Opportunities and Challenges Related to Technology in Supporting Optimal Pharmacy Practice Models in Hospitals and Health Systems

Mark H. Siska; Dennis A. Tribble

Am J Health Syst Pharm. 2011;68(12):1116-1126.

Current State of Medication-use-supporting Technologies

The current state of core medication-use-supporting technologies is best understood by reviewing HIT adoption and integration efforts within the framework of an EHR. The core medication-management-supporting systems representing the essential HIT infrastructure the IOM envisioned include

CPOE and clinical decision-support systems,
Pharmacy systems,
Medication reconciliation systems, and
Medication administration systems.^[1–3]

The overall deployment of CPOE and clinical decision-support systems in U.S. hospitals after nearly 10 years of effort is roughly 15%, with greater penetration occurring in hospitals with more than 200 beds, according to 2007 and 2009 ASHP surveys.^[15,16] A more-recent 2010 report found similar results, with only 14% of hospitals achieving the expected stage one meaningful use requirements; in community hospitals with fewer than 200 beds, the figure was less than 12%.^[17]

The adoption of CPOE or e-prescribing systems in ambulatory care is considerably greater than what has been reported in hospitals, according to a Surescripts 2009 National Progress Report on e-prescribing.^[18] Superscripts, which operates the nation's largest pharmacy health information exchange network, reported that one in four office-based physicians is using e-prescribing systems, nearly 70% of whom use an e-prescribing application within their EHR. The report also projected a steeper rate of growth over the past three years for CPOE adoption in ambulatory care as compared with the inpatient environment.

Purely electronic and integrated medication reconciliation systems are virtually absent from health care organizations today, with only 10.4% of hospitals reporting an entirely electronic process for medication reconciliation.^[15] Most facilities are still reengineering their manual processes in response to The Joint Commission's National Patient Safety Goals and have yet to consider what the ideal functional requirements might be for a multidisciplinary integrated HIT solution.

A 2007 ASHP survey found that nearly all hospitals had implemented a pharmacy information management system; however, only half of them had systems that could inter-operate within the framework of an EHR and other core medication-use systems, a clear indication that pharmacy system integration is still a work in progress.^[15] The remaining half of hospitals still utilized legacy pharmacy systems that were limited in their ability to connect or integrate with ordering and medication reconciliation systems, creating standalone medication management solutions where continuity and communication play an important role in improving medication safety. The majority of hospitals in 2007 had adopted medication distribution and storage technologies (83%) as well as repackaging automation (92%) for unit dose medications.^[15]

Despite significant progress with implementing medication administration systems, including electronic medication administration records, bar-code-assisted medication administration technology, and smart pumps, the integration of these technologies remains low, with only 20% of hospitals having all three systems in place.^[15] Integration of medication administration systems with core medication management technologies including the EHR is limited.

Data on overall EHR usage in U.S. hospitals has suggested levels of adoption ranging from 5% to 60% depending on the definition of EHR.^[19] A study commissioned by the ONC for Health Information Technology found that 8–12% of U.S. hospitals have a basic EHR and only 2% have a system that meets the federal government's meaningful use criteria.^[19] Larger hospitals, urban hospitals, and teaching hospitals were more likely to have electronic records systems.

1 2 3 4 5 6 7 8

Next Section

Cite this: Opportunities and Challenges Related to Technology in Supporting Optimal Pharmacy Practice Models in Hospitals and Health Systems - Medscape - Jun 15, 2011.

Abstract and Introduction
Current State of Medication-use-supporting Technologies
Current Barriers and Challenges
Current State of HIT Quality and Safety Claims
Summary of the Current Context for HIT
Overcoming HIT Challenges
The Preferred Future State of Pharmacy Information Technology
Conclusion
References

Table 1. Preferred Future State of Pharmacy Information Technology: Gap Analysis^a
Future State	Current State	Gap Analysis
Clinical Medication Management
Patient information is contained in a single, addressable longitudinal EHR with a format that can be electronically scanned for actionable situations.	Current EHR deployment is relatively small (less than 20% of institutions) and there are no standards by which the data within those EHR products are represented. Development and standardization of tools that survey them is therefore difficult.	A standardized, best-practice EHR (or at least a view) needs to be developed to permit appropriate implementations of decision support and analysis tools.
Terminology and key metrics are standardized to permit interoperability of data between systems.	Systems are highly proprietary, as are the tools that provide analysis.	Standards for data representation are needed to permit standardized tools to provide decision support.
Operational systems are derived from clinical systems and are constructed in such a way as to drive behavior around clinical care, and not around products.	Pharmacy systems are product-driven with data architectures that are better oriented to product distribution than clinical management.	Clinically oriented systems are needed that can also derive the information necessary to drive drug product selection, preparation and distribution.
Decision-support systems are present that can maintain appropriate context and apply the correct actions to the right person at the right time in the right context.	Current decision support systems are almost exclusively limited to alerts at order entry and operate independent of clinical context. These systems require considerable maintenance and customization at each site.	A richer data set based on best practices that can use clinical information as context to refine systems is needed, as is CDSSs that are more than order-entry alerts. Some standards that permit sharing and creation of best practices are needed to enable more-rapid deployment and more appropriate use.
Provides on-line access to standard references as well as known literature.	While these databases are available, few systems currently take advantage of them and, when they do, they primarily provide on-line representations of published references.	Databases that incorporate contextually relevant links to reference literature are needed so that contextually relevant information can be accessed.
Permits oversight of the medication distribution system.	Current operational systems have little audit trailing and virtually no clinical context with which to evaluate the medication use process.	Operational systems need to be able to trace medication use processes both through the pharmacy distribution process and the clinical medication use process.
Management of Medication Preparation and Distribution
Medication products are marked with a machine-readable code that reliably identifies medication products in terms of their clinically-relevant contents.	Pharmacies currently spend considerable manpower and time maintaining a current list of bar codes/NDCs so that automated systems can identify the products.	Central, reliable database of known codes that systems can query is needed.Database needs to support identification down to the therapeutic entities and amounts in products.
Medication products are packaged in containers that are "automation-friendly" to permit the application of robotics to the selection, preparation and distribution of medication products.	Pharmacies currently spend considerable time and effort packaging doses for machine manipulation. Robotics for i.v.'s have speed issues directly related to the nature of current i.v. containers.	Containers that are easily manipulated by robotics, with labeling that is easily and reliably read, are needed. Containers need to facilitate speedy robotic manipulation.
Medication selection and preparation activities are processed to the extent possible by robotic devices with extensive in-process controls to assure correct and consistent processing steps.	The application of robotics is primarily limited to compounding processes and unit-dose cart filling and tends to stand alone. Robotic devices continue to be expensive to install, manage, and maintain. They tend to be large, multi-purpose devices that attempt to provide a variety of services that, like a Swiss Army Knife, do many things but none of them well or quickly.	Robotic technologies will continue to mature and improve in speed and reliability; integration is needed to further remove reliance on human performance. It remains to be seen whether these devices remain large, multi-functional products or become a collection of tools from which those operating the dispensing function select the most appropriate one.
Where medication selection and preparation activities must be performed by hand, they are governed by processes that provide in-process controls to minimize the opportunities for error. This may include automatic identification of selected products, vision systems that detect appropriate product selection or fluid product measurements, vision systems that can detect the appearance of inappropriate inclusions (e.g., particulates) in liquid doses, intelligent scheduling of activities to ensure that medication doses will be effective (e.g., not expired) when they are finally administered, enforcing use of appropriate containers, etc.	Work flow software is available from a few vendors for the sterile compounding space; more is needed for other manual dispensing activities.	A well-articulated vision for the role of robotics in medication selection, preparation, and delivery is needed, as is the development of technologies to better automate checking activities including smart vision systems for detecting defects and measurement errors.
Medication selection, preparation, and distribution systems that capture significant amounts of data about these processes that can be used to both audit the process and drive various analytical processes and metrics regarding the effectiveness of these processes.	Currently available devices capture extensive logs, but the integration of the data from those logs to view an entire medication use process is not currently practical. Manually performed processes tend to remain largely undocumented, except for some emerging work flow management products.	Metrics need to be defined and devices need to develop support for, and reporting of these metrics.
Medication delivery systems that trace the course of any medication supply to its current location.	There are a few product tracking products on the market as standalone products; integration with clinical systems is poor.	These features will need to become integrated into clinical systems for full traceability.
Medication selection, preparation and delivery systems that limit activities to those users who are trained and, where appropriate, credentialed to perform those activities.	Most systems currently available offer permission-based control of access to critical functions.	These permission-based approaches will need to be continued.
Medication selection, preparation and delivery systems that capture logging data as a byproduct of the performance of the required duties (as opposed to becoming steps someone must remember to do).	Most automated systems do this currently; large amounts of data are captured as a by-product of the normal operation of the devices.	This approach will need to be continued as more automation devices are developed.
Informatics
Informatics practice must embrace data models that drive clinical practice as a primary requirement and distribution practice as a benefit of that clinical knowledge. Informatics practice must start from a well-articulated statement of best practices to be enabled and enforced within the informatics infrastructure.	Informatics tends to accept current offerings as a base point and to request incremental changes to those features. Informatics does not envision what systems should look like unconstrained by "what is" with the result that there is no clear vision of "what should be." Dissatisfaction with current commercial offerings abounds, but there seems to be little consensus regarding what should replace it.	Informatics must develop an ideal operational model for the application of technology, and articulate that vision as a set of requirements to the industry. Informatics must drive requirements as opposed to awaiting the next gift from industry. Focus must move away from incremental modification of "what is" to consideration of "what should be".
Informatics as a practice will know as much or more about the realities of work flow as the providers who participate in it, and will constantly and critically look for ways to apply appropriate automation tools to improve the safety, efficacy, and efficiency of those processes.	Practice is largely uncharacterized in any meaningful way. Informatics tends to layer automation on top of current processes without considering how practice should change.	Informatics needs to be developed as a subspecialty within pharmacy practice that develops new knowledge about both clinical and distributive practice, and guides the development of new technologies that meet the current and future needs of the profession.
The practice must develop a data model for representing drug entities that is designed primarily for clinical purposes but also generates data necessary for drug distribution.	Current formulary data structures are primarily aimed at inventory management and product distribution. This organization often confounds clinical decision-making, which ordinarily occurs at the level of therapeutic entities (drugs) rather than their commercial presentation (products).	One such structure might maintain the concept of a drug as an entity that can be used for specific therapeutic purposes at one level while recognizing that those drugs must be delivered using commercially available or pharmacy-prepared products. Therapeutic decision support would be based on these drug entities and drive the selection and use of specific products that contain these entities.
Therapeutic formularies must operate in the context of a national (if not international) encoding scheme for drugs and products that supports automated maintenance of new drugs and products without requiring the validation of each new product as it comes into the facility. This implies that a coding system (such as the NDC) would be maintained as a nationally available, reliably current data set whose contents were set by an authoritative and neutral body from which any system could reliably identify not only the description of the product for supply chain purposes, but also enough clinical information to determine whether or not that product was suitable to fulfill the requirements of any particular medication order.	Globally, coding is highly fragmented both in structure and data content. Within United States, the NDC represents this coding with a variety of problems: There is no authoritative list of codes and the products to which they refer. The FDA database is updated every 6 months and is not heavily policed. There are multiple permitted formats and code lengths without reliable crosswalks between them. Only the vendor identifier portion of the code is controlled, one cannot look at an NDC and tell from the NDC what drug it represents. Generation of the drug and package code in the NDC is completely at the whim of the vendor and the decision to invent a drug code verus a package code is not uniformly applied by all vendors. NDCs are allowed to be reused, defeating their usefulness in longitudinal health records.	There is a need for a "smarter" numbering system than currently exists, and one whose contents are more centrally controlled than the current NDC. The coding of any particular product must be reliably unique within the longitudinal electronic health record (e.g., codes cannot be reused). The coding should permit a software system to identify the contents of the package irrespective of vendor (e.g., the drug code for a cefazolin 1 gm vial should be the same irrespective of who made it).

^a EHR = electronic health record, CDSS = clinical decision-support system, NDC = National Drug Code.

Comments

Commenting is limited to medical professionals. To comment please Log-in.

Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.