William Truong EST 110 - Session #4 by William Truong

Personalized Overview

williamtruong3 — 2024-08-04 03:50:45 UTC

For this assignment, I decided to explore the potential and existing applications of artificial intelligence within the field of medicine. As someone who currently works in healthcare and wants to pursue further education in the field, I look forward to seeing further applications of artificial intelligence and how it will further assist with efficiency and accuracy in patient outcomes. Although healthcare professionals are generally viewed as competent, there are millions of cases of medical malpractice, misdiagnosis, and issues regarding efficacy in treatments. In addition, clinical research in medicine is very strenous, as there is high regulation, costs, and lengthy timelines for development/approval. By introducing AI, there is potential to improve diagnostic accuracy by analyzing medical images, lab results, and patient histories, detecting early signs of diseases with higher precision than traditional methods. In addition, AI tools can significantly accelerate and enhance medical research by streamlining data analysis, improving patient recruitment for clinical trials, and the simulation of biological systems. Furthermore, AI has the potential to form personalized treatment plans tailored to individual patients' genetic makeup and enviroment, which can lead to more effective diagnosing and treatments. The role of AI can be applied to not only direct patient care, but also administrative tasks such as scheduling, billing, and patient record management. Similar to most application of technology, it is critical that we implement such technologies with ethical and moral responsbility. Although the implentation of artificial intelligence in healthcare can seemingly be beneficial, there have already been cases in which AI has been trained in a malicious manner. In this padlet, I will also provide a general overview of how artificial intelligence can be used to harm patients, by either amplifying historical biases or by denying insurance claims unfairly.

Artificial intelligence used in clinical prediction

williamtruong3 — 2024-08-04 04:19:57 UTC

In a systemic review of 74 research articles that mention AI application in clinical practice, it has been stated that AI tools can be used to accurately enhance diagnostic accuracy, treatment planning, disease prevention, and better patient outcomes. When these AI tools were utilized with real patient data, it has been noted that AI tools are more efficient in the fields of oncology and radiology in more accurately diagnosing in contrast to traditional methods (Khalifa, et. al). As for the specific AI tools used, it does not mention explicitly but rather states machine learning algorithms as a tag. In terms of limitations, AI algorithms can be biased in diagnoses based on their trained dataset.

Personal insight: When it comes to its performance so far, and potential impact, I am surprised in AI's ability to diagnose properly. However, I cannot help but think of how this would affect patient care and job placement. Ultimately, such tools can reduce the need for the number of specialists, such as in the field of radiology in which the AI tools performed the best.

williamtruong3 — 2024-08-04 04:28:17 UTC

How AI is being currently used to accelerate clinical trials

williamtruong3 — 2024-08-04 04:51:07 UTC

When it comes to research, AI tools can be beneficial in clinical trials by optimizing patient recruitment, improving data analysis, and predicting outcomes. In terms of recruitment, AI tools can analyze vast datasets to identify suitable candidates quickly, reducing recruitment time. The source states that “AI can predict whether a trial will succeed, based on the drug molecule, target disease and patient eligibility criteria” (Hutson). These tools can also be used in biological simulation for rare diseases. In general, AI can extract relevant information from medical records, and then interpret medical imaging, enhancing diagnostic accuracy. These technologies facilitate the trial process, cut costs, and improve patient outcomes. However, challenges include data privacy concerns, the need for high-quality data, and regulatory hurdles.

Personal insight: As explored within previous assignments for this course, AI implementation in medical research is not a new phenomenon. Notably, we have seen AI applications in vaccine development for the COVID-19 pandemic. Although its role in research can be beneficial, I strongly believe that there should be oversight concerning what data these AI algorithms are trained on. Historically, medical data has underdiagnosed and misdiagnosed various groups in society, and I wonder how developers of these tools combat this issue.

williamtruong3 — 2024-08-04 04:53:28 UTC

williamtruong3 — 2024-08-04 05:09:56 UTC

In terms of AI integration that I have already seen in use personally, there is the introduction of AI medical scribes, in which machine learning and natural language processing to transcribe physician-patient interactions in real-time. These systems listen to conversations, extract key information, and edit electronic health records automatically. This reduces the administrative burden on physicians, allowing them to focus more on patient care. According to Lindy, benefits include improved documentation accuracy, time savings, and reduced burnout among healthcare providers (Lindy). However, challenges such as ensuring data privacy, maintaining accuracy in diverse clinical settings, and integrating with existing electronic health record systems remain significant hurdles.

Personal insight: Although the company was unspecified, I have personally encountered a physician utilizing an AI scribe to document health information. Although this technology is highly accurate, I do not like how its implementation directly reduces the need for a human scribe to be present. This is an example in which AI tools can be detrimental to the job market, and I am worried about how AI integration will affect the job prospects in medicine.

williamtruong3 — 2024-08-04 05:10:50 UTC

Future Trends and Ethical Considerations

williamtruong3 — 2024-08-04 23:56:09 UTC

The integration of artificial intelligence has such a unique potential in its versatility, efficiency, and accuracy towards analyzing and generating information. When exploring topics regarding AI, there are critical reoccurring themes within this course. Notably, there are strong ethical principles that can potentially be violated if AI integration and growth exceed societal oversight, and currently, we have unfortunately seen AI tools utilized in malicious and harmful forms. When it comes to the field of medicine especially, we are directly dealing with people’s wellbeing, and it is critical that any technology implemented can be controlled, and used only for the greater good of society. Across my three examples of AI integration in the field of medicine, the majority of these tools are capable of assisting the well-being of others. However, similar to what I have mentioned previously, all of them have some unintended consequences. In the case of AI technology assisting in identifying and diagnosing patients with conditions, the algorithms used are created by a large dataset based on past information. As a result, there is heavy potential that these algorithms might have unintended consequences in potential underdiagnosing or mistreatment if the information it is presented with is an area of bias. For example, in black populations, there has been historical mistreatment and late diagnosis of conditions. If the AI inherits this bias and method of treatment, the AI can potentially harm to that population. In addition, the utilization of AI introduces influence on the job market. This is seen in virtually every field that AI is involved in, and the unfortunate reality is that a lot of these AI tools can be used to replace what was once a human job. In this project, I introduced the topic of AI tools in diagnosing patients, as well as being a virtual scribe. In both of those cases, AI systems can be developed and trained to replace some of those jobs within the medical field. Although AI systems are not capable of doing the job as a whole, in a lot of cases, they outperform human counterparts. Since its integration is capable of unintentionally causing harm, we need to have strong ethical and social guidelines, to control such systems and ensure the maximum amount of good, well-mitigating biases and harm.

Societal impact

williamtruong3 — 2024-08-05 00:09:51 UTC

When it comes to these societal impacts shape my perception of AI in healthcare as a double-edged sword. As mentioned previously, the potential for groundbreaking advancements in patient care and efficiency is immense, yet it is tempered by the ethical considerations of employment displacement, privacy risks, and equity concerns. Thoughtful and inclusive policy-making, combined with ongoing dialogue between stakeholders, is essential to harness the benefits of AI while mitigating its adverse effects. When it comes to AI privacy, there needs to be an equal, if not more stringent standard when it comes to data handling, especially as it applies to the medical field. AI systems rely on vast amounts of personal health data to function effectively. Ensuring robust cybersecurity measures and transparent data usage policies is paramount to maintaining public trust. Equity is significantly impacted by AI in healthcare. While AI has the potential to transform healthcare as mentioned above, AI tools can unintentionally increase disparities in technology access. Ensuring that AI technologies are accessible and beneficial to all societal segments, regardless of socioeconomic status, is a critical challenge. Similar to the other section, I am especially interested in how AI tools will continue to influence employment, especially within the healthcare field. AI-driven automation and data analysis can administrative tasks, optimize diagnostic processes, and enhance patient care as mentioned in my technology section. While there are arguments stating that the introduction of these tools increases the need for IT jobs to maintain these systems, I personally believe that it will not be able to counteract the unemployment rates caused by these tools.

Overview Video

williamtruong3 — 2024-08-05 00:53:25 UTC

References

williamtruong3 — 2024-08-05 01:16:51 UTC

1. Al-Antari, M.A. (n.d.). AI algorithms in medical diagnostics. PMC. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9955430/#:~:text=AI%20algorithms%20can%20analyze%20medical,diseases%20more%20accurately%20and%20quickly Rational: Used to understand the extent to which AI tools perform in healthcare.

2. Cheng, Ya-Jian, et al. (2021). AI in arrhythmia diagnosis: New frontiers in cardiology. Heart Rhythm, 21(12). Retrieved from https://www.heartrhythmjournal.com/article/S1547-5271(21)01269-8/fulltext Rationale: Used to see another example in which AI tools can be used in patient care

3. Hutson. (2024). AI in healthcare: The dawn of a new era. Nature. Retrieved from https://www.nature.com/articles/d41586-024-00753-x Rationale: Used to understand how AI tools are used in accelerating clinical trials

4. Khalifa. (2024). AI and personalized medicine: A new frontier in medical imaging. ScienceDirect. Retrieved from https://www.sciencedirect.com/science/article/pii/S2666990024000156#:~:text=By%20thoroughly%20analysing%20medical%20images,the%20advancement%20of%20personalised%20medicine Rationale: Used to understand how AI tools are applied toward the field of radiology

5. King’s College London. (2024, July). AI transforms brain images into realistic synthetic data. Retrieved from https://medicalxpress.com/news/2024-07-ai-brain-images-realistic-synthetic.html

Rationale: Used to see how AI can create clinical conditions, rather an analyze scans

6. Lindy AI. (n.d.). Medical scribe AI. Retrieved from https://www.lindy.ai/medical-scribe

Rationale: Used to give an example in which AI scribes are accessible to healthcare professionals to use curently

7. Napolitano, Elizabeth. (2024). Health insurance and AI: Humana, UnitedHealth use algorithms to optimize care. Retrieved from https://www.cbsnews.com/news/health-insurance-humana-united-health-ai-algorithm/

Rationale: Used to understand how AI is used maliciously in healthcare by insurance companies

8. Tempus. (n.d.). Precision medicine made real: Enhancing genetic testing with AI algorithms. Retrieved from https://www.tempus.com/resources/content/blog/precision-medicine-made-real-enhancing-genetic-testing-with-ai-algorithms/

Rationale: Used to understand how AI tools can play a role in analyzing genetic data and patient information to determine risk

Reflection

williamtruong3 — 2024-08-05 01:32:31 UTC

For this project, I was surprised at how developed some AI tools have been in terms of their ability to perform in a clinical setting. In the majority of the research articles, the technology they have been discussing was shown to be beneficial, and I look forward to learning and working with such technologies in the future. As for challenges faced, I found that finding specific information surrounding companies that utilize/research these AI tools. On their websites, a lot of the information is general, and non-specific about the work that they are doing. As for my knowledge, I feel significantly more informed about how AI has the potential to transform the field of medicine, and the manner in which these tools can be applied non-traditionally. For example, I found it interesting that AI tools were able to produce artificial pictures to how to showcase various conditions present in different humans.