NVIDIA Works With Deloitte to Deploy Digital AI Agents for Healthcare NVIDIA Blog

AI chatbots outperform doctors in empathy and readability for cancer-related questions, study finds

In a big Californian hospital system, researchers followed 9,000 doctors for ten weeks in a pilot test of a digital scribe. The application silently records the conversation between a clinician and a patient (via a phone, tablet or computer microphone, or a dedicated sensitive microphone). “We have a responsibility to harness the power of ‘AI for good’ and direct it towards addressing pressing societal challenges like health inequities,” Nadarzynski said in a statement. “You have to have a human at the end somewhere,”said Kathleen Mazza, clinical informatics consultant at Northwell Health, during a panel session at the HIMSS24 Virtual Care Forum. Get stock recommendations, portfolio guidance, and more from The Motley Fool’s premium services. Recent contract wins and coverage of emerging opportunities have helped power an incredible bull run for the company.

To facilitate effective evaluation and comparison of diverse healthcare chatbot models, the healthcare research team must meticulously consider all introduced configurable environments. By collectively addressing these factors, the interpretation of metric scores can be standardized, thereby mitigating confusion when comparing the performance of various models. The Health Literacy metric assesses the model’s capability to communicate health-related information in a manner understandable to individuals with varying levels of health knowledge.

AI chatbots outperform doctors in empathy and readability for cancer-related questions, study finds

Patients and healthcare professionals alike must be able to trust these intelligent systems to safeguard sensitive information and provide reliable insights. For this, regulators should establish a robust data security framework as well as ethical guidelines for the training and use of these systems. The mean Flesch-Kincaid grade level of physician replies (mean, 10.1) was not significantly different from the third chatbot’s responses (mean, 10.3), although it was lower than that of the first (mean, 12.3) and second chatbots (mean, 11.3). Researchers consistently scored chatbot replies higher concerning empathy, quality, and readability in writing styles.

For instance, the number of parameters in a language model can impact accuracy, trustworthiness, and empathy metrics. An increase in parameters may introduce complexity, potentially affecting these metrics positively or negatively. Conversely, a low parameter count can limit the model’s knowledge acquisition and influence the values of these metrics. The Privacy metric is devised to assess whether the model utilizes users’ sensitive information for either model fine-tuning or general usage42. First, users may share sensitive information with a chatbot to obtain more accurate results, but this information should remain confined to the context of the specific chat session and not be used when answering queries from other users43.

Healthcare professionals can assess the chatbot’s performance from the perspective of the final users, while intended users, such as patients, can provide feedback based on the relevance and helpfulness of answers to their specific questions and goals. As such, these guidelines should accommodate the different perspectives of the chatbot’s target user types. General-purpose human evaluation metrics have been introduced to assess the performance of LLMs across various domains5. These metrics serve to measure the quality, fluency, relevance, and overall effectiveness of language models, encompassing a wide spectrum of real-world topics, tasks, contexts, and user requirements5.

SoundHound AI Named Leader in Healthcare Conversational AI, Taps Into $2.3B Market Opportunity – StockTitan

SoundHound AI Named Leader in Healthcare Conversational AI, Taps Into $2.3B Market Opportunity.

Posted: Thu, 12 Dec 2024 08:00:00 GMT [source]

Success will come to those who approach AI strategically, align it with their mission and focus on solving real problems for their patients and staff. However, successful organizations will approach this proactively and transparently, focusing on how AI can enhance, rather than replace, human work. At Intermountain Healthcare, small tests of change with a subset of users proved powerful. When staff members could tell colleagues that new tools had saved them considerable time, adoption accelerated naturally.

California Health Care Foundation

It envisions a future where healthcare staff excel in both technical skill and emotional intelligence, enabling them to meet the psychological needs of patients with genuine understanding and compassion. AI is proving valuable in drug discovery and the identification of chemical markers from the body, such as those that can signal the presence of cancer. In addition, AI technology like that behind ChatGPT can process medical literature and patient records to help make complex diagnoses. The philosophy underpinning deep phenotyping is to combine this diverse data to enable more accurate and speedy diagnoses, precise and effective treatments, and to advance predictive and preventative medicine strategies. However, the sheer volume and complexity of the collected data pose significant challenges for analysing it all.

The team then calculated that any claim that is worth less than this value could be automatically approved because it is never going to cost more than the bad experience. The private health firm’s use of automation has expanded beyond customer response, to also include processing of customer claims – something that they admit was seen as a bit of a risk initially. Simplyhealth’s tech transformation began when Eddie joined in November 2021 but it accelerated significantly when Nicholls arrived from money management and budgeting platform Snoop a year later. Given that the company was founded more than 150 years ago, it has been, in many respects, a “very traditional” business, explains Eddie. “You’re not just buying today’s product — you’re buying the company and partnership,” Neinstein pointed out.

Another ED physician noted spending hours a day requesting and sifting through several hundred pages of a patient’s records, and now he is able to do so in minutes. In particular, the solution has been exceptional at interpreting handwritten notes and turning them into searchable discrete data. We had multiple waves of provider go-lives at Mile Bluff, and the number of pilot users continued to grow, now surpassing 150 users. Providers were intuitively using the section breakdown within hours of going live to review their provider notes. CMO Dr. Angela Gatzke-Plamann saw the 15 minutes she spent per patient cleaning up problem lists decline dramatically.

ON THE RECORD “One of today’s most important and widely used healthcare technologies, the EHR, has not lived up to its promise,” said Verma in a statement. “Most EHRs were built in the 90s and are ill-equipped to meet the complex security requirements and clinical needs of today’s healthcare networks, practitioners, and patients. That is why we are completely reinventing the EHR. THE LARGER TREND Oracle has been upgrading its various healthcare technologies with new artificial intelligence capabilities since it finalized its $28 billion Cerner acquisition in June 2022. Helping health system clients improve their financial performance while closing gaps in care and succeeding with value-based care models is another stated goal of the new system. The company says the new EHR is also optimized to enable better information sharing between payers and providers and boost patient recruitment capabilities for clinical trials.

Recently, the Australian Health Practitioner Regulation Agency released a code of practice for using digital scribes. Both warn clinicians that they remain responsible for the contents of their medical records. AI is being used inpatient scheduling, and with patients post-discharge to help reduce hospital readmissions and drive down social health inequalities. Thus far, SoundHound AI’s core sources of business have been in categories that include restaurants and automotive.

Building the Connected Hospital: Bridging Operational Gaps Through Technology

However, with so many buzzwords surrounding AI, it’s becoming increasingly difficult to cut through the noise and find which systems will have the greatest impact. Considering providing a positive and consistent patient experience is how health systems can distinguish themselves and meet patients’ expectations. While many organizations in the healthcare domain are bullish on the potential of conversational AI, its widespread adoption still remains hurdled by multiple challenges. The researchers assessed reading comprehension cognitive load using mean dependency distances for syntactic complexities and textual lexical diversities. They offered recommendations to limit the length of the chatbot response to the average physician response word count (125). They conducted a one-way analysis of variance (ANOVA) with post-hoc tests to evaluate 200 readability, empathy, and quality ratings and 90 readability metrics between chatbot and physician replies.

Nabla, one of the providers of such copilots, even uses these notes to generate a set of patient instructions, on behalf of the physician. This capability can be further developed into a gen AI system that sits alongside the doctor and creates personalized treatment and therapy plans based on the current conditions as well as previously recorded parameters, including genetic makeup, health history, and lifestyle. By incorporating machine learning models (ML), deep learning (DL), and natural language processing (NLP), we can transform time consuming manual chemical research into a tool where you can speak or type your queries. AI-powered assistant interfaces can intelligently interpret the questions or instructions and render efficient answers.

He implies that the benefits or value propositions of these AI solutions may need to be more compelling to drive widespread adoption in the market, leading to a lack of commercial viability. Although the WHO states that the AI bot is updated with the latest information from the organization and its trusted partners, Bloomberg recently reported that it fails to include the most current U.S.-based medical advisories and news events. Doctors are told by tech and insurance companies that they must check every summary or letter (and they must). Tired or inexperienced clinicians might think their memory must be wrong, and the AI must be right (known as automation bias). During appointments, clinicians can have their attention divided between good record-keeping and good communication with the patient.

Most recently, the company announced this week its intent to apply for Qualified Health Information Network status, to help its EHR customers more easily take part in information sharing under the TEFCA nationwide interoperability framework. Oracle has been promising for some time to reinvent its core health IT offerings and move “beyond the EHR.” This past month, the company unveiled enhancements to the Oracle Health Seamless Exchange, new improvements for its Oracle Health Ambulatory Referral Management and other advancements for its EHR product. Jackie Rice, vice president and CIO at Frederick Health, will join us in our booth on Wednesday March 13 at 3 p.m.

• Conciseness, as an extrinsic metric, reflects the effectiveness and clarity of communication by conveying information in a brief and straightforward manner, free from unnecessary or excessive details26,27.
• A. In the realm of healthcare, the abundance of misinformation can leave individuals feeling lost and uncertain.
• The first strategy involves scoring after each individual query is answered (per answer), while the second strategy involves scoring the healthcare chatbot once the entire session is completed (per session).
• This is good news for today’s pressed medical workforce, who would prefer to exercise their clinical expertise over appointment scheduling or patient navigation.
• New York-based Hyro offers a conversational AI-enabled call center for providers that allows for automated conversations with their patients.

NVIDIA ACE is a suite of AI, graphics and simulation technologies for bringing digital humans to life. It can integrate every aspect of a digital human into healthcare applications — from speech and translation abilities capable of understanding diverse accents and languages, to realistic animations of facial and body movements. These virtual teammates, built with the NVIDIA AI Enterprise software platform, can have natural, human-like conversations with patients, answer a wide range of questions and provide support prior to preadmission appointments at hospitals. Z.S.H.A. contributed to give guidance, revise critically the paper, and design of the visualizations.

When it comes to retaining current healthcare staff and improving their day-to-day roles, agentic AI plays an important role both on the administrative and clinical front. Data privacy and security is arguably the most important feature to look for in an AI system. When it comes to handling sensitive patient information, the system you choose must provide transparency around data governance, compliance and privacy standards. These systems should also grant robust control to help healthcare leaders build confidence in these systems.

SuperDial acquires MajorBoost to expand voice AI capabilities for US healthcare organizations – GlobeNewswire

SuperDial acquires MajorBoost to expand voice AI capabilities for US healthcare organizations.

Posted: Mon, 13 Jan 2025 08:00:00 GMT [source]

It’s not only important your AI systems meet your organization’s needs today, they also must be able to scale to meet your evolving needs down the road. Systems should offer a track record of seamless tech stack integrations and a broad solutions library to enable this. Appointment scheduling is a critical operation in healthcare facilities but can be a time-consuming operation. Agentic AI can allow patients to easily schedule, reschedule or cancel appointments without the frustrations of a complicated interface or waiting times in a call center.

The approach worked but left physicians overworked as they had to deal with both online and offline patients. With gen AI, healthcare organizations can launch LLM-backed AI assistants to address this. Essentially, they could fine-tune models like GPT-4 on medical data and build assistants that could take basic medical cases and guide patients to the best treatments on the basis of their systems. If any particular case appears more complicated, the model could redirect the patient to a doctor or the nearest healthcare professional.

“While screening postpartum patients, we found 3% of patients have new onset hypertensive disorder of pregnancy, of whom approximately 45% have no symptoms of elevated blood pressure,” she continued. “In addition to blood pressure screening, making it easy for patients to ask questions about symptoms has also allowed us to detect this potentially serious condition.” We spoke with Polyak and Kowalczyk again recently for a more detailed and data-driven conversation about what patients expect from AI in their own care. But according to recent research into patient attitudes on AI, providers should be thinking carefully about how they deploy those tools if they want to preserve patient trust. Escalating technology costs was a sentiment echoed by many digital health leaders last week at HIMSS AI in Healthcare Forum in Boston.

Beyond Boundaries: The Promise Of Conversational AI In Healthcare

The study indicated rapid and widespread adoption of AI within the healthcare and life sciences sector is driving substantial business impacts, with 81% of AI-mature healthcare and life sciences enterprises performing better in 2023, compared with 2022. According to the report, healthcare and life sciences organizations have deployed an average of more than 170 AI models in production, a figure expected to grow to 182 within the next year. “AI is transforming nursing workflows by streamlining administrative tasks, allowing nurses to focus more on patient care,” Corey Miller, vice president of research and development at Epic, said in a statement. “First, a frequently asked question bank was used to generate accurate mapping of questions to the appropriate responses,” Leitner explained. “Second, surveys (standardized conversation templates designed to collect patient data) were created by patients’ clinical characteristics (for example, breast milk versus formula fed).

Various automatic and human-based evaluation methods can quantify each metric, and the selection of evaluation methods significantly impacts metric scores. Automatic approaches utilize established benchmarks to assess the chatbot’s adherence to specified guidelines, such as using robustness benchmarks alongside metrics like ROUGE or BLEU to evaluate model robustness. A significant relationship exists between performance metrics and the other three categories.

• Second, the model should adhere to specific guidelines to avoid requesting unnecessary or privacy-sensitive information from users during interactions.
• Our focus is on establishing an open architecture for openCHA, forging connections with other open health technologies, accessing open-content resources, and shaping future standards for CHAs.
• While these quantitative metrics are important, truly assessing conversational AI success requires an appreciation for the human touch and an understanding of user engagement.
• An AI company with offerings spanning health and wellness as well as gaming, Gaxos.ai Inc.
• And when it comes to delivering effective creative at speed and at scale, EPIC, IPG Health’s tech-enabled end-to-end data platform, transforms healthcare and audience data into actionable insights that fuel the network’s renowned creative output.
• At Intermountain Healthcare, small tests of change with a subset of users proved powerful.

We’re just talking about using images and text, and basically taking that existing data and feeding it into an AI. In other areas, there’s going to be some work that probably needs to be done, a lot more work. For instance, in pathology, those departments, the images may not even have digital images; they may just be using old-school light microscopes.

The most important foundational first steps are establishing the use case you want to start with, ensuring you have set metrics to define success and how you are measuring ROI, and understanding what AI system to choose. Under Rhiannon White’s leadership, Clue is transforming menstrual health tracking into a powerful tool for research and improved reproductive health outcomes. In this interview, NewsMedical speaks with Esther Kieserman and Arvonn Tully from Yokogawa Life Science about how confocal-based high-content imaging is advancing core facility research and improving data reliability. With a $2 million Small Business Innovation Research (SBIR) contract from the National Cancer Institute (NCI) within the National Institutes of Health (NIH), Pieces and MetroHealth will deploy and study how PiecesChat converses with patients. In 2022, a group of researchers published a literature review showing limited diagnostic accuracy for online symptom checkers. Depending on the study in the literature review, diagnostic accuracy ranged from 19% to 37.9%.

Chatbots can help patients obtain their appointments while also answering questions like where patients can park at the clinic or where they can get directions to a certain department in the clinic. Folks who use these tools tend to be younger, female and have higher digital health literacy, posing a risk of creating a steep digital divide. These tools leverage AI chatbot functionality by allowing patients to input their symptoms and producing a list of likely diagnoses. From that list, patients can determine where they should access care (the urgent care versus the emergency department, for example) or identify how to best ride out symptoms on their own, like in the case of displaying cold symptoms. “That would mean people could get access to their everyday healthcare needs while working in Britain pretty much instantaneously and see a healthcare professional either digitally or in-person,” explains Dan Eddie, customer services director at Simplyhealth. There is also a transformative opportunity to rethink efficiency, putting relationships between patients and staff at the core.

In healthcare chatbots, where human inquiries may not precisely align with their underlying issues or intent, robustness assumes paramount importance. Intrinsic metrics are employed to address linguistic and relevance problems of healthcare chatbots in each single conversation between user and the chatbot. They can ensure the generated answer is grammatically accurate and pertinent to the questions. Making sure you choose a system not at risk for hallucinating and able to give patients a timely and accurate response is critical in any industry but especially in healthcare.

In its 2024 Global Healthcare Outlook report, Deloitte describes how the industry’s transformation is being driven by technology, demographic changes and evolving patient needs. This is all happening against the lingering effects of the pandemic, rising costs and labor shortages. Following SoundHound’s acquisition of Amelia, the company has positioned itself among the largest publicly traded conversational AI companies, combining advanced voice AI capabilities with Amelia’s enterprise platform to address complex healthcare use cases. The organizations that will thrive in 2025 aren’t necessarily those with the biggest AI budgets or the most advanced technology.

The broader AI market reached $184 billion in 2024, showing a $50 billion increase from 2023, with projections to exceed $826 billion by 2030. Healthcare remains a critical application area for AI, focusing on improving diagnostics, personalizing treatment plans, and optimizing patient care. Your current data combined with historical knowledge, tossed in a blender with a powerful AI facilitates predictive analysis.

In fact, measuring the success of agentic AI relies on several essential quantitative metrics. Take for example a patient with diabetes who can leverage these systems to track their blood sugar levels, insulin dosages and other related metrics. The assistant can provide personalized feedback on managing an array of health scenarios as well as suggest lifestyle changes or adjusting medication dosages.

This evaluation aids patients with low health knowledge in comprehending medical terminology, adhering to post-visit instructions, utilizing prescriptions appropriately, navigating healthcare systems, and understanding health-related content52. For instance, “pneumonia is hazardous” might be challenging for a general audience, while “lung disease is dangerous” could be a more accessible option for people with diverse health knowledge. Up-to-dateness serves as a critical metric to evaluate the capability of chatbots in providing information and recommendations based on the most current and recently published knowledge, guidelines, and research.