GPT stands for “generative pre-training transformer” which is a coding architecture that is fine-tuned for specific tasks such as language translation, question-answering and text summarization. According to Tesca Fitzgerald, assistant professor of computer science and co-instructor of Computer Science 470: “Artificial Intelligence,” ChatGPT is the most advanced of a new generation of AI systems that can respond to users, generate coherent texts and even produce generated images based on a coded database of online content, digital books and various other media.

“Various beta versions of ChatGPT have existed before,” Fitzgerald said. “For the first time, anyone can access it.”

The previous versions of these “large-language models” were only available to researchers and  used in academic settings. What makes ChatGPT different is its accessibility to anyone with internet access. To use ChatGPT, all users have to do is make an account that is tied to their email address. After joining, users can type various prompts and questions into the chat box, and the bot will respond with sentences that mirror human writing structure.

When the News asked ChatGPT to “explain ChatGPT to a five year old,” the AI responded with the following:

“ChatGPT is like a robot that can talk and understand what you say. It is very good at understanding words and sentences, and it can also make up new sentences that make sense. Just like how you can talk to your friends and tell them stories, ChatGPT can also have a conversation with you and tell you stories. It’s like a magic box that can understand and generate new sentences.”

In addition to asking requests, users can prompt ChatGPT to generate various texts, such as full-length essays and stories. Users could ask ChatGPT to write a 500-word essay about the American Revolution, and ChatGPT would be able to generate a unique response that mirrored human writing patterns.

However, the AI bot’s ability to generate these comprehensive and coherent texts has called into question the effects of the bot on education.

Professor of computer science Dragomir Radev noted that users must  be “very careful when using ChatGPT” due to the “potential for false and misleading information.” According to Radev, although ChatGPT can generate comprehensive texts on various topics, its knowledge is only as strong as its dataset. If the dataset contains false or misleading information, ChatGPT will use that information to generate its response.

Radev also acknowledged the bot’s capabilities. He found the bot’s performance to be “amazing” noting that when he asked the bot to translate some English sentences into Bulgarian, it did so perfectly. Fitzgerald pictures ChatGPT’s role in education as a catalyst for change.

“This will force professors and teachers to re-evaluate how we assess our students’ knowledge of the things we teach,” Fitzgerald said. “I think ChatGPT will foster more critical thinking in the classroom.”

Fitzgerald recounted her own experience playing around with the bot. She asked the bot to generate an academic paper on human-robot interaction, a topic Fitzgerald is greatly familiar with.

ChatGPT generated a comprehensive academic paper, full of citations from authors that were also experts in the field. Fitzgerald even said that the citations were from authors that frequently worked together, and she thought she recognized some of the papers cited.

However, upon further investigation, she found that these citations were completely fake.

“The bot was smart enough to know that these professors had worked together in this area [and] smart enough to generate a title that made sense, but it wasn’t smart enough to actually pull information from existing papers,” Fitzgerald stated.

Fitzgerald’s experience demonstrates the limitations of ChatGPT and its dataset. However, professors have still acknowledged the bot in their classrooms.

Ashley Kwak ’24 told the news that her professor brought up ChatGPT on the first day of class, asking them not to use it.

“The first time a professor mentioned ChatGPT in class was on the first day of school this semester in relation to our writing assignments,” Kwak wrote to the News. “I was definitely surprised.”

At present, despite ChatGPT having a number of limitations, Radev anticipates that it will not be long before another model far surpasses ChatGPT.

“Other companies, such as IBM, Microsoft, Google, you.com, et cetera are all in the game, so [they] should expect major competition,” Radev said. “Government, legal and community regulations are likely to follow.”

ChatGPT was released on Nov. 30, 2022.

The post Yale experts explain ChatGPT appeared first on Yale Daily News.

As part of the Institute’s Global Health Conversation Series, YIGH hosted former CDC directors Jeff Koplan, who served from 1998-2002; Julie Gerberding, who served from 2002-2009; and Tom Frieden, who served from 2009-2017, on a public Zoom and Facebook live attended by nearly 300 virtual audience members. The inaugural director of the Yale Institute of Global Health Saad Omer moderated the discussion and fielded questions from the audience. 

Although Omer directed each audience question to all three panelists and allowed them each to respond, common themes emerged among each former director’s responses.

Koplan emphasized the importance of partnership and relations in public health. He touched on both the CDC’s relationships with legislators and the public’s relationship with scientific institutions. 

Koplan stressed the importance of “engagement among elected officials at the local, state, and federal levels” so that the CDC could “have them as partners moving ahead.” Underscoring the importance of these officials, Koplan asserted that elected officials have had enthusiasm and appreciation for what the CDC has accomplished in past decades. Furthermore, he recognized that some elected officials in the House and Senate have played important roles in public health policies alongside the CDC and expressed optimism about rekindling that partnership. 

In his responses, Koplan also warned of the dangers of the public’s intense scrutiny of the government agency in light of the pandemic. He expressed that the current lack of public confidence in the CDC could stunt the development and improvement of the agency. 

“One of the problems we are dealing with right now is an unfortunate drop in respect for science in general, for scientific institutions, for the people that work in them, [which] is disastrous for the well-being of our country,” Koplan asserted.  “That’s not where growth is.”

U.S. News conducted a survey in May 2020 and again in October 2020 and found that trust in the CDC had dropped 10 percent in just months. 

Frieden echoed many of Koplan’s concerns and highlighted the importance of using science to inform public health policies. 

“It’s not just a question of science,” Frieden told the audience. “It’s a question of science and policy.”

Frieden asserted that the Trump administration’s involvement in CDC matters was “inappropriate” and blurred the lines between science and political interest. He cited the incident in which Trump administration staff posted documents on the CDC website downplaying the effects of the pandemic. 

Frieden also acknowledged the importance of allowing science to inform policy, but he stressed that the CDC does not make the policies.

“When you say follow the science, the science will lead you so far, but the science ends where a community makes a decision,” Frieden said.

Frieden said the onus is on legislators to accept science and create policies that will benefit communities, so the relationship between lawmakers and the CDC is one that needs to be fortified for the greater interest of the country. 

While Frieden and Koplan focused on returning to previous principles of the CDC, Gerberding urged internal change to sustain the perpetuation of the CDC’s mission. She conceded that restructuring can remedy inefficiencies and bring in different kinds of talent, but ultimately foundational changes are the answer to improving government scientific institutions such as the CDC.

“We have to have a much greater degree of harmonization, [including] integration of our data systems, which currently are about as integrated as a children’s sandbox,” Gerberding said. 

Gerberding cited the difficulty of obtaining hospitalization data at the onset of the pandemic due to a lack of centralized data to illustrate the importance of data integration. Before the emergency health declaration, the CDC did not have the authority to require health systems across the nation to report COVID-19 hospitalizations and deaths. Consolidating data from nearly 700 health systems across the country posed a serious challenge to the CDC, and Gerberding cited this as support for a national public health system. 

The lack of data integration caused confusion about death and hospitalization numbers in the first weeks of the pandemic, which opened the CDC to become the subject of many conspiracy theories. Social media posts claimed that the CDC was adding deaths to the national count in order to incite panic amongst the public; however, a lag in reporting and data collection contributed to the unusual and inconsistent numbers. 

Towards the end of the discussion, Omer cited a Feb. 8, 2020 survey conducted by Yale on attitudes toward the pandemic. The survey asked participants who they would like to hear pandemic-related information from. Overwhelmingly, participants stated they wanted to hear from the director of the CDC. Omer asked the former directors what could be done to restore this pre-pandemic level of faith and trust in the CDC.

“The CDC needs to speak regularly to people directly,” Frieden said. “Trust is built a teaspoon at a time and lost buckets at a time.”

Gerberding echoed this sentiment and emphasized that information to the public should come directly from the scientists “so it doesn’t look like something has come down from the mountain and is being disseminated into the valley.”

In response, Koplan again affirmed the importance of collaboration between municipal, state and federal institutions. Moreover, Koplan advocated for the modernization of the agency. He said that individuals who possess novel skills in data analytics must be recruited along with experts in technology and social media in order to effectively reach the public. However, Koplan stressed that these changes would only be possible with consistent and predictable funding, something heavily lacking at the CDC currently. 

“The job of public health can be frustrating, underpaid, tiring,” Omer conceded in closing. “But it is never, ever meaningless.”

The headquarters of the Center for Disease Control and Prevention is located in Atlanta, Georgia. 

The post Former CDC directors explore public health distrust at Yale talk appeared first on Yale Daily News.

Erika Linnander, director of Yale’s Global Health Leadership Initiative, currently leads the Primary Healthcare Transformation Initiative, which is a multi-year effort to create a culture of performance management and accountability in Ethiopia’s district health offices. In a recent paper published in the International Journal of Health Policy and Management, Linnander and her team of researchers examined the effects of zonal-level management practices in the Ethiopian healthcare systems. The paper, a product of ten years of collaboration with the administrative health systems in Ethiopia, demonstrated that investment in zonal-level management practices is associated with significant improvements in lower-level management and performance in primary care systems. 

“The type of management we are focusing on is community and patient-centered,” Linnander said. “Do you meet regularly with your staff to evaluate performance based on data? Do you ask your community what is and isn’t working? Do you coordinate with other sectors that play a role in health determinants?”

The paper’s focus demonstrated the effects that prioritization of management and the strengthening of systems can have on the Ethiopian healthcare system. The Primary Healthcare Transformation Initiative provided professionals (who are nowadays also available at any urgent care near Woodhaven NY) with a master’s level education on health management, created the position of hospital CEO and shifted health management responsibilities from clinicians to professional managers. According to Linnander, this came at a time when Ethiopia had rapidly built its primary care program, so it was critical that proper management systems were put in place.

“I wouldn’t say that clinicians aren’t the best people to step into managerial roles,” Mayur Desai SPH ‘94 GRD ‘97, associate professor of epidemiology and author of the paper, wrote to the News. “Healthcare management and leadership are not taught in medical and nursing schools, so why would we just expect a doctor or a nurse to automatically be an excellent manager?”

Taking clinicians out of these managerial roles proved to not only improve primary care management but also ensured that clinicians could continue their main roles as care providers. 

“When you’re in a context where your healthcare resources are so limited, you don’t want to take your clinicians out of [providing care] and put them in these [managerial] roles,” Linnander explained.

Efforts by the Primary Healthcare Transformation Initiative resulted in improved standards of management practices across three levels –– zonal, district and health facility. For care providers, improving management led to “numerous benefits” such as “increased job satisfaction, greater commitment to the organization and better stewardship of limited resources,” Desai said. For patients, improved healthcare system management resulted in better clinical outcomes. 

“When [management] works well, it’s invisible,” Linnander explained. “If you’re a patient and you’re walking into a health center that’s well managed, it just feels good. It’s clean, they have the supplies that are needed, the clinical teams work together, there’s a medical record that they can access and use. It just makes clinical care better for the patient.”

The Yale School of Public Health is located on 60 College St.

The post Yale global health researchers evaluate Ethiopia Primary Healthcare Transformation Initiative appeared first on Yale Daily News.

VOICES is a digital medical screening tool that utilizes multimedia presentations to educate and empower older adults on elder abuse. Conducted on an iPad, the web-based app guides the patients through video modules which outline the potential outcomes and escalations of elder abuse and describe different behaviors that constitute elder abuse, such as financial mistreatment, emotional abuse and property withholding. The app then asks a series of voiceover questions, prompting the patient to reflect on their own experiences. If the app identifies the user as a victim, a final module urges the user to seek help. 

“Older adults are not willing to speak about [elder abuse],” said Fuad Abujarad, associate professor of emergency medicine and principal investigator of the project. “They feel like it’s their fault, they are ashamed they are getting old, and they don’t want to get their loved ones in legal trouble.”

A number of factors contribute to the underreporting of elder abuse, including the lack of efficient screening mechanisms, hesitancy to come forward and fear of loss of autonomy if the older adult’s care provider faces legal repercussions. To combat this, health care providers are often trained on identifying abuse and giving caretakers tools they can administer to their patients. However, this method can be ineffective, as primary care providers often fail to identify cases without visible signs of abuse. VOICES brings older adults into the process and allows them to be in control of their circumstances. 

Kellen McDonald NUR ’24 has helped recruit study participants for the past year. 

“I just give [the users] the iPad, let them put their headphones in, and give them the privacy to go through the video modules on their own,” McDonald said. “Using the tool really streamlines the process of going from care provider to social worker.”

McDonald said that for an issue so relatively widespread, there are little to no efforts to curb its effects. 

“Stigma plays a huge role in [having little efforts],” McDonald said. “The culture in our country around aging does too. We see [aging] as a tragedy, as a loss, but we don’t think about it in terms of all the life experiences a person has had.”

When the NIH-funded project first launched in 2018, the screening tool was used in the emergency room, the first point of care for many older adults. 

The tool has since been expanded to primary care settings and is being developed to be accessible to older adults with cognitive impairments and Spanish-speaking patients. Maripaz Garcia, a senior lecturer in the Department of Spanish and Portuguese, has been working alongside Abujarad and his team to translate the video modules into Spanish.

Garcia, along with four other translators, worked to create two independent translations and synthesized them into one in order to preserve the integrity of the translation. One translation was done by two translators whose strongest languages were English, and the other was done by two translators whose strongest language was Spanish. 

“Translating is a meticulous work that requires a person with high linguistic skills and paying attention to the details,” Garcia said. “When you hire two people, you can see more linguistic options and decide which one suits better in a particular context.”

Abujarad said that meticulous translation is important when discussing such a sensitive and taboo topic. Cultural considerations have to be taken into account so that older adults feel comfortable answering the questions. 

VOICES will continue to expand and the researchers aim to include more geriatric care facilities as the study continues. 

“There’s a huge stigma around elder abuse,” Abujarad said. “The system has failed to address these problems, and if we want to create scalable intervention, we must bring the older adults in control of their circumstances and care.”

According to the Administration for Community Living, in 2020 there were 74.6 million adults over the age of 60 in the United States. 

The post Yale researchers develop digital tool to identify cases of elder abuse appeared first on Yale Daily News.

Akiko Iwasaki, professor of immunobiology and molecular, cellular and developmental biology at the Yale School of Medicine, is an avid Twitter user. With 140,000 followers, her tweets cover topics such as COVID-19 vaccines, immune responses and discrimination in the sciences. Her active use of Twitter led her to Survivor Corps, a Twitter support group for patients suffering from long COVID. Iwasaki began to closely follow the group and came upon a poll on the Survivor Corps Facebook group, which showed that 40 percent of respondents had experienced an improvement in their long COVID symptoms — which range from insomnia and fatigue to memory loss — after their first dose of the vaccine. The poll prompted Iwasaki to look further into the effects of the vaccine on long COVID patients. She worked alongside Harlan Krumholz, professor of medicine, epidemiology and public health at the Yale School of Medicine. 

“The general design [of the study] is to track changes from before vaccination through the weeks afterward and see if different responses associated with distinctive immune system signatures and evidence of residual virus that might lead to diagnostics and therapeutics,” Krumholz wrote to the News. 

In the ongoing study led by Krumholz’s lab, researchers collect blood and saliva samples from long COVID patients before vaccination. The team then follows up with the patients six and twelve weeks after vaccination to monitor symptoms and recollect samples. 

Iwasaki and Krumholz are investigating two different hypotheses. One is that long COVID is caused by persistent viral infection or remnants of viral protein or RNA. Vaccines may induce strong T-cell and antibody responses, removing persistent viral components associated with the spike protein of the virus. The other is that long COVID is caused by an autoimmune disease. In that case, vaccines could induce interferons, which have immune-modulatory functions and can dampen the pathological impact of autoreactive cells. 

After conducting their small pilot study at Yale, Iwasaki and Krumholz are hoping to expand its scope to better understand the vaccine’s effects on long COVID patients. 

“We are hoping to expand the study nationwide,” said Daisy Massey ’19, postgraduate associate in the Krumholz lab and study coordinator for the project. “We are looking to launch it in the next 2-3 weeks and are hoping to get 100 participants.”

The vaccine’s effects on patients with long COVID may provide the keys to understanding the treatment and origins of the condition. However, the study’s origins on Twitter have sparked a conversation about social media’s role in science. 

“Twitter has been essential in our ability to learn quickly about long COVID, especially from the patients’ perspective,” Iwasaki said. “We collaborated with Patient-Led and Survivor Corps to think through and design the symptom surveys. I also learned about the impact of vaccination on long COVID through tweets from patients.”

Iwasaki values Twitter’s role in science because it allows a direct line of communication between scientists and patients, she said. Scientists, doctors and patients could exchange information in real-time, which was especially important given the rapid evolution of the virus over the last two years. Iwasaki herself engages with patients via Twitter, enabling her to reach thousands of patients and gather their input. 

“[Iwasaki] responds to every long COVID patient that messages her,” Massey said. “She sees her role as a communicator, not just a scientist.”

The research team has not only conducted empirical analyses on long COVID patients, but has also strived to connect patients with resources and support groups to deal with the debilitating condition, in keeping with how the study originated. 

“[Social media] closes the loop,” Massey said. “Rather than allowing information to trickle down through scientific communities, it’s right there on your Twitter feed in real-time.”

If you or a loved one is suffering from long COVID, email covidrecovery@yale.edu to be connected to care clinics and patient support groups or to participate in the study. 

Survivor Corps has 186,000 members.

The post Social media post leads Yale scientist to long COVID discovery appeared first on Yale Daily News.

Implicit biases in health care can have grave consequences, leading to patients being misdiagnosed or turned away more often. These implicit biases are not only associated with poorer patient care but also higher mortality rates among racial minority groups. In a Jan. 21 Rheumatology Grand Rounds talk, Suter shared several suggestions to medical providers to address their own implicit biases. These include acknowledging biases, attempting to join activities with groups that providers might not otherwise meet and implementing standard protocols across departments for hospital screening purposes. 

“I think systemic racism and climate change are the two most important and most urgent issues impacting health care today,” Suter told the News. “Although implicit bias is just one facet of a complex problem that has its underpinnings in deep historical, political and socioeconomic structures, I [feel] it is amenable to change at an individual level and therefore worth our attention.”

In her position as vice chief of diversity, equity and inclusion in the rheumatology section at the School of Medicine, Suter works to bring awareness to implicit biases in medicine by hosting forums to discuss DEI issues and sending a monthly newsletter that covers a range of topics on DEI in medicine.

Suter’s role and actions are a part of a larger mission led by the Department of Internal Medicine. In 2016, the department established a DEI office, appointing Inginia Genao as associate chair for diversity and inclusion. In 2020, the office obtained funding to expand its efforts, which included appointing a vice chief of DEI in each section of the department.

The DEI office helped implement the standardization of care among COVID-19 patients. At the onset of the pandemic, implicit bias in medicine became a more pervasive issue than ever before as health care providers navigated a novel virus without a well-defined treatment protocol. 

“Because of the stress of many patients coming in [to the hospital], there [was] a potential for implicit bias to enter into the decision making of who you admit to the hospital, who you treat, and what treatment you give to each person,” Gary Desir, Yale New Haven Hospital chief of internal medicine, explained. “Very early in the pandemic, we decided to put a group of faculty and staff to figure out what the optimal treatment, at the time, was for treating COVID.”

By April 1, 21 days after the first COVID-19 patient was admitted to Yale New Haven Hospital, Desir and other YNHH staff had developed a standard protocol for admitting patients based on patients’ oxygen saturation. The implementation of this data-driven treatment protocol reduced the burden of decision-making on medical providers but was also developed to reduce implicit biases present when dictating patient care. In order to keep up with the ever-evolving nature of the virus, the protocol was revised 28 times. 

“Whether you were elderly, rich, poor, black, white, everyone got the same treatment,” Desir explained. 

The School of Medicine’s efforts to reduce the effects of implicit biases in medicine have addressed racial disparities in the Yale New Haven Health System; however, implicit biases among medical providers extend beyond the borders of New Haven County. 

“We all have [implicit biases]; the question is which ones do we have and how does it impact the people we provide care for,” Desir said. “If we don’t understand how [implicit biases] affect [providers’] actions with patients, then it becomes a much larger problem.”

Yale students and faculty have previously called for a broader culture of change, such as medical school curriculum reforms to be more inclusive of traditionally underdiagnosed populations or transforming a race-based health care provision system to a race-conscious one. These initiatives attempt to address the negative influence of implicit bias in patient care.

According to Ian Reilly SPH ’22, there are several steps that the health care sector can take to eliminate and mitigate the effect of implicit biases. These include greater representation and racial diversity in leadership positions and providing patients with adequate access to translation services so they can communicate meaningfully with their health care providers.

“Health care providers must recognize their own biases and how it impacts the care they provide if they are to truly provide equitable and compassionate care,” Reilly said. 

Yale New Haven Hospital is located at 20 York St.

The post Yale professor Lisa Suter discusses implicit biases in health care appeared first on Yale Daily News.

In a paper published in the Journal of the American Medical Association, Paul Goldsmith-Pinkham, assistant professor of finance at the School of Management; Jacob Wallace, assistant professor of public health at the School of Public Health; and Karen Jiang ’21 investigated whether Medicare eligibility is associated with a reduction of racial and ethnic gaps in healthcare access. The study, which consisted of over two million participants, assessed disparities in insurance coverage, access to healthcare and self-reported health across different racial demographics. CDC epidemiology data spanning 20 years was also used to assess average mortality rates across races. 

“Conceptually, the idea was to look into the impact of Medicare eligibility on disparities in health by comparing Americans who are on one side of the threshold — 64 years old — with those on the other, who are 65 and older,” Goldsmith-Pinkham told Yale Insights. “Because Medicare is a benefit that universally hits everyone, it gives us a control and a treatment to examine.”

To study the impact of the introduction of Medicare, the team took a three-fold approach to assessing its effects: access to care, insurance coverage and self-reported health.

Rates of coverage, access to regular care, cost barriers and access to flu shots were all used to measure Medicare’s effects on addressing racial disparities that existed before the introduction of the government-sponsored program. After comparing the “control and treatment” populations, the research showed that the gaps in healthcare access and self-reported health began to close after the introduction of Medicare into the patient’s life. 

The positive effect of Medicare on access to care and healthcare coverage was more prevalent among Hispanic and Black populations than White populations. An average of 92 percent of White participants had insurance coverage before the introduction of Medicare, with that number rising to 98 percent after. However, for Black participants, the number rose from 86 percent to 95 percent. Among Hispanic participants, the number increased from 77 percent to 91 percent. 

Additionally, access to Massachusetts Part C Medicare Advantage plans increased the quality of self-reported health responses. The percentage reporting “poor” health decreased from 14.8 percent to 11 percent for Hispanic participants and from 10.3 percent to 7.7 percent for Black participants, while essentially exhibiting no change for White participants. 

After their initial observations, Goldsmith-Pinkham’s team assessed the data geographically to evaluate whether or not specific regions, namely those with more Black and Hispanic Americans, were driving these significant results. Regions with historically high levels of racism and lower funding for state-sponsored insurance were expected to have a disproportionate effect on the final results of the study. However, the team found that racial gaps in healthcare were reduced across all regions. 

The paper draws the conclusion that even decreasing the Medicare eligibility age to 60 could result in a considerable improvement in health outcomes in Hispanic and Black populations; However, it also opens the door to a larger discussion on universal healthcare, Jiang said. 

“If you think about the push toward broader government-sponsored healthcare in general, our work has shown that decreasing the age of eligibility could provide an opportunity to address healthcare inequalities even earlier than age 65. This presents a compelling reason to ask — why wait?” Jiang said.

Although the research began before the onset of the COVID-19 pandemic, the paper was published in June 2021, well after the effects of the pandemic were seen to disproportionately affect communities of color. 

“It was a very timely piece,” Jiang said. “Obviously with the events that happened in the spring of 2020, a theme emerged that it became increasingly apparent to think about race and ethnicity and how those two variables could be further explored.”

In an interview with US News, Wallace suggested that reducing the Medicare eligibility age would be a significant step in improving health equity. 

“Racial and ethnic disparities in our healthcare system are a reflection of longstanding structural racism in the United States,” Wallace said to Yale News. “Identifying policies that reduce these disparities is key to advancing health equity.”

The Medicare program was introduced in 1965 under the Johnson Administration.

The post Medicare could help close racial gaps in healthcare, Yale study finds appeared first on Yale Daily News.