Page 215 - Data Science Algorithms in a Week

Managing Overcrowding in Healthcare using Fuzzy Logic 199

FUZZY SYSTEM ARCHITECTURE

In order to define the fuzzy subsystems, Asplin’s comprehensive ED overcrowding
conceptual model was utilized, which divides emergency care processes into three
interdependent phases: input, throughput and output (Asplin et al., 2003). Each phase in
Asplin’s model contributes significantly to the level of ED crowding, and this research
adapts these phases in Asplin’s conceptual model in developing the ED overcrowding
quantification tool. Many previous studies have taken into consideration three ED
operational aspects: emergency care demand, ED workload, and discharge status in
developing quantitative instruments for crowding. These same operational aspects are
adapted into the framework developed in this study, as shown in Figure 4. By utilizing
fuzzy logic, this study overcomes the limitations of previous studies, by quantifying the
opinion of experts with different perspectives, to reduce the introduction of bias in the
final assessment of crowding.
In addition to the three phases of Asplin’s model, information from ED professionals
and experts is integral to the framework used in this study. This research proposes a
three-level hierarchical fuzzy logic system which is developed based on available
information and knowledge from experts. The purpose of this proposed fuzzy system is to
accurately determine the level of ED crowding. Like the fuzzy system as shown in Figure
3, the proposed fuzzy logic system includes seven inputs, four fuzzy inference systems
(fuzzy subsystems), and one output. The seven inputs of the proposed fuzzy logic system
are developed corresponding to four subsystems, related to Asplin’s three interdependent
phases, and are defined as follows:

Input 1: Patient Demand; Ratio of Waiting Patients to ED Capacity
Input 2: Patient Complexity (Waiting Area)
Input 3: ED Physician Staffing
Input 4: ED Nurse Staffing
Input 5: ED Occupancy Rate
Input 6: Patient Complexity (Emergency Room)
Input 7: Boarding Status; Ratio of Boarded Patients to ED Capacity

Figure 4: Determinants of ED crowding level.