Machine learning prediction models for clinical management of blood-borne viral infections: a systematic review of current applications and future impact.

BACKGROUND: Machine learning (ML) prediction models to support clinical management of blood-borne viral infections are becoming increasingly abundant in medical literature, with a number of competing models being developed for the same outcome or target population. However, evidence on the quality of these ML prediction models are limited. OBJECTIVE: This study aimed to evaluate the development and quality of reporting of ML prediction models that could facilitate timely clinical management of blood-borne viral infections. METHODS: We conducted narrative evidence synthesis following the synthesis without meta-analysis guidelines. We searched PubMed and Cochrane Central Register of Controlled Trials for all studies applying ML models for predicting clinical outcomes associated with hepatitis B virus (HBV), human immunodeficiency virus (HIV), or hepatitis C virus (HCV). RESULTS: We found 33 unique ML prediction models aiming to support clinical decision making. Overall, 12 (36.4%) focused on HBV, 10 (30.3%) on HCV, 10 on HIV (30.3%) and two (6.1%) on co-infection. Among these, six (18.2%) addressed the diagnosis of infection, 16 (48.5%) the prognosis of infection, eight (24.2%) the prediction of treatment response, two (6.1%) progression through a cascade of care, and one (3.03%) focused on the choice of antiretroviral therapy (ART). Nineteen prediction models (57.6%) were developed using data from high-income countries. Evaluation of prediction models was limited to measures of performance. Detailed information on software code accessibility was often missing. Independent validation on new datasets and/or in other institutions was rarely done. CONCLUSION: Promising approaches for ML prediction models in blood-borne viral infections were identified, but the lack of robust validation, interpretability/explainability, and poor quality of reporting hampered their clinical relevance. Our findings highlight important considerations that can inform standard reporting guidelines for ML prediction models in the future (e.g., TRIPOD-AI), and provides critical data to inform robust evaluation of the models.

Routine blood test markers for predicting liver disease post HBV infection: precision pathology and pattern recognition.

BACKGROUND: Early stages of hepatitis B virus (HBV) infection usually involve inflammation of the liver. Patients with chronic infection have an increased risk of progressive liver fibrosis, cirrhosis, and life-threatening clinical complications of end-stage hepatocellular carcinoma (HCC). CONTENT: Early diagnosis of hepatic fibrosis and timely clinical management are critical to controlling disease progression and decreasing the burden of end-stage liver cancer. Fibrosis staging, through its current gold standard, liver biopsy, improves patient outcomes, but the clinical procedure is invasive with unpleasant post-procedural complications. Routine blood test markers offer promising diagnostic potential for early detection of liver disease without biopsy. There is a plethora of candidate routine blood test markers that have gone through phases of biomarker validation and have shown great promise, but their current limitations include a predictive ability that is limited to only a few stages of fibrosis. However, the advent of machine learning, notably pattern recognition, presents an opportunity to refine blood-based non-invasive models of hepatic fibrosis in the future. SUMMARY: In this review, we highlight the current landscape of routine blood-based non-invasive models of hepatic fibrosis, and appraise the potential application of machine learning (pattern recognition) algorithms to refining these models and optimising clinical predictions of HBV-associated liver disease. OUTLOOK: Machine learning via pattern recognition algorithms takes data analytics to a new realm, and offers the opportunity for enhanced multi-marker fibrosis stage prediction using pathology profile that leverages information across patient routine blood tests.

Clinical Validity of a Machine Learning Decision Support System for Early Detection of Hepatitis B Virus: A Binational External Validation Study.

HepB LiveTest is a machine learning decision support system developed for the early detection of hepatitis B virus (HBV). However, there is a lack of evidence on its generalisability. In this study, we aimed to externally assess the clinical validity and portability of HepB LiveTest in predicting HBV infection among independent patient cohorts from Nigeria and Australia. The performance of HepB LiveTest was evaluated by constructing receiver operating characteristic curves and estimating the area under the curve. Delong's method was used to estimate the 95% confidence interval (CI) of the area under the receiver-operating characteristic curve (AUROC). Compared to the Australian cohort, patients in the derivation cohort of HepB LiveTest and the hospital-based Nigerian cohort were younger (mean age, 45.5 years vs. 38.8 years vs. 40.8 years, respectively; p < 0.001) and had a higher incidence of HBV infection (1.9% vs. 69.4% vs. 57.3%). In the hospital-based Nigerian cohort, HepB LiveTest performed optimally with an AUROC of 0.94 (95% CI, 0.91-0.97). The model provided tailored predictions that ensured most cases of HBV infection did not go undetected. However, its discriminatory measure dropped to 0.60 (95% CI, 0.56-0.64) in the Australian cohort. These findings indicate that HepB LiveTest exhibits adequate cross-site transportability and clinical validity in the hospital-based Nigerian patient cohort but shows limited performance in the Australian cohort. Whilst HepB LiveTest holds promise for reducing HBV prevalence in underserved populations, caution is warranted when implementing the model in older populations, particularly in regions with low incidence of HBV infection.

The development of a machine learning algorithm for early detection of viral hepatitis B infection in Nigerian patients.

Access to Hepatitis B Virus (HBV) testing for people in low-resource settings has long been challenging due to the gold standard, enzyme immunoassay, being prohibitively expensive, and requiring specialised skills and facilities that are not readily available, particularly in remote and isolated laboratories. Routine pathology data in tandem with cutting-edge machine learning shows promising diagnostic potential. In this study, recursive partitioning ("trees") and Support Vector Machines (SVMs) were applied to interrogate patient dataset (n = 916) that comprised results for Hepatitis B Surface Antigen (HBsAg) and routine clinical chemistry and haematology blood tests. These algorithms were used to develop a predictive diagnostic model of HBV infection. Our SVM-based diagnostic model of infection (accuracy = 85.4%, sensitivity = 91%, specificity = 72.6%, precision = 88.2%, F1-score = 0.89, Area Under the Receiver Operating Curve, AUC = 0.90) proved to be highly accurate for discriminating HBsAg positive from negative patients, and thus rivals with immunoassay. Therefore, we propose a predictive model based on routine blood tests as a novel diagnostic for early detection of HBV infection. Early prediction of HBV infection via routine pathology markers and pattern recognition algorithms will offer decision-support to clinicians and enhance early diagnosis, which is critical for optimal clinical management and improved patient outcomes.

Hepatitis B virus infection in Nigeria: a systematic review and meta-analysis of data published between 2010 and 2019.

BACKGROUND: Hepatitis B virus (HBV) is an infectious disease of global significance, causing a significant health burden in Africa due to complications associated with infection, such as cirrhosis and liver cancer. In Nigeria, which is considered a high prevalence country, estimates of HBV cases are inconsistent, and therefore additional clarity is required to manage HBV-associated public health challenges. METHODS: A systematic review of the literature (via PubMed, Advanced Google Scholar, African Index Medicus) was conducted to retrieve primary studies published between 1 January 2010 and 31 December 2019, with a random-effects model based on proportions used to estimate the population-based prevalence of HBV in the Nigerian population. RESULTS: The final analyses included 47 studies with 21,702 participants that revealed a pooled prevalence of 9.5%. A prevalence estimate above 8% in a population is classified as high. Sub-group analyses revealed the highest HBV prevalence in rural settings (10.7%). The North West region had the highest prevalence (12.1%) among Nigeria's six geopolitical zones/regions. The estimate of total variation between studies indicated substantial heterogeneity. These variations could be explained by setting and geographical region. The statistical test for Egger's regression showed no evidence of publication bias (p = 0.879). CONCLUSIONS: We present an up-to-date review on the prevalence of HBV in Nigeria, which will provide critical data to optimise and assess the impact of current prevention and control strategies, including disease surveillance and diagnoses, vaccination policies and management for those infected.

One Health Approach: A Data-Driven Priority for Mitigating Outbreaks of Emerging and Re-Emerging Zoonotic Infectious Diseases.

This paper discusses the contributions that One Health principles can make in improving global response to zoonotic infectious disease. We highlight some key benefits of taking a One Health approach to a range of complex infectious disease problems that have defied a more traditional sectoral approach, as well as public health policy and practice, where gaps in surveillance systems need to be addressed. The historical examples demonstrate the scope of One Health, partly from an Australian perspective, but also with an international flavour, and illustrate innovative approaches and outcomes with the types of collaborative partnerships that are required.

Prevalence and antibiotic resistance of Escherichia coli O157:H7 in beef at a commercial slaughterhouse in Moro, Kwara State, Nigeria.

BACKGROUND: Gastroenteritis due to foodborne disease is a leading cause of death in developing countries. In Nigeria, there is an increasing demand for beef. Yet, there is no surveillance for Escherichia coli O157:H7 contamination of raw beef and little is known about the carriage of this pathogen in Nigeria's livestock. METHODS: A total of 415 samples, including 180 cow carcass swabs, 180 caecal content samples, 16 water samples, 25 hand swabs and 14 knife swabs were collected at a large abattoir in the Moro region of Kwara State, Nigeria. The samples were enriched in modified tryptone broth containing novobiocine, and plated onto Sorbitol-MacConkey agar (Oxoid SR0172E) supplemented with 0.05 mg l-1 cefixime and 2.5 mg l-1 potassium tellurite (Oxoid) (CT-SMAC). Indole-producing isolates were confirmed serologically by serotyping with antisera specific for the O157 and H7 antigens. The E. coli O157:H7 isolates were further tested for their susceptibility to antibiotic agents using the disc diffusion method. Commercially available Gram-negative multi-discs (Oxoid) comprising nitrofurantoin (30 µg), ampicillin (5 µg), ceftazidime (30 µg), gentamicin (10 µg), ciprofloxacin (5 µg), augmentin (30 µg), ofloxacin (5 µg) and cefuroxime (30 µg) were tested. RESULTS: Overall, 16 (3.9 %) samples were contaminated with E. coli O157:H7, of which 10 (5.6 %) were isolated from carcass swabs, 4 (2.2 %) from caecal content samples and 2 (12.5 %) from water. All isolates were multidrug-resistant (MDR), with resistance to ampicillin, ceftazidime and cefuroxime being the most common. CONCLUSION: This study provides evidence to suggest that E. coli O157:H7 exists in the beef production chain. The pathogen reveals a high frequency of multidrug resistance, suggesting that consumers and handlers of such meat are at risk of contracting antibiotic-resistant E. coli O157:H7-associated foodborne disease. Routine monitoring of antibiotic resistance is critical to uncovering novel therapeutic strategies that will help inform clinical practice guidelines.