Computerization of the medication circuit: a strategic lever for quality of care and hospital rationalization

The "medication circuit" within healthcare facilities represents a set of complex and interdependent processes that are essential for patient care. Optimizing it is a major challenge because it simultaneously aims to improve the quality of care and streamline the logistics and economics of hospital structures. Faced with this complexity, the computerization of this circuit has become a preferred way to modernize practices and ensure better safety for patients.

Historically, the medication circuit in hospitals and clinics has functioned heterogeneously, often raising questions about the quality of care, with analyses reporting discrepancies of up to 10% or 20% between prescription and actual administration. The Directorate of Hospitalization and Healthcare Organization (DHOS) of the Ministry of Employment and Solidarity thus initiated an in-depth study to promote reflection on good practices in the field of computerization. This study, based on the "value analysis" method, aimed to measure the qualitative and quantitative returns expected by the various stakeholders concerned, whether computerization is total or partial. This article explores the key concepts, economic issues, types of projects that can be considered, their gains and costs, based on the conclusions of this fundamental study.

The context and challenges of computerizing the medication circuit (Mandatory HAS criterion 3.6.2)

The medication circuit encompasses a multitude of steps: prescription, its analysis and validation, preparation, delivery, distribution and administration of the medication, as well as orders by the pharmacy, analysis of activity, and management of expirations and batch withdrawals. These processes require interoperable and communicating hospital information systems for effective cooperation.

The computerization of this circuit is much more than a simple automation of existing procedures. It implies an indispensable reflection on the organizations to be put in place. For approximately five years before the publication of this study (May 2001), multidisciplinary teams—composed of hospital directors, physicians, pharmacists, caregivers, and computer scientists—worked to define the conditions for the success of this computerization, notably by formulating recommendations for the adaptation of hospital information systems. The objective was to have concrete and quantified elements to enlighten decision-makers on the costs and the envisaged gains.

The report from this study is intended for heads of hospitals, information system managers, prescribing physicians, hospital pharmacists, and healthcare professionals. It is intended as a useful tool for initiating reflection on the objectives and modalities of computerization, offering an original methodological approach to the return on investment of this approach, whose main purpose is to ensure the safety of patients.

The value analysis method

The study was based on the "value analysis" method, a technique frequently used in industry for the renewal of products or services, and which has been adapted to administrative and software processes. According to standard X50.150, value analysis is a method of competitiveness, organized and creative, aimed at satisfying the user's need through a specific approach to design that is functional, economic and multidisciplinary. It is used to help imagine and choose solutions, with a dual purpose.

The approach is:

• Functional: It expresses the action of a product or its components in terms of purpose, without reference to solutions.
• Economic: It requires numerical quantities to quantitatively assess the "value." Value is a judgment on the product based on user expectations, increasing with the satisfaction of the need or the decrease in expenses.
• Multidisciplinary: It relies on working groups including operational services, specialized facilitators, and decision-makers.

The study was conducted with the active involvement of three hospitals (a university hospital center, a general hospital center, and a specialized psychiatric hospital center). A multidisciplinary working group, including directors, administrative staff, IT specialists, physicians, pharmacists, and caregivers, was mobilized.

The work process took place in three phases:

1. Phase 1: Identification of functions, levels and issues.
2. Phase 2: Assembly and analysis of solutions.
3. Phase 3: Synthesis of lessons learned.

A detailed functions repository was used to structure the analysis.

Understanding the medication circuit: physical flow and information flow

For a comprehensive understanding, it is important to distinguish between two main circuits within the medication process:

• The physical drug circuit : This concerns the "material flow," that is, the actual movement of the medication. It includes three main stages:
  • The delivery of ordered medications to the pharmacy and their storage.
  • The dispensing of medications in the departments, including preparation and transport.
  • Administering medication to patients, including managing the return of unadministered medications. The physical circuit has two main variations: global distribution (from the pharmacy to the departments) and individual dispensing.
• The drug management information circuit : It concerns the way information circulates around the physical flow. The information system is divided into two subsystems, relatively independent, with a potential interface between them:
  • A system focused on pharmacy drug ordering and storage, often linked to the hospital's economic and financial management system (GEF).
  • A system focused on prescription management and the medication circuit between the pharmacy and the wards, using a specialized product or a ward management tool.

It is important to clarify the terminology:

• Dispensing is defined as all tasks including the analysis, validation, preparation, and delivery of medications.
• Distribution only covers the physical phases of preparation and delivery of medications. A key point is that dispensing is only conceivable if prescription information is known. While distribution can be globalized or nominative, dispensing, if nominative, most likely implies knowledge of prescriptions.

The economic stakes of computerization

The computerization of the medication circuit generates initial expenses (software licenses, hardware, network, services, training, configuration, maintenance). In return, it is supposed to bring improvements, some being "qualitative" and others "quantitative" (or "economic stakes") because they can be quantified.

The study identified three main types of economic stakes:

1. Time savings, or productivity gains, through efficiency in the processing and handling of information. These gains concern several populations:
   • Pharmacy staff: Particularly through the calculation of drug interactions. Without a computerized system, comprehensive control of these interactions would be virtually impossible, requiring the equivalent of several full-time pharmacists in each hospital.
   • Prescribers: Projects based on prescribers entering prescriptions may require an initial investment of their time, but certain system functions may allow them to regain it later.
   • Nurses: Thanks to the elimination of copying and the continuous availability of clear and complete information, they save time in the administration of medications.
   Orders of magnitude of time spent on drug-related tasks have been noted:
   • Up to 50% of nurses' time in non-computerized departments is spent on information processing and handling tasks.
   • Each prescriber spends between 30 minutes and 1 hour per day prescribing medication.
   • In a 20-bed care unit, medication preparation can take up to two hours per day for a team of two nurses.
   • In a 500-bed hospital with individual drug dispensing, the transfer of workload from nurses to pharmacy technicians can represent 15 to 25 Full-Time Equivalents (FTEs).
2. Lower drug costs. Computerization projects can contribute to this reduction in two ways:
   • By promoting the use of less expensive options for a similar volume of consumption.
   • By acting on the overall volume of drug consumption. Decreases in drug expenditure of the order of 10% to 20% have been observed depending on the hospitals, departments and their initial organization. The major principle of this reduction is based on delivering to the departments the drugs that they have actually prescribed, which is made possible by the knowledge of the prescriptions by the pharmacy, whether by individual dispensing (DJIN) or global distribution.
3. The economic consequences linked to the improved quality of care provided to the patient, through the avoidance of incidents and errors. It is essential to distinguish between "medication errors" (prescription, transcription, dispensing, administration errors) and "medication incidents" (adverse events resulting from an error). Computer systems aim to reduce errors and, consequently, incidents, either by preventing them or by intercepting them before they cause an adverse event. The economic consequences are linked to incidents, not errors, although the interception of errors generates a consumption of time. This phenomenon is neither zero nor negligible: a report from the Medicines Agency in 1997 indicates that 10% of hospitalized patients experience an undesirable drug effect, a third of which are serious (death, life-threatening, prolonged hospitalization). The economic consequences of serious incidents include additional days of hospitalization, additional examinations, stays in intensive care, overconsumption of medication, and lost working days for the patient. Studies have estimated the average cost of a serious medication incident between $2,000 and $2,500 (14,000 to 18,000 F) in the United States, and up to 50,000 F for cases treated in intensive care in France. A study in gastroenterology in patients over 65 years of age showed an average cost of 20,602 F per patient for drug poisoning, 12.6% of which was due to interactions. These figures are minimums because they do not take into account lifelong sequelae.

The study applied these concepts to a fictional 800-bed hospital (including 500 short-stay beds), with a budget of 500 million francs/year and a payroll of 350 million francs/year. The orders of magnitude of the annual economic stakes are as follows:

• Time spent on the medication circuit: Approximately 10% of the hospital's energy is dedicated to these tasks, representing a 35 MF/year stake.
• Medication expenditure: This represents approximately 6% of the budget, or 30 MF/year.
• Consequences of serious incidents: For 25,000 admissions per year, approximately 500 incidents occur, including 167 serious ones (1/3 of the incidents). The economic stake is estimated to be between 1.5 and 2 MF/year (based on 10,000 F per preventable serious incident).

To assess the overall economic impact (E1, E2, E3), the study proposes to collect hospital-specific data regarding activity (number of beds, days, admissions, ALOS, budget), time spent on the medication circuit by pharmacy staff, prescribers, and nurses (number of people, payrolls, proportion of time dedicated), medication expenses (purchase budget, distribution), and the consequences of incidents (percentage of patients affected, severity, average cost). Only a portion of these issues is likely to be gained by a specific computerization project.

Types of computerization projects and their benefits

The study classifies computerization projects into different categories, depending on their scope and sophistication, and details the specific gains associated with each.

Projects around pharmacy management (P Projects)

These projects focus on the cycle between the internal management of the pharmacy and external suppliers, including supply, batch recall management, and scope management.

• P1: Renewal of the hospital's economic and financial management system (EFMS).
  • Economic stakes: The major gains are not intrinsic to this project, unless the new system offers significant productivity gains for pharmacy staff or very high-performance specialized modules (discussed in P4). Savings on IT maintenance costs may also be observed.
• P2: Management of lot withdrawals.
  • Economic stakes: A productivity gain is conceivable if the pharmacy can selectively alert the departments concerned via internal messaging, without having to manually manage batch numbers (a tedious task).
  • Valuing gains: The evaluation is done by measuring the workload required to alert all services vs. alerting selectively, and the time saved by not managing batch numbers for all drugs.
• P3: Management of expiration dates.
  • Economic stakes: The goal is to reduce the annual budget for expired products, which can reach tens of thousands of francs. Computerizing the management of expiration dates is one way, but other organizational actions such as low stocks, rapid turnover (FIFO), and the elimination of "wild stocks" in departments are also effective.
• P4: Contribution of additional functions to the GEF system.
  • Economic challenges: These projects bring productivity gains for pharmacy staff. Examples:
    • Barcode readers: Facilitate the management of stock entries and exits, and inventory.
    • Order recommendation: The system calculates stock levels and suggests orders in the GEF system, reducing the time to create and enter orders.
    • Receipt of invoices by EDI (Electronic Data Interchange): Avoids manual re-entry of invoices and accelerates settlement.

Projects on the circuit between the pharmacy and the departments, without prescription management software (A Projects)

These projects aim at improvements without the use of dedicated prescription management software.

• A1: Projects with a purely organizational component.
  • Presentation: Refuse verbal prescriptions, encourage legible writing, create structured and complete prescription forms, develop a therapeutic formulary and hospital therapeutic protocols, implement mechanical distribution systems (elevators, remote-controlled cases), and a system for systematic analysis of medication incidents.
  • Economic stakes: These projects impact the reduction of drug expenditures, the improvement of the quality of care (readability, completeness), and time savings. Although not directly linked to computerization, they can be an essential prerequisite.
• A2: Projects with an IT component, but without prescription management.
  • Presentation: Type prescriptions before signing, use internal messaging between pharmacy and prescribers, distribute the therapeutic formulary and drug databases (Vidal, Thériaque) via the hospital network, and provide access to patient information (civil status, history, allergies).
  • Economic stakes: Typed prescriptions avoid interpretation errors and save time (no need to decipher or request confirmations). Messaging saves time on information exchanges. Access to patient information reduces confusion and errors.
• A3: Make databases on drugs available.
  • Presentation: Put Vidal or Thériaque online, accessible from any workstation connected to the network.
  • Economic challenges:
    • Time savings for consulting information (no need to travel to find a paper copy).
    • Acquisition cost savings of paper copies.
    • Benefit Related to Constant Updating of information, reducing errors due to obsolete data.
    • Reduced Prescription and Administration Errors because prescribers and nurses have constant access to up-to-date information.
  • Valuing gains: The study proposes formulas for estimating these gains, combining purchase savings (G1), time savings (G2), and reduction of iatrogenesis (G3). Online information is the most significant gain, even without direct integration into prescription software.

Projects with prescription entry (B Projects)

These projects represent a very important "step" in terms of computerization, involving the deployment of software for the nominative management of prescriptions.

• B1: Centralized entry of prescriptions upon arrival at the pharmacy.
  • Presentation: The software is deployed only in the pharmacy, where staff enter the prescriptions received (paper, fax).
  • Economic challenges:
    • Pharmacy knowledge of prescriptions: Enables the dispensing of prescribed medications and visibly reduces medication expenses.
    • Automatic Calculation of Drug Interactions: The software can perform systematic and exhaustive checks of interactions, even in deferred time. This gain is considered an injection of "very high quality of care" rather than a productivity gain, as an exhaustive manual check would be inconceivable. Such a check would be equivalent to the effort of 40 to 50 full-time pharmacists for a fictitious hospital.
  • Disadvantages: Requires re-entry of prescriptions by pharmacy staff, a considerable workload (estimated at 7 data entry operators for a 500-bed hospital). The system does not guarantee correct identification of the patient/prescriber, does not offer real-time feedback to the prescriber, and may generate data entry errors.
• B2: Decentralized entry of prescriptions in the departments (specialized software).
  • Presentation: Prescribers develop and enter prescriptions directly using the computer system, which can interact in real time. The prescription immediately reaches the pharmacy via the network.
  • Economic challenges:
    • Time savings and quality of care: Although entry may initially take longer than handwriting (120% to 150% of the time), the benefits outweigh the drawbacks. The prescriber and patient are positively identified, the prescription is complete and legible, reducing time wasted deciphering or searching for missing information. Nurses no longer have to rewrite administration plans. The prescription history is accessible online.
    • Protocols and data-entry shortcuts: considerably shorten data-entry time (15% to 40% reduction in data-entry activity) by automating common prescriptions.
    • Automatic retrieval of administrative information: Saves 4 to 9 minutes per incoming patient.
    • Discharge Orders: Estimated time savings of 15% to 20%.
    • Time Savings Assessment: Efficient software with these improvements can make data entry faster than the handwritten system.
    • Data entry during visit (portable terminals): Contributes to physician adoption and reduces time spent, although costs are higher.
    • Reduction in drug expenditures: Same as B1, with at least a 10% reduction in the hospital drug budget.
    • Reduction in medication-related incidents (iatrogenesis): Up to 50% of iatrogenic medication-related incidents can be avoided with the best information system, approximately half of which can be achieved with a B2 or B3 project. This includes gains related to systematic interaction checks.
• B3: Entry of the prescription in a care unit management software (integrated HIS).
  • Economic stakes: The advantages are the same as for B2, with additional gains linked to complete integration into the Hospital Information System (HIS). This avoids re-entries or copy-pasting between systems, multiple identifiers, and facilitates the consolidation of care plans and patient information. Time savings are greater in the event of successful integration.

Variants in dispensing mode (C Projects)

These projects concern the organization of drug distribution.

• C1 and C2: Global Dispensing (by cumulative prescriptions or cabinet restocking). These variations do not influence the gains, except for C3 and C4 projects.
• C3: Individualized dispensing and individual picking per patient ("DJIN").
  • Economic stakes: The UDDF transfers a significant workload from the departments (nurses) to the pharmacy (technicians), as nurses only have to administer the medications. It is personnel-intensive for the pharmacy, but the cost to the hospital can be seen as a shift in workload. The UDDF requires automatic routing devices to manage emergencies.
• C4: Automated individual picking based on computerized prescriptions ("robotized DJIN").
  • Economic stakes: The robot stores and prepares the distribution according to computerized prescriptions, reducing the preparation workload in the pharmacy compared to manual UDDF. Although experience is limited in France (experiments in progress), these systems can automate part of the preparation tasks (e.g., 50% of the time for the forms processed) and limit the increase in pharmacy staff due to the adoption of UDDF. The impact is estimated at one FTE per 60 to 80 acute care beds.

Variants corresponding to improvements to the basic system (D projects)

These projects involve deploying additional functions that would not be feasible without computerized prescription management. They are grouped into several categories:

• Interfaces with the rest of the HIS (patient identity server, access rights management by the staff system, inventory management). The patient identity retrieval interface is very important to avoid serious errors with consequences.
• Software improvements through automated processing, alerts, and more advanced algorithms (mandatory prescription checks, standard comments, integrated protocols, display of prescription cost, regulatory checks, blocking of editing if prescription not validated by pharmacist, administration alerts, etc.).
  • Prescription valorization: Displaying the daily cost or suggesting cheaper, comparable products can raise prescribers' awareness and reduce drug expenditures (potential gain of 1% of the drug budget).
• Improvements through a greater number of available and manipulated data (integrated drug database, prescription history, monitoring of actual administration and effects of medications).
  • Monitoring of Effective Medication Administration: Allows reporting of unadministered medications and adjustment of stocks but imposes an input burden on nurses. Cabinets managed by the prescription system (opening drawers according to patient/medication/caregiver) reinforce the prescription-administration link and can improve the safety and quality of care.
  • Monitoring the effects of drugs: Entering the observed effects, in particular undesirable ones, facilitates epidemiological studies and can affect expenditure and incidents. More plausible in an integrated B3 project.
• Access to and availability of the tool (entry/consultation on mobile terminals, numerous fixed workstations, real-time access for the pharmacy, etc.).

Computerization project costs

The costs of each IT project are divided into investment and operating expenses. Investments include software licenses, hardware acquisition (workstations, servers, network), user training, integration services (configuration, data recovery, start-up assistance, interfaces), and project management costs (acceptance, project management, coordination). Operating expenses cover operation, hardware and application maintenance.

• Projects focused on pharmacy management (P1, P2, P3, P4) :
  • P1 (EFMS): The cost is often a pro rata of the overall funding for the renewal of the hospital's HIS.
  • P2 (Lot withdrawals): No cost if management does not rely on lot numbers; otherwise, it depends on the acquisition of specific software modules.
  • P3 (Expiration dates): Mainly organizational actions, no specific IT cost.
  • P4 (GEF improvements): The costs of EDI and barcode modules are linked to specific hardware and software modules, which can make their return on investment more difficult. EDI over the Internet is an avenue to explore. The cost of order pre-authorization depends on the availability of the module and the configuration/training efforts.
• Projects without prescription management software (A1, A2, A3) :
  • A1 (Organizational): Costs limited to the workload for implementing organizational actions.
  • A2 (IT without prescription): If the hospital does not have an IT network, deploying microcomputers in all care units makes the project difficult to 'make profitable'. If the network exists, costs are limited to software licenses and project management.
  • A3 (Drug databases): Costs related to the acquisition of software licenses for the database and its distribution on the Intranet. The importance in terms of public health justifies these costs.
• Projects with prescription entry (B1, B2, B3) :
  • B1 (Entry at the Pharmacy): Costs limited to deployment at the pharmacy. For a fictitious hospital, the software license could be between 300 and 400 KF, a dedicated server between 200 and 300 KF, with costs for integration and training. Operating costs are estimated at 12% of investment costs. A significant cost specific to B1 is that of the additional staff required at the pharmacy for the mass entry of prescriptions.
  • B2 and B3 (Entry in the Departments): Costs are divided between those directly attributable to prescription management and those attributable to the general deployment of the HIS. Hardware (microcomputers, network) in the care units is a major expense if not already in place (700 KF to 1 MF for 30 to 50 microcomputers). The software license cost is comparable to B1. Integration services and training are significant but are justified by the annual gains. Return on investment can be less than one year if the hospital is already equipped with a network. For a B3 project (integrated into the HIS), hardware costs are shared with other HIS projects, which is more advantageous.
• Dispensing variants (C3, C4) :
  • C3 (DJIN): Is not a cost in itself, but a shift in workload from the departments to the pharmacy. Budgetary devices can be used to allocate these costs to the care departments.
  • C4 (Unit dose packaging robot): High costs (several million francs) because it involves specific and complex equipment, although potentially profitable by avoiding the total transfer of workload to the pharmacy. These costs are entirely attributable to prescription management projects.
• Improvements to the basic system (Projects D): Their cost is generally not additional, but linked to the license of the chosen software package and the functionalities it includes. The choice must be based on the usefulness of the functions, the gains they bring and the overall cost of the software package.

Summary of key findings and recommendations

Computerization of the medication circuit is an essential approach for healthcare facilities, with significant economic and qualitative stakes.

The three types of economic stakes are time (or productivity) savings, reduced medication expenses, and improved quality of care by avoiding incidents and errors. Added to this are the benefits related to the « piloting » of the drug circuit.

A key finding is that the most significant economic gains are achieved mainly by IT projects that implement prescription management software. Organizational projects or simple IT systems (Projects A) bring improvements, but their gains are not comparable in order of magnitude.

Decentralized prescription entry by prescribers in the departments (Projects B2 or B3) is the recommended principle. Prescriber adherence to this mode of operation is facilitated by software devices that lighten the entry process (shortcuts, protocols) and by the use of hardware allowing entry during the visit (portable terminals). The B1 mode of operation (entry a posteriori at the pharmacy) is less recommended because, although it allows knowledge of prescriptions and calculation of interactions, it generates similar implementation costs and lower gains, in addition to the heavy burden of re-entry for pharmacy staff.

Regarding productivity gains:

• For prescribers, B2/B3 projects do not necessarily bring significant productivity gains at the outset, as data entry may even take longer. However, in the long term, thanks to protocols and shortcuts, computerized data entry can become faster than the handwritten method. Above all, doctors are less solicited for questions of incompleteness or interpretation, which represents a strong advantage.
• For pharmacy staff, the most noticeable gains come from functions related to economic and financial management (barcode readers, order recommendations, EDI), which are often independent projects. Prescription management projects do not show a tangible direct benefit on pharmacy productivity, and may even increase the workload (case of B1) or shift it (case of DJIN). However, the considerable contributions of drug interaction control and calculation functions should be emphasized, even if they do not translate into direct "productivity gains."
• For nurses, the most significant improvements come from the automatic establishment of the Medication Administration Plan (MAP), the continuous availability of complete and legible information, and online access to generic drug information. These productivity gains can be very significant (up to 5% to 10% of the nursing workload), but are often "diluted" and result in time freed up for direct patient care functions rather than tangible budgetary savings. It is important to note that every time saving for healthcare personnel is generally also an improvement in the quality of care.

The reduction in medication expenses is primarily due to the correlation between in-service dispensing and prescriptions, i.e., dispensing to services only the medications they have prescribed. This principle, whether DJIN or global distribution based on knowledge of prescriptions, allows for reductions of approximately 10% to 20%. Other functions also contribute, but to a lesser extent (displaying cost, reducing scope, improving stock management).

The benefits related to improved quality of care and avoidance of medication-related incidents are provided by almost all system functions. The most contributory are the completeness, exhaustiveness, and legibility of prescriptions (B2/B3), systematic pharmacological controls (B1 and beyond), and the linking of patient information with prescribed medications (B2, B3, D). The study estimates that up to 50% of iatrogenic drug-related incidents are avoidable thanks to a high-performance computerized system. Advanced systems (Projects D) offer the highest levels of performance. These functions also improve the operating comfort of healthcare personnel and reduce stress and error factors.

Finally, the study emphasizes that "software" investments are easier to make profitable than specialized "hardware" investments, because software development benefits multiple hospitals and actors simultaneously. Microphones (fixed or portable) are essential for entering prescriptions. Preparation automation robots, although costly, are likely to be profitable.

For the future, it is recommended to continue studies and develop guides to support hospitals in their projects to computerize the medication circuit. Questions such as the correlation between IT functions and types of incidents avoided, the prioritization of deployment according to medical disciplines, or the optimization of alert settings still need to be explored in depth by the hospital community.

Source :

https://sante.gouv.fr/IMG/pdf/circuit1.pdf

‍