoting the adoption of EHR among physicians should focus on the needs and barriers of those most likely to adopt EHR. Given that imminent adopters differ considerably from their peers, current EHR incentive programs that focus on financial barriers only may prove suboptimal in achieving immediate widespread EHR adoption. Primary Funding Source: CMS ●Validity of ICD10 administrative data in recording comorbidity information Hude Quan, M.D. Ph.D., Bing Li, MA, L Duncan Saunders, Ph.D., William A Ghali, M.D. M.P.H. Presented By: Hude Quan, M.D. Ph.D., Assistant Professor, Community Health Sciences, University of Calgary, 3330 Hospital Dr. NW, Calgary, T2N 4N1; Tel: (403) 283 5307; Email: hquan@ucalgary.ca Research Objective: To evaluate the validity of the International Classification of Disease, 10th Version (ICD-10) administrative hospital discharge data and to determine whether there were improvements of the validity compared with ICD-9-CM data. Study Design: We reviewed randomly selected charts at all Alberta teaching hospitals to determine presence or absence of recorded comorbidities and re-coded the same cases using ICD-9-CM. We assessed the agreement between ICD-10 administrative data and chart review data for recording comorbidities constitute the Charlson index and Elixhauser comorbidity measure and then the agreement between recoded ICD-9-CM coded data and chart review data for recording those comorbidities. Finally, we compared the accuracy of ICD-10 data relative to chart data with the accuracy of ICD-9-CM data relative to chart data for those comorbidities. Population Studied: Inpatients discharged during January 1 and June 30, 2003 from four teaching hospitals in two large Canadian cities Principal Findings: We reviewed 4008 inpatient records for assessment of validity to 33 comorbidities. Compared to the chart data, the ICD-9-CM data under reported 30 comorbidities and ICD-10 data under-reported for 31 comorbidities. Six comorbidities had higher prevalence in ICD10 than ICD-9-CM data. Sensitivity a measure of the accuracy of recording presence of comorbidities in administrative data when these were present in chart data ranged 9.3 - 95.7% for ICD-9-CM and 12.7- 93.7% for ICD-10 data. Positive predictive value that determines the extent to which a condition present in the administrative data was also present in the charts ranged 10.5-100% for ICD-9-CM and 11.0 – 100% for ICD-10 data. Specificity to determine the accuracy of reporting absence of these comorbidities in the administrative data when these diseases were absent in the charts and negative predictive value to determine the extent to which a condition absent in the administrative data was truly absent according to the chart data were higher than 85% for both ICD-9-CM and ICD-10 databases. Conclusions: ICD-10 administrative data were coded reasonably well on comorbidity variables although some comorbidities defined based on ICD-10 administrative data tended to be underestimated compared with those defined using patient chart data. Validity of ICD-10 data is similar with that of ICD-9-CM data in recording 33 comorbidities in our sample. Implications for Policy, Delivery, or Practice: Implementation of ICD-10 may not significantly impact on risk adjustment in quality of care assessment relative to ICD-9CM. Primary Funding Source: Canadian Institutes of Health Research ●Security Issues in Outsourcing Ebrahim Randeree, M.B.A. Presented By: Ebrahim Randeree, M.B.A., Ph.D. Candidate, Management Science & Systems, University at Buffalo, 248 Jacobs Management Center, Buffalo, NY 14260; Tel: (716) 2077251; Fax: (716) 645-6117; Email: er4@buffalo.edu Research Objective: The issue of trust and risk in hospital relationships was extended beyond traditional focus as outsourcing models were introduced. As outsourcing service providers evolve, the evolution of IT networks and the adoption of EMRs will increasing information assurance threats. This paper develops a conceptual model of outsourcing adoption in the presence of increased threats; it investigates trust and risk variables that affect the adoption and management of the relationships between hospitals and vendors - specifically focussing on the management of the outsourcing contract and the issue of security. Study Design: Conceptual model with theory review and proposed empirical study Population Studied: Hospitals Principal Findings: None - Theoretical exploration at this time! Conclusions: Pending empirical survey Implications for Policy, Delivery, or Practice: The effect of outsourcing relationships on the way we perceive security and organizational boundaries is changing. The potential risks to both the vendor and the client emerging from adoption can be quite different. The role of IT security and information assurance within hospitals may influence the direction of contracting and vendor management. THis paper explores information assurance concerns under HIPAA and combines the analysis with outsourcing adoption. Primary Funding Source: No Funding ●EMR options in Private & Small Group Practices Ebrahim Randeree, M.B.A. Presented By: Ebrahim Randeree, M.B.A., Ph.D. Candidate, Management Science & Systems, University at Buffalo, 248 Jacobs Management Center, Buffalo, NY 14260; Tel: (716) 2077251; Fax: (716) 645-6117; Email: er4@buffalo.edu Research Objective: The adoption of EMRs has been slow with renewed impetus for small groups and solo physicians to follow larger hospital systems. The options facing physicians may be improving but many are still reluctant to sign-on. This study explored physician concerns and investigates the options available. Study Design: Qualitative w/interviews Population Studied: Private Physicians, Small Group Practices Principal Findings: Qualitative interviews provide depth that is not evident in survey analysis. Physicians express cost concerns as well as lack of IT skills within practices. Conclusions: The challenges faced by this group may create new concerns for IT solutions providers. Prelimenary results provide advocates and technology vendors with the right mix of incentives to increase adoption. Implications for Policy, Delivery, or Practice: This paper presents a research model and investigates the options (partner with a hospital system, use an ASP model, do-ityourself) available to physicians engaged in EMR adoption. Primary Funding Source: No Funding ●User Perceptions of Barcode Medication Administration Systems Julie Sakowski, Ph.D., Thomas Leonard, RN MPA, Jeffrey Newman, M.D. M.P.H. Presented By: Julie Sakowski, Ph.D., Sr. Health Services Researcher/ Health Economist, Sutter Health Institute for Research & Education, 345 California Street Suite 2000, San Francisco, CA 94104; Tel: (415) 296-1808; Fax: (415) 296-1844; Email: sakowsj@sutterhealth.org Research Objective: Electronic barcode medication administration systems (BCMA) have been developed to reduce medication errors and increase patient safety. These systems utilize barcodes placed on medications and patient identification bands to match medications being administered with orders entered in the pharmacy system and alert the clinician if any discrepancies are detected. The ability of these systems to prevent medication administration errors has been documented. However, the impact of these systems on overall patient safety, how the adoption of BCMA changes nursing and other health care provider work flows, and the impact on patient and provider perceptions of quality of care remains unclear. This study fills that gap by evaluating clinician perceptions of BCMA, how it affects the delivery of health care, and its impact on patient safety. Study Design: An anonymous, written survey was distributed to nurses and respiratory care practitioners using BCMA for inpatient medication administration. The survey employed a combination of closed end questions where respondents either indicated their agreement with a statement or rated the amount of change resulting from the adoption of BCMA (i.e. increased, decreased, or had no effect) and open ended questions to provide more detail and context. The survey collected information on respondent characteristics, effectiveness of the BCMA training, BCMA ease of use, quality of the warnings, medication administration process changes, and impact on patient safety and satisfaction. Frequencies and descriptive statistics were calculated for the survey results and variation in responses due to user characteristics was analyzed using logistic regressions. Focus groups were conducted to gain a more detailed understanding of the survey results and explore themes surfaced by the open ended responses. A human subjects review exemption was granted before the survey was distributed. Population Studied: Clinicians using a barcode medication administration system (eMAP) to deliver over 8,500,000 inpatient doses at 9 hospitals affiliated with Sutter Health, a network of not-for-profit community hospitals located in Northern California. Principal Findings: Preliminary analysis of responses from a sample of 57 completed surveys from 4 hospitals indicates users believe the system prevents errors, is easy to use, and increases patient safety. 88% of respondents indicated that they believed the BCMA system increased patient safety. 85% stated the system increased patient perceptions of safety and 61% reported it increased patient satisfaction. 77% rated the system easy to use, but 46% indicated the system provided too many warnings. 45% of the respondents believe that the use of BCMA greatly increased the amount of time it takes to administer medications and 55% have had to change the amount of time dedicated to other patient care tasks. 60% of the respondents indicated using BCMA increased their job satisfaction. Conclusions: Users believe that BCMA systems are easy to use and increase patient safety. Implications for Policy, Delivery, or Practice: BCMA systems can be an effective tool for reducing medication errors and increasing patient safety, but a concerted effort needs to be made to ensure that the warnings the system generates are appropriate, understandable, clinically relevant, and actionable. Primary Funding Source: No Funding ●Does Pay for Performance Affect the Adoption of Quality Enhancing Practices by Child Health Providers (CHPs)? Lisa Simpson, MB, BCh, M.P.H., FAAP Presented By: Lisa Simpson, MB, BCh, M.P.H., FAAP; Email: lsimpso1@hsc.usf.edu Research Objective: Pay for performance programs (PFP) link physician compensation to quality and use of information technology applications. These programs are increasingly being promoted as an effective strategy to improve healthcare quality by child health providers (CHPs). Nationally it is unknown the extent to which these programs affect physicians’ behavior in relation to quality. The purpose of this study was to assess the extent to which CHPs behavior is linked to PFP and whether this is associated with increased adoption and use of health information technology applications including electronic health records, email and personal digital assistants. Study Design: The current study is limited to Child Health Providers (CHPs) who responded to a larger survey of all ambulatory physicians in the state of Florida. The independent variables include physician and practice characteristics (i.e. age, race/ethnicity, specialty, practice size/type and Medicaid volume) and a number of compensation variables (i.e. patient surveys, measures of clinical care, use of clinical information technology, and quality bonus). Analyses included descriptive statistics, univariate analyses and logistic regression to compute adjusted odds ratios for the adoption of health information technology applications. Population Studied: Pediatricians, family practitioners and pediatric sub-specialists received an instrument that included additional questions on quality of care. Specifically, this study analyzed results from all pediatricians, family practitioners serving greater than 20% children, and pediatric subspecialists (n=1,014). Principal Findings: About a third or less of CHPs report that the following are a major or minor factor in their compensation: measures of clinical care (37.6%), patient surveys (31.4%), use of information technology (26.9%) and quality bonus/incentive payments (25.1%). A majority of CHPs report that the following are NOT compensation factors: email with patients (89.6%), phone consultation (86.6%) and group patient visits (86.9%). Results from the logistic regression found that those in the minor category (quality bonus) were 2.71 times more likely to have EHR (p=.023) than those in the “not a factor” category. Those in the major category for (quality bonus) were 3.27 times more likely to have EHR (p=.018) than those in the “not a factor” category. The logistic regression also found that those in the minor category (use of IT for compensation) were 8.35 times more likely to have EHR (p=.001) than those in the “not a factor” category. Conclusions: Pay for performance programs may already be affecting 25-30% of child health providers. The extent to which these are affecting technology adoption will be further examined to determine whether results differ for high volume Medicaid providers. Implications for Policy, Delivery, or Practice: This information may be the first of its kind in the pediatric arena to demonstrate that PFP efforts are promoting the adoption EHRs. A better understanding of the characteristics of these PFP programs is needed to understand their impact on patient care, quality and outcomes. Primary Funding Source: No Funding ●Electronic Office Based Applications Supporting Quality of Care Lisa Simpson, MB, BCh, M.P.H., FAAP Presented By: Lisa Simpson, MB, BCh, M.P.H., FAAP; Email: lsimpso1@hsc.usf.edu Research Objective: The literature provides numerous studies that have identified that office based capabilities may lead to improved quality. However, the extent to which these applications exist in a sample of community based physicians serving children remains unclear. The purpose of this study is to assess the extent to which physicians serving children have electronic and other capabilities known or believed to improve quality of care. Study Design: The current study is limited to Child Health Providers (CHPs) who responded to a larger survey of all ambulatory physicians in Florida. The independent variables include physician and practice characteristics (i.e. age, race/ethnicity, specialty, practice size/type and Medicaid volume). Quality related variables include physician ability to perform certain quality functions (e.g. sending reminder notices), experience with process failures (e.g. missing patient information), and beliefs about the effectiveness of various strategies to improve quality (e.g. involvement in collaboratives). Initial analyses included descriptive statistics and univariate analyses. An index was constructed to characterize a quality enabled practice which then served as the dependent variable in a multivariate analysis of independent factors associated with higher scores on this quality enabled practice index. The index is an un-weighted sum of single points credited for each of several office functions judged to be associated with higher quality such as having an EHR, clinical decision support, weight based dosing or growth charting etc. Population Studied: Pediatricians, family practitioners and pediatric sub-specialists received an instrument that included additional questions on quality of care. Specifically, this study analyzed results from all pediatricians, family practitioners serving greater than 20% children, and pediatric subspecialists (n=1,014). Principal Findings: Less than half of CHPs report being able to perform certain quality related functions (e.g. 44.5% send reminder notices for preventive care, 26.5% receive prompts for special follow up care). CHPs have the ability to create patient lists by diagnosis somewhat or very easily (44.1%) more often than patient lists by medications (15.5%) or laboratory results (17.1%). CHPs report rarely or never observing the following process failures: repeated tests because findings were unavailable (62.2%); or medication errors (82.7%). The logistic regression findings suggest that African American physicians were twice as likely to have practices that ranked in the top 50% of quality enabled practices (p=.017) than White, Hispanic, or Asian child health providers. Additionally, CHPs who are less than 40 or more than 60 years of age are credited with higher scores. Conclusions: Office based capabilities felt to be associated with higher quality have not been widely adopted by child health providers. This suggests that improvements in the quality and safety of care for children and adolescents may continue to lag behind adult populations unless specific attention is paid to the needs of child health providers. Implications for Policy, Delivery, or Practice: Primary Funding Source: No Funding ●The Effect of Information Technology on Quality of Care in the Veterans Health Administration Joanne Spetz, Ph.D., Ciaran Phibbs, Ph.D. Presented By: Joanne Spetz, Ph.D., Associate Professor, Community Health Systems, University of California, San Francisco, 3333 California Street, Suite 410, San Francisco, CA 94118; Tel: 415-502-4443; Fax: 415-476-4113; Email: jojo@alum.mit.edu Research Objective: Computerized patient records and bar code medication systems are gaining favor in the health care industry. These systems are believed to reduce patient care errors and improve the work environment for medical professionals, although there is little objective research on the effects of these systems on patient care and staff morale. In the late 1990s, the Veterans Health Administration (VHA) implemented two major information systems to enhance record-keeping and quality of care. The Computerized Patient Record System (CPRS), phased in over the past decade, consists of a comprehensive electronic patient medical record, including outpatient and inpatient services. The Bar Code Medication Administration (BCMA) system, which was in all VA medical centers by 1999, created a computerized pharmacy ordering, distribution, and administration system for use in the inpatient setting. This paper examines the effects of CPRS and BCMA on hours worked by nursing staff and adverse events experienced by patients. Study Design: Several VA data systems provide data about nurse staffing and patient outcomes. Nurse staffing is measured as hours worked and overtime hours. Patient outcomes are measured using the AHRQ Patient Safety Indicators and Inpatient Quality Indicators. Other patient outcomes measures, such as those used by Needleman and Buerhaus, also are used. All VA facilities were surveyed to obtain exact implementation dates of modules of CPRS and BCMA. The effects of CPRS and BCMA on staffing and outcomes are identified by estimating multivariate regression equations with panel data from the late 1990s through early 2000s. For each equation, the dependent variable is the outcome of interest, with the explanatory variables including whether CPRS or BCMA was implemented, patient characteristics, facility characteristics, and other factors. Population Studied: All facilities in the Veterans Health Administration. Principal Findings: Adoption dates of the information technology systems varied widely across VA facilities. Preliminary results from the quality analysis are expected in February, 2005. Conclusions: There is substantial need for ongoing evaluation of IT systems in healthcare. Implications for Policy, Delivery, or Practice: The lessons learned from this study will be of great importance to the future adoption of health IT systems. Primary Funding Source: RWJF ●Putting Medicaid Providers on the Map: An Evaluation of Address Record Quality in a State-Level Medicaid Provider Database John Stewart, MS, M.P.H., Ana Lopez De Fede, Ph.D., Alina S. Wyatt, BS Presented By: John Stewart, MS, M.P.H., GIS Manager, Institute for Families in Society, University of South Carolina, 1600 Hampton St, Columbia, SC 29208; Tel: (803) 777-5516; Fax: (803) 777-1120; Email: jstewart@gwm.sc.edu Research Objective: Increasingly, patient and medical provider address records are geocoded (geographically located) to estimate small-area disease incidence rates, identify disease clusters, delineate health care provider shortage areas, and calculate travel distance to medical care. Although a number of health researchers have examined positional accuracy issues associated with automated geocoding software, inadequate attention has been paid to the quality of address records in large administrative data sets. This study evaluates address record quality in a state-level Medicaid provider database, identifies common address record errors, and provides solutions for handling multiple address record deficiencies. Study Design: Address records for physicians providing primary or specialty care to South Carolina Medicaid recipients in 2003 were obtained in digital format from the South Carolina Medicaid Management Information System Address record quality was evaluated both manually and with an automated address standardization/error correction software package (ZIP4). Address record errors, including missing data, invalid ZIP Code information, misspellings, and nonstandard abbreviations, were tabulated by error type. Identified address record deficiencies then were corrected using ZIP4 and a variety of Web-based business and professional directories. Population Studied: Stratified sample of physicians providing primary or specialty care to South Carolina Medicaid recipients in 2003 (n = 16,333). Principal Findings: Notably, 12.7% of all physician address records lacked complete street address information. The majority of these records provided only post office box numbers. Approximately 5% of all address records contained obsolete or otherwise invalid ZIP Code information and roughly 3% contained verifiable street misspellings. More than one in ten address records used nonstandard city name abbreviations (e.g., “CHAS” for “Charleston”) and approximately 2% lacked proper directional prefixes/suffixes. Finally, more than 17% of address records employed nonstandard unit (e.g., building, suite) designations. Conclusions: Of the physician address records examined, a substantial proportion provided only post office box information, which does not permit geocoding at the street address level. Health information system initiatives to improve provider address record quality should promote the use of standardized street addresses including valid street numbers, appropriate directional prefixes/suffixes, and unit information. In most instances, address record errors encountered by secondary users of administrative data can be corrected using address standardization software and such Web-based directories as Google, SuperPages, Yahoo Yellow Pages, USPS ZIP Code Lookup, and AMA DoctorFinder. Implications for Policy, Delivery, or Practice: Enhanced address record quality can improve health care provider geocoding results, thereby strengthening state and local health professional shortage area assessment, health care resource allocation, and health policy formation. Primary Funding Source: Other Government ●Impact of Electronic Prescribing on use of Generic and Preferred Medications Christine Vogeli, Ph.D., Michael A. Fischer, M.D. MS, Margret Stedman, M.P.H., Rainu Kaushal, M.D. M.P.H., Timothy G. Ferris, M.D. M.P.H., Joel Weissman, Ph.D. Presented By: Christine Vogeli, Ph.D., Instructor of Medicine, Institute for Health Policy, Massachusetts General Hospital, 50 Staniford Street, 9th Floor, Boston, MA 02114; Tel: 617 7240984; Fax: 617 724-4738; Email: cvogeli@partners.org Research Objective: Controlling spending on prescription drugs remains a critical issue for patients, physicians, insurers, and policymakers. Generic medications represent one important source of potential savings on prescription drugs. Computerized prompts embedded in electronic prescribing (eRx) systems have the potential to increase generic prescribing. We studied the extent to which an eRx system is related to generic medication use. Study Design: In October 2003, two major health insurers, their PBMs, and a large e-prescribing company collaboratively launched a voluntary e-prescribing program in Massachusetts in which prescribers were provided with eRx software and free handheld PDAs. The e-prescribing system allows prescriptions to be written either with a handheld PDA device or on a personal computer linked to the internet. Prescriptions are immediately transmitted to pharmacies electronically or by fax. We studied the impact of an officebased eRx system on the use of generic medications using a pre-post design. We evaluated all prescriptions filled by one of the major health insurance plan during a one-year baseline period (April 2003-March 2004) and one year after the initiation of the eRx system within the plan (April 2004-March 2005). The insurance plan grouped medications into three copayment tiers, with generic medications having the lowest copayments, followed by preferred brand-name and nonpreferred brand-name medications. The eRx system uses color-coded text to identify medication co-payment tiers, with preferred generic medications appearing in green. We used claims for medical services to identify patients under the care of physicians who began using the eRx system during the study period and defined a cohort of patients who had seen physicians who were writing electronic prescriptions. We calculated by month the average proportion of prescriptions in each co-payment tier for patients before and after they had seen an electronic prescribing physician. Population Studied: Patients cared for by physicians practicing in a large sample of community practices in Massachusetts in 2004-2005. Principal Findings: We identified 243,000 patients who had seen a physician using the eRx system, accounting for over 4.1 million prescriptions. During the baseline period, generic medications accounted for an average of 54.4% of filled prescriptions. Over the following year, the proportion of generic medications was 57.0%, an increase of 2.6% (p<0.01). The increase in generic medications corresponded to a decrease in the use of preferred branded medications (second tier) from 33.3% to 32.1% (p=0.04); and a decrease in the use of non-preferred medications (third tier) from 12.3% to 10.9% (p<0.01). Conclusions: Use of generic medications increased slightly after initiation of an eRx system in a sample of communitybased practices, with corresponding decreases in the use of branded medications requiring higher co-payments. Implications for Policy, Delivery, or Practice: For many patients facing sharply tiered co-payment systems, increased use of generic medications offers the potential for considerably reduced out-of-pocket expense. ERx systems are one important tool to promote more widespread use of generic medications in the outpatient setting, even in states that already require generic substitution. Primary Funding Source: AHRQ