Evaluating ChatGPT's Accuracy in Providing Screening Mammography Recommendations among Older Women: Artificial Intelligence and Cancer Communication

Abstract Objective: The U.S. Preventive Services Task Force (USPSTF) recommends biennial screening mammography through age 74. Guidelines vary as to whether or not they recommended mammography screening to women aged 75 and older. This study aims to determine the ability of ChatGPT to provide appropriate recommendations for breast cancer screening in patients aged 75 years and older. Methods: 12 questions and 4 clinical vignettes addressing fundamental concepts about breast cancer screening and prevention in patients aged 75 years and older were created and asked to ChatGPT three consecutive times to generate 3 sets of responses. The responses were graded by a multi-disciplinary panel of experts in the intersection of breast cancer screening and aging . The responses were graded as 'appropriate', 'inappropriate', or 'unreliable' based on the reviewer's clinical judgment, content of the response, and whether the content was consistent across the three responses . Appropriateness was determined through a majority consensus. Results: The responses generated by ChatGPT were appropriate for 11/17 questions (64%). Three questions were graded as inappropriate (18%) and 2 questions were graded as unreliable (12%). A consensus was not reached on one question (6%) and was graded as no consensus. Conclusions: While recognizing the limitations of ChatGPT, it has potential to provide accurate health care information and could be utilized by healthcare professionals to assist in providing recommendations for breast cancer screening in patients age 75 years and older. Physician oversight will be necessary, due to the possibility of ChatGPT to provide inappropriate and unreliable responses, and the importance of accuracy in medicine.

Resilient Adaptation Strategies: Unveiling Older Adults' Coping Dynamics Amidst the COVID-19 Pandemic

Older adults, who are particularly vulnerable to coronavirus disease 2019 (COVID-19), exhibit less stress and greater well-being than their younger peers. However, there have been no in-depth explorations of adaptive coping strategies among this population, nor has the role of frailty status been addressed. The current study examined stress and coping in 30 U.S. older adults (mean age = 81 years, range = 68 to 95 years) amidst the COVID-19 pandemic, uncovering themes of: (1) Pandemic Stresses: stresses experienced during the pandemic centered around social isolation and concern for others' well-being; (2) Resilience: older adults proved highly adaptable, with lifetime experience as a stress buffer; and (3) Silver Linings: older adults reported positive by-products, such as reconnecting with and a renewed appreciation for life and nature. Motivation for change and change itself centered around creating value and meaning in the present, especially around social isolation. Findings challenge existing ageist stereotypes, give insight into interventional design, and highlight the importance of ensuring infrastructural and societal support. [Journal of Gerontological Nursing, 49(12), 32-39.].

Harm and Medication-Type Impact Agreement with Hypothetical Deprescribing Recommendations: a Vignette-Based Experiment with Older Adults Across Four Countries

BACKGROUND

Little is known about what factors are important to older adults when deciding whether to agree with a recommendation to deprescribe.

OBJECTIVE

To explore the extent to which medication type and rationale for potential discontinuation influence older adults' acceptance of deprescribing.

DESIGN

Cross-sectional 2 (drug: lansoprazole - treat indigestion; simvastatin - prevent cardiovascular disease) by 3 (deprescribing rationale: lack of benefit; potential for harm; both) experimental design.

PARTICIPANTS

Online panelists aged ≥65 years from Australia, the Netherlands, the United Kingdom, and the United States INTERVENTIONS: Participants were presented with a hypothetical patient experiencing polypharmacy whose PCP discussed stopping a medication. We randomized participants to receive one of six vignettes.

MAIN MEASURES

We measured agreement with deprescribing (6-point Likert scale, "Strongly disagree (1)" and "Strongly agree (6)") for the hypothetical patient as the primary outcome. We also measured participants' personality traits, perceptions of risk and uncertainty, and attitudes towards polypharmacy and deprescribing.

KEY RESULTS

Among 5311 participants (93.3% completion rate), the mean (M) agreement with deprescribing for the hypothetical patient was 4.71 (95% confidence interval (CI): 4.67, 4.75). Participants reported higher agreement with stopping lansoprazole (n=2656) (M=4.90, 95% CI: 4.85, 4.95) compared to simvastatin (n=2655) (M=4.53, 95% CI: 4.47, 4.58), P<.001. Participants who received the combination rationale (n=1786) reported higher agreement with deprescribing (M=4.83, 95% CI: 4.76, 4.89) compared to those who received the rationales on lack of benefit (n=1755) (M=4.66, 95% CI: 4.60, 4.73) or potential for harm (n=1770) (M=4.65, 95% CI 4.58, 4.72). In adjusted regression analyses (n=5062), participants with a higher desire to engage in health promotion behaviors (b=0.08, 95% CI 0.02, 0.13) or need for certainty (b=0.12, 95% CI 0.04, 0.20) reported higher agreement with deprescribing.

CONCLUSIONS

Older adults across four countries were accepting of deprescribing in the setting of polypharmacy. The medication type and rationale for discontinuation were important factors in the decision-making process.

TRIAL REGISTRATION

ClinicalTrials.gov , NCT04676282, https://clinicaltrials.gov/ct2/show/NCT04676282?term=vordenberg&draw=2&rank=1.

RESILIENCE AND MULTIFACTORIAL STRESSORS AMONG OLDER ADULTS DURING THE COVID-19 PANDEMIC: A QUALITATIVE STUDY

The COVID-19 pandemic represents a complex stressor that is experienced differently across individuals and age strata. The present study explored perceptions and experiences of older adults within the domains of health, social interactions, finances and care of existing chronic medical conditions; and strategies used to cope with these stressors. We recruited 30 people (mean age 81.4 years) stratified by frailty status to complete semi-structured interviews about what changes to the above domains had occurred and what coping strategies were utilized. Using inductive and deductive coding techniques, thematic analysis revealed three overarching themes. The first was Pandemic Experience, which was perceived as stressful, especially in the domains of social isolation from friends and family and concern for others’ well-being. The second domain was Resilience where participants reported highly adaptable and creative ways to connect with others and viewed the pandemic from a lens of lifetime experience, which acted as a stress buffer. The third theme was Silver Linings where participants reported unexpected renewal like reconnecting with family and friends in more meaningful ways and reconnecting with nature. We found no meaningful distinction in experience by frailty status and explore reasons for this. Policy implications including internet access and training and societal aging biases are discussed in the context of aging and coping theories.

Association between willingness to deprescribe and health outcome priorities among U.S. older adults: Results of a national survey

BACKGROUND

It is not known whether older adults' willingness to deprescribe is associated with their health outcome priorities related to medications.

METHODS

A cross-sectional survey was conducted from March-April 2020 using a nationally representative online panel. The survey presented two vignettes: (1) a preventive medicine; and (2) a symptom-relief medicine. Participants were asked whether they would be willing to stop each medicine if their doctor recommended it, and to rate their level of agreement with two health outcome priorities statements: "I am willing to accept the risk of future side effects … to feel better now," and "I would prefer to take fewer medicines, even if … I may not live as long or may have bothersome symptoms sometimes." Ordinal logistic regression was used to examine associations between willingness to stop each medicine, baseline characteristics and health outcome priorities.

RESULTS

Of 1193 panel members ≥65 years invited to participate, 835 (70%) completed the survey. Mean (SD) age was 73 years; 496 (59%) had taken a statin and 124 (15%) a prescription sedative-hypnotic. 507 (61%) were willing to stop preventive medicines; 276 (33%) were maybe willing. 419 (50%) were willing to stop symptom-relief medicines; 380 (46%) were maybe willing. Prioritizing fewer medicines was associated with higher odds of being willing to stop symptom-relief medicines (aOR 1.43 [95% CI 1.02-2.00]) and preventive medicines (aOR 1.52 [95% CI 1.05-2.18]). Prioritizing now over future was associated with lower odds of being willing to stop symptom-relief medicines (aOR 0.62 [95% CI 0.39-1.00]). Current/prior use of statins was associated with lower willingness to stop preventive medicines (aOR 0.66 [95% CI 0.48-0.91]).

CONCLUSIONS

Older adults' health outcome priorities related to medication use are associated with their willingness to consider deprescribing. Future research should determine how best to elicit patients' health outcome priorities to facilitate goal-concordant decisions about medication use.

Innovative Strategies to Facilitate Patient-Centered Research in Multiple Chronic Conditions

Multiple chronic conditions (MCC) are one of today’s most pressing healthcare concerns, affecting 25% of all Americans and 75% of older Americans. Clinical care for individuals with MCC is often complex, condition-centric, and poorly coordinated across multiple specialties and healthcare services. There is an urgent need for innovative patient-centered research and intervention development to address the unique needs of the growing population of individuals with MCC. In this commentary, we describe innovative methods and strategies to conduct patient-centered MCC research guided by the goals and objectives in the Department of Health and Human Services MCC Strategic Framework. We describe methods to (1) increase the external validity of trials for individuals with MCC; (2) study MCC epidemiology; (3) engage clinicians, communities, and patients into MCC research; and (4) address health equity to eliminate disparities.

Assessment of Patient-Preferred Language to Achieve Goal-Aligned Deprescribing in Older Adults

IMPORTANCE

How clinicians communicate about deprescribing, the structured process of reducing or stopping unnecessary, potentially harmful, or goal-discordant medicines, may be associated with the extent to which older adults are willing to do it.

OBJECTIVE

To examine older adults' preferences regarding different rationales a clinician may use to explain why a patient should stop an unnecessary or potentially harmful medication.

DESIGN, SETTING, AND PARTICIPANTS

This cross-sectional survey study was conducted from March 25 to April 19, 2020, among a nationally representative, probability-based online survey panel (KnowledgePanel). KnowledgePanel members aged 65 years and older were recruited by random digit dialing and address-based sampling. Data were analyzed from May 4 to July 8, 2020.

EXPOSURES

The survey presented 2 vignettes involving hypothetical older adults. One described a statin being used for primary prevention by a person with functional impairment and polypharmacy. The second described a sedative-hypnotic, such as zolpidem, being used for insomnia by a person with good functional status.

MAIN OUTCOMES AND MEASURES

After each vignette, participants expressed preferences using a best-worst scaling method for 7 different phrases a clinician may use to explain why they should reduce or stop the medication. Conditional logistic regression was used to quantify respondents' relative preferences.

RESULTS

A total of 1193 KnowledgePanel members were invited, and 835 respondents (70.0%) completed the survey. The mean (SD) age was 73 (6) years, 414 (49.6%) were women, and 671 (80.4%) self-identified as White individuals. A total of 496 respondents (59.8%) had ever used a statin, and 124 respondents (14.9%) had ever used a sedative-hypnotic. For both medications, the most preferred phrase to explain deprescribing focused on the risk of side effects. For statins, this phrase was 5.8-fold (95% CI, 5.3-6.3) more preferred than the least preferred option, which focused on the effort (treatment burden) involved in taking the medicine. For sedative-hypnotics, the phrase about side effects was 8.6-fold (95% CI, 7.9-9.5) more preferred over the least preferred option, "This medicine is unlikely to help you function better."

CONCLUSIONS AND RELEVANCE

These findings suggest that among older adults, the most preferred rationale for deprescribing both preventive and symptom-relief medicines focused on the risk of side effects. These results could be used to inform clinical practice and improve effective communications around deprescribing in older adults.

Using an Agent-based Model to Examine Deimplementation of Breast Cancer Screening

OBJECTIVE

The objective of this study was to examine the potential impact of provider social networks and experiences with patients on deimplementation of breast cancer screening.

RESEARCH DESIGN

We constructed the Breast Cancer-Social network Agent-based Model (BC-SAM), which depicts breast cancer screening decisions, incidence, and progression among 10,000 women ages 40 and over and the screening recommendations of their providers over a 30-year period. The model has patient and provider modules that each incorporate social network influences. Patients and providers were connected in a network, which represented patient-patient peer connections, provider-provider peer connections, connections between providers and patients they treat, and friend/family relationships between patients and providers. We calibrated provider decisions in the model using data from the CanSNET national survey of primary care physicians in the United States, which we fielded in 2016.

RESULTS

First, assuming that providers' screening recommendations for women ages 50-74 remain unchanged but their recommendations for screening among younger (below 50 y old) and older (75+ y old) women decrease, we observed a decline in predicted screening rates for women ages 50-74 due to spillover effects. Second, screening rates for younger and older women were slow to respond to changes in provider recommendations; a 78% decline in provider recommendations to older women over 30 years resulted in an estimated 23% decline in patient screening in that group. Third, providers' experiences with unscreened patients, friends, and family members modestly increased screening recommendations over time (7 percentage points). Finally, we found that provider peer effects can have a substantial impact on population screening rates and can entrench existing practices.

CONCLUSION

Modeling cancer screening as a complex social system demonstrates a range of potential effects and may help target future interventions designed to reduce overscreening.

DEVELOPING A SENSOR-BASED MOBILE APPLICATION FOR IN-HOME FRAILTY ASSESSMENT: A QUALITATIVE STUDY

Frailty is an important concept in the care of older adults, and there is great interest in incorporating user-friendly frailty assessments into research and clinical settings. In-home, sensor-based technologies may provide a more dynamic, sensitive, and accurate assessment of frailty measures. To investigate user perspectives for use of sensor-based technologies and mobile applications, we held five focus groups with community-dwelling older adults (n= 10), their informal caregivers (n=9), and medical professionals (n=8). We used qualitative inductive analysis to organize thematic content. Caregivers and care-recipients viewed the early identification of frailty as beneficial, but highlighted the need for secure data infrastructure and clear demonstration of how frailty assessment would improve care. They also expressed concerns that technology-based communication could reduce in-person interactions. Medical providers noted the utility of objective data for difficult conversations with caregivers of frail patients, but worried about resources for analyses and interpretation of sensor-based health information.