-
Health Care Science Feb 2024Given the strikingly high diagnostic error rate in hospitals, and the recent development of Large Language Models (LLMs), we set out to measure the diagnostic...
BACKGROUND
Given the strikingly high diagnostic error rate in hospitals, and the recent development of Large Language Models (LLMs), we set out to measure the diagnostic sensitivity of two popular LLMs: GPT-4 and PaLM2. Small-scale studies to evaluate the diagnostic ability of LLMs have shown promising results, with GPT-4 demonstrating high accuracy in diagnosing test cases. However, larger evaluations on real electronic patient data are needed to provide more reliable estimates.
METHODS
To fill this gap in the literature, we used a deidentified Electronic Health Record (EHR) data set of about 300,000 patients admitted to the Beth Israel Deaconess Medical Center in Boston. This data set contained blood, imaging, microbiology and vital sign information as well as the patients' medical diagnostic codes. Based on the available EHR data, doctors curated a set of diagnoses for each patient, which we will refer to as ground truth diagnoses. We then designed carefully-written prompts to get patient diagnostic predictions from the LLMs and compared this to the ground truth diagnoses in a random sample of 1000 patients.
RESULTS
Based on the proportion of correctly predicted ground truth diagnoses, we estimated the diagnostic hit rate of GPT-4 to be 93.9%. PaLM2 achieved 84.7% on the same data set. On these 1000 randomly selected EHRs, GPT-4 correctly identified 1116 unique diagnoses.
CONCLUSION
The results suggest that artificial intelligence (AI) has the potential when working alongside clinicians to reduce cognitive errors which lead to hundreds of thousands of misdiagnoses every year. However, human oversight of AI remains essential: LLMs cannot replace clinicians, especially when it comes to human understanding and empathy. Furthermore, a significant number of challenges in incorporating AI into health care exist, including ethical, liability and regulatory barriers.
PubMed: 38939167
DOI: 10.1002/hcs2.79 -
Cureus Jun 2024Autoimmune hepatitis (AIH) is a condition resulting in chronic, inflammatory changes to the liver. Primary biliary cholangitis (PBC) is an autoimmune condition that...
Autoimmune hepatitis (AIH) is a condition resulting in chronic, inflammatory changes to the liver. Primary biliary cholangitis (PBC) is an autoimmune condition that destroys intrahepatic bile ducts. Overlap syndrome with concomitant AIH and PBC comprises a rare subgroup of patients with immune-mediated liver disease, with incidence rates of male patients being exceedingly uncommon in a predominantly female patient population. Our case report investigates a rare case of a 41-year-old male patient diagnosed with overlapping AIH and PBC. He initially presented with symptoms of fatigue, pruritus, and episodes of Raynaud's phenomenon, in addition to findings of persistently elevated liver enzymes despite lifestyle modifications. He had no past medical history, no history of alcohol use disorder, and no family medical history of chronic liver disease. Imaging did not reveal evidence of cirrhosis. Further diagnostic workup was significant for elevated immunologic markers for antinuclear antibodies (ANA) with positive centromere and cytoplasmic patterns, antimitochondrial antibodies (AMA) with F-actin antibodies, anti-smooth muscle antibodies (ASMA), and cytoplasmic antinuclear cytoplasmic antibodies (ANCA C). Liver biopsy showed prominent plasma cells and rare granulomas, consistent with the diagnosis of AIH with a component of PBC, respectively. He was started on ursodeoxycholic acid (UDCA), demonstrating a near-complete clinical response with resolution of symptoms and normalization of liver enzymes. Studies investigating the low incidence of male patients with overlap syndrome are limited, as current research is overwhelmingly based on studies with predominantly female subjects. However, most studies generally recommend treatment with both UDCA and corticosteroids to reduce symptoms and biochemical markers. Our case report highlights a rare case of a male patient documenting excellent biochemical and clinical responses to monotherapy with UDCA. A possible theory is that our patient's early treatment (prior to advanced disease progression) is associated with his near-complete biochemical response and symptomatic resolution on UDCA alone. Further research is needed to fully understand the clinical course and long-term prognosis of male patients with overlap syndrome. Our patient remains in life-long follow-up to monitor if or when he requires treatment with corticosteroids in addition to current monotherapy with UDCA..
PubMed: 38938909
DOI: 10.7759/cureus.63312 -
Health Technology Assessment... Jun 2024Health economic assessments are used to determine whether the resources needed to generate net benefit from an antenatal or newborn screening programme, driven by...
BACKGROUND
Health economic assessments are used to determine whether the resources needed to generate net benefit from an antenatal or newborn screening programme, driven by multiple benefits and harms, are justifiable. It is not known what benefits and harms have been adopted by economic evaluations assessing these programmes and whether they omit benefits and harms considered important to relevant stakeholders.
OBJECTIVES
(1) To identify the benefits and harms adopted by health economic assessments in this area, and to assess how they have been measured and valued; (2) to identify attributes or relevance to stakeholders that ought to be considered in future economic assessments; and (3) to make recommendations about the benefits and harms that should be considered by these studies.
DESIGN
Mixed methods combining systematic review and qualitative work.
SYSTEMATIC REVIEW METHODS
We searched the published and grey literature from January 2000 to January 2021 using all major electronic databases. Economic evaluations of an antenatal or newborn screening programme in one or more Organisation for Economic Co-operation and Development countries were considered eligible. Reporting quality was assessed using the Consolidated Health Economic Evaluation Reporting Standards checklist. We identified benefits and harms using an integrative descriptive analysis and constructed a thematic framework.
QUALITATIVE METHODS
We conducted a meta-ethnography of the existing literature on newborn screening experiences, a secondary analysis of existing individual interviews related to antenatal or newborn screening or living with screened-for conditions, and a thematic analysis of primary data collected with stakeholders about their experiences with screening.
RESULTS
The literature searches identified 52,244 articles and reports, and 336 unique studies were included. Thematic framework resulted in seven themes: (1) diagnosis of screened for condition, (2) life-years and health status adjustments, (3) treatment, (4) long-term costs, (5) overdiagnosis, (6) pregnancy loss and (7) spillover effects on family members. Diagnosis of screened-for condition (115, 47.5%), life-years and health status adjustments (90, 37.2%) and treatment (88, 36.4%) accounted for most of the benefits and harms evaluating antenatal screening. The same themes accounted for most of the benefits and harms included in studies assessing newborn screening. Long-term costs, overdiagnosis and spillover effects tended to be ignored. The wide-reaching family implications of screening were considered important to stakeholders. We observed good overlap between the thematic framework and the qualitative evidence.
LIMITATIONS
Dual data extraction within the systematic literature review was not feasible due to the large number of studies included. It was difficult to recruit healthcare professionals in the stakeholder's interviews.
CONCLUSIONS
There is no consistency in the selection of benefits and harms used in health economic assessments in this area, suggesting that additional methods guidance is needed. Our proposed thematic framework can be used to guide the development of future health economic assessments evaluating antenatal and newborn screening programmes.
STUDY REGISTRATION
This study is registered as PROSPERO CRD42020165236.
FUNDING
This award was funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme (NIHR award ref: NIHR127489) and is published in full in ; Vol. 28, No. 25. See the NIHR Funding and Awards website for further award information.
Topics: Humans; Infant, Newborn; Neonatal Screening; Cost-Benefit Analysis; Female; Pregnancy; Qualitative Research; Technology Assessment, Biomedical; Prenatal Diagnosis; Quality-Adjusted Life Years
PubMed: 38938110
DOI: 10.3310/PYTK6591 -
Muscle & Nerve Jun 2024Total supported abduction, or TSA, is a position for ultrasound evaluations and guided interventions of the upper extremity. It provides optimal access to the medial arm...
Total supported abduction, or TSA, is a position for ultrasound evaluations and guided interventions of the upper extremity. It provides optimal access to the medial arm through the volar wrist and palmar hand for diagnostic evaluations of the median and ulnar nerves as well as procedures including injections for carpal tunnel syndrome, ulnar neuropathy at the elbow, and stenosing tenosynovitis. It enables ease of both ipsilateral and bilateral evaluations/interventions without the need for significant positional changes by the patient or physician. Incorporation of TSA may enhance clinical efficiency by reducing the amount of time, materials, and space required to provide such services.
PubMed: 38938077
DOI: 10.1002/mus.28190 -
Skin Research and Technology : Official... Jul 2024Patients with myotonic muscular dystrophy (MMD) were observed to have numerous basal cell carcinoma (BCC) and abnormal dysplastic nevi (DN) on non-sun exposed skin.... (Review)
Review
OBJECTIVE
Patients with myotonic muscular dystrophy (MMD) were observed to have numerous basal cell carcinoma (BCC) and abnormal dysplastic nevi (DN) on non-sun exposed skin. Simultaneously a large study published in the Journal of American Medical Association (JAMA) illustrated that patients with MMD have "overall" an increased risk for cancer development. Based on these findings, this author in 2010 postulated that dysregulation of RNA binding proteins (RBP), responsible for clinical manifestations of MMD, is also responsible for the development of BCC and melanoma.
METHODS
To report new research elucidating the etiology of melanoma, BCC, MMD-induced cancers, and potentially other environmentally induced malignancies.
RESULTS
Dysregulation of RBP induces aberrant mRNA splicing; recent data indicates that abnormal mRNA splicing not just plays a key role in the pathogenesis of melanoma but is a hallmark of essentially all human malignancies.
CONCLUSION
The author's hypothesis is that ultraviolet (UV) radiation induces DNA damage in intronic regions of a variety of genes. Furthermore, these UV-induced abnormal DNA dimers, repeats and mutations interfere with normal mRNA splicing thus producing abnormal proteins. These abnormal proteins in turn activate oncogenic pathways such as hedgehog, MAP kinase, and WNT.
Topics: Humans; Skin Neoplasms; Melanoma; Carcinoma, Basal Cell; Genetic Predisposition to Disease; Genetic Testing; Myotonic Dystrophy; Ultraviolet Rays
PubMed: 38937899
DOI: 10.1111/srt.13832 -
Journal of Medical Case Reports Jun 2024Since the beginning of the coronavirus disease 2019 pandemic, the most common skin lesions observed due to infection with the severe acute respiratory syndrome...
BACKGROUND
Since the beginning of the coronavirus disease 2019 pandemic, the most common skin lesions observed due to infection with the severe acute respiratory syndrome coronavirus 2 are pseudochilblains (or coronavirus disease toes). However, this pathology remains infrequent and difficult to diagnose, as no specific test exists.
CASE PRESENTATION
Two Caucasian women, 30 and 22 years old, presented to our General Medicine Unit with perniosis lesions on the feet during the first two waves of the coronavirus disease 2019 pandemic. They did not have respiratory or general symptoms of severe acute respiratory syndrome coronavirus 2 infection, the reverse transcription polymerase chain reaction on nasopharyngeal swabs was negative, and the serology was positive only in the first case. The clinical presentation differed for the two cases, as the second patient suffered from swelling and burning after cold application. The diagnosis was based on clinical presentation, temporality, exclusion of other differential diagnoses, and blood test results (positive serology in the first case and high level of CXCL13 and VEGF in the second), supported by current literature. Lesions resolved spontaneously in the first patient. The second case was hospitalized for pain management and received corticosteroid therapy with resolution of the symptoms.
CONCLUSION
These two cases with different clinical presentations illustrate the diagnostic approach to coronavirus disease 2019, a challenging disease with diverse manifestations, including, in some cases, coronavirus disease toes. We present a literature review that illustrates the progression of scientific research. Skin lesions associated with coronavirus disease 2019 infection could be the expression of an important interferon type 1 response and should be considered in the differential diagnosis in a primary care setting.
Topics: Humans; COVID-19; Female; Adult; Toes; SARS-CoV-2; Chilblains; Young Adult; Diagnosis, Differential; Outpatients
PubMed: 38937799
DOI: 10.1186/s13256-024-04626-9 -
Lipids in Health and Disease Jun 2024Traumatic brain injury (TBI) causes neuroinflammation and can lead to long-term neurological dysfunction, even in cases of mild TBI (mTBI). Despite the substantial...
BACKGROUND
Traumatic brain injury (TBI) causes neuroinflammation and can lead to long-term neurological dysfunction, even in cases of mild TBI (mTBI). Despite the substantial burden of this disease, the management of TBI is precluded by an incomplete understanding of its cellular mechanisms. Sphingolipids (SPL) and their metabolites have emerged as key orchestrators of biological processes related to tissue injury, neuroinflammation, and inflammation resolution. No study so far has investigated comprehensive sphingolipid profile changes immediately following TBI in animal models or human cases. In this study, sphingolipid metabolite composition was examined during the acute phases in brain tissue and plasma of mice following mTBI.
METHODS
Wildtype mice were exposed to air-blast-mediated mTBI, with blast exposure set at 50-psi on the left cranium and 0-psi designated as Sham. Sphingolipid profile was analyzed in brain tissue and plasma during the acute phases of 1, 3, and 7 days post-TBI via liquid-chromatography-mass spectrometry. Simultaneously, gene expression of sphingolipid metabolic markers within brain tissue was analyzed using quantitative reverse transcription-polymerase chain reaction. Significance (P-values) was determined by non-parametric t-test (Mann-Whitney test) and by Tukey's correction for multiple comparisons.
RESULTS
In post-TBI brain tissue, there was a significant elevation of 1) acid sphingomyelinase (aSMase) at 1- and 3-days, 2) neutral sphingomyelinase (nSMase) at 7-days, 3) ceramide-1-phosphate levels at 1 day, and 4) monohexosylceramide (MHC) and sphingosine at 7-days. Among individual species, the study found an increase in C18:0 and a decrease in C24:1 ceramides (Cer) at 1 day; an increase in C20:0 MHC at 3 days; decrease in MHC C18:0 and increase in MHC C24:1, sphingomyelins (SM) C18:0, and C24:0 at 7 days. Moreover, many sphingolipid metabolic genes were elevated at 1 day, followed by a reduction at 3 days and an absence at 7-days post-TBI. In post-TBI plasma, there was 1) a significant reduction in Cer and MHC C22:0, and an increase in MHC C16:0 at 1 day; 2) a very significant increase in long-chain Cer C24:1 accompanied by significant decreases in Cer C24:0 and C22:0 in MHC and SM at 3 days; and 3) a significant increase of C22:0 in all classes of SPL (Cer, MHC and SM) as well as a decrease in Cer C24:1, MHC C24:1 and MHC C24:0 at 7 days.
CONCLUSIONS
Alterations in sphingolipid metabolite composition, particularly sphingomyelinases and short-chain ceramides, may contribute to the induction and regulation of neuroinflammatory events in the early stages of TBI, suggesting potential targets for novel diagnostic, prognostic, and therapeutic strategies in the future.
Topics: Animals; Mice; Sphingolipids; Brain; Ceramides; Sphingomyelin Phosphodiesterase; Sphingosine; Disease Models, Animal; Male; Sphingomyelins; Brain Concussion; Mice, Inbred C57BL; Brain Injuries, Traumatic; Lysophospholipids
PubMed: 38937745
DOI: 10.1186/s12944-024-02186-x -
Integrated surveillance systems for antibiotic resistance in a One Health context: a scoping review.BMC Public Health Jun 2024Antibiotic resistance (ABR) has emerged as a major threat to health. Properly informed decisions to mitigate this threat require surveillance systems that integrate... (Review)
Review
BACKGROUND
Antibiotic resistance (ABR) has emerged as a major threat to health. Properly informed decisions to mitigate this threat require surveillance systems that integrate information on resistant bacteria and antibiotic use in humans, animals, and the environment, in line with the One Health concept. Despite a strong call for the implementation of such integrated surveillance systems, we still lack a comprehensive overview of existing organizational models for integrated surveillance of ABR. To address this gap, we conducted a scoping review to characterize existing integrated surveillance systems for ABR.
METHODS
The literature review was conducted using the PRISMA guidelines. The selected integrated surveillance systems were assessed according to 39 variables related to their organization and functioning, the socio-economic and political characteristics of their implementation context, and the levels of integration reached, together with their related outcomes. We conducted two distinct, complementary analyses on the data extracted: a descriptive analysis to summarize the characteristics of the integrated surveillance systems, and a multiple-correspondence analysis (MCA) followed by a hierarchical cluster analysis (HCA) to identify potential typology for surveillance systems.
RESULTS
The literature search identified a total of 1330 records. After the screening phase, 59 references were kept from which 14 integrated surveillance systems were identified. They all operate in high-income countries and vary in terms of integration, both at informational and structural levels. The different systems combine information from a wide range of populations and commodities -in the human, animal and environmental domains, collection points, drug-bacterium pairs, and rely on various diagnostic and surveillance strategies. A variable level of collaboration was found for the governance and/or operation of the surveillance activities. The outcomes of integration are poorly described and evidenced. The 14 surveillance systems can be grouped into four distinct clusters, characterized by integration level in the two dimensions. The level of resources and regulatory framework in place appeared to play a major role in the establishment and organization of integrated surveillance.
CONCLUSIONS
This study suggests that operationalization of integrated surveillance for ABR is still not well established at a global scale, especially in low and middle-income countries and that the surveillance scope is not broad enough to obtain a comprehensive understanding of the complex dynamics of ABR to appropriately inform mitigation measures. Further studies are needed to better characterize the various integration models for surveillance with regard to their implementation context and evaluate the outcome of these models.
Topics: Humans; One Health; Drug Resistance, Microbial; Anti-Bacterial Agents; Population Surveillance
PubMed: 38937706
DOI: 10.1186/s12889-024-19158-6 -
Scientific Reports Jun 2024Accurate screening of COVID-19 infection status for symptomatic patients is a critical public health task. Although molecular and antigen tests now exist for COVID-19,... (Comparative Study)
Comparative Study
Accurate screening of COVID-19 infection status for symptomatic patients is a critical public health task. Although molecular and antigen tests now exist for COVID-19, in resource-limited settings, screening tests are often not available. Furthermore, during the early stages of the pandemic tests were not available in any capacity. We utilized an automated machine learning (ML) approach to train and evaluate thousands of models on a clinical dataset consisting of commonly available clinical and laboratory data, along with cytokine profiles for patients (n = 150). These models were then further tested for generalizability on an out-of-sample secondary dataset (n = 120). We were able to develop a ML model for rapid and reliable screening of patients as COVID-19 positive or negative using three approaches: commonly available clinical and laboratory data, a cytokine profile, and a combination of the common data and cytokine profile. Of the tens of thousands of models automatically tested for the three approaches, all three approaches demonstrated > 92% sensitivity and > 88 specificity while our highest performing model achieved 95.6% sensitivity and 98.1% specificity. These models represent a potential effective deployable solution for COVID-19 status classification for symptomatic patients in resource-limited settings and provide proof-of-concept for rapid development of screening tools for novel emerging infectious diseases.
Topics: Humans; COVID-19; Machine Learning; Cytokines; SARS-CoV-2; Mass Screening; Male; Female; Sensitivity and Specificity; Middle Aged; Adult; Aged
PubMed: 38937503
DOI: 10.1038/s41598-024-63707-3 -
BMJ Case Reports Jun 2024A man in his 70s presented with a sudden onset stabbing back pain radiating to the chest and pre-syncopal symptoms. He underwent urgent investigations, including a CT...
A man in his 70s presented with a sudden onset stabbing back pain radiating to the chest and pre-syncopal symptoms. He underwent urgent investigations, including a CT angiogram aorta which did not reveal any abnormalities within the thorax, abdomen or pelvis and no cause of symptoms was identified. After being discharged, he re-presented 2 days later with syncopal episodes, abdominal pain and a significant drop in haemoglobin levels. This time, a CT mesenteric angiogram showed two hepatic artery pseudoaneurysms and a large haemoperitoneum. Following a hepatic artery embolisation, a workup showed that the likely cause of the pseudoaneurysms was a rare first presentation of polyarteritis nodosa. This case highlights the importance of considering the possibility of an aneurysmal rupture, especially when common causes of an acute abdomen have been excluded, and not relying on previous negative investigations to exclude pathology, as the outcomes can be detrimental.
Topics: Humans; Polyarteritis Nodosa; Aneurysm, False; Male; Hepatic Artery; Aged; Embolization, Therapeutic; Aneurysm, Ruptured; Computed Tomography Angiography; Rupture, Spontaneous; Hemoperitoneum; Abdominal Pain
PubMed: 38937262
DOI: 10.1136/bcr-2023-257411