China
Africa
Explore chapters and articles related to this topic
Solutions Using Machine Learning for COVID-19
Published in Punit Gupta, Dinesh Kumar Saini, Rohit Verma, Healthcare Solutions Using Machine Learning and Informatics, 2023
Muhammad Shafi, Kashif Zia, Jabar H. Yousif
In addition to modeling the spread of the disease to predict its long-term dynamics, models have also been used for short-term predictions. In [48], the authors proposed a model that could predict new cases of COVID-19 for the next six days. Instead of using a state-based system, they used a statistical measure of support vector regression. Further, a pure AI technique of long-short-term-memory (LSTM) deep learning was used to predict when the pandemic would end at the global level [49]. Unfortunately, the end date of June 2020 did not turn out to be true, indicating both the complexity of the phenomenon and the limitations of a model.
Healthcare Disaster Prediction with IoT, Data Analytics, and Machine Learning
Published in Adarsh Garg, D. P. Goyal, Global Healthcare Disasters, 2023
Priyanka Shukla, Akanksha Sehgal, Sonia, Deepika Sherawat
Data modeling is used in healthcare to model the real-world procedure and workflow. If we take an example of hospital admission for that what information is needed: Name of patient, age, DOB, gender, address, etc. If we want to see clinical history in the same hospital, what was the reason? Does the patient come from emergency? and many more. A good data model will cover all the possible queries. Methods to prepare modeling are data mining, ML, statistical analysis, analytical tools.
Clinical and Business Intelligence
Published in Salvatore Volpe, Health Informatics, 2022
In one non-healthcare setting, the weather service uses and presents models regularly. Anytime there is a hurricane, the various models build on the accumulated data and science of meteorology to project the path of the storm. There are often multiple models, and the average drawn from the models is shown by the weather station. The same type of science is being applied in healthcare modeling. Obviously, the storm tracker is not 100 percent accurate; however, it is usually good at determining the general area and time frame for a storm to make land. Healthcare modeling software is evolving just like meteorological storm models did over the years. One difference between the weather service and healthcare modeling, however, is that in healthcare, the model is set to mimic actual processes and then altered to see the effect. As experience is gained, healthcare modeling is becoming stronger and more reliable in predicting the effects of process changes. With many of these advanced concepts, it is often prudent to get expert support when first starting to use these tools.
Improving interventional causal predictions in regulatory risk assessment
Published in Critical Reviews in Toxicology, 2023
This paper has proposed that a valid quantitative risk assessment for predicting how reducing exposure to PM2.5 would affect public health requires four conditions:Valid study design: Use appropriate study designs (e.g. quasi-experiments, intervention studies) to collect data that can support valid causal inferences.Causal analysis: Apply appropriate causal analysis methods to analyze the data and to estimate and validate predicted causal impacts of changes in exposure on changes in health outcome probabilities.Validated models: Check whether the modeling assumptions used are appropriate for the data analyzed and only use models whose key assumptions are not found to be clearly and strongly violated by the data.Control for confounding and biases: Control for important confounding and residual confounding, e.g. by weather variables (on multiple time scales) and income. More generally, use data to refute alternative (non-causal) explanations for estimated exposure-response associations.
Evaluation of U.S. state opioid prescribing restrictions using patient opioid consumption patterns from a single, urban, academic institution
Published in Substance Abuse, 2022
Kortney A. Robinson, Jayson S. Marwaha, Chris J. Kennedy, Brendin R. Beaulieu-Jones, Aaron Fleishman, Justin K. Yu, Larry A. Nathanson, Gabriel A. Brat
Blanket prescription limits are one tool for reducing opioid misuse; however, without additional complementary efforts, prescribing limits fail to incite the optimal use of prescription opioids among surgical patients. Consumption data, as leveraged in this study, can be used to build more precise complementary tools. Evidence-based surgeon education, especially on anticipated population-specific opioid usage and guidance on non-opioid analgesia, is one such tool that has shown promising results.17,18 Other groups have established procedure-specific recommendations for opioid prescriptions after surgery.19 Individual patient-level personalization using predictive modeling techniques is also being used at some institutions.20
Comparative study of using E2SFCA and 3SFCA methods for selected healthcare resources in Jordan during COVID-19 pandemic
Published in International Journal of Healthcare Management, 2022
Nawras Shatnawi, Aslam Al-Omari, Alia Al-Mashaqbeh
Social and healthcare inequalities in society are becoming worse due to inadequate accessibility, limited healthcare, and the concentration of healthcare facilities in a specific area [4]. This suggests the need for healthcare spatial accessibility to be quantified specifically during the latest global COVID-19 pandemic which has reinforced the need for all people, regardless of their backgrounds and neighborhoods, to have equal access to basic medical facilities and services [5]. Mathematical modeling is widely used in health care and for predicting infected people all around the world. For example, Premarathna et al. [6], Moheimani et al [7] and Rafael et al [8] developed several forecasting models to predict both the number of the first and the second waves of COVID-19 cases in Sri Lanka. Data level optimization was introduced Sarkar and Sana [9] as the first stage in the two-step framework for creating an efficient disease decision support system. The second step was to select an optimal data-partition and the best training set for it in tandem. As a result of this, a general prediction model is explored over the obtained information for successful disease identification in step two.