Optimization of Epilepsy Program Procedures With Linear Programming

Introduction

Epilepsy is defined as a tendency to have recurrent, unprovoked seizures, and affects approximately 1% of the world’s population (Gong et al., 2016; Thurman et al, 2011). It contributes to a number of social consequences as part of the disease, including loss of driving privileges, work restrictions, and social isolation. Epilepsy can have such dire consequences as sudden death (Surges and Sander, 2012) and can progressively worsen over time (Nearing, Madhavan, and Devinsky, 2007).

The Nebraska Comprehensive Epilepsy Program is a comprehensive tertiary referral center for patients with seizure disorders. It has a range of services extending from medical therapy to surgical approaches. Integral to the treatment of this patient population is the effective diagnosis of seizure type and brain location responsible for seizure onset. This is accomplished using a diagnostic measure, electroencephalograph or EEG, which is a real-time recording of brain derived electrical activity using electrodes directly applied to the scalp.

EEG is performed at the center using two main study formats: a routine study that takes approximately 2 hours to perform and a more intensive, long-term study that is inpatient based, requiring a stay in the hospital while EEG signals are continuously recorded. The routine studies are typically helpful in capturing interim brain functioning in epilepsy patients or to quickly assess and diagnose changes in mental functioning in other patients who may not have epilepsy. These studies can be critically helpful in certain clinical situations but are typically not the benchmark for most epilepsy diagnoses requiring more in-depth study.

For the diagnosis and/or brain localization of specific seizure types, a long-term EEG recording using video monitoring is the recognized gold standard. In these studies, a patient is admitted to the hospital into a specialized unit where he/she is connected to EEG electrodes that are recording in real time with video recording of patient activity. During these studies medications are often withdrawn in order elicit seizures, thereby allowing the physician an opportunity to capture brain EEG signals and video recordings of the patient while experiencing a seizure. This in turn provides an unparalleled assessment of a patient’s epilepsy and is often used to determine treatment decisions ranging from medication adjustments to brain surgery.

In most epilepsy programs a combination of the two different types of EEG procedures are performed with the bulk of the routine EEGs used for outpatient clinic patients and long-term EEG monitoring performed as an elective inpatient procedure. As expected, revenues associated with these procedures are substantially different. Inpatient monitoring is a far higher reimbursed procedure than a typical, routine EEG. While costs are substantially different between the two, due to the additional expense of an inpatient stay versus an outpatient procedure, the two studies are similar in terms of technology requirements (e.g., the EEG equipment needed to run each study). Because the availability and cost of EEG equipment is an important rate-limiting step in these procedures, an optimization of the ratio of routine versus inpatient EEG procedures is desirable to maximize revenues, thereby allowing for further growth of the epilepsy program including additional purchases of EEG equipment.

The above dilemma is well-known in hospital and healthcare settings. For example, Patient Care Pathways (PCPs) have been developed with the goal of creating sustainable clinical architectures that can be used to plan capital investment for both the short- and long-term (Rechel, et al., 2010). To create such architectures a deep understanding of the bottlenecks of clinical workflows must be understood. The Theory of Constraints (TOC) is often used in healthcare settings when dealing with a finite resource that can potentially act as a bottleneck (Caunhye, Li, and Nie, 2015; Young, et al., 2004). In a TOC model other clinical processes are adjusted to be complementary to the bottleneck, thereby making the bottleneck a critical rate-limiting step. This can in turn create optimal utilization of resource capacity, particularly in situations where resources are limited such that they cannot support full throughput.