The Healthcare Landscape is changing
Raouf Hajji, MD, PhD.
HealthTech Futurist | Professor Assistant of Medicine | Co-Founder & Medical Lead of International Medical Community (IMC) | Chief Academic Officer (CAO) of Supernova Academy Inc.
Neuralink, PacBio, GeneDx, and The UK's National Institute for Health and Care Excellence (NICE) AI adoption and other news in the 31st Edition of Healthcare Present & Future.
Academia:
1.?????Academic Medicine journal has developed a principle-driven approach to the use of AI tools in scholarly publishing. While chatbots cannot be credited as authors, authors are asked to adhere to accountability, disclosure, transparency, and evolution principles. The use of AI tools must be disclosed and described in the manuscript's text. The policy will evolve to align with academic standards.
(DeVilbiss, Mary Beth; Roberts, Laura Weiss MD, MA. Artificial Intelligence Tools in Scholarly Publishing: Guidance for Academic Medicine Authors. Academic Medicine 98(8):p 865-866, August 2023. | DOI: 10.1097/ACM.0000000000005261)
2.?????Generative artificial intelligence (GAI) is rapidly becoming a part of our personal and professional lives, including medical education and scholarship. Chatbots like ChatGPT, which simulate human communication, are particularly useful due to their portability, low cost, and ability to create human-like dialogues. However, there are ethical concerns about their use, and critical thinking should not be outsourced without compromising the integrity of faculty and learners. Guidelines should be implemented to govern the appropriate use of GAI, and transparency must be maintained. GAI-generated content will require critical human review to avoid legal liability and reputational harm.
(van de Ridder, J.M. Monica PhD, MSc; Shoja, Mohammadali M. MD; Rajput, Vijay MD, MACP3. Finding the Place of ChatGPT in Medical Education. Academic Medicine 98(8):p 867, August 2023. | DOI: 10.1097/ACM.0000000000005254)
3.?????AI systems like ChatGPT have the potential to revolutionize medical education by providing personalized learning experiences and real-time feedback on decision-making processes. However, it is essential to note that AI technology cannot replace human expertise and judgment, and there is a risk of bias in AI systems that must be actively monitored and addressed. Overall, ChatGPT can be a powerful tool for medical educators to create effective student learning experiences.
(Feng, Songwei; Shen, Yang MD, PhD. ChatGPT and the Future of Medical Education. Academic Medicine 98(8):p 867-868, August 2023. | DOI: 10.1097/ACM.0000000000005242)
4.?????OpenAI has developed ChatGPT, an artificial intelligence tool that can assist physicians in various ways. It can serve as a medical scribe, create mnemonics, simulate clinical scenarios, and help new interns sharpen their diagnostic skills. However, its performance relies heavily on the quality of data inputted into the system, and biased input will result in a biased response. Implementation of ChatGPT in healthcare must prioritize privacy and security concerns. It is recommended that resident physicians gain a comprehensive understanding of how to use this tool effectively and have strict supervision by supervising consultants.(Munaf, Uzair MBBS; Ul-Haque, Ibtehaj MBBS; Arif, Taha Bin MD. ChatGPT: A Helpful Tool for Resident Physicians?. Academic Medicine 98(8):p 868-869, August 2023. | DOI: 10.1097/ACM.0000000000005250
Biomedical Research:
1.?????Deep learning has been successful in vision tasks, and there is growing interest in its use for diagnostic support in skin-related neglected tropical diseases (skin NTDs). This study aimed to develop deep learning models using clinical images for five skin NTDs and to understand how diagnostic accuracy can be improved using different models and training patterns. Two convolutional neural networks were used to examine performance and validate feasibility in diagnosis. The models were able to correctly predict over 70% of diagnoses, with consistent performance improvement with more training samples. The ResNet-50 model performed better than the VGG-16 model. The study demonstrated that the more images used for training, the more accurate the diagnosis became and that the accuracy of PCR-positive cases of Buruli ulcer improved. AI has the potential to address unmet needs where access to medical care is limited.
(Yotsu RR, Ding Z, Hamm J, Blanton RE. Deep learning for AI-based diagnosis of skin-related neglected tropical diseases: A pilot study [published online ahead of print, 2023 Aug 14].?PLoS Negl Trop Dis. 2023;17(8):e0011230. doi:10.1371/journal.pntd.0011230)
2.?????The Virtual Brain (TVB) modeling uses structural and functional MRI to simulate brain networks in Alzheimer's disease and frontotemporal dementia patients. The results showed disease-specific alterations in brain networks and individual differences in network parameters that correlated with neuropsychological, clinical, and pharmacological profiles. These simulations provide new perspectives for understanding dementia mechanisms and designing personalized therapeutic approaches.
(Monteverdi A, Palesi F, Schirner M, et al. Virtual brain simulations reveal network-specific parameters in neurodegenerative dementias. Front Aging Neurosci. 2023;15:1204134. Published 2023 Jul 28. doi:10.3389/fnagi.2023.1204134)
3.?????A remote patient monitoring (RPM) program was launched in October 2021 to improve the quality and efficiency of care for acute kidney injury (AKI) survivors. Patients enrolled in the program were provided with home monitoring technology and underwent weekly laboratory assessments. Nurses evaluated the data daily and adhered to prespecified protocols for management and escalation of care if needed. Twenty patients were enrolled in AKI RPM in the first five months, with a median duration of program participation of 36 days. Eight patients experienced an unplanned readmission or emergency department visit, half of which were attributed to AKI and related circumstances. Of the nine postgraduation survey respondents, all were satisfied with the RPM program, and 89% would recommend it to other patients with similar health conditions. The RPM program can offer a unique opportunity to bridge the care transition from hospital to home and increase access to quality care for AKI survivors. However, its scalability and generalizability may be limited by the availability of infrastructure and resources.
(Charkviani M, Barreto EF, Pearson KK, et al. Development and Implementation of an Acute Kidney Injury Remote Patient Monitoring Program: Research Letter. Can J Kidney Health Dis. 2023;10:20543581231192746. Published 2023 Aug 11. doi:10.1177/20543581231192746)
4.?????A study developed machine learning models to predict in-hospital mortality, acute renal replacement therapy initiation, and mechanical ventilation in patients with acute heart failure receiving furosemide in intensive care units. The models were trained using an extensive database from a Japanese hospital chain and showed good accuracy for predicting the outcomes of interest. However, the optimal models varied depending on the outcome being predicted. Linear support vector machine classification models had the highest accuracy for predicting in-hospital mortality and mechanical ventilation, while a multi-layer neural network had the highest accuracy for predicting the initiation of acute renal replacement therapy. The study concludes that machine learning models can be useful for predicting clinical outcomes in patients with acute heart failure receiving furosemide. Still, the optimal model may differ depending on the outcome of interest.
(Kamio T, Ikegami M, Machida Y, Uemura T, Chino N, Iwagami M. Machine learning-based prognostic modeling of patients with acute heart failure receiving furosemide in intensive care units. Digit Health. 2023;9:20552076231194933. Published 2023 Aug 11. doi:10.1177/20552076231194933)
5.?????Sleep apnea (SA) is a modifiable risk factor in atrial fibrillation (AF) and is associated with worse outcomes in AF. A cross-sectional study was conducted to assess the prevalence and severity of SA in patients with AF without known SA. Participants underwent four consecutive nights of sleep-recording with the home-monitoring device NightOwl? (NO). Moderate to severe SA by NO was highly prevalent in patients with AF without known SA. The study suggests that a home-monitoring device such as NO could be an easy and feasible SA screening tool in patients with AF.
(Jensen MH, Dalgaard F, Rude Laub R, et al. Prevalence of sleep apnea in unselected patients with atrial fibrillation by a home-monitoring device: The DAN-APNO study.?Int J Cardiol Heart Vasc. 2023;47:101219. Published 2023 May 18. doi:10.1016/j.ijcha.2023.101219)
Clinical practice:
1.?????Study on the association between thiazide diuretics and colorectal cancer risk is conflicting. Researchers aimed to compare thiazide diuretic initiators with dihydropyridine calcium channel blockers (dCCBs) and estimated hazard ratios of colorectal cancer. The study found that thiazide diuretics were not associated with an overall increased colorectal cancer risk compared to dCCBs. However, an increased risk was observed among patients with inflammatory bowel disease and potential polyps. Further research is needed to corroborate these findings.
(Rouette J, McDonald EG, Schuster T, Matok I, Brophy JM, Azoulay L. Thiazide Diuretics and Risk of Colorectal Cancer: A Population-Based Cohort Study [published online ahead of print, 2023 Aug 11]. Am J Epidemiol. 2023;kwad171. doi:10.1093/aje/kwad171)
2.?????Urine analysis for tumor-derived peptides and proteins can offer a noninvasive alternative for clinical oncology diagnosis and screening. A study used tandem mass tags-based multiplexed mass spectrometry to perform comparative quantitative proteomic profiling of urine samples from gastric cancer patients and healthy controls. The study identified 1504 proteins, of which 246 were differentially expressed in gastric cancer cases. Some of the upregulated proteins, such as EFNA1, PGA3, SORT1, and VTN, are known to play crucial roles in gastric cancer progression. The study also discovered other overexpressed proteins, including SHISA5, MUCL1, and LECT2, which had not previously been linked to gastric cancer. The study validated changes in protein abundance using a novel approach for targeted proteomics called SureQuant. The findings provide molecular candidates for biomarker development in gastric cancer and support the potential of urinary proteomics for noninvasive diagnostics and personalized/precision medicine in the oncology clinic.
(Joshi N, Bhat F, Bellad A, et al. Urinary Proteomics for Discovery of Gastric Cancer Biomarkers to Enable Precision Clinical Oncology [published online ahead of print, 2023 Aug 9]. OMICS. 2023;10.1089/omi.2023.0077. doi:10.1089/omi.2023.0077)
3.?????CGM technology has become the standard of care for patients with type 1 and type 2 diabetes using insulin therapy. Hospitals have noticed an increased use of these devices in their hospitalized patients, especially during the COVID-19 pandemic. Studies have demonstrated the efficacy and safety of CGM use in the hospital, leading to clinical practice guideline recommendations. CGM has the potential to become the standard of care for some hospitalized patients, overcoming the limitations of current capillary glucose testing. This review provides a historical overview of the evolution of glycemic monitoring in the hospital and reviews the current evidence, implementation protocols, and guidance for the use of CGM in hospitalized patients.
(Zelada H, Perez-Guzman MC, Chernavvsky DR, Galindo RJ. Continuous glucose monitoring (CGM) for inpatient diabetes management: an update on the current evidence and practice [published online ahead of print, 2023 Aug 1].?Endocr Connect. 2023;EC-23-0180. doi:10.1530/EC-23-0180)
4. A clinical prediction rule for acute appendicitis in children presenting with acute abdominal pain in primary care was developed and validated using data from GP electronic health records. The model included male sex, pain duration, nausea/vomiting, elevated temperature, abnormal bowel sounds, right lower quadrant tenderness, and peritoneal irritation. The rule stratifies the risk of appendicitis into three groups and could improve clinical decision-making and outcomes
Figure 2.?Predicted probability of acute appendicitis by point score the predicted probability of acute appendicitis in the low-risk (green), medium-risk (orange) and high-risk (red) groups, based on the final model, is juxtaposed against the point score. For example, in a boy (2 points) with fever (1 point), vomiting (2 points), pain duration between 24 and 48 h (2 points), absent bowel sounds (1 point), and peritoneal irritation (4 points), but without right lower quadrant tenderness, the point score is 11. As the graph shows, this corresponds to a predicted probability of acute appendicitis of 0.12.
(Blok G, Burger H, van der Lei J, Berger M, Holtman G. Development and validation of a clinical prediction rule for acute appendicitis in children in primary care.?Eur J Gen Pract. 2023;29(1):2233053. doi:10.1080/13814788.2023.2233053)
Emerging technologies:
1.?????Boston Scientific has received FDA approval for its POLARx Cryoablation System, a heart ablation device used to treat atrial fibrillation. The device can accommodate two balloon sizes in one catheter, allowing physicians to tailor care to individual patients. It has been selling well in Europe and Japan, where it is cleared for use. The device uses a balloon catheter to freeze tissue near the pulmonary vein, blocking irregular electric signals that can cause atrial fibrillation. Boston Scientific's electrophysiology sales grew 28% in Q2, driven by the strong international performance of the POLARx and FARAPULSE cardiac ablation devices. Competitors such as Johnson & Johnson and Medtronic are also bringing new cardiac ablation devices to market.
?
2.?????Neuralink, Elon Musk's brain implant company, has raised $280 million in a series D funding round led by Founders Fund, Peter Thiel's venture capital firm. The funds will be used to address regulatory concerns and secure authorization for a clinical trial of the device. Neuralink's technology is being pitched as a treatment for conditions such as blindness and as a way to augment humans to compete against artificial intelligence. The FDA has reportedly rejected Neuralink's first application to test the device in humans over concerns about its wires moving in the brain and its lithium battery failing.
3.?????The UK's National Institute for Health and Care Excellence (NICE) has recommended the use of artificial intelligence (AI) to speed up radiotherapy treatment planning. The move is expected to save time and money, allowing healthcare professionals to focus on complex cases. A trained healthcare professional must still review the contours created by AI before being used in treatment planning. Evidence suggests that using AI is quicker than manual contouring, with similar quality contours produced. The change is not expected to affect patient outcomes. The recommendation is the first of its kind from NICE. More evidence will be generated over the next three years to carry out a full cost/benefit analysis.
4.?????PacBio and GeneDx are collaborating with the University of Washington to study the effectiveness of long-read whole genome sequencing in increasing diagnostic rates for pediatric patients with genetic conditions. The study will compare the diagnostic rates of short- and long-read sequencing platforms and will include WGS sequencing and analysis of samples from 350 people, including 120 enrolled in the SeqFirst WGS study at Seattle Children's Hospital. GeneDx will use the PacBio Revio sequencing system to perform all long-read WGS sequencing and analysis for this study. The study intends to contribute to the scientific understanding of variant prevalence and classification. Google has expanded upon a previous technical collaboration with PacBio to contribute funding for this research study and will fund key enablers for the project.
5.?????Scottish Brain Sciences and Roche Diagnostics are collaborating on a series of major projects to understand the earliest biological changes of neurodegenerative diseases using blood-based biomarkers. The aim is to develop accurate diagnostic and prognostic tests for Alzheimer's disease, which can identify the earliest signs of the disease long before symptoms appear. Early detection of brain changes associated with neurodegenerative diseases will transform the assessment, management, and conceptualization of Alzheimer's disease, leading to better-targeted interventions and even the prevention of late-stage dementia syndromes. The collaboration is a huge vote of confidence in the Scottish life sciences sector and the country's leadership in the brain health space. Roche Diagnostics is driving innovation to deliver a future where an early and accurate diagnosis is available to benefit all individuals and families affected by Alzheimer's disease.
Autodidacte & Polymathe ? Chargé d'intelligence économique ? AI hobbyist ethicist - ISO42001 ? éditorialiste & Veille stratégique - Muse? & Times of AI ? Techno humaniste & Techno optimiste ?
1 年AI Muse? Grenoble ???? ????
Realtor Associate @ Next Trend Realty LLC | HAR REALTOR, IRS Tax Preparer
1 年Thanks for Sharing.