Multimodal machine learning in precision health: A scoping review.

Machine learning is frequently being leveraged to tackle problems in the health sector including utilization for clinical decision-support. Its use has historically been focused on single modal data. Attempts to improve prediction and mimic the multimodal nature of clinical expert decision-making has been met in the biomedical field of machine learning by fusing disparate data. This review was conducted to summarize the current studies in this field and identify topics ripe for future research. We conducted this review in accordance with the PRISMA extension for Scoping Reviews to characterize multi-modal data fusion in health. Search strings were established and used in databases: PubMed, Google Scholar, and IEEEXplore from 2011 to 2021. A final set of 128 articles were included in the analysis. The most common health areas utilizing multi-modal methods were neurology and oncology. Early fusion was the most common data merging strategy. Notably, there was an improvement in predictive performance when using data fusion. Lacking from the papers were clear clinical deployment strategies, FDA-approval, and analysis of how using multimodal approaches from diverse sub-populations may improve biases and healthcare disparities. These findings provide a summary on multimodal data fusion as applied to health diagnosis/prognosis problems. Few papers compared the outputs of a multimodal approach with a unimodal prediction. However, those that did achieved an average increase of 6.4% in predictive accuracy. Multi-modal machine learning, while more robust in its estimations over unimodal methods, has drawbacks in its scalability and the time-consuming nature of information concatenation.

Cost-Effectiveness of Different Strategies for the Prevention of Venous Thromboembolism After Total Hip Replacement in China.

INTRODUCTION: The aim of this study was to evaluate the cost-effectiveness of rivaroxaban and apixaban versus enoxaparin for the universal prophylaxis of venous thromboembolism (VTE) and associated long-term complications in Chinese patients after total hip replacement (THR). METHODS: A decision model, which included both acute VTE (represented as a decision tree) and the long-term complications of VTE (represented as a Markov model), was developed to assess the economic outcomes of the three prophylactic strategies for Chinese patients after THR. Transition probabilities for acute VTE were derived from two randomized controlled studies, RECORD1 and ADVANCE3, of patients after THR. The transition probabilities of long-term complications after acute VTE, utilities, and costs were derived from the published literature and local healthcare settings. One-way and probabilistic sensitivity analyses (PSA) were performed to test the uncertainty concerning the model parameters. The quality-adjusted life years (QALYs) and direct medical costs were reported over a 5-year horizon, and incremental cost-effectiveness ratios (ICERs) were also calculated. RESULTS: Thromboprophylaxis with apixaban was estimated to have a higher cost (US $178.70) and more health benefits (0.0025 QALY) than thromboprophylaxis with enoxaparin over a 5-year time horizon, which resulted in an ICER of US $71,244 per QALY gained and was more than three times the GDP per capita of China in 2014 (US $22,140). Owing to the higher cost and lower generated QALYs, rivaroxaban was inferior to enoxaparin among post-THR patients. The sensitivity analyses confirmed these results. CONCLUSIONS: The analysis found that apixaban was not cost-effective and that rivaroxaban was inferior to enoxaparin. This finding indicates that compared with enoxaparin, the use of apixaban for VTE prophylaxis after THR does not represent a good value for the cost at the acceptable threshold in China; in addition, the cost of rivaroxaban was higher with lower QALYs.

Detection of Anti-Hepatitis B Virus Drug Resistance Mutations Based on Multicolor Melting Curve Analysis.

Detection of anti-hepatitis B virus (HBV) drug resistance mutations is critical for therapeutic decisions for chronic hepatitis B virus infection. We describe a real-time PCR-based assay using multicolor melting curve analysis (MMCA) that could accurately detect 24 HBV nucleotide mutations at 10 amino acid positions in the reverse transcriptase region of the HBV polymerase gene. The two-reaction assay had a limit of detection of 5 copies per reaction and could detect a minor mutant population (5% of the total population) with the reverse transcriptase M204V amino acid mutation in the presence of the major wild-type population when the overall concentration was 104 copies/µl. The assay could be finished within 3 h, and the cost of materials for each sample was less than $10. Clinical validation studies using three groups of samples from both nucleos(t)ide analog-treated and -untreated patients showed that the results for 99.3% (840/846) of the samples and 99.9% (8,454/8,460) of the amino acids were concordant with those of Sanger sequencing of the PCR amplicon from the HBV reverse transcriptase region (PCR Sanger sequencing). HBV DNA in six samples with mixed infections consisting of minor mutant subpopulations was undetected by the PCR Sanger sequencing method but was detected by MMCA, and the results were confirmed by coamplification at a lower denaturation temperature-PCR Sanger sequencing. Among the treated patients, 48.6% (103/212) harbored viruses that displayed lamivudine monoresistance, adefovir monoresistance, entecavir resistance, or lamivudine and adefovir resistance. Among the untreated patients, the Chinese group had more mutation-containing samples than did the Pakistani group (3.3% versus 0.56%). Because of its accuracy, rapidness, wide-range coverage, and cost-effectiveness, the real-time PCR assay could be a robust tool for the detection if anti-HBV drug resistance mutations in resource-limited countries.

Clopidogrel versus aspirin in patients with recent ischemic stroke and established peripheral artery disease: an economic evaluation in a Chinese setting.

BACKGROUND AND OBJECTIVES: Clopidogrel or aspirin are indicated for patients with recent ischemic stroke (IS) or established peripheral artery disease (PAD). We compared the cost effectiveness of clopidogrel with that of aspirin in Chinese patients with recent IS or established PAD. METHODS: A discrete-event simulation was developed to evaluate the economic implications of secondary prevention with clopidogrel versus aspirin. All available evidence was derived from clinical studies. Costs from a Chinese healthcare perspective in 2013 US dollars and quality-adjusted life-years (QALYs) were projected over patients' lifetimes. Uncertainties were addressed using sensitivity analyses. RESULTS: Compared with aspirin, clopidogrel yielded a marginally increased life expectancy by 0.46 and 0.21 QALYs at an incremental cost-effectiveness ratio of $US5246 and $US9890 per QALY in patients with recent IS and established PAD, respectively. One-way sensitivity analyses showed that the evaluation of patients with PAD and recent IS was robust except for the parameter of patient age. Given a willingness-to-pay of $US19,877 per QALY gained, clopidogrel had a probability of 90 and 68% of being cost effective in the recent IS or established PAD subgroups compared with aspirin, respectively. CONCLUSIONS: The analysis suggests that clopidogrel for secondary prevention is cost effective for patients with either PAD or recent IS in a Chinese setting in comparison with aspirin.