Modelling Hospital Materials Management Processes

Materials management is an important issue for healthcare systems because it influences clinical and financial outcomes. Before selecting, adapting and implementing leading or optimized practices, a good understanding of processes and activities has to be developed. In real applications, the information flows and business strategies involved are different from hospital to hospital, depending on context, culture and available resources; it is therefore difficult to find a comprehensive and exhaustive description of processes, even more so a clear formalization of them. The objective of this paper is twofold. First, it proposes an integrated and detailed analysis and description model for hospital materials management data and tasks, which is able to tackle information from patient requirements to usage, from replenishment requests to supplying and handling activities. The model takes account of medical risk reduction, traceability and streamlined processes perspectives. Second, the paper translates this information into a business process model and mathematical formalization. The study provides a useful guide to the various relevant technology-related, management and business issues, laying the foundations of an efficient reengineering of the supply chain to reduce healthcare costs and improve the quality of care.


Introduction
Progressive reduction in public resources -and the subsequent need to restore budgets -means governments are responsible for finding solutions to achieve more operational efficiency in hospital processes. Drug expenditure, in particular, is a relevant factor in the profit and loss accounts of healthcare systems ( [1,2]); hospital pharmacy management is called upon to adopt policies to reduce drug inventory costs and maximize the cost-effective use of personnel and resources [3][4][5].
Awareness of logistics is becoming more widespread and many initiatives and studies dealing with supply chain integration have been undertaken (for example, supply outsourcing strategies [6,7]). However, the internal supply chain (vs. external, i.e., beyond the boundaries of an organization) "remains the sore spot or weak link" [8] in process integration and optimization. This lack of systemic approach to internal supply chain management is reflected in huge costs in materials management and low service quality delivered to patients [9].
Hospital materials management (HMM) also involves the clinical sphere of healthcare service performance. Clinical errors in drug prescription and administration, for example, are always possible, depending on human factors and procedural issues.
Regarding resource use optimization, the difficulties of transferring manufacturing best practice directly to the hospital environment are evident [10], although the first sector of intervention is inventory management (see, for example, [8]). It is conceivable to suppose that the main cause of inefficiency is the existence of hidden stocks to avoid stock-outs [10], which would be more politics-and experience-rather than data-driven [6].
It is clear that a significant percentage of a pharmacist's time is consumed by order entry, verification, clarification, and follow-up activities [11]; the same applies for nurses with prescription transcriptions, stock level control and administration.
In order to achieve a comprehensive image of the HMM process and potential ways of exploiting IT to enable an efficient reengineering of this supply chain, reducing healthcare costs without affecting the quality of care [1], it is fundamental to consider the previously presented logistical and clinical perspectives as both sides of the same coin.
In the literature, there are many collateral references on how materials management works; in real cases analysed, information systems are usually built for fragmented applications, and much information is lost or is not recorded when it flows through processes. This implies losses in traceability and increases in clinical risk, while inventory management techniques and logistics are hardly performed, causing high inventory costs.
The objective of this paper is twofold. First, on the basis of a literature review, analysis of real cases and international guidelines, the paper fills a research gap, providing an integrated and detailed analysis and description of hospital materials management data and tasks (able to tackle information from patient requirements to usage, from replenishment requests to supplying and handling activities) in relation to medical risk reduction, traceability and streamlined processes perspectives. Secondly, the paper translates this knowledge into a business process model and a mathematical formalization, showing the integrated information and physical flows in order to trace and share data among actors with the aim of reducing clinical risk and time-consuming tasks while enabling requirement programmability and, more in general, knowledge management.
The paper is structured as follows: after the materials under examination are defined, the materials, processes and actors are described in section 2. Details about process modelling and formalization are given in section 3, and in section 4 the performance results of a case study are presented. Finally, conclusions and suggestions for possible further research are presented.

HMM elements: materials, processes and actors
The first step in the hospital pharmacy "micro-world" reengineering process is to identify the behaviour of this system, to establish for example what to manage in terms of materials (par. 2.1), processes (par 2.2) and actors (par. 2.3), taking into account information and legal constraints.

Materials and places
Management of materials in healthcare involves two kinds of item clusters: drugs (or medicines) and medical devices, subjected to different regulations harmonized by countries according to international guidelines. The properties of medicines in a hospital information system may be mandatory or optional depending on the contextual workflow [12]. A fundamental "identifier" is the ATC (Anatomic -Therapeutic -Chemical) classification, internationally accepted and maintained by the World Health Organization. In addition to commercial drugs, drug administrations can also refer to galenics, such as personalized medicines prepared as a "mixture" of commercialized products at the bedside, in hospital pharmacy or in another defined medical unit. In parallel, medical devices, such as surgical kit and apparatus, can be managed as single or grouped items by pharmacists. The item list (in other words, the set of medicines or medical devices that can be administered/dispensed to or implanted in patients in a healthcare system) changes from hospital to hospital, depending not only on the healthcare services managed, but also on the physician's expertise and preferences, and following pharmaco-economics principles [13].
Regarding places of materials management and dispensing, it is possible to distinguish between clinical and managerial perspectives. From a clinical point of view, the usefulness of having centralized patientoriented pharmacy services to deliver professional services to the patient has been recognized [14,15] The active substance(s), as stated above, is usually a key element because it can permit evaluations about pharmaco-economics and availability (for example, to distribute the cheapest brand name stored in satellite or central pharmacies or to prescribe what is stored in the local warehouse), but the clinical consideration can imply a more rigid selection.
In the case of surgical intervention, some materials have to be explicitly prescribed (for example, an orthopaedic prosthesis and its size or surgical kit).
Considering the data related to inventory management activities, the output of this phase is one or a list of records of prescriptions (p) for each patient, formalized as follows.
In particular, having defined: � � �� drug with the attributes: Nf: number of units contained in a package of f LTf: supplier lead time, time between the ordering and the delivery of a supply order � � �, Medical Unit belonging to a hospital with the attribute: TTm: time needed to transport orders from the pharmacy to the medical unit m.
The classical prescription record for a patient is composed of: � � �, hospital prescription for a single drug f, with: tsp: time bucket at which the physician states the prescription p tap: scheduled administration time bucket for the prescription p DRp: drug prescribed f, chosen from � Qp: administration quantity MUp: medical unit where the patient is hospitalized m, belonging to M.

Exams prescription (a.1.2)
The physician can prescribe some diagnostic examinations to patients, such as laboratory exams (e.g., blood, RX) or other activities that require materials.

Pharmacy Prescription Validation (a.2)
After a prescription, prescription information may be made available to pharmacists for a pharmacy validation. It can be advisable but not mandatory, so that many organizations tend to jump this step. This implies no double reviews and, hence, no possibility of medical error reduction [13].
A detected problem can be a supply issue (suspended medication, out-of-stock, etc.), a legal issue (medication recalled or not allowed under certain conditions), or a medical issue (redundancy, interaction, contra-indication, ICA, etc.).

Materials delivery to patient (a.3)
In the K frame of Figure 2, the administrations dependent by the prescriptions placed are evaluated in terms of dose availability at the administration time tap and possibly activate the (b) process.
The possibility of managing urgent deliveries caused by unavailability of administration materials is also presented. Delivery urgency means lead time (TTm or LTf) compression and, consequently, higher delivery costs, not excluding being behind schedule. However, deliveries may be easily evaluated by the physician before being activated. Indeed, a time can be given to the physician to confirm the feasibility of the administration behind the schedule on the basis of the delivery scheduled time.
In the M frame of Figure 2, the cancellation of a prescription or the discharge of an inpatient, are transmitted to process (b). While

Preparation, administration or implant (a.3.1)
This refers to the delivery of materials inside the hospital at the tap occurrence, depending on a prescription and carried out by a nurse.
Many hospitals do not have an information tool or do not have an information system at all, so nurses are in charge of transcribing prescriptions on provided diaries, checking them to know when to administer. This implies risk of errors in transcription and administration execution (right time, person, etc.).
Preparation is the phase in which nurses take drugs from the stock and prepare them to be administered to patients.
Galenics are a special case because the preparation, depending on the organization, may involve pharmacists and requires an appropriate recording of all actions and items used (information about drug ID, lot number, quantity used, etc.) for traceability reasons. For example,

Dispense to pa
In some hea can also be care (by outp These materi inpatients, bu than a time b The outside unit dispens material is t given to the p travenous inj m chloride so d, oncological medical unit an and galenic while ensuring [12], the ass ent is a "dispen s the instant in ications are u filled trolley taining the m rs for specific [21]  In particular, events involving materials consumption are ranked from the least traceable to the most (the independent variable) as follows: low cost and genericgoods usage (by definition untraceable), materials employed for examinations (traceable if the information system provides this feature), materials dispense (not traceable when it comes to administration because it is performed at the patient's home) and materials administration/implant (traceable until the end of the process, because the usage of an identified material, with its characteristics, expiration date, batch number, etc., can be fully recorded). The other axis shows the materials requirements programmability, which goes from the possibility only to forecast the consumption of some materials in a period of time, to the highest probability that a particular consumption is going to take place in a specific moment. The diagram shows activities concerning patient care.
Given that "a plan must cover a period at least equal to the time required to accomplish it" [22], the programming interval �� � is fundamental data related to prescriptions (when a batch consignment for each medical unit is chosen, PI is the maximum number of time buckets available for medical unit replenishments or supply activities): On the other hand, genericgoods usage is predictable with low uncertainty, as, because of the big volumes, it is easily forecasted by time series techniques. Almost the same applies for examinations demand. Outpatient dispense is in the middle of this plane because it concerns a prescription kept by the pharmacy for a given period while the �� � is not known.

Medical Unit Inventory Management process (b)
The second identified macro-process deals with the materials management at the medical unit level. The process starts when an administration is planned. This passage is particularly important because, without defining a provisioning politics, it is able to introduce the real requirements data in the definition of an order, allowing the medical unit (and hospital pharmacy, as a consequence) to manage its stocks facing patient requirements when they emerge.
The actors involved in this process are the nurse manager and the pharmacist, and the two main activities are: (b.1) Medical unit stock management and replenishment (handling of incoming and wasted materials, dispensing and keeping the warehouse management system up-to-date, defining replenishment requests), (b.2) Pharmacy requirements assessment.
The first element to consider when modelling the (b) process (and later (c) too) is the management policy adopted. Basically, the two fields of intervention which need to be globally optimized are medical unit replenishments and pharmacy supplies. Look-back (for example Re-Order Level, Re-Order Cycle, Just in Time), look-ahead (for instance Material Requirements Planning) or mixed (for example Vendor Managed Inventory) approaches can be used.
As Nicholson et al. [6] claim, the most traditional servicing approach in the pharmacy management field is the periodic review par level (or order-up-to level), that requires setting the review interval and the optimal security stock (base stock level). While the second depends on therapeutic and medical constraints set by taking into account demand variability, the first has to be defined according to the involved resources.
Some examples of re-ordination strategies are given by Kalmeijer et al. [23], who promote the extensive use of information systems to manage requirements considering the default medication database as the local stock. Not-stock items are automatically ordered from pharmacy, instead.
In general, the look-back approach is more popular than the others, but brings higher inventory levels. Moreover, forecasts on aggregate data about consumptions recorded by pharmacy are influenced by medical unit management techniques [6]. Look-ahead methods, instead, usually need careful and punctual information about requirements forecasts.
While the BPMN representation of this process is a subset of the one shown for the process (c), the mathematical formalization of the involved inventory management procedures is as follows: Our integrated approach to the HMM process, oriented to the information management and traceability, allows the adoption of Materials Requirement Planning (MRP) techniques [22]. In particular, starting from the patient prescriptions collected in sub-process (a), it is possible to plan medical unit replenishment actions and supplier order releases.
The following is first defined: (2) The replenishment policy adopted depends, among other things, on [24]: • lot sizing policy (lot for lot, Economic Order Quantity, Fixed Order Period, etc.), • lead time offsetting (at least equal to �� � ).
Consequently, the decision maker (usually the medical unit nurse manager) can release an order to the hospital pharmacy for the transportation quantity (��� ����� � , planned order releases for the medical unit m, the drug f at the time bucket� � ) to satisfy the requirements of the medical unit m at the k th time bucket (where � � � � � �� � ), dependent on the policy and at least equal to the net requirement rounded to the nearest integer of Nf.
The aggregate information may be directly used by the centralized inventory management process in order to plan medical unit distributions and supplier orders, freeing medical unit staff from managerial activities and making more aggregated and powerful data available about medical unit requirements and hospital pharmacy availability over the time.

Centralized Inventory Management (c)
The centralized Inventory Management process (Figure 4) is triggered by a medical unit requirement and can be interpreted as substitutive with respect to the medical unit inventory management, or in line with it, as already explained in the previous section.
The involved sub-processes are: (c.1) Pharmacy stock management, order disposition and supplying activities, (c.2) Internal distribution, (c.3) Materials admission, quality control and payment.

Pharmacy stock management, order dispositions and supplying activities (c.1)
Based on pharmacist order assessment, the stock management and supplying activities take place. Operationally speaking, the tasks carried out by pharmacy unit actors are (N frame of Figure 4): • incoming and expired materials handling, • warehouse management system updating, • budget reconciliation assessment (each cost centre/medical unit typically has its own budget to manage for each expenditure class; the same goes for materials belonging to tenders, which have their specification budget. Materials have to be transferred according to these; otherwise, a budget integration has to be requested from the Superintendence and Treasurer's office reporting quantities and the amount of extra funding needed), • stock levels control and authorization to dispense, • supplying activities.
As with medical unit stock management, different inventory management policies can be implemented in order to take frequency/quantity supplying decisions.
This "Central of Purchasing" collects the orders and forwards them to the suppliers, feed backing information such as order acceptance or delivery due date to the ordering point, with the exception of life-saving drug stock-out, which is managed by urgent procedures.
Finally, the O frame of Figure 4 illustrates the design of the urgent deliveries management process described in the sub-process (a).
In terms of order releases, the same considerations and formalization already expressed for the process (b) can be repeated. In this case, the medical unit under examination is the central pharmacy,with the exception of the Gross requirement. This is given by: with m=h occurring when the particular medical unit is the central pharmacy.
This time, the results of the calculations are the array of orders to suppliers, which are decided, again, according to the policy chosen and the feasibility evaluation of transportation to the medical unit (in this last case, some urgent procedures may be needed in order to avoid stock-outs).

Internal
The interna pharmacist's replenishmen supply order frame of F preparation a  Figure 4).

Expected p
The propose (par. 4.1) and ralized Inventor distribution (c.
al distributio authorizatio nts, and is e r (also transien Figure 4, wh and delivery t ital pharmacy ital service qu n inventory oriented ones e, and differe outing proble n which tran nnel and phys es taken into a reparation, per also virtually for a medica thod is app le with all ma considering m task, delivery immediately (these may b 's nursing staf erials admissio ried out by ph performances ed HMM has d economic con ry Management

.2)
on process b on triggered eventually com nt materials), hich is com to medical uni y delivery sys uality [25]. Th problem [7]  With this aim, we developed a detailed discrete event simulation model (using Arena Rockwell©) able to compare two scenarios: 1. the "AS IS" scenario, (the "traditional way to manage hospital materials"), a periodic review par level servicing approach, set on the basis of past consumption, keeping distributed (among ward warehouses) safety stocks to prevent stocks-out, 2. the "TO BE" scenario, which involves a delivery scheduling based on patient requirements (MRP technique). The performance evaluations are carried out through an inventory cost function (�) that takes into account the overall cost sustained by a hospital in adopting an inventory management policy; in particular: The simulation model has been validated and verified using data from an Italian regional university acute hospital with 700 beds and 26 specialist medical areas.
Drug daily demand distribution was extracted from the hospital information system for the year 2012, a selection was made of the most expensive (26.6% of the total purchasing costs) items in the main medical units (3) by means of the Pareto analysis, based on 55,000 medical unit consumption records related to 900 items. The drug demand distributions were used to rebuild the periodic medical prescription placings of physicians.
The cost parameters belonging to the function were calculated by performing an Activity Based Costing analysis [26] on the data collected on the field, excluding the actual resources' saturation due to the particular management policy adopted in the hospital.   Nevertheless, modelling the data sharing and the integration and coordination among actors and activities, the proposed HMM gives the guidelines for the design of a common and distributed information platform that can be also be used to collect data, identify key performance indicators and compare them in different management technique scenarios, carrying out wider performance evaluations.