Cognitive Computing
Submitted by Ravi Janumpally
Contents
Introduction
Cognitive computing is a method of using computerized models to simulate human thought process in complex situations where the answers might be ambiguous and uncertain [1]. This is similar to the grey area of human thought rather than concrete black and white process that apply to many workflows in clinical information systems. It can involve artificial intelligence, machine learning and Natural language processing (NLP)[2]. Cognitive computing systems are often based on artificial neural networks which are inspired by the human brain and are able to learn from data and improve their performance over time [3]. For a detailed lesson in various AI models, see Advances in Artificial Intelligence.
Cognitive Computing in Healthcare
Systems using cognitive computing can aid in diagnosing diseases, developing new treatments and managing patient care[4]. With cognitive computing, enormous health-related data can be harnessed including from personal fitness trackers, mobile apps, electronic medical records and genomic and clinical research[5]. IBM Watson defines this era of cognitive computing in healthcare as the era of cognitive health which brings together individual data with social and medical data from diverse sources including research and medical records. Cognitive systems can also help health information sharing for better patient outcomes. Cognitive computing can rapidly and intelligently parse through disparate data to help coordinate care. It is important to note that cognitive computing involves a category of analytics that are taught or learned rather than being programmed or rules-based analytics and computing. For example, in Epic a lot of analytics and computing is based on conditional rules with various logic clauses. On the other hand, there is now some cognitive computing in Epic that is used in risk scoring as well as in generative AI. For example, Microsoft OpenAI will be used for a generative AI run inbox messaging system that should save provider pajama time[6].
Cognitive Computing used in Wearables and Mobile Apps
Wearable devices, such as smartwatches and activity trackers are being used by more and more patients in their everyday lives to manage their health and well-being[7]. These devices collect and analyze long-term continuous data which can provide clinicians with a more comprehensive view of a patient's health compared with sporadic data collected at office and hospital visits. There are currently 3 types of wearables: consumer-grade, medical-grade or research-grade. Consumer-grade wearables, which consist of most smartwatches like fitbit, apple watch, etc are available for purchase in the general market and are intended for fitness, well-being or entertainment. On the other hand, medical-grade wearables may require a doctor's prescription and are intended for medical applications only. Research-grade wearables are used in clinical trials as well as in industry trials are aren't available for purchase in the general marketplace. Of course, these lines aren't clearly defined as the fitbit is available in the general marketplace but is used in research, medical and consumer applications. Medical-grade devices are subject to FDA oversight. The enormous volume of information from wearable devices often requires intelligent algorithms and computational power to generate meaningful information. ML(Machine Learning) and DL(Deep Learning) are used to extract useful information from the raw data collected. Wearables have been shown to improve medical management in various cardiovascular diseases, such as atrial fibrillation and heart failure[8,9,10]. For example, the Zio patch, which is developed by iRhythm Technologies of San Francisco, CA, improved the practicality of ambulatory cardiac monitoring as the patch was smaller and less burdensome than traditional Holter monitors and was able to capture 14 days of data versus the typical 2-3 days with a holter. Some cohort studies showed that the patch was better at detecting arrhythmias and led to better anticoagulation and health care utilization[11,12]. This patch uses proprietary AI algorithms involving deep neural networks in its operation. AI and ML technologies are faced with various challenges such as fairness and bias, transparency and accountability and data drift. This can lead to disparities and biases, particularly in underserved and underrepresented communities. Also, AI and ML algorithms tend to assume that data from the past will be representative of future data, which is always not the case in real-world settings, especially in medicine, which is not as predictable as banking or retail. Still, AI and ML and DL can be used to augment human expertise and diagnosis in medicine rather than replacing it, thus eliminating a lot of the bias and lack of accountability.
Cognitive Computing and Public Health
Public health data often requires massive amounts of data capture, data analysis and data transfer. It can be even said that public health data is "big data"[13}. Big Data involves information that is structured, semi-structured or unstructured. A lot of public health data is big data in that it comes from various sources including clinics, hospitals and community organizations. It is fairly disparate data that can even come from social media such as google search, twitter and facebook. AI and especially DL with its neural networks can process the big data of public health to predict epidemics and pandemics in real-time. The CDC is harnessing AI/ML/DL in surveillance, outbreak response and covid-19 vaccine safety monitoring. They are using MedCoder, an NLP/ML system to code national vital statistics and are also using AI/ML for forecasting trends in opoid overdose, foodborne outbreaks and syndromic surveillance.
References
1. Coccoli, M., Maresca, P., & Stanganelli, L. (2017). The role of big data and cognitive computing in the learning process. J. Vis. Lang. Comput., 38, 97-103. http://dx.doi.org/10.1016/j.jvlc.2016.03.002
2. Norden, A. D., Dankwa-Mullan, I., Urman, A., Suarez, F., & Rhee, K. (2018). Realizing the Promise of Cognitive Computing in Cancer Care: Ushering in a New Era. JCO clinical cancer informatics, 2, 1–6. https://doi.org/10.1200/CCI.17.00049
3. Esteva A, Kuprel B, Novoa RA, et al: Dermatologist-level classification of skin cancer with deep neural networks. Nature 542:115-118, 2017. https://doi.org/10.1038/nature21056
4. Bajwa, J., Munir, U., Nori, A., & Williams, B. (2021). Artificial intelligence in healthcare: transforming the practice of medicine. Future healthcare journal, 8(2), e188–e194. https://doi.org/10.7861/fhj.2021-0095
5. Ahmed, M.N., Toor, A.S., O'Neil, K., & Friedland, D. (2017). Cognitive Computing and the Future of Health Care Cognitive Computing and the Future of Healthcare: The Cognitive Power of IBM Watson Has the Potential to Transform Global Personalized Medicine. IEEE Pulse, 8, 4-9.
7. Hughes, A., Shandhi, M. M. H., Master, H., Dunn, J., & Brittain, E. (2023). Wearable Devices in Cardiovascular Medicine. Circulation research, 132(5), 652–670. https://doi.org/10.1161/CIRCRESAHA.122.322389
8. Sana F, Isselbacher EM, Singh JP, Heist EK, Pathik B, Armoundas AA. (2020) Wearable devices for ambulatory cardiac monitoring: JACC state-of-the-art review. J Am Coll Cardiol, 75, 1582–1592. doi: 10.1016/j.jacc.2020.01.046
9. Dendale P, De Keulenaer G, Troisfontaines P, Weytjens C, Mullens W, Elegeert I, Ector B, Houbrechts M, Willekens K, Hansen D. (2012) Effect of a telemonitoring- facilitated collaboration between general practitioner and heart failure clinic on mortality and rehospitalization rates in severe heart failure: the TEMA-HF 1 (Telemonitoring in the Management of Heart Failure) study. Eur J Heart Fail, 14, 333–340. doi: 10.1093/eurjhf/hfr144
10. Anand IS, Tang WH, Greenberg BH, Chakravarthy N, Libbus I, Katra RP. (2012) Design and performance of a multisensor heart failure monitoring algorithm: results from the Multisensor Monitoring in Congestive Heart Failure (MUSIC) study. J Card Fail, 18, 289–295. doi: 10.1016/j.cardfail.2012.01.009
11. Turakhia MP, Hoang DD, Zimetbaum P, Miller JD, Froelicher VF, Kumar UN, Xu X, Yang F, Heidenreich PA. (2013) Diagnostic utility of a novel leadless arrhythmia monitoring device. Am J Cardiol, 1125, 20–524. doi: 10.1016/j.amjcard.2013.04.017
12. Steinhubl SR, Waalen J, Edwards AM, Ariniello LM, Mehta RR, Ebner GS, Carter C, Baca-Motes K, Felicione E, Sarich T, Topol EJ. (2018) Effect of a homebased wearable continuous ECG monitoring patch on detection of undiagnosed atrial fibrillation: the mSToPS randomized clinical trial. JAMA, 320, 146–155. doi: 10.1001/jama.2018.8102
13. Benke, K., & Benke, G. (2018). Artificial Intelligence and Big Data in Public Health. International journal of environmental research and public health, 15(12), 2796. https://doi.org/10.3390/ijerph15122796
14. Centers for Disease Control. Artificial Intelligence and Machine Learning. Published July 3, 2023, accessed May 2, 2024. https://www.cdc.gov/surveillance/data-modernization/technologies/ai-ml.html
