Part one of a two-part series (Access part two here):
There is too much good stuff in the recently released report, Computational Technology for Effective Health Care: Immediate Steps and Strategic Directions, to do it all justice in just one Health IT Strategist-length article, evidenced by an interview with one report co-editor, William Stead.
On a call for an interview that was scheduled for 15 minutes, we spoke for an hour and still didnt cover everything. So well be doing a two-part series based on the 138-page report and the discussion with Stead prompted by it.
Stead, a physician who is associate vice chancellor for strategy/transformation and the chief information officer of the Vanderbilt University Medical Center in Nashville, worked with Herbert Lin as co-editors of the report, prepared by the Committee on Engaging the Computer Science Research Community in Health Care Informatics, a committee of the Computer Science and Telecommunications Board of the National Research Council of the National Academies.
The report was funded by HHS, the National Science Foundation, Vanderbilt University, 10-hospital Partners HealthCare System, Boston, the Robert Wood Johnson Foundation and the Commonwealth Fund. Stead is a member of the Computer Science and Telecommunications Board and Lin is its chief scientist. Corporate members of the board include representatives from Google, IBM Corp., Microsoft Corp. and Yahoo! Research.
The study had two goals: to identify how computer usage might be applied more effectively to healthcare, and how the limitations of current technologies and approaches might be overcome through additional research and development. The study group focused on the information technology usage of major healthcare organizations, which its authors conceded is a limitation, noting the majority of healthcare is delivered in small-practice settings (of two to five physicians) that lack significant organizational support. (Actually, about 37% of office-based physicians are in solo practice, according to National Center for Health Statistics survey data.) Still, the authors say they hoped their efforts would lay the groundwork for future efforts of exploring unanswered questions raised by this study.
The study group visited eight hospital organizations deemed leaders in the use of health IT, including government, not-for-profit and for-profit organizations where many of the important innovations in IT would be found. They were the Palo Alto (Calif.) Medical Foundation; the 642-bed UCSF Medical Center, San Francisco; 18-hospital Intermountain Healthcare, Salt Lake City; 12-hospital Partners HealthCare System, Boston; 833-bed Vanderbilt University Medical Center, Nashville; TriStar Health System, Nashville; 291-bed Veterans Affairs Medical Center, Washington; and 12-hospital UPMC, Pittsburgh. In addition to site visits, the committee also leaned heavily on previous work by the Institute of Medicine, particularly its 2001 report, Crossing the Quality Chasm, as well as a review of other literature and the committee members own experience.
Part of the problem with health IT, according to the authors, is that the systems appear to have been designed more with an eye to replicate existing paper-based forms and functions than aiming to seize the opportunity to innovate and improve care, noting that today's clinical IT systems provide little support for the cognitive tasks of clinicians or the workflow of the people who must actually use the system(s). These computer systems are poorly designed and do not take advantage of human-computer interaction principles, which sometimes lead to more work and the increased chance of error.
The authors insist that healthcare reform should follow IOM recommendations about care being patient-centered, that it improve safety, efficiency and equity, but none is more important than the effective use of information.
The authors said the overarching, grand research challenge of the computer science research community is developing patient-centered cognitive support. Computer systems need to build virtual models of a patients status, models that depict and simulate a theory about interactions going on in the patient similar to those models going on in the head of a physician working unaided by a computer.
What links the raw data in the computer to these abstract models the authors describe as medical logic or computer-based tools that examine raw data relevant to a specific patient and suggest their clinical implication given the context of the models and abstractions. Such a vision invokes many challenging computer science research problems, such as the challenge to create decision-support systems that could accommodate multiple clinicians involvement with a single patient as well as serve as decision support for the patient himself or herself.
The eight sites of healthcare organizations visited by the committee were deemed to be the health IT equivalents of shining cities on hills because of their successful use of IT, but the authors faintly praised their successes as merely encouraging. According to the report, these renowned organizations also fall far short, even in the aggregate, of what is needed to support the IOMs vision of quality healthcare.
IT systems that in many cases took decades to implementa problem in itselfwere rarely well-integrated into clinical practice and were rarely used to provide clinicians with evidence-based decision support and feedback. Clinician attitudes about IT also were an issue, as care providers said they spend a lot of time documenting patient care with a computer, but said it was to comply with regulations or to defend against lawsuits but evidenced little faith the information would ever be used to improve clinical care.
Yes, the places that we visited have made a significant step forward, but that gap between where they are and what is envisioned in the chasm is still wide, Stead said during the interview.
There is something about seeing all this in so many places one after another; it just hit us in the face the fact that the patient is hidden behind a sea of transaction data and there is very little help for the clinician, (a system) that helps them think, Stead said. Im not sure thats been recognized clearly. If were really going to get the benefit by widespread (IT) adoption, we have to go after a notion of IT that can really close that gap.
While acknowledging that federal and state governments play important roles in supporting research, the report authors specifically warn that the federal government should shun support for adoption of technology for its own sake, as if the technology were an end unto itself, but rather explicitly embrace the ends rather than the means, that is, measurable healthcare quality improvement as the driving rationale for its health IT adoption efforts.
The report included several recommendations to the federal government along these lines, including the following:
- Any government incentives should be for clinical performance, not IT acquisition per se. These incentives should reward one-foot-at-a-time improvements in quality of care using an iterative process of software and system development.
- The government should encourage the development of performance standards and measures for decision support.
- It also should encourage interdisciplinary research into the design of healthcare systems processes and workflow, computable knowledge structures and models for medicine and human-computer interaction in a clinical setting.
- And the government should at the least not impede, but at best, encourage the aggregation of healthcare data, processes and outcomes subject to appropriate protection of privacy and confidentiality.