User:Shawndouglas/sandbox/sublevel12

Sandbox begins below

Title: What role does systems integration play in the laboratory and why is this important to address?

Author for citation: Shawn E. Douglas

License for content: Creative Commons Attribution-ShareAlike 4.0 International

Publication date: February 2024

Interoperability and systems integration

In order to answer this question, we first must discuss the concept of "interoperability," of which integration of other informatics systems is just one component. Interoperability is defined by the Healthcare Information and Management Systems Society (HIMSS) as “the ability of different information systems, devices and applications (‘systems’) to access, exchange, integrate and cooperatively use data in a coordinated manner, within and across organizational, regional and national boundaries” to, in the case of a laboratory, ensure timely, portable, and accurate analytical results (the "deliverable" of most laboratories). While HIMSS' definition is focused on the clinical realm, their definition is robust enough that it, at least in part, can be applied to laboratory-based organizations serving most industries.

The why and how of laboratory integration

Why

Why should labs focus on interoperability and systems integration? Let's look at a few industries.

1. Clinical diagnostic and research labs: In the realm of clinical laboratories, improving interoperability among clinical informatics systems is recognized as an important step towards improving health outcomes. The National Academies of Sciences, Engineering, and Medicine had much to say on this topic in their 2015 publication Improving Diagnosis in Health Care:

Improved interoperability across different health care organizations—as well as across laboratory and radiology information systems—is critical to improving the diagnostic process. Challenges to interoperability include the inconsistent and slow adoption of standards, particularly among organizations that are not subject to EHR certification programs, as well as a lack of incentives, including a business model that generates revenue for health IT vendors via fees associated with transmitting and receiving data.

In particular, the National Academies discussed an additional concern, one that still causes issues today: interfaces between electronic health records (EHR) and the laboratory and other clinical information systems that feed medical diagnostic information into the EHRs. In particular, they found "the interface between EHRs and laboratory and radiology information systems typically has limited clinical information, and the lack of sufficiently detailed information makes it difficult for a pathologist or radiologist to determine the proper context for interpreting findings or to decide whether diagnostic testing is appropriate." EHR integration was also a problem at the peak of the COVID-19 pandemic. In early April 2020, a report from Nature revealed that academic research laboratories wanting to assist with COVID-19 testing efforts had at times been stymied by the incompatibility between academic informatics systems and hospital EHRs. Not only were hospitals using EHRs of differing types, but many of those EHRs were not designed to talk to other EHRs, let alone to academic and research laboratories' informatics systems. Combine this with strict account procedures and the costs of developing interfaces on-the-fly, more than a few medical systems turned away the offer of help from academic and research labs during the height of the pandemic. Had there been greater systems integration across these two essentially disparate lab types, it's possible even more academic laboratories with the necessary testing equipment could have assisted with running patient-based clinical testing.

While this constitutes an extreme example, it's possible that a push for improved interoperability across the systems used in commercial clinical diagnostic labs and more academic clinical research labs could have other benefits, for example with improving the state of interdisciplinary research, diagnosis, and treatment of cancer. A similar case can be made for clinical diagnostic systems and academic researchers seeking to conduct translational research using de-identified clinical patient data found in EHRs. However, haphazardly throwing technology at dynamic, real-time scheduling won't work, particularly in part due to how integration problems can quickly emerge.

2. Pharmaceutical manufacturing quality control labs: In the realm of manufacturing, laboratories play an important role in ensuring the safety and quality of produced goods, as well as participating in their development and optimization. Here we'll use a 2020 article published in Computers in Industry as an example, where Coito et al. present their middleware platform for intelligent automation, as applied to the quality control (QC) laboratories of the pharmaceutical manufacturing industry. The authors note that in these labs, "every drug must be sampled and tested to ensure it meets all safety and quality requirements," and that due to the "dynamic scheduling problems" inherent to the labs' six major activities (i.e., "system preparation, system suitability, sample preparation, analytical run, data processing, and review"), how laboratory personnel and equipment is managed and optimized in real-time is essential. The authors not that "the current level of data integration required to develop an intelligent automation system in real-time settings can be very complex, having to fetch data from many different sources while dealing with poor data quality." This makes a focus on interoperability essential and itself a difficult challenge to solve. Their middleware automation solution incorporated "two different industrial identification solutions to demonstrate the interoperability of the system," with a strong focus on the OPC Unified Architecture (UA) information modelling framework as a more future-proof solution to addressing interoperability for both legacy devices and future devices. The end result or "why" of their approach is that the increasingly complex pharmaceutical QC lab benefits from improved interoperability and integration through the optimization of resource utilization, material preparation, and workflow throughput and efficiency.

How

While there are viable options for labs (including laboratory information systems [LIS] and laboratory information management systems [LIMS] capable of extensive instrument and data system integration), the "how" of interoperability and integration in today's labs remains challenging. A 2019 article in the American Association for Clinical Chemistry's CLN Stat addressed remaining roadblocks, including lack of standards development, data quality issues, clinical data matching, lack of incentivizing health IT optimization, text-based reporting formats, differences in terminology, and HL7 messaging issues. They add that proposals from the Office of the National Coordinator for Health Information Technology (ONC) and the Centers for Medicare and Medicaid Services include possible fixes such as standardized application programming interfaces (API). They also note that middleware may pick up the slack in connecting more laboratory devices, rather than depending on the LIS to handle all the interfacing.

Many challenges emerge in attempting interoperability and integration through an intelligent automation solution (whether it's middleware or some other laboratory informatics solution) in this environment:

The current level of data integration required to develop an intelligent automation system in real-time settings can be very complex, having to fetch data from many different sources while dealing with poor data quality ... Among the challenges in the development of intelligent automation solutions we consider features such as: interoperability, as the capacity of one system to be seamlessly integrated with others; responsiveness, as the ability to acquire information, analyze and deliver insights immediately; digitalization, as the process of converting information into the digital format, including digital twins and simulation models used to simulate and analyze the behavior of complex systems; traceability, tracking resources and products over their entire life-cycle; decentralization, related to decision and where it is made, through the use of DSSs; flexibility, to ensure the integration of new modules or the adaptation of the existing ones when there are changes in the requirements; security, regarding intellectual property and fail safe mechanisms; interface, as the way information is visualized and communicated; scalability, as the capacity of maintain the level of performance under an increasing workload; and finally, the data characteristics we are dealing with...

As it turns out, the implementation of HL7- and other standard-based interfaces in LIS and LIMS historically has been expensive for many vendors to implement.