How AI is Already Being Used in Healthcare
From drug discovery to stroke diagnosis — AI is rapidly altering the healthcare system
Artificial intelligence (AI) technology is being adopted in many different areas of healthcare and is quickly reshaping medical practices.
AI is already benefiting our healthcare in more ways than you may think. Here are some examples of AI in healthcare that you might not know about.
Emergency Treatment
Sepsis Detection
Sepsis is a life-threatening complication of infection. It is the number one cause of death in non-cardiac intensive care units¹. Diagnosis of sepsis is rather difficult as the symptoms are similar to those of many other illnesses.
However, diagnosis of sepsis is time-sensitive — minutes matter. You can think of it kind of like a house fire. It starts off small and isolated, but if it is not resolved quickly, it can spread and cause significant damage and fatalities.
HCA Healthcare created an artificial intelligence tool called SPOT to detect sepsis development¹. SPOT monitors a patient’s vital signs and alerts healthcare staff when changes occur that indicate sepsis. Using this tool, sepsis mortality decreased by almost 23% from 2017–2018 in HCA hospitals.
Diagnostics
Stroke Diagnosis
As with detecting sepsis, diagnosing stroke patients is time-sensitive. There is a window of approximately 90 minutes from the time of the stroke to diagnose and treat a patient for the best chance of survival and recovery².
Conventional stroke diagnosis requires a CT scan and the results need to be analyzed by a specialist. This usually takes around 30 minutes — a significant chunk of the small amount of time doctors have to treat stroke patients.
AUTOStroke is an artificial intelligence technology that detects signs of ischemic and hemorrhagic strokes from CT scans in 30 seconds. It consolidates the results of the scan into a summary and notifies the specialist of any abnormalities indicating a stroke².
This allows stroke patients to be diagnosed earlier, increasing their chances of survival and recovery.
Breast Cancer Diagnosis
Breast cancer is the second most fatal disease in women³. The conventional detection and diagnosis methods include a mammogram, which is analyzed by an oncologist, followed by a biopsy if an abnormality is found.
However, manual reviews of mammogram results have a high false detection rate. This means that many women undergo unnecessary biopsies³. This is obviously not ideal for either the patient or the healthcare institution.
Deep learning systems have been developed to aid in the detection of breast cancer by automatically discerning features of cancerous masses from a mammogram³. However, this technology still needs a lot of improvement and currently cannot be solely relied on; it must be used in conjunction with a specialist diagnosis.
General Diagnosis
Artificial intelligence technologies such as IBM’s Watson⁴ are currently being used to aid in medical diagnosis.
The use of artificial intelligence technology like Watson helps clinicians come to a diagnosis faster and create a patient-specific treatment plan, without having to spend hours combing through the literature.
Watson stores more medical information (such as medical notes and clinical research) than a human brain would ever be able to. It can scan the biomedical literature and aid medical professionals in finding the right diagnosis and the best, evidence-based treatment.
Drug Discovery
Understanding the interaction between a possible drug and the target molecule in the body is incredibly important in drug development. However, it is very time-consuming. DEEPScreen is a tool that predicts such interactions using convolutional neural networks⁵.
DEEPScreen learns the complex features of a 2D structural representation of components of a potential drug. It then accurately predicts the drugs interactions with biomolecules⁵.
Using this kind of artificial intelligence technology in the early stages of drug development will shorten the production cycle, making drug development faster, more efficient, and cheaper.
Condition Monitoring and Alerts at Home
Arrhythmia Detection
Arrhythmia describes the improper beating of the heart, whether it is too fast, too slow, or irregular. Complications of arrhythmias include strokes and heart failure. An echocardiogram (ECG) is a simple test that detects arrhythmias; however, they have to be run by a specialist.
AliveCor has developed an artificial intelligence technology that is used to check your heartbeat without a cardiologist present. It’s very simple to use and can detect abnormalities in your heartbeat as well as changes to your heart rate. It compiles a report of your heart rhythm data which you can then use for self-monitoring and can assist your cardiologist in diagnosis and treatment⁶.
Seizure Alerts
Seizure alert bracelets and smartwatches allow people with epilepsy to prepare for a seizure before it occurs. These devices monitor electrodermal activity, temperature, pulse, and body motion to detect oncoming seizures.
The devices vibrate when a seizure is detected and the wearer can respond. Designated friends or family will also be notified of the seizure if the wearer does not respond to the alert. This could potentially save someone’s life as seizures can be life-threatening⁷.
Conclusion
Artificial intelligence is already reshaping the healthcare system. As this technology continues to be improved we’re sure to see exciting developments in medical solutions and improvements in clinical care.
References
- HCA Healthcare Using Algorithm Driven Technology to Detect Sepsis Early and Help Save 8,000 Lives. HCA Healthcare. May 16, 2019. Accessed January 5, 2023. https://investor.hcahealthcare.com/news/news-details/2019/HCA-Healthcare-Using-Algorithm-Driven-Technology-to-Detect-Sepsis-Early-and-Help-Save-8000-Lives/default.aspx
- Rapid Stroke Diagnosis a Reality in UK with AI-Assisted Triage Tool. Canon Medical Systems LTD. March 24, 2022. Accessed January 5, 2023. https://uk.medical.canon/rapid-stroke-diagnosis-a-reality-in-uk-with-ai-assisted-triage-tool/#:~:text=Stroke%20automates%20diagnosis%20with%20zero,and%20speed%20up%20diagnostic%20reporting
- Madani M, Behzadi MM, Nabavi S. The Role of Deep Learning in Advancing Breast Cancer Detection Using Different Imaging Modalities: A Systematic Review. Cancers (Basel). 2022;14(21):5334. Published 2022 Oct 29. doi:10.3390/cancers14215334
- How AI is Impacting Healthcare. IBM. Accessed January 5, 2023. https://www.ibm.com/resources/watson-health/artificial-intelligence-impacting-healthcare/
- Rifaioglu AS, Nalbat E, Atalay V, Martin MJ, Cetin-Atalay R, Doğan T. DEEPScreen: high performance drug-target interaction prediction with convolutional neural networks using 2-D structural compound representations. Chem Sci. 2020;11(9):2531–2557. Published 2020 Jan 8. doi:10.1039/c9sc03414e
- Kardia by AliveCor. AliveCor. Accessed January 5, 2023. https://www.kardia.com/
- Herrera-Fortin T, Bou Assi E, Gagnon MP, Nguyen DK. Seizure detection devices: A survey of needs and preferences of patients and caregivers. Epilepsy Behav. 2021;114(Pt A):107607. doi:10.1016/j.yebeh.2020.107607
