The Impact of Clinical Information Systems and Automation to Health System Pharmacies
Recently, the American Society of Health-System Pharmacists published the results of a survey1 employing a stratified random sample of pharmacy directors at 1,101 U.S. general and children's medical-surgical hospitals. There were 514 surveys returned, with response rate of 46.7 percent. The results showed that although health systems have been slow to adopt computerized prescriber order entry systems (7 percent) and bar-code technology at the patient's bedside (1.5 percent), the majority of pharmacy departments have implemented automated dispensing devices.
During the first part of this decade, these pharmacies implemented two types of automated systems. The majority (58 percent) employed pharmacy automation that totally redesigned the medication-management system from the traditional unit-dose-dispensing system to a decentralized system utilizing automated dispensing machines (ADMs) located in the patient care areas. For hospitals with 300 or more staffed beds, 83 percent used ADMs. Although only 7.8 percent of the surveyed pharmacies employed centralized robotic automation that automates the daily individual patient-medication-preparation and pharmacist-verification activities of the unit-dose system; this type of automation occurred in 24 percent of the hospitals with greater than 300 beds.
This data verifies that most pharmacies that implemented centralized robotics also incorporated some decentralized automation to manage federal- and state-controlled medications, floor-stock medications, and as-needed medications (PRNs), since these are difficult to manage and are inefficient to dispense using centralized robotic automation.
Automated Dispensing — Streamlining Pharmacy and Nursing Processes
The driving forces behind adoption of this technology were the positive impact of automation on cost savings for pharmacy and nursing in labor and medications; the quality improvements to patient care; the nurse and pharmacist labor shortage; and impressive returns on investment. Figure 1 illustrates the impact and results of these systems. The ROI to health care facilities was usually between two and three years. In some cases, the ROI was less than a year. These systems have proven to be both cost- and quality-effective, and therefore, their adoption rate has been quite remarkable.
Initial robotics and automated dispensing machines were designed to replace or support the unit-dose drug distribution system. With the exception of total parenteral nutrition formulators, the preparation of medications for administration by the intravenous route has been performed manually within pharmacy departments. Very recently, a new company2 entered the marketplace offering a fully automated device to prepare small-volume parenteral products. This device, PARxD IV™, prepares injectable medications from vials or reservoir bags into syringe-based dosage forms. It can store up to 50 different medications of commercially available vials with a capacity of 1,500 doses. Operation speeds range from 60 doses per hour to over 300 doses per hour, depending upon the medication source container selected. The device is a fully enclosed class-100 sterile environment, with built-in routines for quality control and product sterility monitoring. Projected advantages are:
- Reductions in medication errors
- Labor savings for nursing and pharmacy personnel
- Significant medication-acquisition cost savings
- Prepares bar-coded products for use with bedside scanning devices
- Readily retrievable record of doses prepared
Market acceptance of this product has been strong to date, with 12 devices sold, one installed, and a second undergoing installation.
Bedside Scanning Systems — Don't Underestimate the Impact to Pharmacy
In contrast to pharmacy automated-dispensing systems, the adoption of bar-code technology at the patient's bedside has been adopted by only 1.5 percent of health care facilities. Given a choice of implementing either a computerized order-entry system or a bedside scanning system, many believe bedside scanning to be the easier and most cost-effective choice, and that it will significantly reduce adverse drug events. Why then, hasn't the health care community raced to adopt this technology, and why have nearly 5 times as many hospitals implemented CPOE systems?
There are many issues with bedside scanning technology that the health care industry is trying to resolve. They include the need for wireless networks; the lack of availability of bar-coded, unit-dose medications from the pharmaceutical manufacturers; the inability for pharmacy information systems to produce bar-coded intravenous labels; the lack of electronic interfaces to scanners, reliable handheld scanners, and electronic interfaces to electronic medication documentation systems. Many issues on this list have a negative impact on the pharmacy in terms of the cost of labor, materials, and medications.
For bedside scanning technology to improve patient safety related to medication, blood administration, and laboratory tests, it must function in real time. Therefore, a wireless network is essential. Although wireless technology is readily available, there are still concerns about information security and reliability in addition to the added costs of purchase, installation, and maintenance.
One of the largest barriers to implementing bedside scanning technology is the lack of bar-coded medications available to scan. Only about 30 to 35 percent of unit-dose packaged medications are bar-coded by the manufacturer; this translates into approximately 20 percent of the total medications administered to patients. For the past several years, nearly all professional organizations have requested that pharmaceutical manufacturers voluntarily initiate bar-coded unit-dose packaging, with little result. With minimal action coming from the manufacturing sector, these same organizations have turned to the Food and Drug Administration, asking for mandated bar-coded medications in unit-dose form. The FDA has been slow to respond but is presently drafting legislation. The proposed legislation will have to go through the normal process of allowing for public comment, revisions, legislation, and, finally, permitting the manufacturers time to implement. This process is expected to take two to three more years at least, and it may be longer before the manufacturers can be expected to regularly produce bar-coded medications.
Therefore, to implement bedside scanning technology today, the pharmacy must provide the bar-code on medications. This has been problematic. There are four basic funding needs that the pharmacy must address to provide bar-codes on medications:
-
1. Capital costs for medication packaging machines that provide bar-codes on
labels.
2. Costs for software programs and label printers to place bar-coded labels on non-repackaged medications.
3. Costs associated with modifying different pharmacy information systems to enable bar-coded intravenous medication labels.
4. Increased labor costs (both pharmacist and technician) to repackage the medications.
Not only do these needs add cost to the health system, but they also come at a time when there is an acute pharmacist labor shortage. Faced with these shortages, and the high costs of labor, equipment, and packaging material, pharmacists are hesitant to accept the responsibility for the accurate repackaging of medications. One mistake in the repackaging process could lead to many drug errors before the error is detected. There are some companies, such as Cardinal Health, that offer packaging machines that provide safeguards to avoid these errors, with systems that can be integrated with both centralized robotic and decentralized automated-dispensing machines to further streamline the dispensing process.
Yet, purchasing the best packaging machines and modifying pharmacy information systems to enable bar-coded labels for intravenous medications may not be enough; the pharmacy will still have to extemporaneously bar-code approximately 20 to 30 percent of the remaining medication dosages that are to be administered. The risk of an error occurring during the bar-coding process is much higher, and quality policies and procedures must be developed and followed to minimize this risk.
From our experience implementing bedside scanning technologies, we have discovered several other internal processes that must be addressed for a successful implementation. The first of these is to ensure that the pharmacy enters all medication orders into the pharmacy information system. At times, we have discovered pharmacies that don't enter selected types of medication orders; for example, medications that are provided by a "floor stock" drug distribution system, such as controlled substances. Medications that are not entered into the pharmacy information system cannot appear on the bedside scanning device and therefore, no drug safety checks are available. If the pharmacy is not entering all medication orders, then more labor needs to be added to perform this function.
A second process that needs to be addressed is to ensure the synchronization of drug-dosing times between pharmacy and nursing. In many cases, nurses are administering a drug according to times that are different from the times the pharmacy has scheduled the medication to be administered. This usually occurs because of a time delay between when the order was written and when the pharmacy had time to process it. When these times are not synchronized, nursing receives warnings related to administering the drug too early or too late. This slows the nursing process and creates frustration for both pharmacy and nursing staff.
It is obvious that nurses are impacted by bedside scanning technology since they administer nearly 100 percent of all medications. Nurses find handheld scanners somewhat difficult to use. The text used on handheld scanner screens is restricted in size and amount by the limitations of the scanner screen size. Furthermore, an aging nurse workforce has more difficulty reading the screens. Cardinal Health has developed Pyxis PatientStation™, a product that places a computer at the bedside for bedside scanning and computerized prescriber order entry. Security for the clinical functions is provided by finger-scan technology to approved caregivers.
Although there are costs associated with implementing bedside scanning technologies, it is believed that these costs will be more than offset by the reduction in medication-adverse events, since this technology will prevent many mistakes at the final checkpoint. The cost savings associated with managing one adverse-medication event is projected at $4,700. Thirty-eight percent of medication errors occur at the bedside, and perhaps 60 percent of these could be prevented by this technology.
Computerized Prescriber Order Entry — Improving Patient Care at the Source
Computerized physician order entry (CPOE) with decision support and results reporting is one of the three initiatives supported by the Leapfrog Group to reduce medical errors.3 Evidence exists in the literature that 60 to 70 percent of adverse events can be prevented by such clinical systems. Adoption of CPOE by health care systems has been slow primarily due to design, low rates of physician acceptance, and the cost of these systems. Persuading physicians, who are "volunteer" medical staff, has been difficult, because these systems are not viewed by them as saving time and improving quality. It is self-evident, however, that forces such as the Leapfrog Group's report cards publicly comparing health care systems' quality will keep promoting the adoption of these clinical systems. How will CPOE impact pharmacy services?
A quality CPOE system improves the workflow not only of the pharmacy, but all ancillary departments. A completed CPOE medication order, guided by patient-specific decision support, is immediately transferred electronically to the pharmacy information system. This process eliminates two major sources of medication errors - errors related to paper medication orders not reaching the pharmacy and transcription errors by the pharmacist when handwritten orders are interpreted and input into the pharmacy computer system. This not only improves quality but lowers both pharmacist and technician labor associated with the manual processes. Turn-around-time from prescriber medication ordering with CPOE systems to nurse access to the medication from an ADM, can now be reduced to a few minutes compared to a few hours under a manual system. Therefore, patients obtain needed medications faster.
CPOE systems provide pharmacists and other caregivers with access to the data necessary to monitor patients' medication therapy. Everyone with privileges to access patient information has access to all the needed information present in the results-reporting component. Panic alerts can be established for values outside the range established for specific results, allowing clinical staff to intervene to protect the patient from further decompensation. Monitoring for medication effectiveness is a requirement of the Joint Commission on the Accreditation of Healthcare Organizations and is difficult to perform in manual systems because needed information is located in many different places. With CPOE, it's "one-stop shopping."
The actual process of implementing CPOE will require much work in the pharmacy. Most of it will be related to developing and testing the interface between the CPOE system and the pharmacy information system. When physicians write a drug order, they think in terms of drug, dose, route, and frequency of administration. They do not think, and often don't know, which various products the pharmacy has to choose from to dispense a medication for a specific drug order. Often pharmacists can choose from three or four products. Therefore, specific orders from CPOE have to be mapped to the various drug product choices in the pharmacy information system, and the pharmacist can then select the best product to complete that order.
The Cost Impact of Clinical Systems and Dispensing Automation — Can it Be Measured?
There is a way to measure the cost (labor and equipment) of clinical systems such as CPOE, bedside scanning, and dispensing automation on the medication management system. One parameter that could be calculated to give a true assessment of the costs of implementing these systems is "the total cost of administering a medication." The information needed to calculate this parameter must come from the various sources involved with each step of the medication-use process (prescribing, transcribing, preparation, dispensing, administration, and documentation) and should include equipment costs (lease, purchase, and maintenance), and labor costs in terms of hours per caregiver (pharmacist, pharmacy technician, nurse, and physician) within each of these functions.
The exciting aspect of this approach is that once the baseline measurements are determined for this measurement parameter, only those components of the medication management system that are expected to be eliminated, modified, or added due to the system changes would need to be re-measured. For example, if the decision were made to implement a bedside scanning system, what portions of the medication management system would need to be re-measured to assess the changes to the cost of a administering a medication? In this example, the processes of prescribing, transcribing, and dispensing are not affected by a bedside scanning system, so they would not have to be measured. As shown in Figure 2, only the processes involved with preparation, administration, and documentation would be affected. A management engineer using work-sampling methods can easily accomplish this task.
Knowing these costs, the number of doses administered can usually be obtained from the billing system. The calculation can then be made by applying an average salary for each professional discipline to their respective hours in the process, and the final calculations performed. Of course, quality parameters and other expected outcomes (such as revenue generation from the bedside computer model, reduction in adverse drug events, and so on) could also be monitored to assess the overall impact to patient care and the health care facilities' bottom line.
Conclusion
The impact of clinical systems and automation to a pharmacy department can be both positive and negative related to the costs of labor and equipment. However, adding costs in the pharmacy may have greater returns for other components of the health care system, such as nursing and risk management. Systems can be devised to measure the impact of such technologies to the health care system.
