The complexity of cancer care has increased significantly over the years. Those who were once considered the only diseases are divided into many subtypes that require different treatment plans, each based on clinical guidelines, each.
Difficulty
This has created a growing difficulty for oncologists who managed a wide range of cancer types and to keep up with rapidly changing best practices.
Another major challenge in oncology today is the volume and complexity of developing clinical guides. National organizations such as the National Comprehensive Cancer Network, the American Clinical Oncology Association and the American Cancer Association regularly update their suggestions based on new clinical research data, caused treatments and developing treatment paradigms hundreds of times a year.
These guidelines are not always standardized among organizations, and individual cancer centers often make it more difficult for clinicians to monitor and practice the best practices by adding their own specialty layers.
He is also an assistant professor at the University of California in San Francisco. Travis Zack said that access to private oncologists is difficult.
“Many regions face the problems of oncologists, forcing general practitioners to take more responsibility for their first cancer work and treatment planning.” “However, GPs often lack time or special training to be fully updated in the latest oncology guidelines, which can lead to inconsistencies and delays in treatment.
“There is also time to collect and review this information in accordance with the updated treatment guidelines to make the basic difficulty of unconfaced patient data and the best possible suggestions for the patient.”
The University of California in San Francisco, which recognizes these challenges, has tried to develop AI technology that can automate the latest clinical guides for oncologists with all information about patients.
“The aim was to create a decision support system that could integrate national directives and patient data with the best local corporate best practices, and enabled each patient to get the most up -to -date, evidence -based maintenance without adding additional cognitive burden to clinicians.”
“This is the main challenge-while optimizing the doctor’s time, to enable them to have a quick and reliable access to current, evidence-based suggestions-pushed the first-class oncology expertise to explore AI-oriented systems that can make more accessible, efficient and scalable.”
OFFER
The artificial intelligence system would combine a large language model informed by the national and local institutional directives in force with transparent logic, so that clinicians could fully see how and why the AI made suggestions.
To achieve the purpose Each oncology consultancy started with a full, configured and current data set, reduced the information gaps and optimized the duration of the doctor to complete the patient studies.
To achieve this, Zack explained that AI was designed with two basic functions:
-
Collection and configuration of clinical data – The system attracts and regulates the relevant patient information from electronic health records to create a comprehensive appearance of the patient’s condition. If the critical data such as biopsy results such as molecular test or staging screening are missing, AI flags before oncology counseling to prevent unnecessary delays.
-
Integrate national and local clinical guides -AI includes both standard guides (sources such as NCCN, ACS and Asco) and institutional protocols, which allows doctors to offer the most important, current treatment suggestions adapted to the patient’s specific condition.
“For example, if a patient is referred to suspicious lung cancer, the system can automatically evaluate whether all the necessary diagnostic steps have been taken.” “If a turnkey is missing, the referral pushes the referral to order to order before the oncology visit. During the consultation, AI provides a evidence -based framework for decision -making, reduces the cognitive load on the doctor while complying with the best practices.
“The inclusive aim was not to replace the human judgment, but to improve it – allowing oncologists to focus on personalized treatment decisions instead of spending valuable time to get information and verify the information.”
Meet the struggle
. AI technology was deployed in oncology workflows to support both practitioners and oncologists, and enabled each step in the patient journey to be directed by extensive, evidence -based insights.
For the published UCSF study, the color clinicians of the Health CT and Clinical Services Company analyzed the patient cases defined by UCSF-50 for breast cancer and 50 for 50 and 50 for colon cancer. Each case included two registration sets: Treatment records covering diagnostic records and treatment records containing all the information available until the date of diagnosis, but not included in the treatment date.
In order to evaluate artificial intelligence, color clinicians committed these cases in two stages:
-
Diagnostic Run Type: 100 patient cases (50 breasts, 50 columns) uses existing records until the diagnosis date.
-
Treatment Type: Until the date of starting treatment, but not beyond the date of starting, 100 patient cases with records (50 memes, 50 columns).
Zack, “a colorful primary step physician AI has reviewed the output and made adjustments when necessary.” He said. “The performance of the system was evaluated by monitoring the number of modifications made in three key areas: the accuracy of the decision factors, the interest of the proposed studies with the condition of the patient and the completeness of the related studies.
“The AI system was integrated with electronic health records and other medical databases to facilitate access and interpretation of patient information.” “Patient data is defined to maintain confidentiality. The system is also integrated into various technical flows to understand and evaluate updated clinical guides for breast and colon cancer types.”
So how did he work in practice? Like this:
-
Data collection and configuration. Before an oncology consultation, AI automatically compiled all clinical information from the patient’s records and determined the missing diagnostic steps.
-
Guide -based recommendations. At the point of maintenance, the system has made special suggestions based on national directives and institutional policies.
-
Continuous learning and updates. Artificial Intelligence, dynamically included the latest clinical research and guide updates, allowing doctors to work with the most up -to -date evidence.
“By reducing the time spent on administrative duties and eliminating inconsistencies in care, AI allowed the focus on patient interactions and treatment planning with the intention of faster and more effective cancer care.” He said.
Consequences
. The application of artificial intelligence in oncology workflows has led to measurable improvements in productivity and decision -making. One of the most remarkable results was a significant decrease when oncologists spent patient records and clinical guidelines before making treatment decisions.
“This process may take a to two hours for complex cases that require the review of comprehensive medical past and developing guidance suggestions,” Zack, “Previously.” “While the AI system has a system, this time it has been reduced to about 10 to 15 minutes in most cases. By configuring data collection and relevant clinical information, the system allows oncologists to focus on making decisions rather than manual data intake.
“Another important finding was a high level of harmony between the suggestions produced by AI and the suggestions made by oncologists.” “In a comparative study, there was a 95% harmony between AI’s treatment suggestions and clinical decisions based on standard guides by oncologists.”
This shows that the AI system effectively synthesized and implemented national and institutional directives to support clinical decisions. Although human supervision continues to be the basis, this level of agreement shows that AI can be a reliable tool to strengthen evidence -based care.
Zack, “Also, the system contributed to the on -time improvements.” Delays in ordering basic diagnostic tests such as “biopsies or genomic tests can sometimes extend the time between diagnosis and treatment up to weeks or months.
“The artificial intelligence system helped to reduce these delays by determining the missing but necessary studies at the beginning of the process, allowing patients to progress on time on time.” “Considering that premature intervention is critical in oncology, this decrease in delays represents an important improvement in patient care.”
In general, these consequences suggest that AI may play a significant role in efficiency, standardization and time development in oncology care, especially in environments where access to specialized expertise may be limited.
Recommendation for others
A strategic and structured approach is essential for health institutions that wish to integrate artificial intelligence into oncology or other expertise.
“One of the primary issues is to ensure that the AI system has access to comprehensive and accurate patient data.” He said. “AI -oriented decision -making tools rely on a full set of data to create clinically significant suggestions.
“However, coexisiness difficulties between electronic health records and other data sources may result in incomplete clinical paintings that may affect the reliability of AI outputs.” “Considering these gaps through effective data integration and standardization should be a priority before implementation.”
Another important factor, AI -guided suggestions and clinical judiciary is the balance.
“Artificial intelligence should be seen as a tool to support instead of oncologists and other health service providers.” “Organizations should ensure that clinicians continue to interpret insights produced by AI and ensure that they can invalidate or change their suggestions when necessary.
“To facilitate this, AI systems should provide transparent and explanable decision -making ways and allow users to understand how suggestions are created.” “A visible visibility to basic logic is confident in the AI -supported decision -making process and encourages adoption among clinicians.”
Follow Bill’s hit in LinkedIn: Bill Siwicki
Send him e -mail: bsiwicki@himss.org
Healthcare It News is a HIMSS media broadcast.
Watch now: Seattle Children’s chef AI Officer gives better results through technology
