2026 Eastern States Conference

Authors: Kaitlyn Healy, PharmD; Siu Yan (Amy) Yeung, PharmD, BCCCP; Clement Ng, PharmD, CAHIMS; Brian Grover, PharmD, BCPS; Hyunuk Seung, MS 
Learning Objective: Describe the feasibility of utilizing a large language model (LLM) to analyze clinical interventions completed by pharmacists and assign appropriate categorization.   
Background/Objective: The goal of this study is to use a LLM software to sort pharmacist interventions into predefined categories determined by the research team and compare the interrater reliability of sorting completed by the LLM compared to that of a pharmacist.
Methods: This retrospective study evaluated pharmacist interventions documented within the electronic health record (EHR). Completed pharmacist interventions include both structured fields and free-text documentation. Interventions were extracted from the EHR, de-identified, and a random sampling was selected for inclusion. Selected interventions were categorized by both the research team and a large language model (LLM). The LLM used in this study, GPT-5 accessed via the Microsoft Copilot™ interface, was adapted using structured prompt engineering to simulate pharmacist clinical reasoning. Retrieval-augmented generation (RAG) was incorporated to support predefined categorization of interventions. Inter-rater agreement between the research team and the LLM will be assessed using weighted Cohen’s kappa analysis. A total of 852 interventions will be analyzed based on a power calculation targeting a kappa of 0.70.  
Results: Overall, 889 interventions were included. All 13 categories were used by both the LLM and the pharmacist reference standard. Interventions received a mean of 1.23 labels and a median of 1, though 185 pharmacist-reviewed and 168 LLM-reviewed interventions were assigned multiple labels. Primary category agreement between the LLM and the reference standard was substantial (κ=0.70; 95% CI: 0.67–0.74), closely aligning with inter‑pharmacist agreement (κ=0.72; 95% CI: 0.69–0.76). Multi‑label agreement was similarly strong, with Jaccard similarity of 0.76 for LLM vs reference and 0.77 for pharmacist vs pharmacist, alongside low Hamming loss and no consistent pattern found regarding category‑level variability.
Conclusions:  LLM performance in classifying pharmacist interventions aligned with human classification, with substantial agreement seen on primary categorization and multilabel alignment. These findings support the potential use of LLMs for pharmacist intervention classification tasks. 
Self-Assessment Question:  

• Which of the following describes the purpose of using a large language model in this study?  

• To replace pharmacist clinical decision making.  

• To generate new pharmacist interventions.  

• To categorize completed pharmacist interventions.  

• To standardize documentation of pharmacist interventions.  

Authors: Seo Young Chun, PharmD; Melissa Chaung, PharmD, BCPS; Michael A. Wynd, PharmD, BCPS

Learning Objective: At the conclusion of my presentation, the audience will be able to compare the allograft function outcomes of early vs late conversion from tacrolimus to belatacept in kidney transplant recipients. 

Background/Objective:
Assess the impact of timing of conversion from tacrolimus to belatacept on allograft function in kidney transplant recipients (KTRs).

Methods:
This was a single-center, retrospective cohort study evaluating first kidney-alone, adult (age ≥ 18 years at the time of conversion), Epstein-Barr virus IgG seropositive KTRs who converted from tacrolimus to belatacept between January 1, 2013, and December 31, 2024. Recipients of multi-organ transplants or patients receiving belatacept for < 30 days were excluded. Outcomes at 1-year post-conversion were compared between KTRs who converted early (< 6 months post-transplant) vs late (≥ 6 months post-transplant). The primary endpoint was the change in estimated glomerular filtration rate from baseline. Secondary endpoints included patient and allograft survival, biopsy-proven acute rejection, opportunistic viral infections, malignancy, discontinuation-rate of belatacept, and adverse events. Demographic data, within-cohort, and between-cohort comparisons were analyzed using appropriate statistics.

Results:
Of 177 patients screened, 76 patients were included, with 67 patients in the early conversion group and 9 patients in the late conversion group. Baseline demographics were similar between groups, although pre-transplant donor-specific antibodies were more common in the late conversion group (44.4% vs 11.9% [p = 0.04]). Median eGFR at conversion was lower in the early conversion group compared to the late conversion group (20 vs 45 mL/min/1.73m2 [p = 0.01]). At 1 year post-conversion, the early conversion group demonstrated a greater change in eGFR (23 vs 7 mL/min/1.73m2 [p = 0.0083]). No significant differences were observed in patient and allograft survival, rejection, opportunistic infections, and belatacept discontinuation between groups. 

Conclusion:
Early conversion from tacrolimus to belatacept was associated with greater improvement in eGFR, likely reflecting lower baseline eGFR and greater opportunity for recovery at the time of conversion. While interpretation is limited by the disproportionately smaller late conversion cohort, this study provides descriptive real-world data on institutional belatacept conversion practices and outcomes across different conversion strategies.

Self-Assessment Questions:
Which of the following is a common reason for patients to convert from tacrolimus to belatacept?

1. Nephrotoxicity
2. Gastrointestinal side effects
3. Neurotoxicity
4. All of the above

Title: Evaluation of guideline-directed venous thromboembolism prophylaxis in high-risk ambulatory cancer patients
Authors: Srivishnu Vardhan Parasaram, PharmD; Julie Shupp, PharmD, BCOP; Jorge Aguilera, PharmD, BCPS
Learning Objective: Audience members will be able to evaluate institutional adherence to NCCN/ASCO guideline-directed venous thromboembolism (VTE) prophylaxis in ambulatory cancer patients.
Background/Objective: Cancer-associated thrombosis is a significant cause of morbidity and mortality in oncology patients. The National Comprehensive Cancer Network (NCCN) and American Society of Clinical Oncology (ASCO) guidelines recommend consideration of VTE prophylaxis for high-risk ambulatory cancer patients initiating systemic chemotherapy, defined by a Khorana Score of 2 or greater. This study aimed to determine the rate of appropriate VTE prophylaxis prescribing at John R. Marsh Cancer Center within Meritus Health and to evaluate associated clinical outcomes.
Methods: This retrospective observational study evaluated adults aged 18 and older presenting with a new cancer diagnosis and initiating chemotherapy at John R. Marsh Cancer Center between January 1, 2024 and December 31, 2024. Baseline Khorana Scores were calculated to stratify VTE risk. Patients were excluded if they were receiving therapeutic anticoagulation for a pre-existing condition (e.g., atrial fibrillation, secondary prophylaxis), had a documented absolute contraindication to anticoagulation, were pregnant, or received treatment exclusively at an outside facility. The primary outcome was the rate of appropriate VTE prophylaxis prescribing based on NCCN guidelines. Secondary outcomes included the incidence of VTE events, major bleeding events, and all-cause mortality, stratified by Khorana Score, management appropriateness, and metastatic status.
Results: A total of 167 patient charts were included for analysis. Guideline-adherent VTE prophylaxis management was observed in 67.1% of patients (n=112), with 55 patients (32.9%) managed inappropriately based on their baseline Khorana Score. Of the entire cohort, only 3 patients (1.8%) were prescribed pharmacologic thromboprophylaxis, reflecting a very low overall prescribing rate despite the high proportion of high-risk patients. VTE events occurred in 9.8% (n=11) of appropriately managed patients compared to 25.5% (n=14) of inappropriately managed patients, a statistically significant difference (p=0.015; RR 2.60, 95% CI 1.27–5.33). Patients with a Khorana Score ≥2 experienced VTE at a rate of 24.6% compared to 10.0% in low-risk patients (p=0.023; RR 2.46, 95% CI 1.19–5.06). Major bleeding events occurred in 3.0% of the overall cohort (n=5); no statistically significant differences were observed across any subgroup, though the analysis was limited by the low event count. All-cause mortality occurred in 24.0% of the cohort (n=40), with significantly higher rates observed in patients with a Khorana Score ≥2 (36.8% vs. 17.3%, p=0.009) and in inappropriately managed patients (38.2% vs. 17.0%, p=0.005).
Conclusion: Adherence to NCCN/ASCO VTE prophylaxis guidelines at our institution was 67.1%, with an overall pharmacologic prophylaxis prescribing rate of only 1.8%, suggesting that guideline-concordant management in this population is largely driven by appropriate withholding of prophylaxis in low-risk patients rather than active prescribing in high-risk patients. Inappropriately managed patients experienced significantly higher rates of VTE, demonstrating the clinical consequences of non-adherence. All-cause mortality differences across subgroups likely reflect underlying cancer burden and disease severity rather than VTE-attributable death, as cause of death was not captured. While provider hesitance regarding anticoagulation is understandable given bleeding risk and potential treatment delays, this study highlights a meaningful quality improvement opportunity. The occurrence of VTE in low-scoring patients and variable outcomes among high-scoring patients also suggest the Khorana Score alone may have limited precision in this population, underscoring the need for prospective evaluation and potentially more individualized risk stratification tools.