Improving the science of healthcare delivery and informatics using modeling approaches

Abstract

The science of health care delivery and informatics is to be viewed as mutually connected disciplines and, therefore, the problems in both streams often occur concomitantly. For example, inefficiencies or ineffectiveness of Healthcare Information Technologies (HIT) quickly translates into poor healthcare delivery or vice versa. Research focusing on various decision making or decision support aspects of HIT and healthcare delivery is still sparse. We identify six key thematic areas that require significant attention from various researchers and stakeholder communities and emphasize on ways research in this space could further developed. We provide examples from the articles appearing in the special issue on healthcare modeling.

Introduction

Healthcare represents an important sector of any nation's economy. US Department of Health & Human Services reports that healthcare spending comprises about 17% of US Gross Domestic Product and projects that healthcare spending growth is expected to outpace the GDP growth [17]. However, our health care system is fragmented and currently designed to support billing rather than care coordination and continuity [7], [8]. A part of the reason is that healthcare delivery is complex and marred by low efficiencies, poor technology interfaces, communication gaps, and high costs. Complexity in care environments also results from team oriented clinical task management and decision making where interactions among several care providers and stakeholders such as patients, physicians, nurses, residents, fellows, insurance companies, and federal agencies are involved. Such interactions often require the streamlining of the flow of information, work, material, and patients through the use of well-designed technology. It is known that the use of well-designed technology and good decision making models result in improved safety, efficiency and reduced costs [9], while the use of ill-designed clinical systems, poor integration of workflow with systems and lack of clinical data utilization for decision models often translates into reduced efficiencies, increased medical errors, and poor decision making. These in turn can lead to numerous unintended consequences such as errors, slow adoption of technology, etc. [2], [6], [14].

Various methodological approaches can be applied to understand the problems associated with currently deployed systems and identifying mechanisms to improve them, or in some cases develop entirely new systems of decision models. We recognize that there is an ample need to utilize different methodological approaches that will make significant contributions toward a) identifying and developing decision models that can harness the power of big clinical data, b) understand and develop new or improved models of healthcare delivery from provider and patient-centric perspectives, and c) improve safety, quality and efficiency while curtailing costs. Based on the foundational reasoning philosophies, various scientific approaches that can be applied to vigorously pursue research under various healthcare themes may be grouped into three categories: data analytic modeling, simulation modeling, and behavioral modeling. These could be utilized either independently or as in multi-method approaches where a combination of two or more of these modeling approaches can lead to further insightful findings. Multi-method approaches often result in more insightful findings [12]. These approaches could be inductive (theory development), deductive (theory testing) or a combination of inductive and deductive approaches such as in simulation modeling [3].

We identify six key research areas within healthcare domain that we recognize as important streams for conducting further research and can have tremendous impact on improving the process of healthcare delivery and associated outcomes (Fig. 1). These research themes have overlapping components and ought to be viewed using systems approaches, i.e., it is important to recognize the interrelationships among two or more of these research themes. For example, various characteristics of HIT (Healthcare Information Technology), Computerized Physician Order Entry systems, Patient Portal and Clinical Decision Support Systems will have an impact on the clinical workflow and cognitive processing of care providers. The cognitive processing style of care providers, on the other hand, will influence how such systems are used in clinical environment.