The challenges involved in interfacing pharmacy systems with core clinical information systems are propelling IT leadership to seek an enterprise approach and move towards integration.
The demands of meaningful use under the federal American Reinvestment and Recovery Act/Health Information Technology for Economic and Clinical Health (ARRA-HITECH) Act for transforming clinical data into usable information are compelling CIOs and other healthcare IT leaders towards a more integrated, enterprise-wide approach to clinical information systems.
IF YOU DON'T HAVE A PHARMACY SYSTEM THAT IS FULLY TIED INTO THAT [ENTERPRISE SYSTEM], THAT PHARMACY SYSTEM CANNOT TAKE ADVANTAGE OF THE CLINICAL ALERTS THAT WOULD FIRE, SO YOU'D FIND YOURSELF HAVING TO INTERFACE ALERTS. -RICK SCHOOLER
This is especially true around the issue of pharmacy systems. Increasingly, the demands involved in interfacing standalone pharmacy systems with core EMRs-interfacing two separate systems, separate formularies, and separate allergy lists-are becoming unsupportable for the IT leaders of many hospital-based organizations. In recent years, according to a 2009 report from the Orem, Utah-based KLAS Research, patient safety concerns and frustration with system interfacing have motivated healthcare organizations to integrate their pharmacy systems with their core clinical solutions.
INTERFACING DISPARATE SYSTEMS
Rick Schooler, vice president and CIO of the 1,780-bed Orlando Health, knows the difficult challenges involved in interfacing pharmacy and core clinical systems. His hospital is four months away from replacing its standalone McKesson (San Francisco) pharmacy IS that has been in place for 15 years, with the Allscripts Sunrise pharmacy component from the Chicago-based Allscripts, to match the health system's core Allscripts EHR. After signing the contract eight years ago for Eclipsys Sunrise Pharmacy, Schooler says his organization has been waiting for Allscripts (formerly Eclipsys) to build the product so Orlando Health can make the final switch. Eclipsys's pharmacy unit has been dogged with being seen as the least integrated on the market and as a weak link in its offering, according to 2009 KLAS report “Pharmacy Information Systems: In the Age of Integration.”
Many difficulties can result from integrating pharmacy with core clinical systems. Not only can an interface “introduce a stop in the information flow” if and when it goes down, according to Schooler, standalone pharmacy systems create duplicative work for clinicians having to enter the same information into two systems. Schooler also notes that maintaining two different formularies with drugs in one that are not in the other can be problematic, as well as interfacing drug-food, drug-drug, and drug-allergy alerts. “If you don't have a pharmacy system that is fully tied into that [enterprise system], that pharmacy system cannot take advantage of the clinical alerts that would fire, so you'd find yourself having to interface alerts,” he says.
[INTERFACING] CAN BE DONE EFFECTIVELY, BUT IT'S VERY TIME CONSUMING, AND I THINK IT'S FRAUGHT WITH A LOT OF OPPORTUNITIES FOR ERRORS OR ISSUES IN KEEPING THOSE SYSTEMS EXACTLY IN SYNC. -CINDY WILLIAMS, R.PH
Furthermore, Schooler says, CIOs need to take a real “buyer beware” approach when it comes to balancing out the natural desire to select a particular vendor product based on feature-function characteristics with the need to integrate systems. “For all the feature-function you think you're going to get in the pharmacy department [when purchasing a standalone pharmacy IS], you're going to now require a lot of interfaces of workflow that you won't need if you're integrated,” he says.
Jeffrey Firlik, R.Ph, principal consultant with the Health Delivery Group at the Falls Church, Va.-based CSC, says interfacing systems can be done, but it is a big challenge that might not always go smoothly. He remembers one assessment he did three years ago on an institution pondering whether to switch to an integrated system, which emphasized how cumbersome interfacing systems can be. In this organization's pharmacy, pharmacists had two different computers on their desks, one that accessed the pharmacy system, and one that ran the EHR. “Even though those orders were supposed to flow back and forth, they just didn't,” Firlik says.
“[Interfacing] can be done effectively, but it's very time consuming, and I think it's fraught with a lot of opportunities for errors or issues in keeping those systems exactly in sync,” says Cindy Williams, R.Ph, system director of pharmacy at the five-hospital Newport News, Va.-based Riverside Health System. She cites one opportunity for error when it comes to accessing different vendors' drug databases, which don't always contain the same drug information-for example, the Kansas City-based Cerner Corp.'s Multum drug database, versus that of First DataBank (which Siemens and other vendors use).
Beyond limiting the amount of interfacing, implementing an integrated pharmacy system can provide an added measure of patient safety. When Orlando Health's Schooler was quizzing pharmacists on his informatics team, they all cited the possibility for data entry errors when there are separate systems for ordering and prescribing. What can happen often is interfaces don't gel 100 percent, and additional data entry is needed at the ordering and verification phases, “which could open up opportunities for inconsistent information,” says Schooler. He notes the benefit of having one complete record of the patient that allows pharmacists to know what was ordered and dispensed. He also points out that nurses can act on a change in a physician's note, and unless the pharmacist has access to that clinician point-of-care system, the pharmacist won't be able to see how the drug was administered to the patient.
Firlik brings up another patient safety point: having integrated systems unites drug order catalogs, which gives the pharmacy more control to influence prescribing and improving outcomes. “The pharmacy can influence formulary compliance, and use more drug use evaluation data to influence prescribing through order sets and algorithms,” Firlik says. This helps create a closed loop medication strategy, which is enhanced with an integrated system. “The two greatest places where errors occur are ordering and administration,” he adds. “With the barcode medication administration, it can reduce administration errors, so patient safety is huge.” Firlik gives the example of both pharmacists and prescribers taking advantage of built-in closed looped lab alerts for drugs to be aware of any dosing, allergy, or drug-interaction problems.
Having an integrated pharmacy/core clinical system also introduces many efficiencies for the pharmacists' daily activities, according to those interviewed. Schooler mentions that not having data in one consistent place makes decisions more difficult when a pharmacist has to check two separate systems to seek information. He also knows that many healthcare organizations' pharmacy departments are “staffed to bone,” so 30 seconds here or a minute there checking multiple sources for relevant information really takes time away from lean pharmacy departments.
Williams' organization, the Riverside Health System, has become a reference site for customers of the Malvern, Pa.-based Siemens' Soarian. In 2003 it replaced the old Siemens Invision pharmacy unit with the newer Siemens Pharmacy. But in doing so, Riverside also contemplated the Cerner Millennium PharmNet product, as it had both laboratory and radiology products from the company. After doing its due diligence, Riverside decided to take an enterprise approach with Siemens. “I still remember the day we really committed to aligning with the Siemens Soarian platform, and that would be the path we would take to achieve CPOE [computerized physician order entry],” Williams says.
In the last 12 to 18 months, Williams has observed other Soarian customers, who have spent time at Riverside, dealing with the difficulties of interfacing systems. “I think that once folks look at the reality of CPOE and meaningful use, I think it's really driving people towards an integrated system,” says Williams. “My impression is that you're seeing more and more organizations really moving toward an enterprise selection.”
However, no integration is perfect. Johnanne Ross, Pharm.D, director of pharmacy IT automation at the 20-hospital University of Pittsburgh Medical Center (UPMC) health system, notes that the Cerner Millennium PharmNet system, which is linked with the hospital's Cerner EHR, is a separate application that requires clinicians to have to push a button that flips over to other applications. Lab results and historical medications do flow into the system, but when a patient is taken to a temporary location, for instance after surgery, a pharmacist can't see that patient information.
IF WE LOOK TO MEANINGFUL USE TO ADD MORE AND MORE REQUIREMENTS, THE MORE YOU CAN HAVE THAT INFORMATION INTEGRATED WITHIN A COMMON DATABASE SO IT CAN CLEARLY TALK TO ITSELF, VERSUS TALKING A DIFFERENT LANGUAGE TO THE OUTSIDE [THE BETTER]. - CINDY WILLIAMS, R.PH
From what Ross hears from designers, the system will become more integrated, so she can have more of a prescriber perspective within the pharmacy. “There are so many things a pharmacy can get by the medications being so integrated, so I don't see so many standalone [systems] in the future,” Ross says. “But it just depends on what phase that institution is at and what's developed that people can buy.”
MEANINGFUL FOR MU
As many of the requirements for meaningful use coming out of the ARRA-HITECH Act focus on medication orders, drug-drug checking, and maintaining active medication lists, industry experts agree that an integrated system makes those tasks much simpler.
“I think, especially with all the requirements, there will be more people going to an enterprise model,” says UPMC's Ross. “Some of the vendors offer to do suites of builds to make it easier because it is very difficult to build these systems.”
ePrescribing: Biometric Authentication
Today, Ohio is the only state currently requiring two-factor authentication for electronic prescribing. This edict from the Ohio State Board of Pharmacy spurred one organization, the eight-hospital, Columbus-based OhioHealth to embark on a system-wide project to streamline the authentication process with biometric identification.
Currently, 220 physicians at 35 OhioHealth practices use the Advanced ePrescribing solution for the GE Centricity Electronic Medical Record (EMR) (from the Barrington, III.-based GE Healthcare) on an outpatient basis. When the prescriber chooses to send the prescription electronically the solution presents a pop-up window initiating a finger scan. On the inpatient side, physicians use a phone key code authentication process within its EHR (from the San Francisco-based McKesson), provided by Imprivata (Lexington, Mass.).
In complying with the requirements of the Health Insurance Portability and Accountability Act (HIPAA), OhioHealth had been down the two-factor authentication road before by using proximity ID badges and RSA security tokens. “Most of our physicians found them cumbersome and unpredictable,” says Mrunal Shah, M.D., vice president, Physician Technology Services, OhioHealth Information Services. “We didn't think that was the future of technology, and quite frankly biometric access and single sign-on is where we think everybody is going; so we've made that investment throughout our entire organization,” he says.
OhioHealth began implementation of its inpatient biometric authentication in 2008 as the 73-bed Dublin Methodist Hospital, the system's new showcase digital facility, was going live on CPOE. After a few stuttering stops in 2007 with ambiguities in the Ohio Board of Pharmacy mandate, the project restarted in earnest in 2009 and was approved in the summer of 2010. Now, about 500 physicians use order entry and advanced ePrescribing on the inpatient side.
With the federal Drug Enforcement Administration (DEA) regulation that went into effect on June 1, 2010, aimed at enabling secure electronic prescribing of Schedule II-V Controlled Substances (narcotics), Shah believes more organizations will be moving toward biometric authentication. “With the DEA requirements coming through, having dual processes in any environment will be very challenging,” he says. “So, I have a feeling that most states will start to adopt the same level of positive identification requirements as the DEA, so they're consistent. I do feel that will be far more widespread.”
Despite the challenges involved in clinician training and re-outfitting physician offices with new USB readers or keyboards with integrated readers, the benefits of biometric authentication have been keenly felt by OhioHealth. Shah says that his organization is inherently conservative and likes to implement security in all its initiatives. “It future orients us towards any other positive identification requirements that may come down the line, perhaps if there's a new requirement for all lab studies to require positive identification,” he adds. “The beauty is the technology has been built, and the hardware is there. It's just a matter of turning it on and training our users.”
In the future, Shah wants to implement a more robust phone or biometric solution to allow physicians to ePrescribe from a remote access platform.
Just recently, Genesis HealthCare, based in Zanesville Ohio, started implementing the Wall, N.J.-based BIO-key's biometric identification software with Lumidigm (Alburqurque, N.M.) fingerprint sensors so staff can quickly and securely establish identity when ordering, verifying, or administering medications.
THE TWO GREATEST PLACES WHERE ERRORS OCCUR ARE ORDERING AND ADMINISTRATION. WITH THE BARCODE MEDICATION ADMINISTRATION, IT CAN REDUCE ADMINISTRATION ERRORS, SO PATIENT SAFETY IS HUGE. -JEFFREY FIRLIK, R.PH
Ross gives many examples of how her organization's integrated system allows her staff to meaningfully use records, whether it is for drug alerts or order sets. As she's worked with both standalone and integrated pharmacy systems, Ross notes that with an integrated system she can find data that she couldn't get at before. “If a physician puts in a problem that the patient is pregnant, it's much easier for me to set off pregnancy alerts when certain medications are ordered,” she says. “You can't do that necessarily in a pharmacy system if you're not getting that kind of lab data or problem list data into that system.”
She also notes that if the pharmacy unit is part of the EMR, then all the labs, radiology, medication orders, and nursing tasks for patient care can be used to make the order sets much more robust, and clinical information updates can be set up much easier.
A further synergy that is made from UPMC's integrated pharmacy system is its ability to download order sets from its EMR. The ePractice team, which is made up of multi-disciplinary representatives from each hospital, gathers weekly to review data and create new order sets. Then every Sunday night UPMC downloads the new order sets from the EMR so clinicians can search, find, and use them.
Riverside Health's Williams believes that with the latter stages of meaningful use requiring more integration and codified data to flow within healthcare organizations, an integrated system will make it easier to obtain meaningful use. It's also much easier to control data elements, she adds. “If we look to meaningful use to add more and more requirements, the more you can have that information integrated within a common database so it can clearly talk to itself, versus talking a different language to the outside [the better],” says Williams.
In addition, certification for meaningful use might be easier by having a pharmacy system from one's core clinical vendor. In a previous interview with this magazine, Pamela McNutt, senior vice president and CIO at Methodist Health System in Dallas, and chair of CHIME's Policy Steering Committee, noted there were a few options if organizations do not pursue a single-source clinical IT strategy. They can ask their core-clinical vendor to seek certification for their core clinical product, minus use of the pharmacy component (something no vendor so far has pursued on behalf of its customer organizations), and not use the pharmacy component; can seek separate certification for their separate pharmacy system; can pursue self-certification separately for the pharmacy system; or can “rip and replace” and dump their pharmacy system altogether and replace it with their core-clinical vendor's pharmacy system.
MOVING TOWARDS INTEGRATION
Many steps need to be thought through before a standalone pharmacy system is replaced with an enterprise system. All who were interviewed for this story agreed that pharmacy IT leadership needs to be involved from the start. “The most complicated and clearly the most dangerous piece of implementing CPOE is the medications,” says Williams. “Having a very active pharmacy involvement, not only in the selection, but in the planning and implementation [helps], because otherwise it will be very difficult.”
Orlando Health's Schooler agrees that pharmacy leadership needs to buy into the project, as the CIO can't manage clinician resistance and the transition by himself. “There is a steering committee for every project we do, the appropriate sign-offs and designs are done to ensure that we are designing the system properly and people get their preferences made,” he says.
Since this type of project transitions pharmacy to be part of a larger platform-the organization's EMR-pharmacist voices need to be heard. “If those resources [in the pharmacy] don't report into CIO, then there's going to be some resistance to moving toward an integrated platform,” Schooler adds. He also advises that this type of project needs to be adequately resourced and funded, while not being planned in the middle of other large-scale initiatives.
If the selection of the system happens in concert with choosing the EMR, then making sure that the pharmacy system is equally as robust as the electronic medical record is key. CSC's Firlik notes that the organization also needs to be mindful of regulatory issues regarding medications and administration validation. He gives an example of one state, Ohio, which has a strong rule about two-factor authentication for medication orders (see sidebar on p. 28).
Firlik recommends that IT leadership think about a pharmacy system from a financial point of view, as well as a clinical perspective. “The drug budget at a hospital is huge. If they can have a system that can optimize charging, billing, and inventory management that will be optimal as well,” he says.
Ross says the institution needs to look at the strength of the pharmacy system, especially the medication process, before purchasing and how flexible is it going to be able to meet the organization's changing practice patterns. “How robust are they, what kind of sites do they support now? How integrated and safe is the medication process now?” she asks.
Ross says that well before UPMC transitioned to the PharmNet system, the staff looked through formularies and what they called the “big 10,” which were high process, high risk medication protocols to understand how they were dispensed, administered, and monitored, so that they could be correctly implemented in the new system. Then internal pharmacy processes and workflows were mapped out to figure out what needed to be changed in the new system, and discussions were had to address any gaps. When CPOE was implemented, design sessions were held for everyone, and the IT staff had weekly follow-ups and on-site meetings.
“[It's important to] recognize that it is critical for the pharmacy and the CPOE to be completely in sync because there is so much interplay between those two and because the meds are so complicated,” Williams concludes.
Healthcare Informatics 2011 April;28(4):24-30