Faculty Bibliography

Ensuring reliability of Large Language Models (LLMs) in clinical tasks is crucial. Our study assesses two state-of-the-art LLMs (ChatGPT and LlaMA-2) for extracting clinical information, focusing on cognitive tests like MMSE and CDR. Our data consisted of 135,307 clinical notes (Jan 12th, 2010 to May 24th, 2023) mentioning MMSE, CDR, or MoCA. After applying inclusion criteria 34,465 notes remained, of which 765 underwent ChatGPT (GPT-4) and LlaMA-2, and 22 experts reviewed the responses. ChatGPT successfully extracted MMSE and CDR instances with dates from 742 notes. We used 20 notes for fine-tuning and training the reviewers. The remaining 722 were assigned to reviewers, with 309 each assigned to two reviewers simultaneously. Inter-rater-agreement (Fleiss' Kappa), precision, recall, true/false negative rates, and accuracy were calculated. Our study follows TRIPOD reporting guidelines for model validation. For MMSE information extraction, ChatGPT (vs. LlaMA-2) achieved accuracy of 83% (vs. 66.4%), sensitivity of 89.7% (vs. 69.9%), true-negative rates of 96% (vs 60.0%), and precision of 82.7% (vs 62.2%). For CDR the results were lower overall, with accuracy of 87.1% (vs. 74.5%), sensitivity of 84.3% (vs. 39.7%), true-negative rates of 99.8% (98.4%), and precision of 48.3% (vs. 16.1%). We qualitatively evaluated the MMSE errors of ChatGPT and LlaMA-2 on double-reviewed notes. LlaMA-2 errors included 27 cases of total hallucination, 19 cases of reporting other scores instead of MMSE, 25 missed scores, and 23 cases of reporting only the wrong date. In comparison, ChatGPT's errors included only 3 cases of total hallucination, 17 cases of wrong test reported instead of MMSE, and 19 cases of reporting a wrong date. In this diagnostic/prognostic study of ChatGPT and LlaMA-2 for extracting cognitive exam dates and scores from clinical notes, ChatGPT exhibited high accuracy, with better performance compared to LlaMA-2. The use of LLMs could benefit dementia research and clinical care, by identifying eligible patients for treatments initialization or clinical trial enrollments. Rigorous evaluation of LLMs is crucial to understanding their capabilities and limitations.

BACKGROUND:Because of advances in medical treatment of heart failure, patients are living longer than in previous eras and may approach the need for advanced therapies, including heart transplantation, at older ages. This study assesses practices surrounding heart transplant in older adults (> 70 years) and examines short- and medium-term outcomes. METHODS AND RESULTS/RESULTS:This study is a retrospective analysis using the United Network for Organ Sharing (UNOS) database from 2010 to 2021. The absolute number of older adults being transplanted is increasing. Older adults were more likely to have had a prior malignancy or ischemic cardiomyopathy and less likely to be on extra-corporeal membrane oxygenation or have a high UNOS status prior to transplant. Mortality at 1-year was higher for older adults (27.8% vs. 23.4%), but at 5 years there was no significant difference (22.3% vs. 19.4%.). Older adults were more likely to die of malignancy or infection. Adults under 70 were more likely to die of cardiovascular causes or graft failure. There was less rejection in older adults. Mortality has not changed for older adults transplanted before versus after the 2018 UNOS allocation change. CONCLUSIONS:Carefully selected older adults may be considered for heart transplantation, given similar intermediate-term mortality.

BACKGROUND:Some of the most promising strategies to reduce hospital readmissions in heart failure (HF) is through the timely receipt of home health care (HHC), delivered by Medicare-certified home health agencies (HHAs), and outpatient medical follow-up after hospital discharge. Yet national data show that only 12% of Medicare beneficiaries receive these evidence-based practices, representing an implementation gap. To advance the science and improve outcomes in HF, we will test the effectiveness and implementation of an intervention called Improving TRansitions ANd OutcomeS for Heart FailurE Patients in Home Health CaRe (I-TRANSFER-HF), comprised of early and intensive HHC nurse visits combined with an early outpatient medical visit post-discharge, among HF patients receiving HHC. METHODS:This study will use a Hybrid Type 1, stepped wedge randomized trial design, to test the effectiveness and implementation of I-TRANSFER-HF in partnership with four geographically diverse dyads of hospitals and HHAs ("hospital-HHA" dyads) across the US. Aim 1 will test the effectiveness of I-TRANSFER-HF to reduce 30-day readmissions (primary outcome) and ED visits (secondary outcome), and increase days at home (secondary outcome) among HF patients who receive timely follow-up compared to usual care. Hospital-HHA dyads will be randomized to cross over from a baseline period of no intervention to the intervention in a randomized sequential order. Medicare claims data from each dyad and from comparison dyads selected within the national dataset will be used to ascertain outcomes. Hypotheses will be tested with generalized mixed models. Aim 2 will assess the determinants of I-TRANSFER-HF's implementation using a mixed-methods approach and is guided by the Consolidated Framework for Implementation Research 2.0 (CFIR 2.0). Qualitative interviews will be conducted with key stakeholders across the hospital-HHA dyads to assess acceptability, barriers, and facilitators of implementation; feasibility and process measures will be assessed with Medicare claims data. DISCUSSION/CONCLUSIONS:As the first pragmatic trial of promoting timely HHC and outpatient follow-up in HF, this study has the potential to dramatically improve care and outcomes for HF patients and produce novel insights for the implementation of HHC nationally. TRIAL REGISTRATION/BACKGROUND:This trial has been registered on ClinicalTrials.Gov (#NCT06118983). Registered on 10/31/2023, https://clinicaltrials.gov/study/NCT06118983?id=NCT06118983&rank=1 .

BACKGROUND/UNASSIGNED:Pursuing initial invasive or conservative management of chronic coronary disease (CCD) is a preference-sensitive decision that should include shared decision-making. Communicating the benefits of either approach is challenging, as individual patients rarely achieve the population-averaged outcomes reported in clinical trials. Our objective was to develop a patient decision aid (PDA) with patient-specific estimates of outcomes for initial invasive versus conservative management of CCD, based on the ISCHEMIA trial (International Study of Comparative Health Effectiveness With Medical and Invasive Approaches). METHODS/UNASSIGNED:This was a multiphase mixed-methods study using focus groups of outpatients with CCD, caregivers, clinicians, and researchers. Focus groups were held in Kansas City, MO and New York City, NY between September 2021 and June 2022. Patients with CCD were included if they had a positive stress test within 1 year. Phase 1 focused on patient priorities for outcomes to guide treatment decisions. Phase 2 involved PDA development and refinement. Phase 3 involved further refinement and member checking. Key themes involving shared decision-making and treatment preferences were elicited from focus groups using a deductive approach to develop a PDA representing the outcomes most important to patients. RESULTS/UNASSIGNED:Of 46 patient and caregiver participants, the mean age was 63.5 years, 53% were female, 61% were White, 24% were Black, and 9% were Hispanic. When deciding between treatments, participants valued shared decision-making but generally deferred decisions to clinicians. The outcomes most important to participants were survival and quality of life, followed by physical functioning and symptoms. To represent these outcomes, participants favored simple visualizations, such as a speedometer or health meter. When deciding between treatment options, participants preferred to use the PDA collaboratively with a clinician instead of as a stand-alone tool. CONCLUSIONS/UNASSIGNED:Our novel, patient-centered approach to developing a PDA for CCD with patient-specific outcomes has the potential to rapidly translate clinical trial results to individual patients and support shared decision-making.

BACKGROUND:Frailty is associated with adverse cardiovascular outcomes independent of age and comorbidities, yet the independent influence of frailty progression on cardiovascular outcomes remains uncertain. METHODS:To determine whether frailty progression is associated with adverse cardiovascular outcomes, independent of baseline frailty and age, we evaluated all Medicare Fee-for-Service beneficiaries ≥65 years at cohort inception with continuous enrollment from 2003 to 2015. Linear mixed effects models, adjusted for baseline frailty and age, were used to estimate change in a validated claims-based frailty index (CFI) over a 5-year period. Survival analysis was used to examine frailty progression and risk of adverse health outcomes. RESULTS:There were 8.9 million unique patients identified, mean age 77.3 ± 7.2 years, 58.7% female, 10.9% non-White race. In total, 60% had frailty progression and 40% frailty regression over median follow-up of 2.4 years. Compared to those with frailty regression, when adjusting for age and baseline CFI, those with frailty progression had a significantly greater risk of incident major adverse cardiovascular and cerebrovascular events (MACCE) (hazard ratio [HR] 1.31, 95% confidence interval [CI] 1.31-1.31), all-cause mortality (HR 1.34, 95% CI 1.34-1.34), acute myocardial infarction (HR 1.08, 95% CI 1.07-1.09), heart failure exacerbation (HR 1.30, 95% CI 1.29-1.30), ischemic stroke (HR 1.14, 95% CI 1.14-1.15). There was also a graded increase in risk of each outcome with more rapid progression, as well as significantly fewer days alive at home (DAH) with more rapid progression compared to the slowest progression group (270.4 ± 112.3 vs. 308.6 ± 93.0 days, rate ratio 0.88, 95% CI 0.87-0.88, p < 0.001). CONCLUSIONS:In this large, nationwide sample of older Medicare beneficiaries, frailty progression, independent of age and baseline frailty, was associated with fewer DAH and a graded risk of MACCE, all-cause mortality, myocardial infarction, heart failure, and ischemic stroke compared to those with frailty regression.

BACKGROUND:This study aimed to examine the prevalence of inappropriate tight glycemic control in older adults with type 2 diabetes and other chronic conditions in New York City, and to identify factors associated with this practice. METHODS:We conducted a retrospective cohort study using the INSIGHT Clinical Research Network. The study population included 11,728 and 15,196 older adults in New York City (age ≥ 75 years) with a diagnosis of type 2 diabetes, and at least one other chronic medical condition, in 2017 and 2022, respectively. The main outcome of interest was inappropriate tight glycemic control, defined as HbA1c <7.0% (<53 mmol/mol) with prescription of at least one high-risk agent (insulin or insulin secretagogue). RESULTS:The proportion of older adults with inappropriate tight glycemic control decreased by nearly 19% over a five-year period (19.4% in 2017 to 15.8% in 2022). There was a significant decrease in insulin (27.8% in 2017; 24.3% in 2022) and sulfonylurea (29.4% in 2017; 21.7% in 2022) medication prescription, and increase in use of GLP-1 agonists (1.8% in 2017; 11.4% in 2022) and SGLT-2 inhibitors (5.8% in 2017; 25.1% in 2022), among the total population. Factors associated with inappropriate tight glycemic control in 2022 included history of heart failure (adjusted odds ratio [aOR] 1.38), chronic kidney disease ([aOR] 1.93), colorectal cancer ([aOR] 1.38), acute myocardial infarction ([aOR] 1.28), "other" ([aOR] 0.72) or "unknown" ([aOR] 0.72) race, and a point increase in BMI ([aOR] 0.98). CONCLUSIONS:We found an encouraging trend toward less use of high-risk medication strategies for older adults with type 2 diabetes and multiple chronic conditions. However, one in six patients in 2022 still had inappropriate tight glycemic control, indicating a need for continued efforts to optimize diabetes management in this population.

BACKGROUND:Sarcopenia has been associated with adverse postoperative outcomes in older age cohorts, but has not been assessed in older adults with inflammatory bowel disease (IBD). Further, current assessments of sarcopenia among all aged individuals with IBD have used various measures of muscle mass as well as cutoffs to define its presence, leading to heterogeneous findings. METHODS:In this single-institution, multihospital retrospective study, we identified all patients aged 60 years and older with IBD who underwent disease-related intestinal resection between 2012 and 2022. Skeletal Muscle Index (SMI) and Total Psoas Index (TPI) were measured at the superior L3 endplate on preoperative computed tomography scans and compared through receiver operating characteristic curve. We then performed multivariable logistic regression to assess risk factors associated with an adverse 30-day postoperative outcome. Our primary outcome included a 30-day composite of postoperative mortality and complications, including infection, bleeding, cardiac event, cerebrovascular accident, acute kidney injury, venous thromboembolism, reoperation, all-cause rehospitalization, and need for intensive care unit-level care. RESULTS:A total of 120 individuals were included. Overall, 52% were female, 40% had ulcerative colitis, 60% had Crohn's disease, and median age at time of surgery was 70 years (interquartile range: 65-75). Forty percent of older adults had an adverse 30-day postoperative outcome, including infection (23%), readmission (17%), acute kidney injury (13%), bleeding (13%), intensive care unit admission (10%), cardiac event (8%), venous thromboembolism (7%), reoperation (6%), mortality (5%), and cerebrovascular accident (2%). When evaluating the predictive performance of SMI vs TPI for an adverse 30-day postoperative event, SMI had a significantly higher area under the curve of 0.66 (95% CI, 0.56-0.76) as compared to 0.58 (95% CI, 0.48-0.69) for TPI (P = .02). On multivariable logistic regression, prior IBD-related surgery (adjusted odds ratio [adjOR] 6.46, 95% CI, 1.85-22.51) and preoperative sepsis (adjOR 5.74, 95% CI, 1.36-24.17) significantly increased the odds of adverse postoperative outcomes, whereas increasing SMI was associated with a decreased risk of an adverse postoperative outcome (adjOR 0.88, 95% CI, 0.82-0.94). CONCLUSIONS:Sarcopenia, as measured by SMI, is associated with an increased risk of postoperative complications among older adults with IBD. Measurement of SMI from preoperative imaging can help risk stratify older adults with IBD undergoing intestinal resection.

BACKGROUND:As the inflammatory bowel disease (IBD) patient population is aging, the prevalence of polypharmacy is rising. However, data exploring the prevalence, risk factors, and clinical outcomes associated with polypharmacy among older adults with IBD are limited. AIMS/OBJECTIVE:To determine (i) prevalence of polypharmacy (≥5 medications) and potentially inappropriate medication (PIM) utilization in older adults with IBD, (ii) changes in medications over time (iii) predictors of polypharmacy, and (iv) the impact of polypharmacy/PIMs on one-year hospitalization rates. METHODS:We conducted a retrospective single-center study of older adults with IBD from September 1st 2011 to December 31st 2022. Wilcoxon-signed rank and McNemar's tests were used to assess changes in polypharmacy between visits, with ordinal logistic regression and Cox proportional hazards models used to determine risk factors for polypharmacy and time to hospitalization, respectively. RESULTS:Among 512 older adults with IBD, 74.0% experienced polypharmacy at initial visit, with 42.6% receiving at least one PIM. Additionally, severe polypharmacy (≥10 medications) was present among 28.6% individuals at index visit and increased to 38.6% by last visit (p<0.01). Multivariable analysis revealed that age ≥70 years, BMI ≥30.0 kg/m2, prior IBD-related surgery, and the presence of comorbidities were associated with polypharmacy. Moreover, severe polypharmacy (adjHR 1.95, 95%CI 1.29-2.92), as well as PIM use (adjHR 2.16, 95%CI 1.37-3.43) among those with polypharmacy, were significantly associated with all-cause hospitalization within a year of index visit. DISCUSSION/CONCLUSIONS:Severe polypharmacy was initially present in more than 25% of older adults with IBD and increased to 34% within 4 years of index visit. Severe polypharmacy, as well as PIM utilization among those with polypharmacy, were also associated with an increased risk of hospitalization at one-year, highlighting the need for deprescribing efforts in this population.

We characterized comorbidity profiles and cardiometabolic risk factors among older adults with multiple chronic conditions (MCCs) in New York City using an intersectionality approach. Electronic health record data were obtained from the INSIGHT Clinical Research Network on 367,901 New York City residents aged 50 years or older with MCCs. Comorbidity profiles were heterogeneous. The most common profile across sex and racial and ethnic groups was co-occurring hypertension and hyperlipidemia; prevalence of these 2 conditions differed across groups (4.7%-7.3% co-occurrence alone, 65.1%-88.0% with other conditions). Significant sex and racial and ethnic differences were observed, which may reflect accumulated disparities in risk factors and health care access across the life course.

BACKGROUND:There is growing recognition that healthcare should align with individuals' health priorities; however, these priorities remain undefined, especially among older adults. The Rehabilitation Using Mobile Health for Older Adults with Ischemic Heart Disease in the Home Setting (RESILIENT) trial, designed to test the efficacy of mobile health cardiac rehabilitation (mHealth-CR) in an older cohort, also measures the attainment of participant-defined health outcome goals as a prespecified secondary endpoint. This study aimed to characterize the health priorities of older adults with ischemic heart disease (IHD) using goal attainment scaling-a technique for measuring individualized goal achievement-in a sample of 100 RESILIENT participants. METHODS:The ongoing RESILIENT trial randomizes patients aged ≥65 years with IHD (defined as hospitalization for acute coronary syndrome and/or coronary revascularization), to receive mHealth-CR or usual care. For the current study, we qualitatively coded baseline goal attainment scales from randomly selected batches of 20 participants to identify participants' cardiac rehabilitation outcome goals and their perceptions of barriers and action plans for goal attainment. We used a deductive framework (i.e., 4 value categories from Patient Priorities Care) and inductive approaches to code and analyze interviews until thematic saturation. RESULTS:This sample of 100 older adults set diverse health outcome goals. Most (54.6%) prioritized physical activity, fewer (17.1%) identified symptom management, fewer still (13.7%) prioritized health metrics, mostly comprised of weight loss goals (10.3%), and the fewest (<4%) were related to clinical metrics such as reducing cholesterol or preventing hospital readmission. Participants anticipated extrinsic (access to places to exercise, time) and intrinsic (non-cardiac pain, motivation) barriers. Action plans detailed strategies for exercise, motivation, accountability, and overcoming time constraints. CONCLUSIONS:Using goal attainment scaling, we elicited specific and measurable goals among older adults with IHD beginning cardiac rehabilitation. Priorities were predominantly functional, diverging from clinical metrics emphasized by clinicians and healthcare systems.