The Value Of Health Care

In this paper we assess the value of electronic health care information exchange and interoperability (HIEI) between providers (hospitals and medical group practices) and independent laboratories, radiology centers, pharmacies, payers, public health departments, and other providers. We have created an HIEI taxonomy and combined published evidence with expert opinion in a cost-benefit model. Fully standardized HIEI could yield a net value of $77.8 billion per year once fully implemented. Nonstandardized HIEI offers smaller positive financial returns. The clinical impact of HIEI for which quantitative estimates cannot yet be made would likely add further value. A compelling business case exists for national implementation of fully standardized HIEI. A ttent ion to the use of information technology (IT) in health care is intensifying rapidly, with President George W. Bush calling for widespread adoption of electronic medical records (EMRs) within the next ten years.1 In addition to digitizing the information that providers use to care for their patients within organizations, clinicians, patients, and policymakers are looking ahead to sharing appropriate information electronically among organizations. David Brailer, newly appointed national health information technology coordinator, recently called for expansion of such interoperability to the flow of clinical and other administrative data, citing its importance for encouraging health care IT investment and facilitating health care reform.2 To explore the qualitative and economic implications of health care information exchange and interoperability (HIEI), we studied the value of electronic data flow between providers (hospitals and medical group practices) and other providers, and between providers and five stakeholders with which they exchange information most commonly: independent laboratories, radiology centers, pharmacies, payers, and public health departments. We hypothesized that the clinical benefits of electronic data exchange would be substantial H e a l t h T r a c k i n g W 5 1 0 1 9 J a n u a r y 2 0 0 5 DOI 10.1377/hlthaff.W5.10 ©2005 Project HOPE–The People-to-People Health Foundation, Inc. The authors are with the Center for Information Technology Leadership, Partners HealthCare System, in Boston, Massachusetts. Jan Walker ([email protected]) is its executive director; Eric Pan, associate fellowship director and a senior analyst; Douglas Johnston, a senior analyst; Julia Adler-Milstein, a research analyst; David Bates, a member of the executive committee; and Blackford Middleton, chairman. Bates is also chief of the Department of General Internal Medicine at Brigham and Women’s Hospital (Boston) and director, clinical and quality analysis, Partners HealthCare System. Middleton is also corporate director, clinical informatics research and development, at Partners. on S etem er 9, 2017 by H W T am H ealth A fairs by http://conealthaffairs.org/ D ow nladed fom and that financial benefits would outweigh costs. In this paper we report on the results of our analysis. Study Data And Methods We used a range of methods to gather evidence, including literature reviews, expert interviews, and estimates by an expert panel. We focused our efforts on analyzing published sources for data but, where these were lacking, turned to experts to fill critical gaps. We then created a cost-benefit model to project value to organizations and to the country. A full project report that contains a detailed description of the methods we employed is forthcoming.3 Literature review. We worked with a medical librarian to complete a systematic review of the U.S. literature addressing the clinical, financial, and organizational value of HIEI in these interorganizational relationships, and we also searched trade press, general press, and online sources. Not surprisingly, given the lack of real-world implementation of interoperable systems in health care, we found few sources targeting HIEI value specifically. Experts. We convened a panel of nationally known experts to advise us throughout this project. They brought expertise in regional data-sharing initiatives, economics, public health, payment systems, informatics, and public policy. With relatively little research and literature on the value of HIEI, the panelists played an important role, participating in structured telephone interviews, a oneday meeting, e-mail polling, and discussions. We also interviewed more than twenty other experts, including provider information systems executives working with various facets of interoperability and directors of regional datasharing initiatives. The panelists and other experts helped identify data sources and estimated key data points that were not available in published sources. Analytic framework. We devised a conceptual framework describing how health care entities share information and created a functional taxonomy reflecting the amount of human involvement required, the sophistication of IT, and the level of standardization. The taxonomy has four levels. Level 1: Nonelectronic data—no use of IT to share information (examples: mail, telephone). Level 2: Machinetransportable data—transmission of nonstandardized information via basic IT; information within the document cannot be electronically manipulated (examples: fax or personal computer [PC]–based exchange of scanned documents, pictures, or portable document format [PDF] files). Level 3: Machine-organizable data—transmission of structured messages containing nonstandardized data; requires interfaces that can translate incoming data from the sending organization’s vocabulary to the receiving organization’s vocabulary; usually results in imperfect translations because of vocabularies’ incompatible levels of detail (examples: e-mail of free text, or PC-based exchange of files in incompatible/proprietary file formats, HL-7 messages). Level 4: Machine-interpretable data—transmission of structured messages containing standardized and coded data; idealized state in which all systems exchange information using the same formats and vocabularies (examples: automated exchange of coded results from an external lab into a provider’s EMR, automated exchange of a patient’s “problem list”). Software model. Using Analytica software (version 3.0.1) from Lumina Decision Systems Inc. (Los Gatos, California), we created the analytic model as an influence diagram. This software allowed us to depict complex factor relationships graphically, to consider many factors simultaneously, to incorporate probability distributions to represent uncertainties, and to test the sensitivity of projections to variations in key inputs. Although we cannot include full model specifications in this brief paper, we reference important data sources in each topic area. Projections of costs. We projected costs for the interfaces required by each participating organization’s computers for communicating with external computers and for internal HIEI-capable systems for providers. To calculate national costs, we allocated relevant costs to relevant organizations. The only exception to this approach was for providerM a r k e t W a t c h H E A LT H A F F A I R S ~ W e b E x c l u s i v e W 5 1 1 on S etem er 9, 2017 by H W T am H ealth A fairs by http://conealthaffairs.org/ D ow nladed fom payer costs, which were taken directly from the Health Insurance Portability and Accountability Act’s (HIPAA’s) Final Impact Analysis.4 Interfaces are programs that enable different systems to communicate with one another. We estimated Level 3 interface development costs based on expert opinion, assigning $50,000 per interface for hospitals, labs, radiology centers, pharmacies, and public health departments, and $20,000 per interface in grouppractice offices. Experts were divided on whether Level 3 or Level 4 interfaces would cost more; we assumed that they would cost the same. Level 3 requires a unique interface to each external organization, and we assumed from eight to twenty interfaces per provider, depending on the provider’s size. Level 4 HIEI requires one interface to each type of external organization—for example, one interface to all external laboratories, totaling five per provider. For both Level 3 and Level 4, each external organization requires an interface to providers, and we assumed one per laboratory, radiology center, and pharmacy and two per local public health department—one to hospitals and one to office practices. Relatively few providers currently have broad and mature clinical information systems.5 Thus, we assumed that all U.S. providers would install new systems, using the Institute of Medicine’s (IOM’s) definition of minimal functional specifications for the electronic health records that would be required for HIEI Levels 3 and 4.6 To estimate the national costs of these systems, we applied Christian Birkmeyer’s cost estimates to hospital providers and our earlier estimates for advanced ambulatory systems to outpatient providers.7 Acquisition costs include initial licenses, hardware, implementation, and training. For both interfaces and provider systems, we assigned annual maintenance costs equal to 17.5 percent of the initial acquisition costs to cover ongoing license fees, system upgrades, and hardware replacement costs. Projections of benefits. We searched for evidence about information flows between organizations and asked the expert panel to estimate the impact of each level of HIEI on these flows. The model calculated benefits to organizations and to the country as a whole by combining published quantitative evidence, national provider statistics, other data points, and expert-panel estimates of HIEI impact. As an example, Exhibit 1 illustrates the projection of benefits from Level 4 HIEI between outpatient providers and independent laboratories. As in our other calculations, we asked expert panelists to estimate the impact on participating organizations once they are connected at each HIEI level. To simplify our analysis, we assumed that this was effective 100 percent of the time. The model first estimates baseline total lab test costs: a combination of fees billed by laboratories and administrative costs incurred by providers in handling the paper and phone calls associated with tests. Then it estimates the proportion of tests (and costs) that are redundant and avoidable with HIEI. For the remaining tests, it estimates the impact of HIEI on the administrative portion of test costs. Finally, the model sums these cost savings and applies them to recent population statistics to calculate national benefit. National rollout scenario. To allocate benefits and costs over time, we developed a ten-year national implementation scenario. We assumed that 20 percent of organizations would install systems in each of the first five years, incurring all acquisition and start-up costs in year 1, and maintenance costs in years 1 through 10. We postulated that each organization would accrue 50 percent of potential benefits in year 1, and that benefits would increase by 10 percent each year. We did not attempt to account for inflation, discounting, or changes in utilization from changes in the national population. Therefore, amounts are in 2003 dollars and reflect current care patterns and population figures. Again, provider-payer costs are an exception, as they were amortized over three years to be consistent with HIPAA’s Final Impact Analysis.