Hospital processes within an integrated system view: a hybrid simulation approach

Processes in hospitals or in other healthcare institutions are usually analyzed and optimized isolated for enclosed organizations like single hospital wards or certain clinical pathways. However, many workflows should be considered in a broader scope in order to better represent the reality, i.e., in combination with other processes and in contexts of macro structures. Therefore, an integrated view is necessary which enables to combine different coherences. This can be achieved by hybrid simulation. In this case, processes can be modeled and simulated by discrete simulation techniques (i.e., DES or ABS) at the meso-level. However, holistic structures can be comfortably implemented using continuous methods (i.e., SD). This paper presents a theoretical approach that enables to consider reciprocal influences between processes and higher level entities, but also to combine hospital workflows with other subjects (e.g., ambulance vehicles).

Improvements in healthcare delivery are important to master global challenges in future, which are particularly triggered by demographic changes and increasing costs. Hospital managers and those of other healthcare providers must solve many trade-off problems, i.e., a better service quality versus higher resource usage and operational costs. Process improvements performed by simulation and modeling (SaM) allow to better handle budgets and to increase the organizational performance prior to their cost-intensive real-world implementation. However, the success of such improvement processes crucially depends on the applied SaM method.

Process analysis by Business Process Modeling (BPM) plays a significant role in the management of companies or divisions (Giaglis 2001). This modeling technique enables to structure the process knowledge of single domain experts in common models that can be used as a basis for further discussions and studies. Together with simulation, BPM offers an efficient and comprehensive toolset for managers and other decision-makers (Harrison et al. 2007, Laguna and Marklund 2013, Weske 2012). This combination helps to detect bottlenecks and problems at an early stage, and to evaluate different scenarios and solutions prospectively. The main focus of process modeling is the (graphical) representation of different paths and process steps, but resources and organizational structures can also be considered. The process analysis aims to achieve an understanding of a certain process and to identify weaknesses and potential improvements (Rebuge and Ferreira 2012). Therewith, simulation is capable to calculate the expected outcome for both the currently existing process and for a prospectively adjusted process that does not exist yet. In this case, management decisions, especially at the strategic level, could be more objective by limiting biases of individual perceptions.

BPM and DES are primarily used for studies in the inpatient sector (Vera and Kuntz 2007), but they also can be applied in cases when abstract clinical pathways for certain diseases are evaluated. In recent time these methods receive a growing attention in situations where hospital services are reimbursed through flat-rate or fee-per-case charges, such as the Diagnosis Related Groups (DRGs). In this case, a hospital will receive a predefined and agreed reimbursement value for each patient. This value is dependent on diagnosis and not on the real costs of provided services. Consequently, hospitals are extremely interested to keep the costs per case below this value, but also to provide an adequate service quality in a competitive market. That means, process improvements have significant impacts on the operating result and organizational performance.

Process modeling is widely used and already established in hospitals across different application areas. On the one hand, the medical treatment process of a certain disease (i.e., clinical pathways) and thus the quality of care are in focus (Ronellenfitsch et al. 2012), on the other hand, the profitability is a major business target figure. In this case, e.g., the process structure can be modeled and underlaid with cost data. This approach serves as a basis for an activity based costing and enables to generate important information for an effective management (Oker and Ozyapici 2013, Hada et al. 2014, Kaplan et al. 2014, Cannavacciuolo et al. 2015). However, the prospective analysis of processes and their realization in practice vary considerably. Furthermore, such techniques are often not used holistically, but rather for specific diseases or for a particular hospital ward. One reason is, that experts investigate their research scope focusing on one particular disease (Akhavadan 2016), henceforth, the willingness to model foreign processes does not exist. Another reason is, that in economic evaluations cost centers are mostly considered in separate studies (Ibrahim et al. 2014).

An overall view containing bordering influences typically does not exist. In particular, different related processes in the same institution and even the broader environment of the hospital should be considered when analyzing a system. For example, if a patient requires an imaging examination, the radiology department usually interacts with other medical wards during a treatment process. The same also occurs in other supporting processes like laboratory tests or occupancy management. Ideally, the whole hospital organization should be considered, but without over-complicating a model. Going one step further, even external influences can affect a considered process, i.e., the surrounding population structure, economical factors, political influences, or technological developments. Using BPMand DES, a modeler has to represent the whole environmental complexity by very detailed process models. However, in most cases it is sufficient to represent less important influences by more abstract or aggregated models. In this paper we present a hybrid simulation approach that allows to develop detailed process models by discrete simulation techniques using Discrete-Event Simulation (DES) and Agent-Based Simulation (ABS), as well as to represent the process environment and bordering processes by more abstract high level System Dynamics (SD) models.

System dynamics example (Casual-Loop-Diagram)

An example scenario modeled by a Casual-Loop-Diagram.

Related posts