AI against hypertension

Using AI to redefine our fight against hypertension

Visual representation of AI model in detecting hypertension. Image provided by the Journal and the platform of ScienceDirect

Worldwide hypertension rates have skyrocketed the world due to our massive consumer culture and society. In fact around 1.28 billion in the world have hypertension, with many of them unknowing of their condition. What is more interesting is that ⅔ of those with hypertension live in low and middle income countries (LMIC), having inadequate public healthcare systems and guidance on urbanization and dietary consumption. It is predicted that the mismanagement of public health systems worldwide is going to cost healthcare systems around the world 3 trillion in 2030, and skyrocketing to 18 trillion when the year 2026 reaches. A common factor in the outbreak of non-communicable diseases like hypertension is centered around diet, excessive and uncontrolled consumption, faulty guidelines, and a lack of awareness on nutrition are mainly to blame for this outbreak. However a possible solution with Artificial intelligence is emerging with its role in healthcare, being able to identify disease including its causes and correctly prescribing treatments. In fact during the Covid-19 pandemic AI was a model that effectively analyzed epidemiological and clinical data that helped to create Covid 19 treatments. With AI proving itself to improve public health overall, AI has the potential to be an effective weapon against combating non communicable diseases that have plagued life on earth revolutionizing the fields of public health and medicine. 

AI has revolutionized hypertension treatment in a variety of methods ranging from everyday life to a tool for medical professionals to assess and carry out their job to the best of their abilities. The journal frontiers in public health states “Wearable devices help the consumer and their medical professionals have 24 hour access to their blood pressure monitors to keep track of oxygen saturation, pulse waveforms, masked and white coat HTN(hypertension)” to offer more accurate diagnosis and treatment. AI has proven to be an accurate risk predictor for patients with hypertension and other related cardiovascular diseases. Research has greatly benefited from AI in that it is used to research further on the pathogenesis and pathophysiology of HTN. AI has made great advancements in personalized care by predicting patients risks and necessary therapy adjustments according to their disease progression or how successful their therapy treatment has been. AI can lay out a plan that changes medications and lifestyle to control blood pressure effectively and its effectiveness in creating combination therapies. When making a diagnosis it is very crucial to get the most amount of information and that is where large generative models comes in, where the National Heart, Lung, Blood institute writes  “Large generative models offer a pathway to more efficiently process complex data and improve electronic healthcare records, and it can analyze a multitude of data from different sources to relay back to the healthcare professional” to make the best decision when treating a patient. With this new proposed solution, public health systems, universities, and pharmaceutical companies should consider AI as a viable option that offers reliable and accurate predictions, with advanced personalized treatment options.

AI based systems monitor BP in wearable technologies from the photophlethymsopgraph(PPG) signal, by producing pulse waveforms. An advantage of PPG is that it can be measured in electronics that are miniature, affordable and wearable. AI can help in areas of detection such as the machine called “HyMNet” combining fundus images with demographic information to improve hypertension detection capabilities. Since this machine “HyMNet” relies on diabetes as a huge factor to make its predictions it further solidifies itself as reliable as diabetes is the main disease that leads to a host of other cardiovascular related diseases such as stroke and coronary artery disease. “HyMNet” is part of a larger learning model known as MMDL and it has shown considerable progress in other health fields accurately predicting dementia and Alzheimer's disease rates as it has features such as magnetic resonance imaging scans. Using AI techniques is an efficient and non-invasive method to predict future complications. AI has been found especially in pregnancy to predict and manage hypertension effectively as it enhances the decision making process, the therapy behind, and ensures the best treatment which in turn can result in less of a need to do continuous screening tests. AI models like ML have reportedly found success in predicting pre-eclampsia, which is a pregnancy compilation coming with hypertension and excess protein in the urine at a rate of 97.3%, as “Women with hypertension in their first pregnancy face a 1.5–2.7 times higher risk of developing heart diseases later in life compared to those with normal blood pressure levels during pregnancy”. In many healthcare settings AI tools can suggest treatment protocols for hypertensive patients, and analyze and connect information between sophisticated datasets to create personalized treatment and care. With the development of prognostic data gathering that can be used in different stages of hypertension, individuals can learn more about the state of health including their potential risks for cardiovascular disease. Early lifestyle modifications can be made by physicians as they better understand a patient’s risks as it says “Physicians, with better understanding of individuals' risk through the AI models, could make early lifestyle modifications in prevention of hypertension”.  TransfoRhythm framework is an attention based DDN architecture for accurate blood prediction. According to the Bland-Altman diagram that plots two measurement differences, the study results and the optimal prediction done for blood pressure done by the TransfoRhythm model was found to be very identical and close to each other. In a quantitative evaluation graph the MAE(Mean absolute error) and the RMSE(Root mean squared error) were found to be very low indicating the high accuracy and reliability of TransfoRhythm as an AI prediction machine.

Through this research it is evident how AI can help everyone and the whole world in the global fight against hypertension from patients to professionals in different sectors and fields. Technology sectors can adopt PPG signaling into their everyday devices running from smartphones and smartwatches, to constantly monitor blood pressure so that the patient can live their best life while the physician can treat their patients to the best of their ability and resources. Without constant tracking physicians would not gather the best data leaving faulty and inefficient diagnosis and treatment plans. The machine “HyMNet” performs well in making patient predictions for professionals, here it uses advanced data about a patient and the types of diseases they have. By using this machine medical professionals can detect signs of a cardiovascular disease very early on, saving many years of life and money wasted fighting a disease. AI is also an important tool to make sure the health of pregnant women is at its best, using AI models like ML(Machine learning) have been found to manage hypertension and excess urine protein otherwise known as pre-eclampsia. With these AI models having a 97% success rate in predictions it is paramount to install these in the healthcare system so that the next generation of mothers and children can be protected and safeguarded. With AI comes the possibility of personalized medications, according to the prognostic data by AI, physicians can use advanced data and link it to create a tailored lifestyle and diet plan for a person with a high risk or already have hypertension. Lastly, machines that measure blood pressure are paramount in the fight against hypertension such as the TransfoRhythm model that had error percentages at a very low 1 to 2%, making sure that physicians have a reliable method to make accurate diagnoses for every patient especially hypertension.

Conclusion: In conclusion the planet and humanity have suffered the devastating toll hypertension has taken, infecting and killing many worldwide. In order to free humanity from this destruction the path to victory is found in utilizing AI to help make accurate public health and medical decisions to identify hypertension, decrease hypertension rates, and lower the overall population risk of hypertension.