There's an inevitable clash between the goals of a CPOE and the realities required for its clinical use. Conceptually, the intent is for the provider to govern quality and cost through direct interaction with the machine, without interference. But in reality, most providers need help when keying in the data.
CPOE, therefore, has multiple potential redeeming values. How may of those actually come to fruition in real ED life? Let's look...
In theory, on the positive side, we have:
- Legibility - No handwriting to decipher
- Decision Support - Cross-checking allergies, drug-drug interactions, dosing accuracy, order sets- enabling smoother initiation and real0time modification of protocols for common yet complex problems like Sepsis and ACS.
- Real-time Alerts - Identify issues such as improper dosing, safety in pregnancy... warnings should avoid pitfalls, and thereby aid, not hinder, patient flow.
- Accurate Recording - An honorable CPOE system should create and accurate record of events, as the system gathers the information input with timestamps of all communications, instructions, and actions. This is good for quality measures, physician incentives and charge capture.
- Facilitates Read-through prior to submittal - Ideally, the provider would have the opportunity to consider all the orders as to their practical clinical and economic value, functioning not unlike the final checkout review before an placing an Amazon order.
- Efficiency - By eliminating duplication, unnecessary steps and personnel, CPOE has great potential to not only speed up processing of orders to nurses, pharmacy, lab, and x-ray, but also to remove vulnerable points in the process that add to risk.
- Auditing - Such tedious data tasks as verifying compliance various local and governmental requirement (think Core Values PQRS Clinical Support) would, again theoretically, support capture of data necessary to assure hospital financial incentives are received for compliance with various guidelines and regulations.
All these anticipated "pipe-dream" positives are balanced by the inevitable presence of unintended consequences. CPOE, as many efforts in the realm of HIT, suffers from the old axiom: "The Devil is in the Details". We would think that when CPOE implementation successfully keeps it simple stupid (KISS), addressing the goals noted in the positives, it can in most cases, be easily utilized, and with success. Yes, one would think...Right??
Actually what happens... First, there are the realities of HIT development in the modern world. Often, the development team has difficulty retaining the focus and goal of more simple modules and projects. As a result, the addition of new wish-lists from those beta-testing the rougher drafts, accumulate and grow. In that way, many teams feel they can accomplish more than they actually can. And just because the team's desire is there, that does not translate into superbly usable final product. Commonly a development team can get lulled into trying to expand the conceptual boundaries of a module, like CPOE, and make an all-inclusive "problem solver". But in reality, that extension of purpose often adds layers of complexity, and winds up impairing usability for the clinician. Moreover, since CPOE implementations are often hospital-wide, there frequently arises a "one-size fits all" mentality, that encourages uniformity, instead of specialized diversity, and tries to cater to every specialty with the same structure. Thus, many attempts to create facility wide CPOE have turned into cumbersome, user-unfriendly, mammoths that demand and the record voluminous transactions, failing to truly assist each individual physician user in a fashion appropriately specific to his or her needs.
Since the CPOE record is necessarily interwoven with the various documents of care, these extraneous elements become part of the chart, and in real-time, influence the tracking board, hence flow and efficiency with the equivalent of sludge in an engine. Here are some design principles to help that not happen.
- CPOE design should optimally incorporate modules that are directed, not to the entire institution, but to the are setting in which it is used.
- The CPOE should be broken apart by interfaces to provide the needs of a specific department. The Emergency department, or other care setting, usually has finite definable order needs 98% of the time. Modular templates of care enable rapid development of a focused order plan. When more are needed, the whole database can be searched.
- All CPOE modules should automatically facilitate input of test and other clinical data into the chart. findings from x-ray reports, labs, consultations, vital signs, and other ordered assessments should be added automatically, and the provider should be notified when these are complete or when a significantly abnormal result is obtained.
- The hart should reflect significant orders that change status or care setting. For example, a CPOE order to admit should be reflected in the chart, and a note to admit on the chart should appear in the CPOE screen.
- A good system will enable the clinician to choose a way of ordering common tests, and replicating that with simple screen or keyboard input (KISS approach).
- Ordered medications to be given while the patient is still in house should be clearly distinct, perhaps allotted a separate section from medications provided as outpatient prescriptions. Failure to clearly make this distinction can add-significantly to complexity and potential confusion. Care must be given not to complicate medication and/or prescription ordering to the extent that it affects patient flow. Pediatric dosing should not need a time and resource consuming pharmacy consult.
- Recognizing that the ED probably uses only about 50-60 standard P.O and IV medications, some complex medications included in high-risk protocols, oxygen, need treatments, IV fluids, will keep the formulary selection simple. The outpatient prescription writer is unlikely to need more than 50 common drugs that all staff physicians agree upon.
In conclusion: A cohesive, concise CPOE would allow the provider to practice medicine in a manner that is interactive with the machine, but is not overburdened with data entry. It would support the provider in making make intelligent choices that optimize patient care, and consciously utilize resources.
Work-flow Commentary by Donald Kamens, MD, FACEP
Some footnotes on physician-machine interaction, and how to get the most efficiency from the most expensive person:
Do not allow current absence of good usability to put a barrier between the physician and the machine. That could happen if- for the sake of workflow- unskilled, non-clinical workers are inserted into the doctor-patient relationship. Any barrier between the physician and the machine is a symptom of poor usability- PERIOD!
A well-functioning user interface should NEVER slow down a physician, but should speed him or her up. Persistent work-flow impairments that result from new implementations should--after a customary break-in period-- be deemed to result from a poor interface.
It is important to analyze productivity in terms of evolution of the technological capacity of the providers. The coming generations not only don't want other persons in the communication flow which they control (think Facebook, Twitter), but also see these as interfering obstructions in environment they are quite comfortable within.
Any attempt to insert individuals into the direct communication link between provider and ultimate enactor (nurse, lab tech, x-ray tech, etc.) is doomed. These communications should be open and free... and CPOE is a good place to start.
An important element in all this is the cognitive environment. That is, in what space does a provider best think? There are two: 1) at the bedside, and 2) at the machine (eliminating the bathroom, that is!). Optimally the CPOE device will be employed at the bedside. Tidying up, communications, and arranging disposition are okay times for sitting down. But the ED of the future, will need to eliminate the necessity of sitting down for long periods AFTER seeing the patient, to compose thoughts and write orders. There is not, nor will there be, enough time for such digression from actually seeing and caring for patients with dynamic tools like CPOE that are best executed at the bedside.
Thus, the most efficient physician, will no longer sit down at a desk to compose his/her chart. That will be completed (for the most part) while interviewing and examining the patient, and updated later as labs come in, treatments and re-assessments are done, and disposition is arranged. The result will be a much more efficient (albeit more physically tired) physician, who can go from room to room easily.
If our children can walk though their lives with eyes glued to a small screen, and thumbs ablaze, we should expect no less from the next generation of docs.
