It is a metric that measures the ratio of correct predictions to the total number of predictions. Accuracy is calculated using the formula:
π΄πππ’ππππ¦ = ππ’ππππ ππ πΆππππππ‘ πππππππ‘ππππ πππ‘ππ πππππππ‘ππππ
Accuracy is commonly used when dealing with classification problems, where the goal is to categorize data into different classes.
2.2.7.2. Loss
It is a metric that measures the difference between predicted values and actual values.
The goal is to minimize the value of the loss function to enable the model to learn from the data. Loss functions are often chosen based on the specific type of problem, such as Cross-Entropy Loss for classification.
2.2.7.3. Validation Accuracy
It is the accuracy calculated on a validation dataset, which is an independent dataset not used for training the model. Using a validation dataset helps assess the model's
generalization ability on new data.
2.2.7.4. Validation Loss
It is the value of the loss function calculated on a validation dataset. Similar to validation accuracy, validation loss helps evaluate the performance of the model on data it has not seen during training.
2.2.7.5. Privacy and Data Protection
The European Union (EU) has introduced several General Data Protection Regulations (GDPR), which have influenced changes in privacy regulations globally, including in the United States and Canada. These regulations require that all personal and operational data, whether from the public or foreign companies, be processed by data processors or
25
controllers within the EU to ensure the complete security of citizens' information.
Meanwhile, the United States has enacted the Genetic Information Nondiscrimination Act (GINA), which prohibits employers from making discriminatory decisions based on an individual's genetic health information. Notably, artificial intelligence (AI) in healthcare plays a significant role in analyzing consumer health data and medical device images, enhancing diagnostic outcomes, and advancing health research activities.
Vietnam is a developing country, so the application of artificial intelligence is not yet widely popular. However, the Vietnamese government has built and established many large health data centers such as the Health Management Information System (HMIS - Health Management Information System), whose main goal is to collect, Manage and analyze medical data from basic to mid-level medical facilities. HMIS helps improve health policy decisions and monitor and evaluate the effectiveness of health programs.
There are also other systems such as the National Infectious Disease Reporting System (NIDR), the National Immunization Information System (NIVTS), and the Group Data Management System. blood and blood compatible.
Social media, as part of AI, plays a crucial role in disseminating health news and medical advice, particularly during pandemics. A notable example is the utilization of social networks by Vietnamese authorities during the COVID-19 pandemic to communicate policies, home healthcare methods, and pandemic-related information, including case numbers, fatalities, and recoveries. However, despite these positive aspects, ensuring the safety of patient data remains a significant concern when utilizing AI. Several challenges and concerns related to patient data and AI in healthcare include:
- Insufficient Legal Protection: Current laws and policies do not comprehensively safeguard citizens' health records.
- Security Threats: Clinical data collected can be compromised and misused, posing threats to privacy and security.
- Data Collection by Social Networks: Some social networks clandestinely collect and retain substantial amounts of user data, including individual mental health data, for potential marketing, advertising, and sales benefits.
- Illegal Sale of Consumer Data: Certain healthcare organizations may unlawfully sell consumer data to pharmaceutical and biotechnology corporations without obtaining proper patient consent.
2.2.7.6. Informed Consent and Autonomy
The ethical responsibility surrounding the use of medical records, especially in the context of emerging technologies like AI, is paramount. Patients and healthcare professionals must engage in discussions regarding the purposes of using medical records, such as turning them into reference documents or selling them to pharmaceutical companies. Ethical principles emphasize the patients' right to receive comprehensive information about various aspects of their healthcare. Key considerations include:
26
- Informed Consent: Patients have the right to explicit and specific information about the use of their medical data. Informed consent is crucial for each intended use of their data, particularly with the growing use of AI in healthcare applications.
- Right to Information and Inquiry: Before any medical operations or treatments, individuals have the right to receive information and make inquiries about their diagnosis, treatment plan, results, costs, health insurance, and other medical details.
- Protection of Genetic Information: Genetic information obtained from testing should be protected. Patients should be informed about the treatment process, associated risks, abnormalities in data capture, programming errors, data privacy, and access control.
- Right to Reject Care: Patients have the right to reject any care that medical professionals deem necessary. Informed consent plays a crucial role in respecting patients' autonomy and decision-making.
- Accountability for Robotic Medical Devices: Patients have the right to know who is responsible for malfunctions or errors in robotic medical devices. Clearly defining accountability is essential for upholding patient rights and maintaining the integrity of the healthcare industry.
2.2.7.7. Social Gaps and Justice
The development of artificial intelligence (AI) brings forth the challenge of the social gap, contributing to increased vulnerability to social injustice and inequality on a global scale. While AI enhances access to knowledge across various domains, it also exacerbates social disparities. Here are some examples illustrating the impact of AI on social inequality:
- Growing Disparities Between Nations: Automation and advancements in economies have widened the gap between developed and underdeveloped nations. The unequal distribution of AI technologies and benefits has contributed to increased global economic disparities.
- Job Losses Due to Automation: the increasing sophistication of robots has resulted in the loss of numerous jobs. This is particularly evident in sectors where automation replaces human tasks, leading to unemployment for various roles.
- Impact on Management and Bookkeeping Jobs: The proliferation of automated systems poses a threat to jobs in management and bookkeeping. Individuals in these roles may face unemployment, and those still employed may experience significant declines in wages.
- Challenges in Healthcare Employment: The healthcare industry is transforming with the introduction of robotic nurses and surgical robots. This evolution threatens employment prospects for surgeons and nurses as automation takes over certain responsibilities in patient care.
27
Addressing the social implications of AI necessitates a holistic approach. Strategies should focus on promoting equal access to AI technologies, implementing educational initiatives, and ensuring the ethical deployment of AI to prevent the widening of existing social inequalities. This approach aims to harness the benefits of AI while mitigating its potential negative impacts on various communities.
2.2.7.8. Medical Consultation, Empathy, and Sympathy
The integration of artificial intelligence (AI) into all facets of healthcare is perceived as a formidable and unattainable task due to the emotional void it entails. It seems improbable that patients would willingly opt for a "machine-human" medical interaction over a "human-human" one. The therapeutic journey of patients is profoundly influenced by the compassionate and empathetic care delivered by healthcare professionals.
Unfortunately, replicating such a connection with robotic nurses and physicians appears unfeasible. Collaborating with automated healthcare providers risks depriving patients of empathy, kindness, and appropriate conduct, as these machines lack essential human qualities, particularly compassion. This stands out as a major drawback to implementing artificial intelligence in the medical field.
Several notable limitations arise from the absence of human qualities in AI-powered healthcare:
- Lack of Empathy and Compassion in Obstetrics and Gynecology: Clinical examinations in obstetrics and gynecology inherently demand empathy and compassion, attributes that robotic doctors are incapable of providing.
- Impact on Children's Emotional Well-Being: Children often experience fear or anxiety when interacting with healthcare professionals. Introducing robotic medical systems may escalate their aggressive, withdrawn, and uncooperative behaviors, posing challenges in healthcare settings.
- Negative Effects on Patients with Mental Health Issues: Patients grappling with severe mental health challenges may face adverse consequences with the deployment of medical robots in psychiatric hospitals, potentially exacerbating their conditions.
Navigating the ethical terrain of AI in healthcare necessitates a delicate balance between technological advancement and preserving the human-centric qualities crucial for effective and compassionate patient care.