This study suggests that HIT adoption remains limited and variable across key stakeholders. Use of HIT appears to be predominantly driven by financial functions, as reflected by the relatively ubiquitous use of electronic claims submission checking. While there is increased interest in the adoption of technology to improve the quality of care, significant adoption challenges remain, particularly in the area of EHRs and physician-patient communication.
Organizations face enormous financial challenges in adopting HIT. Our informants consistently discussed the large upfront investments necessary to deploy these systems, and only a small fraction of institutions, predominantly large institutions such as IDNs, could afford comprehensive versions of these systems. Ironically, the low penetration of HIT may itself contribute to its high cost, as each vendor must charge its small customer base a relatively high fee in order to recuperate research and development costs. These fees might put HIT out of the financial reach of small organizations, particularly small physician practices.
Misalignment of incentives is an important barrier to HIT adoption. Initial capital expenditures of HIT are high and payers do not directly compensate institutions or providers for its use, or for the resulting higher quality and safer care. This coupled with a hard-to-assess return on investment makes it difficult for institutions to risk making even necessary purchases[12
]. Decision makers tend to invest in areas that can readily be seen to be directly financially beneficial to the institution, which under the current reimbursement scheme are areas like MRI scanners and new buildings rather than IT infrastructure. While financial benefits can be realized from network investment, these benefits are seen across many IT cost centers and are hard to measure.
Organizations face enormous financial challenges in adopting HIT. Our informants consistently discussed the large upfront investments necessary to deploy these systems, and only a small fraction of institutions, predominantly large institutions such as IDNs, could afford these systems. Small organizations, particularly small physician practices, find it hard to purchase and maintain these systems. These practices are highly risk-averse and are justifiably fearful about the possibility of implementation failures and are therefore less likely to take the initiative to deploy these systems. While these institutions are small, they deliver the bulk of medical care in the US[34
]. Therefore, their financial barriers to the implementation of HIT have significant implications for the nation's quality improvement agenda in healthcare.
Apart from the issues of cost, significant concerns remain regarding the impact of HIT initiatives on productivity. Several institutions we interviewed cited the up-front loss of productivity during the transition from paper-based to computer-based systems. When the income of health-care providers is directly tied to their productivity but not to their quality, this resistance to change could be particularly difficult to overcome in the era of decreasing reimbursement. However, research seems to suggest that the negative impact of EHR implementation on productivity is modest and may diminish over time[36
]. Such discordance between provider perception in the community and findings from these recent studies performed at academic institutions may be attributable to several factors. First, usability and clinical decision support among vendor products vary significantly. Second, fewer resources may be available to train physicians in the community setting. Third, many small practices may not be able to afford the fees vendors might charge to customize the HIT products to fit the local workflow. Therefore, future research, possibly in collaboration with vendors, needs to focus on ways to design applications that would be intuitive to even new users and adaptable to the different workflow patterns in small practices.
While other industries, such as the financial and computer industries, have long established uniform standards for the interchange and ordering of parts and data, health care has lagged behind. The lack of data standards makes it difficult to manage the myriad of existing homegrown and vendor systems each organization might own. This, in turn, makes it very costly to invest in more advanced HIT capabilities. The lack of interoperability between different data sources also undermines the usefulness of HIT to the clinician, who may be asked to use a variety of paper-based and electronic methods to retrieve and enter data even for the same patient. The challenges associated with using these non-interoperable HIT systems may negatively impact workflow and productivity, which in turn contribute to clinicians' resistance to adopt these systems[24
Within the markets we studied, the use of HIT varied considerably across the stakeholders. For example, while several IDNs in both Boston and Denver are making major investments in HIT, most nursing homes and rehabilitation hospitals lag behind significantly. Small physician practices are highly risk-averse and are justifiably fearful about the possibility of implementation failures and are therefore less likely to deploy HIT. This variation in the use of HIT across stakeholders is noteworthy from several standpoints. Since patients often transition from acute-care settings to non-acute care settings, the improvement in quality gained through HIT investments in acute-care hospitals may be attenuated by the under-investment in chronic care institutions and physician practices. While these non-acute facilities are often small in size, they deliver the bulk of medical care in the US[34
]. Therefore, the financial barriers that they face have significant implications for the nation's quality improvement agenda in healthcare. Furthermore, our informants at chronic care institutions and many of the physician practices indicated that this underinvestment in HIT would continue at these organizations. If that is indeed the case, the investment gap between acute care and non-acute care facilities would likely widen, potentially leading to greater disparities in quality. From a healthcare policy perspective, our findings point towards the urgent need to understand ways to overcome the barriers to diffusing HIT in chronic-care facilities and small physician practices.
The findings of this study should be interpreted in light of its limitations. First, we acknowledge that the qualitative methods used in this study may yield less precision than quantitative methods, although collecting quantitative data in all the stakeholders covered in our study would be very difficult. Second, while there is good agreement between the estimates of HIT adoption in the two markets and the estimates provided by the expert panel, we acknowledge that our expert panel was probably influenced by the market analysis we performed. Third, the selection of Boston and Denver were based our access to contacts in those markets, and not all of the potential informants we contacted participated in the interviews. Therefore, the results of our qualitative analysis might have been subject to selection and responder biases. Both forms of biases would likely have caused the research team to over-estimate the level of HIT adoption, and the actual levels of adoption may be even lower. Fourth, the national adoption estimates derived from our expert panel were likely influenced by their personal biases. Fifth, our limited resources prevented us from studying in depth the adoption of HIT in more than two markets, and regional variations in HIT adoption would have been difficult to discern.
In summary, HIT adoption in the US remains in its infancy, and the significant disparities in adoption among key stakeholders will likely worsen unless incentives for HIT adoption can be realigned to reward quality rather than quantity of care. Several levers, including quality-based financial incentives and adoption of standards by payors, will likely represent significant facilitators of this process. Despite the daunting challenges that lie ahead, US policy makers and health care institutions should take heart in the fact that other industrialized nations have been successfully deploying HIT to improve the quality of care[18