Literature Review

Machine learning for targeted advance care planning in cancer patients: A quality improvement studyJournal of Pain and Symptom Management; by Mihir N Patel, Alexandria Mara, Yvonne Acker, Jamie Gollon, Noppon Setji, Jonathan Walter, Steven Wolf, S Yousuf Zafar, Suresh Balu, Michael Gao, Mark Sendak, David Casarett, Thomas W LeBlanc, Jessica Ma; 12/24Objectives: Examine a quality improvement mortality prediction algorithm intervention's impact on ACP documentation and EOL care. Conclusion: Identifying patients with cancer and high mortality risk via machine learning elicited a substantial increase in documented ACP conversations but did not impact EOL care. Our intervention showed promise in changing clinician behavior. Further integration of this model in clinical practice is ongoing.

Large language model influence on diagnostic reasoning-A randomized clinical trialJAMA Network Open; Ethan Goh, MBBS, MS; Robert Gallo, MD; Jason Hom, MD; Eric Strong, MD; Yingjie Weng, MHS; Hannah Kerman, MD; Joséphine A. Cool, MD; Zahir Kanjee, MD, MPH; Andrew S. Parsons, MD, MPH; Neera Ahuja, MD; Eric Horvitz, MD, PhD; Daniel Yang, MD; Arnold Milstein, MD; Andrew P. J. Olson, MD; Adam Rodman, MD, MPH; Jonathan H. Chen, MD, PhD; 10/24Large language models (LLMs) have shown promise in their performance on both multiple-choice and open-ended medical reasoning examinations, but it remains unknown whether the use of such tools improves physician diagnostic reasoning. In this trial, the availability of an LLM to physicians as a diagnostic aid did not significantly improve clinical reasoning compared with conventional resources. The LLM alone demonstrated higher performance than both physician groups, indicating the need for technology and workforce development to realize the potential of physician-artificial intelligence collaboration in clinical practice.

With the rise of AI, is a Chief Automation Officer necessary? HR Daily Advisor Newsletter - Technology; by Lou Bachenheimer; 11/11/24 In the age of digital transformation and advanced technology like AI, automation is a critical element for organizations to stay competitive. That’s why many C-suites are considering incorporating a Chief Automation Officer (CAO) into the fabric of their organization to provide senior executive oversight of automation initiatives. But is a whole new role in the c-suite needed and necessary to successfully implement AI and automation into your organization? ... Instead of pouring new resources and energy into new roles that are created off the back of hype, what is needed is executive sponsorship and leadership of advanced automation programs at the highest and most influential levels, underpinned by the appropriate business knowledge and network to be able to drive real change.

Adopting tomorrow’s tech: 5 steps to effective AI training in hospice Hospice News; by Mick Stahlberg; 11/7/24 In today’s hospice environment, providers are continually asked to do more with less. Staffing shortages, rising costs and growing patient demands make efficiency a critical component of success. Enter artificial intelligence (AI). From automating administrative tasks to improving scheduling and optimizing resource allocation, AI is poised to significantly enhance operational efficiency at every level. In turn, proper training will play a monumental role in helping teams get comfortable incorporating this technology into their clinical and administrative workflows from day one. Providers will be best served to follow 5 key steps to effective AI training as they work with their technology partners to articulate a comprehensive strategy based on their organization and its unique needs. These steps incorporate:

Top 10 Governance, Risk & Compliance (GRC) tool eSecurity Planet; by Claire dela Luna; 10/28/24 In today’s global economy, even small businesses operate across multiple regions and markets, each with its own regulatory landscape. Without effective GRC tools, organizations can quickly fall out of compliance, exposing themselves to security risks and operational disruptions. GRC software solutions enable businesses to manage these risks by automating policies, tracking controls, and providing real-time compliance monitoring across international borders. .. Here’s a guide to help you make the best choice.

Change Healthcare cyberattack impacts 100 million people Becker's Health IT; by Naomi Diaz; 10/25/24 The Feb. 21 ransomware attack on UnitedHealth Group subsidiary Change Healthcare has impacted 100 million individuals. The number of impacted individuals was posted on the Office for Civil Rights Breach Portal, which is used for reporting breaches of unsecured protected health information under HIPAA. Previously, UnitedHealth said that the data stolen by hackers likely covered a "substantial proportion of people in America." The cyberattack crippled financial operations for hospitals, insurers, pharmacies and medical groups nationwide. In July, the organization began sending out breach notification letters to individuals affected by the attack.  

Perspectives on artificial intelligence–generated responses to patient messagesJAMA Network Open; Jiyeong Kim, PhD, MPH; Michael L. Chen, BA; Shawheen J. Rezaei, MPhil; April S. Liang, MD; Susan M. Seav, MD; Sonia Onyeka, MD; Julie J. Lee, MD, MPH; Shivam C. Vedak, MD, MBA; David Mui, MD, MBA; Rayhan A. Lal, MD; Michael A. Pfeffer, MD; Christopher Sharp, MD; Natalie M. Pageler, MD, MEd; Steven M. Asch, MD, MPH; Eleni Linos, MD, DrPH; 10/24Generative artificial intelligence (AI) has the potential to assist clinicians in responding to patients’ messages. Satisfaction was consistently higher with AI-generated responses than with clinicians overall and by specialty. However, satisfaction was not necessarily concordant with the clinician-determined information quality and empathy. For example, satisfaction was highest with AI responses to cardiology questions while information quality and empathy were highest in endocrinology questions. Interestingly, clinicians’ response length was associated with satisfaction while AI’s response length was not. The findings suggest that the extreme brevity of responses could be a factor that lowers satisfaction in patient-clinician communication in EHR.

"Hospice Care Could Be a Compassionate Choice": ChatGPT Responses to Questions About Decision Making in Advanced CancerJournal of Palliative Medicine; by Meghan McDarby, Emily L Mroz, Jessica Hahne, Charlotte D Malling, Brian D Carpenter, Patricia A Parker; 9/24Objective: To examine the content of ChatGPT responses to a hypothetical patient question about decision making in advanced cancer... ChatGPT responses (N= 96) were coded for mentions of: hospice care, palliative care, financial implications of treatment, second opinions, clinical trials, discussing the decision with loved ones, and discussing the decision with care providers... Responses more frequently mentioned clinical trials for vignettes describing 45-year-old patients compared with 65- and 85-year-old patients. When vignettes mentioned a preexisting recommendation for hospice, responses more frequently mentioned seeking a second opinion and hospice care... ChatGPT responses to questions about advanced cancer decision making can be heterogeneous based on demographic and clinical characteristics. Findings underscore the possible impact of this heterogeneity on treatment decision making in patients with cancer.

Florida health system rolls out virtual nursing tech to 1K rooms Becker's Health IT; by Giles Bruce; 10/17/24 Fort Myers, Fla.-based Lee Health is expanding virtual nursing technology to nearly 1,000 patient rooms, representing about half its beds. The four-hospital system is rolling out dual-camera virtual care devices from telehealth company Caregility to scale its virtual nurse offering. The remote fleet management tools come with artificial intelligence capabilities and cloud-based updates. "Our virtual nursing program introduces innovative possibilities in patient care workflows by connecting the care in the complex patient journey," Lee Health Chief Nursing Officer Jennifer Higgins, DNP, RN, said in an Oct. 16 statement. "From admissions and discharges to continuous patient monitoring, patients benefit from an additional team member dedicated to their care."

Online plain language tool and health information quality-A randomized clinical trialJAMA Network Open; Julie Ayre, PhD; Carissa Bonner, PhD; Danielle M. Muscat, PhD; Erin Cvejic, PhD; Olivia Mac, MPH; Dana Mouwad, MHSc (Nsg); Heather L. Shepherd, PhD; Parisa Aslani, PhD; Adam G. Dunn, PhD; Kirsten J. McCaffery, PhD; 10/24National and international policies recognize that health literacy—a person’s capacity to access, understand, and act on health information—is a critical source of inequity in our health systems. Low health literacy contributes to higher mortality, morbidity, rates of hospitalization, emergency department visits, and medication errors independently of other social determinants of health, such as age, education, and socioeconomic disadvantage. In this randomized clinical trial, the Health Literacy Editor helped users simplify health information and apply health literacy guidelines to written text. The findings suggest the tool has high potential to improve development of health information for people who have low health literacy.

Racial differences in pain assessment and false beliefs about race in AI modelsJAMA Network Open; Brototo Deb, MD, MIDS; Adam Rodman, MD, MPH; 10/24Physicians undertreat Black patients’ pain compared with White patients, irrespective of setting and type of pain, likely from underassessment of pain and undertreatment of pain on recognition. Large language models (LLMs) encode racial and ethnic biases and may perpetuate race and ethnicity–based medicine. Although LLMs rate pain similarly between races and ethnicities, they underestimate pain among Black individuals in the presence of false beliefs. Given LLMs’ significant abilities in assisting with clinical reasoning, as well as a human tendency toward automation bias, these biases could propagate race and ethnicity–based medicine and the undertreatment of pain in Black patients. Mitigating these biases involves many strategies during dataset preparation, training, and posttraining stages.

AI and change management: Navigating the human side of AI Method; by Jaon Rome; 10/14/24 ... An effective AI change management strategy must consider multiple employee groups — including front-, middle-, and back-office workers — whom AI implementation most affects. How you talk about AI and introduce AI tools to your employees influences how your people perceive AI, which affects their acceptance and adoption of it. ... Your AI investments’ success depends on your people. They must buy into AI’s value, see themselves as innovating with AI rather than training their replacements, and use the AI tools you’ve built in their day-to-day work.