-
The Korean Journal of Internal Medicine Sep 2019About 21% of adults with osteoarthritis (OA) are diagnosed with concomitant depression in addition to chronic pain. Duloxetine, an anti-depressant medication, has been... (Meta-Analysis)
Meta-Analysis
About 21% of adults with osteoarthritis (OA) are diagnosed with concomitant depression in addition to chronic pain. Duloxetine, an anti-depressant medication, has been recently approved for managing Knee OA. We performed a systematic review to ascertain the efficacy and safety of duloxetine for OA. We searched MEDLINE, EMBASE, Web of Science, Google Scholar, and the Cochrane Database from inception to December 2018. Randomized clinical trials (RCTs) assessing the efficacy and/or safety of duloxetine versus placebo in OA patients were included. Data extraction and quality assessment were undertaken by two independent reviewers. Seven RCTs (n = 2,102 participants) met our inclusion criteria, and five RCTs (n = 1,713) were eligible for meta-analysis. The results of our analyses indicate that duloxetine has statistically significant, moderate benefits on pain, function, and quality of life in knee OA patients for up to 13 weeks. Reported incidences of gastrointestinal adverse events were three to four times higher in participants who received duloxetine versus placebo. Duloxetine may be an effective treatment option for individuals with knee OA, but use of the drug is associated with a significantly higher risk of adverse events. Patient preferences and clinicians' judgment must be considered before the initiation of duloxetine.
Topics: Aged; Antirheumatic Agents; Duloxetine Hydrochloride; Female; Humans; Male; Middle Aged; Osteoarthritis, Knee; Quality of Life; Remission Induction; Risk Factors; Treatment Outcome
PubMed: 30871298
DOI: 10.3904/kjim.2018.460 -
Diagnostics (Basel, Switzerland) Jan 2024Artificial intelligence (AI) has emerged as a transformative force in various sectors, including medicine and healthcare. Large language models like ChatGPT showcase... (Review)
Review
Artificial intelligence (AI) has emerged as a transformative force in various sectors, including medicine and healthcare. Large language models like ChatGPT showcase AI's potential by generating human-like text through prompts. ChatGPT's adaptability holds promise for reshaping medical practices, improving patient care, and enhancing interactions among healthcare professionals, patients, and data. In pandemic management, ChatGPT rapidly disseminates vital information. It serves as a virtual assistant in surgical consultations, aids dental practices, simplifies medical education, and aids in disease diagnosis. A total of 82 papers were categorised into eight major areas, which are G1: treatment and medicine, G2: buildings and equipment, G3: parts of the human body and areas of the disease, G4: patients, G5: citizens, G6: cellular imaging, radiology, pulse and medical images, G7: doctors and nurses, and G8: tools, devices and administration. Balancing AI's role with human judgment remains a challenge. A systematic literature review using the PRISMA approach explored AI's transformative potential in healthcare, highlighting ChatGPT's versatile applications, limitations, motivation, and challenges. In conclusion, ChatGPT's diverse medical applications demonstrate its potential for innovation, serving as a valuable resource for students, academics, and researchers in healthcare. Additionally, this study serves as a guide, assisting students, academics, and researchers in the field of medicine and healthcare alike.
PubMed: 38201418
DOI: 10.3390/diagnostics14010109 -
The Cochrane Database of Systematic... Sep 2021Monoclonal antibodies (mAbs) are laboratory-produced molecules derived from the B cells of an infected host. They are being investigated as a potential therapy for... (Review)
Review
BACKGROUND
Monoclonal antibodies (mAbs) are laboratory-produced molecules derived from the B cells of an infected host. They are being investigated as a potential therapy for coronavirus disease 2019 (COVID-19).
OBJECTIVES
To assess the effectiveness and safety of SARS-CoV-2-neutralising mAbs for treating patients with COVID-19, compared to an active comparator, placebo, or no intervention. To maintain the currency of the evidence, we will use a living systematic review approach. A secondary objective is to track newly developed SARS-CoV-2-targeting mAbs from first tests in humans onwards. SEARCH METHODS: We searched MEDLINE, Embase, the Cochrane COVID-19 Study Register, and three other databases on 17 June 2021. We also checked references, searched citations, and contacted study authors to identify additional studies. Between submission and publication, we conducted a shortened randomised controlled trial (RCT)-only search on 30 July 2021.
SELECTION CRITERIA
We included studies that evaluated SARS-CoV-2-neutralising mAbs, alone or combined, compared to an active comparator, placebo, or no intervention, to treat people with COVID-19. We excluded studies on prophylactic use of SARS-CoV-2-neutralising mAbs.
DATA COLLECTION AND ANALYSIS
Two authors independently assessed search results, extracted data, and assessed risk of bias using the Cochrane risk of bias tool (RoB2). Prioritised outcomes were all-cause mortality by days 30 and 60, clinical progression, quality of life, admission to hospital, adverse events (AEs), and serious adverse events (SAEs). We rated the certainty of evidence using GRADE.
MAIN RESULTS
We identified six RCTs that provided results from 17,495 participants with planned completion dates between July 2021 and December 2031. Target sample sizes varied from 1020 to 10,000 participants. Average age was 42 to 53 years across four studies of non-hospitalised participants, and 61 years in two studies of hospitalised participants. Non-hospitalised individuals with COVID-19 Four studies evaluated single agents bamlanivimab (N = 465), sotrovimab (N = 868), regdanvimab (N = 307), and combinations of bamlanivimab/etesevimab (N = 1035), and casirivimab/imdevimab (N = 799). We did not identify data for mortality at 60 days or quality of life. Our certainty of the evidence is low for all outcomes due to too few events (very serious imprecision). Bamlanivimab compared to placebo No deaths occurred in the study by day 29. There were nine people admitted to hospital by day 29 out of 156 in the placebo group compared with one out of 101 in the group treated with 0.7 g bamlanivimab (risk ratio (RR) 0.17, 95% confidence interval (CI) 0.02 to 1.33), 2 from 107 in the group treated with 2.8 g (RR 0.32, 95% CI 0.07 to 1.47) and 2 from 101 in the group treated with 7.0 g (RR 0.34, 95% CI 0.08 to 1.56). Treatment with 0.7 g, 2.8 g and 7.0 g bamlanivimab may have similar rates of AEs as placebo (RR 0.99, 95% CI 0.66 to 1.50; RR 0.90, 95% CI 0.59 to 1.38; RR 0.81, 95% CI 0.52 to 1.27). The effect on SAEs is uncertain. Clinical progression/improvement of symptoms or development of severe symptoms were not reported. Bamlanivimab/etesevimab compared to placebo There were 10 deaths in the placebo group and none in bamlanivimab/etesevimab group by day 30 (RR 0.05, 95% CI 0.00 to 0.81). Bamlanivimab/etesevimab may decrease hospital admission by day 29 (RR 0.30, 95% CI 0.16 to 0.59), may result in a slight increase in any grade AEs (RR 1.15, 95% CI 0.83 to 1.59) and may increase SAEs (RR 1.40, 95% CI 0.45 to 4.37). Clinical progression/improvement of symptoms or development of severe symptoms were not reported. Casirivimab/imdevimab compared to placebo Casirivimab/imdevimab may reduce hospital admissions or death (2.4 g: RR 0.43, 95% CI 0.08 to 2.19; 8.0 g: RR 0.21, 95% CI 0.02 to 1.79). We are uncertain of the effect on grades 3-4 AEs (2.4 g: RR 0.76, 95% CI 0.17 to 3.37; 8.0 g: RR 0.50, 95% CI 0.09 to 2.73) and SAEs (2.4 g: RR 0.68, 95% CI 0.19 to 2.37; 8.0 g: RR 0.34, 95% CI 0.07 to 1.65). Mortality by day 30 and clinical progression/improvement of symptoms or development of severe symptoms were not reported. Sotrovimab compared to placebo We are uncertain whether sotrovimab has an effect on mortality (RR 0.33, 95% CI 0.01 to 8.18) and invasive mechanical ventilation (IMV) requirement or death (RR 0.14, 95% CI 0.01 to 2.76). Treatment with sotrovimab may reduce the number of participants with oxygen requirement (RR 0.11, 95 % CI 0.02 to 0.45), hospital admission or death by day 30 (RR 0.14, 95% CI 0.04 to 0.48), grades 3-4 AEs (RR 0.26, 95% CI 0.12 to 0.60), SAEs (RR 0.27, 95% CI 0.12 to 0.63) and may have little or no effect on any grade AEs (RR 0.87, 95% CI 0.66 to 1.16). Regdanvimab compared to placebo Treatment with either dose (40 or 80 mg/kg) compared with placebo may decrease hospital admissions or death (RR 0.45, 95% CI 0.14 to 1.42; RR 0.56, 95% CI 0.19 to 1.60, 206 participants), but may increase grades 3-4 AEs (RR 2.62, 95% CI 0.52 to 13.12; RR 2.00, 95% CI 0.37 to 10.70). 80 mg/kg may reduce any grade AEs (RR 0.79, 95% CI 0.52 to 1.22) but 40 mg/kg may have little to no effect (RR 0.96, 95% CI 0.64 to 1.43). There were too few events to allow meaningful judgment for the outcomes mortality by 30 days, IMV requirement, and SAEs. Hospitalised individuals with COVID-19 Two studies evaluating bamlanivimab as a single agent (N = 314) and casirivimab/imdevimab as a combination therapy (N = 9785) were included. Bamlanivimab compared to placebo We are uncertain whether bamlanivimab has an effect on mortality by day 30 (RR 1.39, 95% CI 0.40 to 4.83) and SAEs by day 28 (RR 0.93, 95% CI 0.27 to 3.14). Bamlanivimab may have little to no effect on time to hospital discharge (HR 0.97, 95% CI 0.78 to 1.20) and mortality by day 90 (HR 1.09, 95% CI 0.49 to 2.43). The effect of bamlanivimab on the development of severe symptoms at day 5 (RR 1.17, 95% CI 0.75 to 1.85) is uncertain. Bamlanivimab may increase grades 3-4 AEs at day 28 (RR 1.27, 95% CI 0.81 to 1.98). We assessed the evidence as low certainty for all outcomes due to serious imprecision, and very low certainty for severe symptoms because of additional concerns about indirectness. Casirivimab/imdevimab with usual care compared to usual care alone Treatment with casirivimab/imdevimab compared to usual care probably has little or no effect on mortality by day 30 (RR 0.94, 95% CI 0.87 to 1.02), IMV requirement or death (RR 0.96, 95% CI 0.90 to 1.04), nor alive at hospital discharge by day 30 (RR 1.01, 95% CI 0.98 to 1.04). We assessed the evidence as moderate certainty due to study limitations (lack of blinding). AEs and SAEs were not reported. AUTHORS' CONCLUSIONS: The evidence for each comparison is based on single studies. None of these measured quality of life. Our certainty in the evidence for all non-hospitalised individuals is low, and for hospitalised individuals is very low to moderate. We consider the current evidence insufficient to draw meaningful conclusions regarding treatment with SARS-CoV-2-neutralising mAbs. Further studies and long-term data from the existing studies are needed to confirm or refute these initial findings, and to understand how the emergence of SARS-CoV-2 variants may impact the effectiveness of SARS-CoV-2-neutralising mAbs. Publication of the 36 ongoing studies may resolve uncertainties about the effectiveness and safety of SARS-CoV-2-neutralising mAbs for the treatment of COVID-19 and possible subgroup differences.
Topics: Adult; Antibodies, Monoclonal; COVID-19; Cause of Death; Humans; Middle Aged; Randomized Controlled Trials as Topic; SARS-CoV-2
PubMed: 34473343
DOI: 10.1002/14651858.CD013825.pub2 -
Obesity Reviews : An Official Journal... Jan 2023This narrative systematic review examined effectiveness of interventions during pregnancy and up to 2 years of age in improving energy balance-related behaviors or... (Review)
Review
The effectiveness of interventions during the first 1,000 days to improve energy balance-related behaviors or prevent overweight/obesity in children from socio-economically disadvantaged families of high-income countries: a systematic review.
This narrative systematic review examined effectiveness of interventions during pregnancy and up to 2 years of age in improving energy balance-related behaviors or prevent overweight/obesity in children from families experiencing socio-economic disadvantage. We identified 24 interventions, from 33 articles, since 1990. Overall, despite their heterogeneity and variability in internal and external validity, there was some evidence of beneficial impact of interventions on obesity risk (4/15), and associated behaviors, e.g.: breastfeeding (9/18), responsive feeding (11/16), diet (7/8), sedentary (1/3) and movement (4/7) behaviors, and sleep (1/2). The most effective interventions aimed at promoting breastfeeding commenced antenatally; this was similar for the prevention of obesity, provided the intervention continued for at least 2 years postnatally and was multi-behavioral. Effective interventions were more likely to target first-time mothers and involve professional delivery agents, multidisciplinary teams and peer groups. Among ethnic/racial minorities, interventions delivered by lay agents had some impact on dietary behavior but not weight outcomes. Co-creation with stakeholders, including parents, and adherence to theoretical frameworks were additional ingredients for more pragmatic, inclusive, non-judgmental, and effective programs. The growing body of evidence on obesity prevention interventions targeting families experiencing socio-economic disadvantage is promising for reducing early inequalities in obesity risk.
Topics: Child; Pregnancy; Female; Humans; Overweight; Pediatric Obesity; Developed Countries; Diet; Breast Feeding
PubMed: 36394375
DOI: 10.1111/obr.13524 -
The Annals of Thoracic Surgery Aug 2022The United Kingdom National Institute for Health and Care Excellence guidelines recommend that patients and professionals make shared decisions between surgery and... (Review)
Review
BACKGROUND
The United Kingdom National Institute for Health and Care Excellence guidelines recommend that patients and professionals make shared decisions between surgery and stereotactic ablative radiotherapy (SABR) when treating early-stage non-small cell lung cancer (NSCLC). Variation by center suggests treatment decisions may be disproportionately influenced by clinician judgment and treatment availability rather than by patient preference. This systematic review critically evaluates studies of patient and clinician preferences for treatment of early-stage NSCLC.
METHODS
Primary empirical research up to April 30, 2020, was identified from searches of MEDLINE, Embase, PsycInfo, and Web of Science databases. Data extracted included study characteristics and methods, preferences for NSCLC treatment, and involvement in decision making and risk of bias using the Mixed Methods Appraisal Tool. Findings were synthesized using descriptive data and narrative synthesis.
RESULTS
Included in the review were 23 studies, of which 18 measured patient preferences, 4 clinician preferences, and 1 both clinician and patient preferences. Patients and clinicians were both most likely to prefer a collaborative role in treatment decisions. Most patients did not recall there being a choice between surgery or SABR options and thus experienced minimal decisional conflict.
CONCLUSIONS
For professionals to support patients in making informed, value-based decisions about NSCLC treatments, better quality evidence is needed of the clinical and quality of life trade-offs for both surgery and SABR.
Topics: Carcinoma, Non-Small-Cell Lung; Decision Making; Decision Making, Shared; Humans; Lung Neoplasms; Neoplasm Staging; Quality of Life; Small Cell Lung Carcinoma
PubMed: 33581150
DOI: 10.1016/j.athoracsur.2021.01.046 -
BMC Medical Education Jul 2022There is significant variability in the performance and outcomes of invasive medical procedures such as percutaneous coronary intervention, endoscopy, and bronchoscopy....
BACKGROUND
There is significant variability in the performance and outcomes of invasive medical procedures such as percutaneous coronary intervention, endoscopy, and bronchoscopy. Peer evaluation is a common mechanism for assessment of clinician performance and care quality, and may be ideally suited for the evaluation of medical procedures. We therefore sought to perform a systematic review to identify and characterize peer evaluation tools for practicing clinicians, assess evidence supporting the validity of peer evaluation, and describe best practices of peer evaluation programs across multiple invasive medical procedures.
METHODS
A systematic search of Medline and Embase (through September 7, 2021) was conducted to identify studies of peer evaluation and feedback relating to procedures in the field of internal medicine and related subspecialties. The methodological quality of the studies was assessed. Data were extracted on peer evaluation methods, feedback structures, and the validity and reproducibility of peer evaluations, including inter-observer agreement and associations with other quality measures when available.
RESULTS
Of 2,135 retrieved references, 32 studies met inclusion criteria. Of these, 21 were from the field of gastroenterology, 5 from cardiology, 3 from pulmonology, and 3 from interventional radiology. Overall, 22 studies described the development or testing of peer scoring systems and 18 reported inter-observer agreement, which was good or excellent in all but 2 studies. Only 4 studies, all from gastroenterology, tested the association of scoring systems with other quality measures, and no studies tested the impact of peer evaluation on patient outcomes. Best practices included standardized scoring systems, prospective criteria for case selection, and collaborative and non-judgmental review.
CONCLUSIONS
Peer evaluation of invasive medical procedures is feasible and generally demonstrates good or excellent inter-observer agreement when performed with structured tools. Our review identifies common elements of successful interventions across specialties. However, there is limited evidence that peer-evaluated performance is linked to other quality measures or that feedback to clinicians improves patient care or outcomes. Additional research is needed to develop and test peer evaluation and feedback interventions.
Topics: Bronchoscopy; Endoscopy; Feedback; Humans; Peer Review, Health Care; Percutaneous Coronary Intervention; Prospective Studies; Reproducibility of Results; Surgical Procedures, Operative
PubMed: 35906652
DOI: 10.1186/s12909-022-03652-9 -
European Journal of Medical Research Sep 2021Pregnant women are at high risk for severe influenza. However, maternal influenza vaccination uptake in most World Health Organization (WHO) European Region countries... (Review)
Review
BACKGROUND
Pregnant women are at high risk for severe influenza. However, maternal influenza vaccination uptake in most World Health Organization (WHO) European Region countries remains low, despite the presence of widespread national recommendations. An influenza vaccination reduces influenza-associated morbidity and mortality in pregnancy, as well as providing newborns with protection in their first months. Potential determinants of vaccine hesitancy need to be identified to develop strategies that can increase vaccine acceptance and uptake among pregnant women. The primary objective of the systematic review is to identify the individual determinants of influenza vaccine hesitancy among pregnant women in Europe, and how to overcome the hesitancy.
METHODS
Databases were searched for peer-reviewed qualitative and quantitative studies published between 2009 and 2019 inclusive. Databases included PubMed via MEDLINE, Cochrane Central Register for Controlled Trials, PsycINFO, SAGE Journals, Taylor and Francis and Springer nature. These covered themes including psychology, medicine, and public health. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) approach, 11 studies were eligible and analyzed for significant determinants of influenza vaccine hesitancy among pregnant women in Europe.
RESULTS
The most commonly reported factors were psychological aspects, for example concerns about safety and risks to mother and child, or general low risk perception of becoming ill from influenza. Doubts about the effectiveness of the vaccine and a lack of knowledge about this topic were further factors. There was also influence of contextual factors, such as healthcare workers not providing adequate knowledge about the influenza vaccine or the pregnant lady stating their antivaccine sentiment.
CONCLUSION
Health promotion that specifically increases knowledge among pregnant women about influenza and vaccination is important, supporting a valid risk judgment by the pregnant lady. The development of new information strategies for dialogue between healthcare providers and pregnant women should form part of this strategy.
Topics: Female; Health Knowledge, Attitudes, Practice; Humans; Influenza A virus; Influenza, Human; Pregnancy; Pregnant Women; Vaccination
PubMed: 34583779
DOI: 10.1186/s40001-021-00584-w -
Journal of Medical Internet Research Oct 2022When investigating voice disorders a series of processes are used when including voice screening and diagnosis. Both methods have limited standardized tests, which are... (Meta-Analysis)
Meta-Analysis Review
BACKGROUND
When investigating voice disorders a series of processes are used when including voice screening and diagnosis. Both methods have limited standardized tests, which are affected by the clinician's experience and subjective judgment. Machine learning (ML) algorithms have been used as an objective tool in screening or diagnosing voice disorders. However, the effectiveness of ML algorithms in assessing and diagnosing voice disorders has not received sufficient scholarly attention.
OBJECTIVE
This systematic review aimed to assess the effectiveness of ML algorithms in screening and diagnosing voice disorders.
METHODS
An electronic search was conducted in 5 databases. Studies that examined the performance (accuracy, sensitivity, and specificity) of any ML algorithm in detecting pathological voice samples were included. Two reviewers independently selected the studies, extracted data from the included studies, and assessed the risk of bias. The methodological quality of each study was assessed using the Quality Assessment of Diagnostic Accuracy Studies 2 tool via RevMan 5 software (Cochrane Library). The characteristics of studies, population, and index tests were extracted, and meta-analyses were conducted to pool the accuracy, sensitivity, and specificity of ML techniques. The issue of heterogeneity was addressed by discussing possible sources and excluding studies when necessary.
RESULTS
Of the 1409 records retrieved, 13 studies and 4079 participants were included in this review. A total of 13 ML techniques were used in the included studies, with the most common technique being least squares support vector machine. The pooled accuracy, sensitivity, and specificity of ML techniques in screening voice disorders were 93%, 96%, and 93%, respectively. Least squares support vector machine had the highest accuracy (99%), while the K-nearest neighbor algorithm had the highest sensitivity (98%) and specificity (98%). Quadric discriminant analysis achieved the lowest accuracy (91%), sensitivity (89%), and specificity (89%).
CONCLUSIONS
ML showed promising findings in the screening of voice disorders. However, the findings were not conclusive in diagnosing voice disorders owing to the limited number of studies that used ML for diagnostic purposes; thus, more investigations are needed. While it might not be possible to use ML alone as a substitute for current diagnostic tools, it may be used as a decision support tool for clinicians to assess their patients, which could improve the management process for assessment.
TRIAL REGISTRATION
PROSPERO CRD42020214438; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=214438.
Topics: Algorithms; Humans; Machine Learning; Supervised Machine Learning; Voice Disorders
PubMed: 36239999
DOI: 10.2196/38472 -
Medicines (Basel, Switzerland) Nov 2020: Polymyalgia Rheumatica (PMR) is one of the most frequent rheumatologic immune-related adverse effects (IRAEs) in cancer patients following therapy with immune... (Review)
Review
Identification and Classification of Polymyalgia Rheumatica (PMR) and PMR-Like Syndromes Following Immune Checkpoint Inhibitors (ICIs) Therapy: Discussion Points and Grey Areas Emerging from a Systematic Review of Published Literature.
: Polymyalgia Rheumatica (PMR) is one of the most frequent rheumatologic immune-related adverse effects (IRAEs) in cancer patients following therapy with immune checkpoint inhibitors (ICIs). Atypical findings in many patients often lead to diagnosing PMR-like syndromes. : The aim of our research was to review reported diagnoses of PMR and PMR-like syndromes following ICIs therapy, and assess whether they can be redefined as adverse drug reaction (ADR). In line with PRISMA guidelines, we carried out a systematic search on three main bibliographic databases, based on a combination of subject headings and free text. We included all studies and case-reports published after 2011 (when FDA approved the use of the first ICI) describing the association of PMR or PMR-like syndromes with all types of ICIs therapy. We excluded reviews, conference abstracts, comments, secondary articles, and non-English language studies. : We reviewed data from seven studies and eight case-reports, involving a total of 54 patients. Limitations included: the small size of all studies; only one retrospective study used validated criteria for PMR; most reports assessed IRAEs by clinical judgment only and did not seek validation through assessment scales. To date, it remains a conundrum whether IRAEs-PMR is identical to the idiopathic form of the disease, or whether it should be considered a subset of the disease or a new entity. Our review indicates that the relationship between PMR and ICIs therapy is yet to be clearly understood and defined and that future research should remedy the current limits in study design.
PubMed: 33153016
DOI: 10.3390/medicines7110068 -
Journal of Orthopaedic Surgery and... Dec 2022In the emergency room, clinicians spend a lot of time and are exposed to mental stress. In addition, fracture classification is important for determining the surgical...
BACKGROUND
In the emergency room, clinicians spend a lot of time and are exposed to mental stress. In addition, fracture classification is important for determining the surgical method and restoring the patient's mobility. Recently, with the help of computers using artificial intelligence (AI) or machine learning (ML), diagnosis and classification of hip fractures can be performed easily and quickly. The purpose of this systematic review is to search for studies that diagnose and classify for hip fracture using AI or ML, organize the results of each study, analyze the usefulness of this technology and its future use value.
METHODS
PubMed Central, OVID Medline, Cochrane Collaboration Library, Web of Science, EMBASE, and AHRQ databases were searched to identify relevant studies published up to June 2022 with English language restriction. The following search terms were used [All Fields] AND (", "[MeSH Terms] OR (""[All Fields] AND "bone"[All Fields]) OR "bone fractures"[All Fields] OR "fracture"[All Fields]). The following information was extracted from the included articles: authors, publication year, study period, type of image, type of fracture, number of patient or used images, fracture classification, reference diagnosis of fracture diagnosis and classification, and augments of each studies. In addition, AI name, CNN architecture type, ROI or important region labeling, data input proportion in training/validation/test, and diagnosis accuracy/AUC, classification accuracy/AUC of each studies were also extracted.
RESULTS
In 14 finally included studies, the accuracy of diagnosis for hip fracture by AI was 79.3-98%, and the accuracy of fracture diagnosis in AI aided humans was 90.5-97.1. The accuracy of human fracture diagnosis was 77.5-93.5. AUC of fracture diagnosis by AI was 0.905-0.99. The accuracy of fracture classification by AI was 86-98.5 and AUC was 0.873-1.0. The forest plot represented that the mean AI diagnosis accuracy was 0.92, the mean AI diagnosis AUC was 0.969, the mean AI classification accuracy was 0.914, and the mean AI classification AUC was 0.933. Among the included studies, the architecture based on the GoogLeNet architectural model or the DenseNet architectural model was the most common with three each. Among the data input proportions, the study with the lowest training rate was 57%, and the study with the highest training rate was 95%. In 14 studies, 5 studies used Grad-CAM for highlight important regions.
CONCLUSION
We expected that our study may be helpful in making judgments about the use of AI in the diagnosis and classification of hip fractures. It is clear that AI is a tool that can help medical staff reduce the time and effort required for hip fracture diagnosis with high accuracy. Further studies are needed to determine what effect this causes in actual clinical situations.
Topics: Humans; Artificial Intelligence; Hip Fractures; Machine Learning; Databases, Factual; Emergency Service, Hospital
PubMed: 36456982
DOI: 10.1186/s13018-022-03408-7