Faculty Bibliography

The adoption of large language models (LLMs) in healthcare demands a careful analysis of their potential to spread false medical knowledge. Because LLMs ingest massive volumes of data from the open Internet during training, they are potentially exposed to unverified medical knowledge that may include deliberately planted misinformation. Here, we perform a threat assessment that simulates a data-poisoning attack against The Pile, a popular dataset used for LLM development. We find that replacement of just 0.001% of training tokens with medical misinformation results in harmful models more likely to propagate medical errors. Furthermore, we discover that corrupted models match the performance of their corruption-free counterparts on open-source benchmarks routinely used to evaluate medical LLMs. Using biomedical knowledge graphs to screen medical LLM outputs, we propose a harm mitigation strategy that captures 91.9% of harmful content (F1 = 85.7%). Our algorithm provides a unique method to validate stochastically generated LLM outputs against hard-coded relationships in knowledge graphs. In view of current calls for improved data provenance and transparent LLM development, we hope to raise awareness of emergent risks from LLMs trained indiscriminately on web-scraped data, particularly in healthcare where misinformation can potentially compromise patient safety.

Solid organ transplant recipients (SOTRs) suffer more frequent and more severe infections due to their compromised immune responses resulting from immunosuppressive treatments designed to prevent organ rejection. Pharmacological immunosuppression can adversely affect immune responses to vaccination. A cohort of kidney transplant recipients (KTRs) received their third dose of ancestral, monovalent COVID-19 vaccine in the context of a clinical trial and antibody responses to the vaccine strain, as well as two Omicron variants BA.1 and BA.5 were investigated and compared with healthy controls who also received a third dose of mRNA vaccine (HCs). Total IgG and live virus neutralizing antibody titers were reduced in KTRs compared with controls for all variants. KTRs displayed altered IgG subclass switching, with significantly lower IgG3 antibodies. Responses in KTRs were also very heterogeneous, with some individuals showing strong responses but a significant number showing no Omicron-specific neutralizing antibodies. Taken together, immune responses after COVID-19 vaccination in KTRs were not only lower than HCs but highly variable, indicating that simply increasing the number of vaccine doses alone may not be sufficient to provide greater protection in this population. These findings underscore the need for tailored vaccination strategies for immunosuppressed populations, such as KTRs. Alternative formulations and doses of COVID-19 vaccines should be considered for people with severely compromised immune systems, as more frequent vaccinations may not significantly improve the response, especially regarding neutralizing antibodies.IMPORTANCEThis study addresses the challenges faced by kidney transplant recipients (KTRs) in mounting effective immune responses against COVID-19. By evaluating the antibody responses to a third dose of monovalent mRNA COVID-19 vaccine and its effectiveness against Omicron subvariants (BA.1 and BA.5), this study reveals significant reductions in both binding and neutralizing antibodies in KTRs compared with healthy controls. The research highlights altered IgG subclass switching and heterogeneous responses within the KTR population. Reduced recognition of variants, coupled with differences in IgG subclasses, decreases both the quality and quantity of protective antibodies after vaccination in KTRs. These findings underscore the need for tailored vaccination strategies for immunosuppressed populations, such as KTRs. Alternative formulations and doses of COVID-19 vaccines should be considered for people with severely compromised immune systems, as more frequent vaccinations may not significantly improve the response, especially regarding neutralizing antibodies.

Human leukocyte antigen-level matching in US kidney allocation has been deemphasized due to its role in elevating racial disparities. Molecular matching based on eplets might improve risk stratification compared to antigen matching, but the magnitude of racial disparities in molecular matching is not known. To assign eplets unambiguously, we utilized a cohort of 5193 individuals with high-resolution allele-level human leukocyte antigen genotypes from the National Kidney Registry. Using repeated random sampling to simulate donor-recipient genotype pairings based on the ethnic composition of the historical US deceased-donor pool, we profiled the percentage of well-matched donors available for candidates by ethnicity. The prevalence of well-matched donors with 0-DR/DQ eplet mismatch was 3-fold less racially disparate for Black and Asian candidates and 2-fold less for Latino candidates compared to 0-ABDR antigen mismatches. Compared to 0-DR antigen mismatch, 0-DR eplet mismatch was 1.33-fold more racially disparate for Asian and 1.28-fold more for Latino, with similar disparity for Black candidates, whereas 0-DQ eplet mismatch reduced disparities, showing 1.26-fold less disparity for Black, 1.14-fold less for Latino, but 1.26-fold higher for Asian candidates. The prevalence of well-matched donors for candidates of different ethnicities varied according to which molecules were chosen to define a low-risk match.

PURPOSE/OBJECTIVE:In October 2018, the OPTN changed adult heart transplant (HT) allocation policy, increasing the number of adult candidates that had higher priority than pediatric candidates, potentially disadvantaging pediatric waitlist registrants. METHODS:To understand the impact of this policy change, we used SRTR data to identify 1469 pre-policy (7/2016-9/2018) and 2901 (10/2018-12/2022) post-policy pediatric (< 18 years) HT registrants. We quantified mortality and transplant risks using weighted cause-specific hazard models, and then using weighted competing risks regression. We further stratified these analyses by age to understand risks for those in direct competition with adults for organs (≥ 12 years). RESULTS:, p = 0.02). Post policy, 1-year transplant rate did not change in those < 12years (68.2%-71.0%, p = 0.77), but in those ≥ 12years, transplant rate increased (77.3%-81.0%, p = 0.003). CONCLUSIONS:Mortality on the waitlist decreased and access to HT for pediatric registrants did not decline following the 2018 policy change. The decreased mortality rate may reflect changes in patient casemix and/or improved patient care. Continued surveillance is important in ensuring equity in pediatric, and adult, HT.

BACKGROUND:Highly effective modulator therapies (HEMT) including ivacaftor (IVA) and elexacaftor/tezacaftor/ivacaftor (ETI) have transformed treatment for people with cystic fibrosis (pwCF). However, non-HEMT-responsive mutations are more common in pwCF of non-White race/ethnicity; introduction of HEMT might have exacerbated racial/ethnic disparities in CF care. METHODS:Using the Scientific Registry of Transplant Recipients, we identified all lung transplant candidates and recipients 05/2005-12/2022 and categorized them by diagnosis (CF/non-CF), race/ethnicity (non-Hispanic White/Black/Hispanic) and era [Pre-HEMT (2005-1/30/2012), IVA (1/31/2012-10/30/2019), ETI (10/31/2019-12/31/2022)]. We compared the percentage of patients listed, delisted/died, or transplanted by race/ethnicity and era. RESULTS:34,659 lung transplants were performed: 10,521 pre-HEMT, 15,944 in IVA era, and 7,888 in ETI era. Over the three eras, the percentage of lung recipients with CF of White race decreased (94.5 % to 92.4 % to 78.4 %) and of Black race (1.7 % to 2.4 % to 5.7 %) or Hispanic ethnicity increased (3.5 % to 4.6 % to 14.2 %; p < 0.001). Similarly, among candidates listed for CF over the three eras, the percentage that were of White race decreased (82.0 % vs. 78.6 % vs. 71.0 %) and of Black race (9.2 % vs. 10.0 % vs. 10.3 %) or Hispanic ethnicity increased (6.4 % vs. 8.6 % vs. 13.6 %; p < 0.001). CONCLUSION/CONCLUSIONS:The introduction of HEMT appears to have benefitted CF lung transplant candidates and recipients of Black race or Hispanic ethnicity less than those of White race. This is likely due to the higher prevalence of HEMT-ineligible CFTR mutations among Black and Hispanic patients and underscores the need for therapies aimed at non-HEMT-responsive mutations prevalent in these racial/ethnic populations.

RATIONALE & OBJECTIVE/OBJECTIVE:Despite substantial growth in the population of older adults with kidney disease, there remains a lack of evidence to guide clinical care for this group. The Kidney Disease and Aging Research Collaborative (KDARC) conducted a Delphi study to build consensus on research priorities for clinical geriatric nephrology. STUDY DESIGN/METHODS:Asynchronous modified Delphi study. SETTING & PARTICIPANTS/METHODS:Clinicians and researchers in the US and Canada with clinical experience and/or research expertise in geriatric nephrology. OUTCOME/RESULTS:Research priorities in geriatric nephrology. ANALYTICAL APPROACH/METHODS:In the first Delphi round, participants submitted free-text descriptions of research priorities considered important for improving the clinical care of older adults with kidney disease. Delphi moderators used inductive content analysis to group concepts into categories. In the second and third rounds, participants iteratively reviewed topics, selected their top 5 priorities, and offered comments used to revise categories. RESULTS:Among 121 who were invited, 57 participants (47%) completed the first Delphi round and 48 (84% of enrolled participants) completed all rounds. After 3 rounds, the 5 priorities with the highest proportion of agreement were: 1) Communication and Decision-Making about Treatment Options for Older Adults with Kidney Failure (69% agreement), 2) Quality of Life, Symptom Management, and Palliative Care (67%), 3) Frailty and Physical Function (54%), 4) Tailoring Therapies for Kidney Disease to Specific Needs of Older Adults (42%), and 5) Caregiver and Social Support (35%). Health equity and person-centricity were identified as cross-cutting features that informed all topics. LIMITATIONS/CONCLUSIONS:Relatively low response rate and limited participation by private practitioners and older clinicians and researchers. CONCLUSIONS:Experts in geriatric nephrology identified clinical research priorities with the greatest potential to improve care for older adults with kidney disease. These findings provide a roadmap for the geriatric nephrology community to harmonize and maximize the impact of research efforts.

Developing real-world evidence from electronic health records (EHR) is vital to advance kidney transplantation (KT). We assessed the feasibility of studying KT using the Epic Cosmos aggregated EHR dataset, which includes 274 million unique individuals cared for in 238 U.S. health systems, by comparing it with the Scientific Registry of Transplant Recipients (SRTR). We identified 69,418 KT recipients transplanted between January 2014 and December 2022 in Cosmos (39.4% of all US KT transplants during this period). Demographics and clinical characteristics of recipients captured in Cosmos were consistent with the overall SRTR cohort. Survival estimates were generally comparable, although there were some differences in long-term survival. At 7 years post-transplant, patient survival was 80.4% in Cosmos and 77.8% in SRTR. Multivariable Cox regression showed consistent associations between clinical factors and mortality in both cohorts, with minor discrepancies in the associations between death and both age and race. In summary, Cosmos provides a reliable platform for KT research, allowing EHR-level clinical granularity not available with either the transplant registry or healthcare claims. Consequently, Cosmos will enable novel analyses to improve our understanding of KT management on a national scale.