Use Machine Learning to Predict Acute Kidney Injury (AKI) and its sequalae
Appropriately standardized decision support tools in clinical practice could predict significant clinical events, prompt and specific therapeutic interventions, prioritize resource utilization as a precursor to improve overall patient outcomes. Acute kidney injury (AKI) adversely impacts the lives of 317 per 100,000 population in general care and 179 per 100,000 population in intensive care units. No surprise caregivers are actively looking for ways to drastically lower these statistics. Several researchers recently completed studies that looked at using machine learning and statistical analysis to improve accuracy in predicting AKI in perioperative patients.
Problems with Conventional Prediction Models
In conventional approaches, the prediction models used for AKI rates were unable to be accurately applied to larger groups of patients as they ignore vast physiological time-series data. These limitations prevented them from being of any significant clinical value who are still spending a lot of time and resources trying to treat patients with post-surgical AKI.
How Machine Learning Improves AKI Prediction
The previously mentioned study included a group of 2,911 adults who had surgery at the University of Florida Health sometimes from 2000 to 2010. By using a combination of statistical analysis and machine learning algorithms, the researchers were able to create a perioperative model that was capable of attempting to predict the risk of these patients developing AKI during the course of their illness.
The results were segmented and analyzed as three separate time periods. The first was within three days after the date of the surgery, the second time period was within seven days of the surgery, and the third time period was from the date of the surgery to as far as the data went. One of the biggest differences in these models when compared to other existing ones was the inclusion of intraoperative physiological time-series variables.
By using this new prediction model, the researchers were able to achieve much higher accuracy when compared to the previous models across all three time periods. In total, the machine learning algorithm correctly reclassified 40 percent of patients who had originally received a false negative. These patients were now being classified as high-risk for developing AKI, which would have allowed them to get the preemptive treatment they needed.
When someone develops AKI, it makes them much more susceptible to developing future cases of AKI and also increases the likelihood of them developing morbidity, heart disease, a stroke, and much more. Therefore, the increased accuracy of this machine learning prediction algorithms means that significant amounts of money and man-hours will be saved trying to treat AKI and its related future health issues.
The results of this study indicate that with a machine learning-based prediction model, future predictions of AKI risk will soon be far more accurate in nation-wide predictions for any group of patients. To get more up-to-date news regarding AI and machine learning use in healthcare, make sure to follow our blog at RediMinds.
How AI Can Predict the Spread of Coronavirus and Other Serious Diseases
Coronavirus is currently dominating headlines all over the world. The virus’s impact has been felt not only in the travel industry and stock market, but also the healthcare industry as well. This is because hospitals are trying to stock up on supplies like facemasks, gowns, and hand sanitizer, as well as making sure that their personnel is undergoing the proper training. Unfortunately, this is causing approximately 86 percent of American hospitals to be concerned about their current limited supply of personal protective equipment.
But while these medical facilities attempt to deal with this new strain of Coronavirus, they can take comfort in knowing that artificial intelligence (AI) systems are currently being developed and optimized to be able to better predict the spread of the Coronavirus and other deadly diseases.
How Coronavirus Was Flagged By AI
The sooner that medical professionals can become aware of serious diseases, the sooner that they can begin to take steps to better prepare themselves in order to handle it. That is why it is so crucial to have systems like BlueDot’s, which used its AI system to identify Coronavirus as a lingering health crisis over a week before the World Health Organization (WHO) alerted the public about the new strain of the disease.
Its natural language processing and machine learning algorithm was able to do so by automatically reading through over 100,000 articles every single day. Therefore, it becomes aware of a potential disease outbreak when it is still in its very early stages. In fact, the system is able to accurately track over 100 infectious diseases simultaneously, one of which happened to be the new strain of Coronavirus.
Even once the disease was publicly acknowledged as a health crisis, AI systems continue to be used in order to predict how it is going to spread. The perfect example of this is the Healthmap system, which automatically aggregates data from various chat rooms, news reports, and many other online sources in order to create a current database of confirmed cases of various diseases. It is then able to generate a visual map representing how the diseases have spread from the beginning and how they are likely to continue to spread. Not only that, AI algorithms are also helping to identify early signs of hemodynamic instability, a common COVID-19 complication.
How AI Will Be Used in the Future to Predict Serious Diseases
It is not just the Coronavirus that AI systems can be used to predict. They are also able to be applied to nearly every type of serious disease as long as they are given the right data to be trained with, which is where initiatives like the Global Virome Project come in handy.
This research effort is attempting to develop a very detailed ecological and genetic database of current animal viruses that could possibly infect humans. By having this information readily available, it can be provided to new and current AI algorithms to help them better predict which diseases are more likely to spread to the public.
In order to stay up-to-date on news regarding the application of AI and machine learning in healthcare, make sure to follow our blog at RediMinds today.
Three Ways AI could revolutionize intensive care of critically ill patients
The management of a patient in an intensive-care unit (ICU) is cost intensive and poses a major financial burden on the hospital systems. In the United States, every year 5 million ICU patients are admitted. Besides heavy expenditure incurred, the intensive care teams and the patient families are deeply stressed due to uncertain clinical course of a critically ill patient. Currently the research focus is on the application of artificial intelligence (AI) to judiciously utilize scarce resources, to improve the predictability of the treatment actions and to contain medical errors.
Predicting Sepsis
In an intensive care environment, the sepsis is associated with high morbidity and mortality and difficult to treat complication. Patients who have developed sepsis are a major part of what makes ICU patient treatment so expensive. In fact, the cost of treating sepsis has risen by $1.5 billion in the past three years alone. This is partially due to the fact that conventional prediction models are only able to predict sepsis with a maximum accuracy of 0.725 area under the curve (AUCs).
However, through the use of deep learning models, tests have shown that it is able to predict which children would develop sepsis as early as 12-hours before onset and maintain an AUC of 0.83. This means that doctors will be able to effectively preempt the occurrence of sepsis and treat the same with appropriately selected antibiotics and save thousands or even millions of dollars in the process.
Predicting Mortality Rates
Understanding the probability of mortality with any extreme accuracy is something that has remained challenging due to the complexity of ICU patients in each medical facility. But with the use of Deep Learning Models, researchers have been able to track lab results, physiologic observations, and any administered drugs i.e the entire patient data in order to create mortality probability predictions that are far superior to previous models.
In fact, using the new algorithms achieved an AUC of 0.93 and allows doctors to more accurately understand the risk of mortality associated with each ICU patient and take the appropriate life-saving and cost-saving measures.
Predicting Proper Mechanical Ventilator Removal
The use of mechanical ventilation requires an eventual ventilator removal, which can significantly increase the risk of mortality if done too early or too late. Prolonged ventilation is associated with significant complications as well as it involves huge opportunity cost of blocking a scarce resource. Whereas premature weaning of ventilation increases reintubation and readmission rates. Hence predicting optimal time of weaning the patients off the ventilation is a challenging task for the clinicians in ICU.
A study done using an AI algorithm has demonstrated that this predictive model is able to more accurately predict when extubation should take place. With an AUC of 0.82, which has significantly improved the predicting power of the intensivists.
But it is important to note that this algorithm is limited to the fact that it can help to predict when a specific time is okay to perform extubation. However, it is not able to indicate whether it also would have been fine to perform the extubation earlier or later than when it was. Therefore, improved AI algorithms with larger datasets are needed to help further the accuracy of optimal extubation prediction.
These three areas are only the start of how AI systems are expected to revolutionize the treatment of ICU patients and hospital operations in general. To get more industry news regarding the use of artificial intelligence in clinical care, make sure to subscribe to our blog at RediMinds.
How AI based Clinical Decisioning Tools Can Lower Premature Deaths From Chronic Diseases
Chronic diseases are responsible for a significant portion of the time and money that the healthcare industry spends on an ongoing basis. In fact, it is estimated that approximately 60 percent of American adults suffer from at least one chronic disease.
n a new study exploring AI predictions of premature patient deaths that was conducted by the University of Nottingham, a number of machine learning and cox regression models were compared in terms of their prediction performance.
Machine Learning Algorithms Outperform Traditional Approaches
Models developed took into account the demographics, biometrics, and lifestyle factors such as the dietary habits for each individual. After making the prediction, the results were tracked and directly compared against the 2016 death record information provided by other national sources. This allowed them to calculate how accurately the ML algorithms could predict premature deaths caused by chronic diseases. Deep learning model and Random Forest model out performed multivariate cox regression model which was focused on limited number of variables.
How AI Identifies the Risk of Chronic Diseases More Effectively
When compared against standard human-developed prediction models, which in this case was a Cox regression and a multivariate Cox model, the random forest and deep learning AI models were shown to be significantly more accurate. In fact, they were approximately 10 percent more accurate than the Cox regression and over 3 percent more accurate than the multivariate Cox model.
Even this slight increase in predictive accuracy can have significant impacts in care delivery throughout the United States and with active feedback loops, these models can be improved to reduce the 23 to 25 million premature deaths that the U.S. suffers every year.
By replacing the traditional approaches for model development with Deep Learning architectures and algorithms the next generation of clinical decisioning tools can be prospectively deployed into clinical settings. These more promising AI models can potentially detect and save millions of lives that are prematurely dying from chronic diseases. This mechanism can provide the clinicians an observation window wherein an adverse event is foreseen, and risk level assessed with an opportunity to administer appropriate preventative care and potentially avert the complications and premature death.
Through further testing, this team of healthcare data scientists will be able to determine just how successful this AI predictive model would be with other population groups. With enough progress, these models could soon be successfully used by doctors and other medical professionals in the treatment of chronic diseases as part of routine healthcare.Of course, the prediction of premature deaths in chronic disease patients is only one of many white space opportunities where AI can be used to improve clinical care. To get more industry news regarding the use of artificial intelligence in clinical care, make sure to subscribe to our blog at RediMinds.
In the midst of this robotic revolution, now and then some real game-changing technologies emerge that accelerate the pace of innovation.
The introduction of Generative Adversarial Networks (GANs) into the wild world of AI is the perfect example of one of those game-changers. GANs function by approximating generative models. At a high level, GANs is an approach to machine learning involving two algorithms – a “generator” and a “discriminator” – that are pitted against each other to improve the abilities of the overall system.
Let’s dive into exactly how this works:
Venturing into a new era of GANs, researchers from UC Berkeley published an amazing work on motion retargeting that can make fake videos showing amateurs dancing like professional dancers.
They used custom GANs long with other video generation and smoothening techniques to transpose dance moves to amateurs with no advanced skills from a source video.
Video source: Caroline Chan
This is just the beginning of what’s possible with Generative Adversarial Networks (GANs).
How Does This AI Dancing Process Work?
In the research paper, the authors state:
“With our framework, we create a variety of videos, enabling untrained amateurs to spin and twirl like ballerinas, perform martial arts kicks or dance as vibrantly as pop stars.” AI dancing is the process of taking a video of a professional dancer (Dancer A). Deep learning video generation detects the pose of dancer A and transposes this onto a source video of an amateur dancer (Dancer B).
Essentially, making anyone a pro dancer on video through machine learning and deep learning applications.
Differentiating Between AI, Deep Learning & Machine Learning
Artificial Intelligence (AI)
Artificial intelligence is the general category that all of these terms belong to. It’s a reference to any intelligence performed by a machine that results in an optimal or suboptimal resolution to a problem.
Deep Learning: The Human Connection
Inspired by the function and structure of the brain (specifically the networks of interconnected neurons) – deep learning features algorithms that mimic the biological structure of the brain.
Machine Learning: Cognition Signs
Machine learning is the use of algorithms to interpret data, draw conclusions from it, then go on to make predictions or determinations about something in the world.
Where Did GANS Originate?
What started as a night out with fellow students in 2014 ended with the development of a whole new avenue that would shape the future of artificial intelligence.
While Ian Goodfellow celebrated the recent graduation of a fellow doctoral graduate, he was asked by a peer to help him with a project that would enable a computer to create photos independently.
At that time, technology that used neural networks did exist – where algorithms were modeled similarly to neurons in the brain – to create “generative” machines able to produce plausible data independently. But the results of this process were still very primitive and prone to errors like blurred faces or missing ears.
To find a solution for the project the other students were working on, massive amounts of number crunching would be required, which would not work.
Goodfellow returned home and pondered on this challenge which led to an idea. He wondered what would happen if you pitted two neural network systems against each other, essentially competing to solve the problem.
So, he gave it a shot and coded his software.
It worked the very first time. And the concept of Generative Adversarial Networks (GANs) was born.
Taking these methods a step further, the “do as I do” motion transfer system performs video generation while preserving intricate details like the subjects’ faces.
Recent advancements in image generation and general image mapping frameworks have laid the groundwork and enabled the authors to create a mapping from pose to target subject.
Many more purposes exist for GANs, especially in image generation because of their remarkable ability to replicate detailed, high-quality images.
Future Applications of GANs
Room for Improvement
Upon observation, you will notice that the subject dancer isn’t perfect. You will catch some waviness and choppy motions.
Nonetheless, “Everyone Can Dance” is a shining example of what modern AI is capable of. Some fine-tuning and a little time behind this development will render methods indistinguishable from the real subjects.
Causes for Concern
Obvious scenarios render the concern of GANs and motion transfer technology being used to implicate people in false scenarios.
In recent years, Reddit cracked down on deepfake pornography and the political sphere is concerned about false information spreading via fake videos. Furthermore, down to a very fundamental level, endless scenarios exist as to how these methods can cause interpersonal conflicts.
Future Developments
Many applications of machine learning have been facilitated by the emergence of deep learning methods. Collectively, these improvements are advancing the AI field.
The technological ability to break down tasks into ways that utilize various machine assists will facilitate future developments, including exciting developments like:
Researchers are even asking questions about how GANS and motion re-targeting can be exploited to solve some healthcare problems. For instance: Can this technology be transferred to perform motion re-targeting from an expert surgeon to a novice surgeon?
We believe the technologies needed to codify an expert surgeon’s proficiency into an AI system are here and we will see them in action soon. Such tech will empower novice surgeons and provide care at the same level as a skilled expert with years of experience.
Needless to say the AI spectrum is evolving at lightning speed and there’s no telling how far this is going to take us and how fast. Remember: this is only the beginning.
“Deep Learning is a superpower. With it, you can make a computer see, synthesize novel art, translate languages, render a medical diagnosis, or build pieces of a car that can drive itself. If that isn’t a superpower, I don’t know what is.” – Andrew Ng, Founder of deeplearning.ai
RediMinds are working on numerous exciting AI projects right now, helping businesses and healthcare provides change the face of their industries. Stay up to date with our latest projects and exciting AI news by subscribing to our monthly newsletter.
