New Survey Documents the Headaches of Computerized Medicine
02/25/11 Portland, Ore.
An OHSU-led survey of 176 hospitals shows that workflow shifts and a host of adverse unintended consequences are common, but the end result is far fewer medical errors.
Clinical information technology systems - especially those known in the health care industry as computerized provider order entry (CPOE) systems - promise to improve health outcomes, reduce medical errors and increase cost efficiency, but hospitals adopting them must plan for "immense" workflow issues and a host of other unanticipated consequences that come with them or face potentially crippling problems, concluded a study led by researchers at Oregon Health & Science University.
The researchers found in a survey of 176 hospitals where CPOE systems have been integrated into daily operations that unintended adverse consequences were virtually universal. CPOE systems are those that require a physician or other health care professional to enter prescriptions and other medical orders directly into a computer database.
For six out of eight previously defined categories of unintended consequences, more than 70 percent of the institutions ranked the level of impact on operations as "moderately to very important." Those were issues involving alterations in workloads, workflow, communication patterns, never ending system demands, emotions and system overdependence that led to havoc during system failures. Doctors, for example, were spending much more time at the computer inputting prescriptions and other orders. The two remaining categories were the generation of new kinds of errors and changes in the hospital power structure which fewer than half of the respondents ranked as high in importance All eight categories - identified in an extensive preliminary study at five hospitals of varying sizes - were encountered by most of the hospitals surveyed regardless of how long they had been using CPOE.
The survey was the first to quantify the breadth and importance of CPOE's unintended consequences. The survey results were described in a research paper published in the July/August issue of the peer-reviewed Journal of the American Medical Informatics Association ["The Extent and Importance of Unintended Consequences Related to Computerized Provider Order Entry" at http://www.jamia.org/content/vol14/issue4/]. The survey was conducted over during the last six months of 2005.
"The workflow issues are immense and they affect nearly every hospital staff member," the survey found. "More work and new work are inevitable, as are increasing system demands…Unless we make a concerted effort to avoid, manage, and/or overcome unintended consequences, the implementation of clinical information systems may lead to detrimental results," the paper's authors said.
"We were told that CPOE systems are stopping far more errors than they are causing," said Joan S. Ash, Ph.D., M.L.S., M.S., M.B.A., the lead investigator, who is an associate professor and vice chair, department of medical informatics and clinical epidemiology, OHSU School of Medicine. "Handwritten prescriptions, for one thing, are no longer the problem they were before the systems were adopted," she said, "but we're also seeing new kinds of errors crop up. The one we've seen over and over again is physicians trying to enter orders for the wrong patient. Usually they're caught, most often by the pharmacist or the nurse, but sometimes they're not."
Ash, in a separate article published in the "News and Views" section of the June 30, 2007 issue of the prestigious British Medical Journal headlined "How to avoid an e-headache," provided additional insight into why implementation of health information technology systems is proving so difficult. "It is because we are transforming health care through information technology rather than simply automating old processes. Workflow and work life must change, which means people must adapt. Such change is deeply disruptive. The related personal and organizational challenges are enormous, yet efforts to manage change receive inadequate attention and funding…Boldness breeds occasional blunders, which can teach us much about what is required for eventual success."
At least 400,000 preventable drug-related injuries occur in hospitals each year, the Institute of Medicine of the National Academies reported last year, and illegible handwritten prescriptions figure in a significant share of them. Prescriptions ordered electronically are safer and, combined with decision support tools, automatically alert prescribers to possible interactions, allergies and other potential problems, the Institute said, and urged that all health care providers have electronic systems in place by 2010.
Fears that the CPOE "cure" might be worse than the disease likely are impeding the diffusion of CPOE throughout hospitals in the United States, the authors of the JAMIA paper asserted. Those fears gained credence when a pediatrics hospital in Pittsburgh attributed a higher mortality rate to its CPOE system - mistakenly, it later turned out - and when Cedars-Sinai Medical Center in Los Angeles, hospital of the stars, shelved its $34 million system after a staff revolt.
"An important goal of the current work," said the authors, "is to identify the types of unintended consequences so that they can be monitored and managed: once unintended consequences are predicted or detected, their management can knowingly involve tradeoffs. An example of a CPOE-related tradeoff is the degree to which the extra time physicians must exert to use CPOE (an undesirable consequence) is offset by the increased information physicians find available at the point of care via CPOE - making patient visits more effective. The ability to maintain control over consequences may give hospital decision makers more confidence when making the determination to implement CPOE.
With respect to the lower level of importance the survey respondents assigned to power structure shifts and the generation of new kinds of errors, the authors cautioned that the survey methodology could have produced biases because those responding to the survey tended to be information technology staff with a clinical background. The power shifts in question tended to be away from physicians and to administration or information technology staff, "so respondents may not recognize their own gain in power," the authors wrote. Also: "New kinds of errors may seem minor to these interviewees who seem confident that errors are most often caught before harm is done."
Authors of the JAMIA paper besides Ash were Dean F. Sittig, Ph.D.; Kenneth Guappone, M.D.; Emily Campbell, R.N., M.S.; and Richard H. Dykstra, M.D., M.S., all affiliated with OHSU, and Eric G. Poon, M.D., M.P.H., of Brigham and Women's Hospital and Harvard Medical School. Sittig also serves as director of applied research in medical informatics for Northwest Permanente, PC.
The research was supported by a grant from the National Library of Medicine of the National Institutes of Health.