Empowering US healthcare delivery organizations: Cultivating a community of practice to harness AI and advance health equity.

Healthcare delivery organizations (HDOs) in the US must contend with the potential for AI to worsen health inequities. But there is no standard set of procedures for HDOs to adopt to navigate these challenges. There is an urgent need for HDOs to present a unified approach to proactively address the potential for AI to worsen health inequities. Amidst this background, Health AI Partnership (HAIP) launched a community of practice to convene stakeholders from across HDOs to tackle challenges related to the use of AI. On February 15, 2023, HAIP hosted an inaugural workshop focused on the question, "Our health care delivery setting is considering adopting a new solution that uses AI. How do we assess the potential future impact on health inequities?" This topic emerged as a common challenge faced by all HDOs participating in HAIP. The workshop had 2 main goals. First, we wanted to ensure participants could talk openly without reservations about challenging topics such as health equity. The second goal was to develop an actionable, generalizable framework that could be immediately put into practice. The workshop engaged 77 participants with 100% representation from all 10 HDOs and invited ecosystem partners. In an accompanying Research Article, we share the Health Equity Across the AI Lifecycle (HEAAL) framework. We invite and encourage HDOs to test the HEAAL framework internally and share feedback so that we can continue to refine and maintain the set of procedures. The HEAAL framework reveals the challenges associated with rigorously assessing the potential for AI to worsen health inequities. Significant investment in personnel, capabilities, and data infrastructure is required, and the level of investment needed could be beyond reach for most HDOs. We look forward to expanding our community of practice to assist HDOs around the world.

Strengthening the use of artificial intelligence within healthcare delivery organizations: balancing regulatory compliance and patient safety.

OBJECTIVES: Surface the urgent dilemma that healthcare delivery organizations (HDOs) face navigating the US Food and Drug Administration (FDA) final guidance on the use of clinical decision support (CDS) software. MATERIALS AND METHODS: We use sepsis as a case study to highlight the patient safety and regulatory compliance tradeoffs that 6129 hospitals in the United States must navigate. RESULTS: Sepsis CDS remains in broad, routine use. There is no commercially available sepsis CDS system that is FDA cleared as a medical device. There is no public disclosure of an HDO turning off sepsis CDS due to regulatory compliance concerns. And there is no public disclosure of FDA enforcement action against an HDO for using sepsis CDS that is not cleared as a medical device. DISCUSSION AND CONCLUSION: We present multiple policy interventions that would relieve the current tension to enable HDOs to utilize artificial intelligence to improve patient care while also addressing FDA concerns about product safety, efficacy, and equity.

Development and preliminary testing of Health Equity Across the AI Lifecycle (HEAAL): A framework for healthcare delivery organizations to mitigate the risk of AI solutions worsening health inequities.

The use of data-driven technologies such as Artificial Intelligence (AI) and Machine Learning (ML) is growing in healthcare. However, the proliferation of healthcare AI tools has outpaced regulatory frameworks, accountability measures, and governance standards to ensure safe, effective, and equitable use. To address these gaps and tackle a common challenge faced by healthcare delivery organizations, a case-based workshop was organized, and a framework was developed to evaluate the potential impact of implementing an AI solution on health equity. The Health Equity Across the AI Lifecycle (HEAAL) is co-designed with extensive engagement of clinical, operational, technical, and regulatory leaders across healthcare delivery organizations and ecosystem partners in the US. It assesses 5 equity assessment domains-accountability, fairness, fitness for purpose, reliability and validity, and transparency-across the span of eight key decision points in the AI adoption lifecycle. It is a process-oriented framework containing 37 step-by-step procedures for evaluating an existing AI solution and 34 procedures for evaluating a new AI solution in total. Within each procedure, it identifies relevant key stakeholders and data sources used to conduct the procedure. HEAAL guides how healthcare delivery organizations may mitigate the potential risk of AI solutions worsening health inequities. It also informs how much resources and support are required to assess the potential impact of AI solutions on health inequities.

Artificial intelligence and frozen section histopathology: A systematic review.

Frozen sections are a useful pathologic tool, but variable image quality may impede the use of artificial intelligence and machine learning in their interpretation. We aimed to identify the current research on machine learning models trained or tested on frozen section images. We searched PubMed and Web of Science for articles presenting new machine learning models published in any year. Eighteen papers met all inclusion criteria. All papers presented at least one novel model trained or tested on frozen section images. Overall, convolutional neural networks tended to have the best performance. When physicians were able to view the output of the model, they tended to perform better than either the model or physicians alone at the tested task. Models trained on frozen sections performed well when tested on other slide preparations, but models trained on only formalin-fixed tissue performed significantly worse across other modalities. This suggests not only that machine learning can be applied to frozen section image processing, but also use of frozen section images may increase model generalizability. Additionally, expert physicians working in concert with artificial intelligence may be the future of frozen section histopathology.

Alinity hq platelet results are equivalent with the international reference method in thrombocytopenic samples.

INTRODUCTION: Accurate and precise platelet (PLT) count is critical for the appropriate management of patients with thrombocytopenia. This study evaluated the performance of PLT counting with the Abbott Alinity hq hematology analyzer, which utilizes multi-dimensional optical technology. METHODS: Imprecision, linearity, and accuracy were assessed per CLSI guidelines. Alinity hq PLT results were compared to the international flow cytometry reference method (IRM) in the concentration range of 6.3 to 103.0 × 109 /L. Additional comparisons were made with Sysmex XN-3000 PLT counts: impedance (PLT-I), optical (PLT-O), and optical fluorescent (PLT-F) methods. RESULTS: The average within-run %CV was 4.7% on patient samples with PLT concentrations ranging from 13.1 to 41.7 × 109 /L, and the within-laboratory %CV was 3.6% at the level of 68.2 × 109 /L. Linearity evaluation indicated a maximum deviation of 3.1% from the linear fit in the range of 0.1 to 316.8 × 109 /L. Comparison between Alinity hq and the IRM PLT counts yielded a correlation coefficient of 0.99 and predicted bias of 0.0 and -0.5 × 109 /L at 10.0 and 20.0 × 109 /L transfusion thresholds, respectively. Alinity hq PLT counts also correlated well with Sysmex PLT counts, with strongest correlation obtained with PLT-F and PLT-O (r = .99) methods. CONCLUSION: This study demonstrated excellent analytical performance of Alinity hq PLT counting in thrombocytopenic samples, equivalency with the IRM and strong agreement with Sysmex PLT-F and PLT-O methods. The Alinity hq multi-dimensional optical PLT count is available with every CBC without additional reagents and may help promote efficiency in clinical laboratories.

Photonic crystals: emerging biosensors and their promise for point-of-care applications.

Biosensors are extensively employed for diagnosing a broad array of diseases and disorders in clinical settings worldwide. The implementation of biosensors at the point-of-care (POC), such as at primary clinics or the bedside, faces impediments because they may require highly trained personnel, have long assay times, large sizes, and high instrumental cost. Thus, there exists a need to develop inexpensive, reliable, user-friendly, and compact biosensing systems at the POC. Biosensors incorporated with photonic crystal (PC) structures hold promise to address many of the aforementioned challenges facing the development of new POC diagnostics. Currently, PC-based biosensors have been employed for detecting a variety of biotargets, such as cells, pathogens, proteins, antibodies, and nucleic acids, with high efficiency and selectivity. In this review, we provide a broad overview of PCs by explaining their structures, fabrication techniques, and sensing principles. Furthermore, we discuss recent applications of PC-based biosensors incorporated with emerging technologies, including telemedicine, flexible and wearable sensing, smart materials and metamaterials. Finally, we discuss current challenges associated with existing biosensors, and provide an outlook for PC-based biosensors and their promise at the POC.

Advances in biosensing strategies for HIV-1 detection, diagnosis, and therapeutic monitoring.

HIV-1 is a major global epidemic that requires sophisticated clinical management. There have been remarkable efforts to develop new strategies for detecting and treating HIV-1, as it has been challenging to translate them into resource-limited settings. Significant research efforts have been recently devoted to developing point-of-care (POC) diagnostics that can monitor HIV-1 viral load with high sensitivity by leveraging micro- and nano-scale technologies. These POC devices can be applied to monitoring of antiretroviral therapy, during mother-to-child transmission, and identification of latent HIV-1 reservoirs. In this review, we discuss current challenges in HIV-1 diagnosis and therapy in resource-limited settings and present emerging technologies that aim to address these challenges using innovative solutions.

Flexible Substrate-Based Devices for Point-of-Care Diagnostics.

Point-of-care (POC) diagnostics play an important role in delivering healthcare, particularly for clinical management and disease surveillance in both developed and developing countries. Currently, the majority of POC diagnostics utilize paper substrates owing to affordability, disposability, and mass production capability. Recently, flexible polymer substrates have been investigated due to their enhanced physicochemical properties, potential to be integrated into wearable devices with wireless communications for personalized health monitoring, and ability to be customized for POC diagnostics. Here, we focus on the latest advances in developing flexible substrate-based diagnostic devices, including paper and polymers, and their clinical applications.

Advances in addressing technical challenges of point-of-care diagnostics in resource-limited settings.

The striking prevalence of HIV, TB and malaria, as well as outbreaks of emerging infectious diseases, such as influenza A (H7N9), Ebola and MERS, poses great challenges for patient care in resource-limited settings (RLS). However, advanced diagnostic technologies cannot be implemented in RLS largely due to economic constraints. Simple and inexpensive point-of-care (POC) diagnostics, which rely less on environmental context and operator training, have thus been extensively studied to achieve early diagnosis and treatment monitoring in non-laboratory settings. Despite great input from material science, biomedical engineering and nanotechnology for developing POC diagnostics, significant technical challenges are yet to be overcome. Summarized here are the technical challenges associated with POC diagnostics from a RLS perspective and the latest advances in addressing these challenges are reviewed.

Functionalized Polymer Microgel Particles Enable Customizable Production of Label-Free Sensor Arrays.

Probe molecule immobilization onto surfaces is a critical step in the production of many analytical devices, including labeled and label-free microarrays. New methods to increase the density and uniformity of probe deposition have the potential to significantly enhance the ultimate limits of detection and reproducibility. Hydrogel-based materials have been employed in the past to provide a 3D protein-friendly surface for deposition of antibodies and nucleic acids. However, these methods are susceptible to variation during polymerization of the hydrogel scaffold and provide limited opportunities for tuning deposition parameters on an antibody-by-antibody basis. In this work, a versatile hydrogel nanoparticle deposition method was developed for the production of label-free microarrays and tested in the context of antibody-antigen binding. Poly(N-isopropylacrylamide) nanoparticles (PNIPAM) were conjugated to antibodies using an avidin/biotin system and deposited onto surfaces using a noncontact printing system. After drying, these gel spots formed uniform and thin layers <10 nm in height. The conjugates were characterized with dynamic light scattering, scanning electron microscopy, and atomic force microscopy. We tested this format in the context of tumor necrosis factor-alpha (TNF-α) detection via arrayed imaging reflectometry (AIR), a label-free protein microarray method. This method of probe molecule deposition should be generally useful in the production of microarrays for label-free detection.

Enhancing the detection limit of nanoscale biosensors via topographically selective functionalization.

Nanoscale biosensors have remarkable theoretical sensitivities but often suffer from suboptimal limits of detection in practice. This is in part because the sensing area of nanoscale sensors is orders of magnitude smaller than the total device substrate. Current strategies to immobilize probes (capture molecules) functionalize both sensing and nonsensing regions, leading to target depletion and diminished limits of detection. The difference in topography between these regions on nanoscale biosensors offers a way to selectively address only the sensing area. We developed a bottom-up, topographically selective approach employing self-assembled poly(N-isopropylacrylamide) (PNIPAM) hydrogel nanoparticles as a mask to preferentially bind target to only the active sensing region of a photonic crystal (PhC) biosensor. This led to over an order of magnitude improvement in the limit of detection for the device, in agreement with finite element simulations. Since the sensing elements in many nanoscale sensors are topographically distinct, this approach should be widely applicable.

Selective virus detection in complex sample matrices with photonic crystal optical cavities.

Rapid, sensitive, and selective detection of viruses is critical for applications in medical diagnostics, biosecurity, and environmental safety. In this article, we report the application of a point-defect-coupled W1 photonic crystal (PhC) waveguide biosensor to label-free optical detection of viruses. Fabricated on a silicon-on-insulator (SOI) substrate using electron-beam (e-beam) lithography and reactive-ion-etching, the PhC sensing platform allows optical detection based on resonant mode shifts in response to ambient refractive index changes produced by infiltration of target biomaterial within the holes of the PhC structure. Finite difference time domain (FDTD) calculations were performed to assist with design of the sensor, and to serve as a theoretical benchmark against which experimental results could be compared. Using Human Papillomavirus virus-like particles (VLPs) spiked in 10% fetal bovine serum as a model system, we observed a limit of detection of 1.5 nM in simple (buffer only) or complex (10% serum) sample matrices. The use of anti-VLP antibodies specific for intact VLPs with the PhC sensors provided highly selective VLP detection.