Physician computer order entry in a hospital setting.

In a story told at medical meetings that bears on health care costs, the presenter asks the audience to name the most expensive item of medical equipment. The audience will normally respond with various items of advanced technology, such as radiology scanners or laser surgical equipment. After they are done guessing, the presenter removes a pen from his or her pocket and states that it is the most expensive piece of equipment, because physicians use it to order all the other tests.

At the present time, in most instances, a physician caring for a hospitalized patient will order tests and treatments by writing on the physician order form in the medical record. However, in some institutions physicians are ordering through direct entry by computer. What I would like to do is discuss what has already been done and think about where this might lead.

Stefanichik has coined the term "point of care" and draws the analogy to "point of service" in industry. He also states that, between his surveys in 1986 and 1990, the question of installing this technology has gone from an "if" to a "when" decision. [1] When will be determined by the ability to balance capital requirements with the expected return on investment. It will occur when the costs of the systems decrease or more benefits are obtained for the same amount of dollars. The more physician order entry functions that can be incorporated on the system, the more rapidly the benefits and savings can be expected to accrue.

The addition of order entry functions has proceeded incrementally. An early attempt involved efforts at the Medical College of Virginia in 1986 to utilize physician order entry in the Department of Pharmacy. The department was among the first to experience and report that physicians were not willing to use an order entry system that took more time to use than the present one. Prior to making the system useful, the department had to make it faster. It was able to do this by installing order sets and utilizing a light pen format.[2] It was only after it removed this obstacle that the department was able to start achieving some of the economies and improvements that it had anticipated. Once it started using computer order entry, the department began to influence physician prescribing habits. The process for prescribing formulary items was much faster than for nonformulary items. This encouraged physicians to prescribe from the formulary, which was the behavior that the pharmacy wished to encourage. In addition, it was able to provide cost information to physicians when orders were entered, and their experience was that this would reduce the cost of medications.

Michael Reese Hospital, Chicago, Ill., had a similar experience. It was interested in influencing the prescribing behavior of physicians treating pneumonia. It felt that two antibiotics were equally effective but that one was less costly, and it wished to encourage the use of the latter.

The first step was to add a paragraph on the order entry screen that explained the benefit of using the preferred antibiotic. The result was not totally satisfactory, so the next step was to have the screen default to the preferred medication. This produced the desired behavior, because it required several more steps in order entry to prescribe medications other than the preferred item. However, there was some physician concern that there were other acceptable options and the final modification allowed the physician to select from any of three antibiotic choices without having to do extra steps in order entry. The three choices were the preferred antibiotic and two less expensive, slightly less effective alternatives.[3]

Other advantages of pharmacy order entry have been reported. The computer makes it possible to rapidly cancel orders that have already been ordered by another physician and save pharmacy time and effort. It also makes it possible to rapidly disseminate information in the event that a drug has been recalled. The computer enhanced the hospital's ability to prevent drug reactions. Its experience was that it was finding three possible drug reactions a day. The potential cost of each drug reaction avoided was estimated at $6,000 of additional care.[4]

Another clinical area that has shown potential for physician order entry is the ordering of laboratory tests. Besides the already-stated advantages of faster input into the system, the potential for expert advice is seen as a strong incentive to add laboratory test ordering to any menu. A problem in manual order entry systems is the potential for the same test to be ordered by several physicians involved in the care of the patient and unaware of what their colleagues have ordered. At the University of Minnesota Laboratory, an embedded system in the physician order entry program assists physicians in making better clinical decisions. Within the embedded system is the ability to inform the physician when he or she may have ordered a redundant or nonvalue-added test. The system also has the capability to block routine ordering of a test that appears to be inappropriate for the diagnosis. To order these tests requires the physician to acknowledge that they may be inappropriate in a general context but are needed in this particular situation. This has required the hospital to develop a system to provide for periodic updating as knowledge is advanced in the field. It also developed the capability for an interface between lab and pharmacy. This allows it to create dosage adjustments in patients with impaired organ function and an inability to metabolize the drug in a normal manner.[5]

As the science of computerized physician order, entry advances, new problems come to light. Most systems to date have been developed and used in teaching hospitals where they were thought to have an inherent advantage, in that physician ordering would be done by residents and faculty physicians. It was also thought to be an advantage that resident physicians were younger and therefore more likely to be computer literate. Although both assumptions were true to some degree, they were balanced against the unanticipated psychological hurdle that order entry had been done traditionally by the unit clerk, seen by the resident physicians as a low-status job. They were now being asked to perform a task that they saw beneath their status and ability. It took some time before their awareness of other benefits within the system was able to overcome this initial difficulty.[6]

Another problem is the nature of work in a hospital. This has been described as an "unbounded system," an organization where the lines of authority are not well drawn and where the decision making process is ill-defined. At the University of Virginia, the clinical-administrative boundary was unclear. It was only after awareness of this issue developed and it was addressed that progress was possible. It happened when clinical decision makers were involved in the process and felt ownership independent of the medical information systems.[7]

This process showed the need to transfer the ability to see the benefits of the new system from administration, which made the decision to proceed with the system, to the clinical staff, which would have to work with the system. To do this, administration had to explain the advantages of the system. Now the physician would be able to order from any terminal within the institution and not have to be present on the unit. The computer would now provide prompts that suggested dosages and timing of drug administration. The computer incorporated standard order sets for recurrent problems, which reduced the time it took physicians to order tests and treatments. They received relevant feedback on lab tests when it was appropriate to ordering medications.

A more recent installation of a physician computer order entry system has occurred at Brigham and Women's Hospital, Boston, Mass. As expected, the hospital built on what was already known. It was interested in reducing order entry problems by eliminating transcription errors in handwritten orders. It anticipated improved reimbursement by being better able to effectively capture charges. By moving orders electronically, it anticipated being able to move orders around the hospital faster than when they were carried by hand. The hospital accommodated initial staff reluctance about the system by providing a variety of options for entry. It had an assisted mode where prompts appeared on the screen and also used a text mode that better approximated the traditional way of order entry.[8]

Brigham was among the first institutions to create flexibility in the way physicians could adapt standard order sets to the occasionally variable nature of how an individual patient responds to treatment. It broke multiday protocols into individual days. If the patient process varied from the routine, the hospital could provide the flexibility to advance or postpone the protocol.[9] They encountered the same problem seen elsewhere: the need to accommodate physicians' need for speed in the order entry process and the need for flexibility as orders are written. In exchange for accommodating these requirements, the hospital has documented significant savings and faster response to patient needs.

At the present time, medicine is at a crossroads on computer order entry. The rate at which knowledge expands continues to accelerate. Even as more specialty knowledge is needed, systems thinkers are calling for the return of generalists. Can embedded expert knowledge systems help bridge this gap? The other central issue is the cost of health care. Any technologies that will lower rather than raise costs will be favored. The computerized physician order entry process offers a method to reduce nonvalue-added tests and treatments. The computer offers the possibility of standardized order sets, elimination of redundant tests, and the possibility of avoiding treatments with minimal or no benefit. It offers a rapid means of disseminating new knowledge and incorporating best practices or benchmarks into current practice. It allows all members of the health care delivery team access to current state-of-the-art expertise.

References

[1.] Stefanchik, M. "Patient Care Systems. Point-of-Care Revolution." Computers in Healthcare 12(4):19-20,24, April 1991. [2.] Schroeder, C., and Pierpaoli, P. "Direct Order Entry by Physicians in a Computerized Hospital Information System." American Journal of Hospital Pharmacy 43(2):355-9, Feb. 1986. [3.] Kawahara, N., and Jordan, F. "Influencing Prescribing Behavior by Adapting Computerized Order-Entry Pathways." American Journal of Hospital Pharmacy 46(9):1798-801, Sept. 1989. [4.] Morrissey, J. "Workstations That Work." Modern Healthcare 24(41):40-2, Oct. 10, 1994. [5.] Connelly, D. "Embedding Expert Systems in Laboratory Information Systems." American Journal of Clinical Pathology 94(4 Suppl 1):S7-14, Oct. 1990. [6.] Massaro, T. "Introducing Physician Order Entry at a Major Academic Medical Center: II. Impact on Medical Education. Academic Medicine 68(1):25-30, Jan. 1993. [7.] Massaro, T. "Introducing Physician Order Entry at a Major Academic Medical Center: I. Impact on Organizational Culture and Behavior." Academic Medicine 68(l):20-5, Jan. 1993. [8.] Teich, J., and others. "Design of an Easy-to-Use Physician Order Entry System with Support for Nursing and Ancillary Departments." Proceedings of Annual Symposium on Computer Applications in Health Care, 1993. p. 99. [9.] Teich, J., and others. "Response to a Trial of Physician-Based Inpatient Order Entry." Proceedings ofo Annual Symposium on Computer Applications in Health Care, 1993, p. 316.

Robert H. Fabrey II, MD, is Vice President Medical Affairs, St. Joseph's Hospital, Asheville, N.C. He may be reached at 428 Biltmore Ave., Asheville, N.C. 28801, 704/255-3100, FAX [phone].