Jefferson Investigates: ChatGPT in Sleep Apnea, Distress in Cancer Care and Hunger for Moms in Opioid Treatment
African American men are more affected by cancer, ChatGPT may help patients understand sleep apnea, and looking at the basic needs of mothers in treatment for opioids use.
Image Credit: © Adobe Stock nimito/416739915
Racial disparities in cancer diagnoses and outcomes are an ongoing issue in the United States. African American women have the highest mortality rate for breast cancer despite having a lower incidence rate than white women. African American men have the highest incidence and mortality rates for colon, lung and prostate cancers. They also have higher cancer incidence and mortality rates than all other races and sexes overall.
Socioeconomic disparities, increased risk factors like drinking and smoking, healthcare discrimination and distrust of the medical system due to historical racism all play a role in worsening outcomes and distress during treatment. In a new study in The American Journal of Men’s Health, researchers led by Stephanie Kjelstrom, a PhD student in Population Health at Thomas Jefferson University, explored distress during cancer treatment to see how African American men compared to other populations.
The team evaluated questionnaires designed by the National Comprehensive Cancer Network to measure distress for 770 Philadelphia cancer patients, where distress is defined as “an unpleasant experience of a mental, physical, social or spiritual nature that can affect the way you think, feel or act and is normal during cancer care, but it may make it harder to cope with cancer.” They found that African American men were more likely to experience higher levels of distress than white men.
“There is a real need for increased attention on how we detect and treat cancer in African American men,” says population health researcher Sharon L. Larson, PhD, senior author of the study.
Dr. Kjelstrom hopes that future studies can drill down into the factors that cause higher distress during cancer treatment for African American men. Dr. Larson notes that initiatives like the Voices Project at Philadelphia’s Main Line Health Center for Population Research, which she leads, reach out to patients and community members to get their perspective on health concerns, could help. The hope is that more detailed studies will reveal potential avenues for intervention.
“In my opinion, we've spent 125 years describing the problem of racial health disparities,” Dr. Larson says. “It's now time to do something about it.”
By Rachel Feltman
Image Credit: © Adobe Stock Amy Walters/30371272
While many individuals seek out medical information from physicians, countless others will look for answers on the internet. Increasingly, classic search engine tools like Google and WebMD are giving way to chatbots like ChatGPT, which use artificial intelligence (AI), or machine learning, could become the next frontier in armchair diagnosis.
In a recent study in the Journal of Clinical Sleep Medicine, clinicians at Jefferson Health put one of the web’s newest tools to the test. Playing the part of curious sufferers of obstructive sleep apnea, the team queried ChatGPT to test the reliability of its answers about the condition. Obstructive sleep apnea (OSA), a disorder in which a person frequently stops breathing while asleep, affects at least 24 million people in the United States alone.
“People have been using Doctor Google for years. It can take you a lot of places, some good and some bad,” says study senior author and otolaryngologist Maurits S. Boon, MD. Junior resident Daniel J. Campbell, MD, first author on the study, suggested testing the prowess of ChatGPT.
Drs. Boon, Campbell and their colleagues crafted 24 questions on OSA that represented a reasonable range of queries a layperson might ask about the condition. Questions were posed in an order intended to simulate a logical flow of conversation, as ChatGPT’s answers are built upon previous inputs.
Two reviewers graded the answers based on medical accuracy and clinical appropriateness using the following scale: incorrect, partially correct, correct, correct with either a statistic or citation referenced or correct with both a statistic and citation reference.
The good news, Dr. Boon says, is that ChatGPT’s answers to questions about OSA symptoms and treatments were “almost scary in their accuracy,” hitting correct responses 71.9% of the time and providing citations and statistics when asked. While the AI tool slightly missed the mark on a handful of queries, he says — getting points for partial correctness in all but one interaction — it was never far off enough to warrant concern. The study concluded that all responses remained above accepted recommendations for presenting medical information to patients.
Dr. Boon’s team has already finished a follow-up study to see if ChatGPT is useful in educating patients with insomnia. He says he expects to see lots more work on the clinical implications of AI in the coming years.
“We’re just scratching the surface,” Dr. Boon says. “This has the potential to help clinicians identify patients who are at risk for all sorts of conditions.”
By Rachel Feltman
Image Credit: © Adobe Stock Mediaphotos/494420942
For women in opioid addiction treatment, becoming a mother can be a blessing and a challenge. With the added stresses of parenthood, new mothers are at their highest risk of relapse or overdose, making recovery especially difficult. Now, research from Thomas Jefferson University finds another hurdle facing new mothers in opioid treatment is food insecurity.
“Food insecurity is a huge issue in the United States and can have profound effects on families,” says Meghan Gannon, PhD, MSPH, an assistant professor in the College of Nursing. Past research has shown that people in opioid treatment often face major barriers to finding, affording and preparing nutritious food, but there has been limited research to date on how food insecurity affects mothers in this group.
In a study published in the American Journal of Health Promotion, Dr. Gannon surveyed new mothers in opioid treatment in the Maternal Addiction Treatment, Education & Research (MATER) program. She found that two-thirds of the women reported low or very low food security, leading to outcomes like skipping meals to feed their family or only eating one meal a day. Less than half of the women felt they had enough food to support breastfeeding, and women facing food insecurity were more likely to also suffer from mental health issues such as depression and anxiety.
The women said some of the barriers they faced to food security included not having enough money to buy food, not having stores that sold healthy food nearby, not being able to travel to stores on public transportation and not having supplies or utilities at home to cook with.
Kimberly McLaughlin, PhD, LPC, CAADC, program director of MATER and a co-author on the study, says this research is a foundational step in providing better care to new mothers in opioid treatment.
“We can't expect them to do the hard work of diving into their trauma work or strengthening their parenting skills if they're hungry,” says Dr. McLaughlin.
By Marilyn Perkins
