Saloni Kumar Maharaj

All Publications

• Asking the Right Questions: Benchmarking Large Language Models in the Development of Clinical Consultation Templates. Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing McCoy, L. G., Wu, D., Khemani, S., Maharaj, S. K., Pahwa, A., Rosengaus, L., Giang, L., Jee, O., Goh, E., Haredasht, F. N., Chopra, K., Wu, D. J., Conteh, A., Ravi, V., Weng, Y., Li, K. Z., Shirvani, D., Chen, J. H. 2026; 31: 400-416

  Abstract

  This study evaluates the capacity of large language models (LLMs) to generate structured clinical consultation templates for electronic consultation. Using 145 expert-crafted templates developed and routinely used by Stanford's eConsult team, we assess frontier models-including o3, GPT-4o, Kimi K2, Claude 4 Sonnet, Llama 3 70B, and Gemini 2.5 Pro-for their ability to produce clinically coherent, concise, and prioritized clinical question schemas. Through a multi-agent pipeline combining prompt optimization, semantic autograding, and prioritization analysis, we show that while models like o3 achieve high comprehensiveness (up to 92.2%), they consistently generate excessively long templates and fail to correctly prioritize the most clinically important questions under length constraints. Performance varies across specialties, with significant degradation in narrative-driven fields such as psychiatry and pain medicine. Our findings demonstrate that LLMs can enhance structured clinical information exchange between physicians, while highlighting the need for more robust evaluation methods that capture a model's ability to prioritize clinically salient information within the time constraints of real-world physician communication. Limitations include reliance on Stanford-specific templates and concordancebased grading, which may not capture all clinically reasonable outputs.

  View details for DOI 10.1142/9789819824755_0028

  View details for PubMedID 41758156

• Automated Evaluation of Large Language Model Response Concordance with Human Specialist Responses on Physician-to-Physician eConsult Cases. Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing Wu, D. J., Haredasht, F. N., Wu, D., Ravi, V., McCoy, L. G., Weng, Y., Chopra, K., Everett, S. S., Nageeb, G., Chen, W., Ma, S. P., Maharaj, S. K., Tran, J., Rosengaus, L., Giang, L., Jee, O., Goh, E., Chen, J. H. 2026; 31: 372-387

  Abstract

  Specialist consults in primary care and inpatient settings typically address complex clinical questions beyond standard guidelines. eConsults have been developed as a way for specialist physicians to review cases asynchronously and provide clinical answers without a formal patient encounter. Meanwhile, large language models (LLMs) have approached human-level performance on structured clinical tasks, but their real-world effectiveness requires evaluation, which is bottlenecked by time-intensive manual physician review. To address this, we evaluate two automated methods: LLM-as-judge and a decompose-thenverify framework that breaks down AI answers into verifiable claims against human eConsult responses. Using 40 real-world physician-to-physician eConsults, we compared AI-generated responses to human answers using both physician raters and automated tools. LLM-as-judge outperformed decompose-then-verify, achieving human-level concordance assessment with F1-score of 0.89 (95% CI: 0.750, 0.960) and Cohen's kappa of 0.75 (95% CI 0.47,0.90) , comparable to physician inter-rater agreement κ = 0.69-0.90 (95% CI 0.43-1.0).

  View details for DOI 10.1142/9789819824755_0026

  View details for PubMedID 41758154

• Transforming Culture: Postoperative Venous Thromboembolism Prophylaxis in a Transgender Patient on Estrogen. Journal of general internal medicine Maharaj, S. K., Yu, A., Singh, A., Khanijow, K. 2025

  Abstract

  A 40-year-old transgender patient (assigned male at birth, pronouns: they/them) with paroxysmal atrial fibrillation presented for elective transoral chondrolaryngoplasty. They had been on gender-affirming hormone therapy with intramuscular estradiol valerate and oral progesterone for 16 months and their atrial fibrillation was managed with flecainide and aspirin. The otolaryngology team recommended holding estrogen postoperatively to reduce venous thromboembolism (VTE) risk, but the patient expressed concern about withdrawal-related gender dysphoria and vasomotor symptoms.Transfeminine patients on estrogen are at higher risk of VTE compared to the general population Goodman M, Zhang Q. Stroke and Blood Clot Risk in Transgender Women Taking Hormones. Patient-Centered Outcomes Research Institute (PCORI); 202 Accessed November 18, 2024. http://www.ncbi.nlm.nih.gov/books/NBK593552/ . There has historically been concern about perioperative VTE in these patients in the context of gender affirming vaginoplasty, which is considered to have risk factors associated with VTE (lengthy pelvic surgery with prolonged bedrest) Coleman et al. in Int J Transgender Health. 23:S1-S259, 2022. Previously, it was routine practice to hold gender affirming estrogen perioperatively for vaginoplasty. However, newer literature has shown no increased risk with estrogen maintenance during vaginoplasty Coon et al. in Plast Reconstr Surg Glob Open. 11:e5033, 2023;Lee and Spiegel in The Laryngoscope. 132:918-919, 2022;, with the World Professional Association for Transgender Health (WPATH) guidelines now recommending the continuation of estrogen perioperatively Kozato et al. in J Clin Endocrinol Metab. 106:e1586-e1590, 2021 to avoid symptoms of estrogen withdrawal including gender dysphoria with increased risk of suicide, and vasomotor symptoms. Unfortunately, the practice of holding estrogen therapy perioperatively persists, and has been applied to many surgeries outside of vaginoplasty, even those with little VTE risk, such as this patient's facial feminization surgery Badreddine et al. in Eur Rev Med Pharmacol Sci. 26:2511-2517, 2022.Despite evidence and WPATH guidelines, discontinuation of estrogen therapy perioperatively remains common Lee and Spiegel in The Laryngoscope. 132:918-919, 2022. For this patient, the low bleeding and VTE risk allowed for a tailored approach. After discussing risks, the team opted for 2-week VTE chemoprophylaxis with Enoxaparin 40 mg daily. The patient recovered well without bleeding or thromboembolic events. This case underscores the need to update surgical practices based on current evidence and prioritize patient-centered care.

  View details for DOI 10.1007/s11606-025-09668-3

  View details for PubMedID 40646333

  View details for PubMedCentralID 10226616

• TAKE MY BREATH AWAY - A CASE OF POSTOPERATIVE HYPOXIA Maharaj, S. K. SPRINGER. 2024: S383

  View details for Web of Science ID 001433572700616