About

Joydeep Banerjee is an Adjunct Professor of Data Sciences and Operations at USC Marshall School of Business. He is a member of the Senior Technical Leadership for Multicloud Management at Red Hat, where his focus is on Observability, Far Edge, and Scalability. Prior to this, he worked at The Walt Disney Studios, partnering with the VP of Strategic Initiatives to drive the digital transformation of a key application by breaking it into microservices and migrating it to the Cloud. With many years of experience at IBM, Joydeep has led key projects and worked across various technologies in Distributed Computing, Streaming Analytics, and J2EE. He is a strong advocate of DevOps transformation, Agile Methodology, and Design Thinking. His research interests include applying AI/ML success stories to the area of Observability.

Research topics

• Internal medicine
• Medicine
• Surgery
• Emergency medicine
• Nursing

Selected publications

• Enhancing endocrine training through virtual simulation: addressing social determinants of health, health inequities and multiple long-term conditions

  Endocrine Abstracts · 2026-02-13

  article
  PublisherDOI

• Primary Care Lifestyle Redesign ^® Clinical Trial: Diabetes Outcomes, Healthcare Utilization, and Costs

  Canadian Journal of Occupational Therapy · 2026-02-25

  articleOpen access

  Background: In primary care (PC) clinical, healthcare utilization, and financial outcomes associated with Lifestyle Redesign ® Occupational Therapy (LR-OT), a framework that incorporates healthy habits into daily routines, are not well documented in the United States. Purpose: (a) Compare clinical outcomes, healthcare utilization, and associated cost differences between patients who received LR-OT and a control group. (b) Assess the investment required integrate LR-OT into PC. Method: LR-OT was integrated into a safety-net PC clinic. In a hybrid effectiveness-implementation clinical trial (#NCT03293914), clinical (glycated hemoglobin A1c [HbA1c], blood pressure, cholesterol, and body mass index) and utilization (outpatient, emergency department [ED], hospital visits) data were extracted from medical records of provider-referred adults with diabetes (HbA1c ≥ 9.0%) who were willing to make lifestyle changes. Results: Of 155 referred patients, 142 were randomized to LR-OT ( n = 73) or a control group ( n = 69), and 42 completed eight LR-OT sessions within 6 months. Underpowered clinical outcomes indicated no significant change, with aggregated trends suggesting LR-OT supports sustained cardiometabolic improvement. Savings of $250,518.96 is attributed to LR-OT, driven by significant decreases in ED visits ( p = .03) and hospital days ( p = .03), despite nonsignificant increases in outpatient utilization. The projected cost-benefit of integrating a full-time LR-OT in PC is $352,326.72 with a 236% return on investment. Conclusion: Clinical trends and significant reductions in acute care services drove healthcare-wide cost savings, demonstrating clinical and financial value of integrating LR-OT into PC.

  PublisherDOI

• Interpreting Blood Culture Results as Early Guidance for Infective Endocarditis

  JAMA Network Open · 2025-05-01 · 3 citations

  articleOpen access

  Importance: Few bedside tools with defined accuracy have been described that are useful to alter bayesian prior probability for infective endocarditis (IE) in patients with bacteremia. Objective: To evaluate the accuracy of simple blood culture parameters to guide pretest probability of IE. Design, Setting, and Participants: This multicenter, retrospective case-control study of blood culture data in adults with IE vs without IE was conducted at 3 acute care public hospitals in the Los Angeles County Department of Health Services between December 2018 and August 2022. Patients were individuals aged 18 years or older who had positive blood cultures and met the inclusion criteria, including cases who met the Duke criteria for definite or possible IE, and control cases who did not have concern for endocarditis. Exposures: Positive blood cultures for methicillin-susceptible Staphylococcus aureus, methicillin-resistant S aureus, Enterococcus faecalis, low-risk Streptococcus species, or high-risk Streptococcus species. Main Outcomes and Measures: The primary outcome was the negative likelihood ratio (LR) of having endocarditis based on the number of positive blood cultures on admission. Positive LRs, evaluation of 2 or more of 4 bottles positive on admission, bacteremia lasting for at least 2 days, and combination groups were secondary outcomes. Results: A total of 252 eligible patients with IE (182 male [72%]; median [IQR] age, 54 [38-65] years), including 164 definite and 88 possible IE cases, and 455 controls (321 male [71%]; median [IQR] age, 53 [41-63] years) were identified. The negative LR point estimates for having IE with only 1 of 4 positive blood culture bottles on admission ranged from 0.05 (95% CI, 0.01-0.37) for E faecalis to 0.12 (95% CI, 0.03-0.49) for methicillin-susceptible S aureus. Sensitivity analysis of cases restricted to definite IE found similar results. Blood culture clearance by day 2 also had modestly helpful negative LRs for methicillin-resistant S aureus (0.24; 95% CI, 0.13-0.42) and Enterococcus species (0.34; 95% CI, 0.21-0.56). If patients had 4 of 4 bottles positive on admission, positive LRs were helpful for Enterococcus species (LR, 4.21; 95% CI, 2.53-7.02) and high-risk streptococci (LR, 5.35; 95% CI, 3.39-8.42) and for all organisms with persistent bacteremia (with LRs ranging from 1.78 [95% CI, 1.36-2.34] to 9.60 [95% CI, 3.43-44.60]). If both were true, the positive LRs ranged from 1.63 (95% CI, 1.17-2.28) to 8.59 (95% CI, 3.43-21.55) for all organisms. Conclusions and Relevance: In this case-control study of patients with and without IE, the number of initial positive blood culture bottles and days to culture clearance were helpful to adjust pretest probability of IE. These findings may help guide diagnostic and therapeutic decisions around bacteremia early during hospitalization.

  PublisherOA PDFDOI

• Provision of Temporary Access to Inpatient Hemodialysis to Uninsured Patients Initiating Hemodialysis

  JAMA Network Open · 2025-11-18

  articleOpen access1st author

  Importance: Uninsured patients who initiate hemodialysis (HD) typically cannot be placed in outpatient HD centers until their insurance applications are processed. Allowing these patients temporary access to an inpatient HD unit as outpatients could shorten their hospital length of stay (LOS). Objective: To evaluate the association of hospital LOS with a first-in-state model in which an inpatient HD unit received regulatory program flex approval to provide temporary, transitional outpatient access to uninsured patients. Design, Setting, and Participants: This quasi-experimental, pre-post quality improvement study was conducted at 1 intervention hospital and 2 control hospitals within the safety net of the Los Angeles County Department of Health Services among uninsured inpatients who newly initiated HD from January 1, 2016, through December 31, 2024. Exposure: Transitional outpatient HD access to an inpatient unit was implemented in February of 2020 at the Los Angeles General Medical Center (LA General). Control hospitals began to use their emergency departments for outpatient HD access during the intervention period. Main Outcomes and Measures: The primary outcome measure was the difference in LOS before vs after transitional HD implementation. Secondary outcomes included all-cause 30-day readmission and mortality rates. Results: Overall, 951 patients were included in this study. In the preintervention period, LA General treated 200 uninsured inpatients newly initiating HD (mean [SD] age, 52.0 [11.7] years; 130 men [65.0%]) vs 241 patients (mean [SD] age, 52.9 [11.1] years; 171 men [71.0%]) in the postintervention period. The mean (SD) LOS at LA General was significantly shorter after vs before the intervention (7.6 [6.6] vs 13.0 [17.5] days; P < .001). At control hospitals, 234 patients (mean [SD] age, 52.4 [13.6] years; 164 men [70.1%]) were treated in the preintervention period, and 276 patients (mean [SD] age, 52.7 [12.5] years; 209 men [75.7%]) were treated in the postintervention period. The mean (SD) LOS at control hospitals was significantly shorter after vs before the intervention (9.1 [9.4] vs 12.5 [15.3] days; P = .002). Difference-in-difference analysis did not demonstrate a reduction in LOS at LA General vs control hospitals (-2.0 days; P = .23). However, run charts demonstrated immediate, sustained reductions in LOS at LA General but variations in LOS after the intervention at the control hospitals, such that the parallel trends assumption for difference-in-difference validity was likely not met. Conclusions and Relevance: In this quality improvement study, regulatory approval of an inpatient HD unit to provide transitional, outpatient services to uninsured patients newly initiating HD was associated with a significantly reduced LOS. These results suggest that this model could be a viable solution for other hospitals facing similar difficulties with expeditious discharge planning for uninsured patients undergoing HD.

  PublisherOA PDFDOI

• Health Economic Analysis of an All-Virtual, At-Home Acute Care Model

  JAMA Network Open · 2025-06-23 · 2 citations

  articleOpen accessSenior author

  Importance: An all-virtual, at-home acute care model, called Safer@Home, was found to enable an average 4-day reduction in hospital length of stay. The program is not currently reimbursed. Objective: To estimate costs and savings associated with the Safer@Home program from a hospital and payer perspective. Design, Setting, and Participants: This retrospective, economic evaluation analyzed costs associated with the Safer@Home program at a large, academic, public, level I trauma hospital near downtown Los Angeles, California, between September 2022 and August 2023. Patients with 1 of 10 protocolized diagnoses were eligible for the program. Data analysis occurred from January to July 2024. Exposure: Patients who presented to the hospital and were enrolled in an all-virtual, at-home acute care program called Safer@Home were compared with matched controls with similar diagnoses who received entirely in-hospital care. Main Outcome and Measures: The primary outcome was estimated net hospital and payer cost with the program vs without. Revenue from third-party payers was compared with hospital variable costs. Results: A total of 876 patients receiving care in the Safer@Home program (541 male [61.8%]; mean [SD] age, 54 [15 years]; mean [SD] expected mortality, 1.6% [4.7%]; mean [SD] case mix index, 1.27 [0.66]) were compared with 1590 matched control patients (901 male [56.7%]; mean [SD] age, 52 [20] years; mean [SD] expected mortality, 1.9% [5.9%]; mean [SD] case mix index, 1.26 [0.59]). Safer@Home enabled net hospital savings of $5.60 million, calculated as variable costs saved minus revenue lost, for the enrolled patients. Overall savings were due to net savings for Medicaid ($8380 per patient) and unfunded patients ($10 934 per patient), but net losses were due to significant loss of revenue for Medicare (-$4143 per patient) and commercially insured patients (-$25 999 per patien). Modeling demonstrated that revenue based on payer mix, rather than avoided variable hospital costs, was the primary factor of net hospital savings and losses. Absent reimbursement, the program was cost-saving to payers in all modeled scenarios. Creating reimbursement rates of 50% to 60% of hospital costs would enable the program to be cost-saving to both the hospital and payers, across payer mixes. Conclusions and Relevance: In this economic evaluation study, an all-virtual, at-home acute care program was associated with both hospital and payer savings; however, in the absence of reimbursement, it was only cost-saving to hospitals for Medicaid-funded or uninsured patients. These findings suggest that payer reform is needed to enable program generalization.

  PublisherOA PDFDOI

• Virtual Home Care for Patients With Acute Illness

  JAMA Network Open · 2024-11-26 · 12 citations

  articleOpen access1st author

  Importance: Recent evolutions in clinical care and remote monitoring suggest that some acute illnesses no longer require intravenous therapy and inpatient hospitalization. Objective: To describe outcomes of patients receiving care in a new, outpatient, virtual, home-based acute care model called Safer@Home. Design, Setting, and Participants: This retrospective cohort analysis, conducted from September 1, 2022, through August 31, 2023, included 2466 patients treated at a safety net hospital in Los Angeles County for 10 core illnesses and 24 other acute illnesses for which patients are commonly hospitalized. Exposure: Outpatient, home-based, acute care with virtual monitoring and clinic visits in lieu of inpatient or in-home care. Main Outcomes and Measures: The primary measure was hospital length of stay. Secondary measures included all-cause mortality, 30-day readmission, return urgent care visit rates, and return emergency department (ED) visit rates. Results: Safer@Home provided care to 876 patients (mean [SD] age, 54.0 [14.5] years; 541 men [61.8%]) during the study period, compared with a cohort of 1590 patients (mean [SD] age, 52.3 [19.6] years; 901 men [56.7%]) with matching diagnoses who received standard, hospital-based care. Safer@Home patients had significantly shorter mean (SD) lengths of inpatient stay than the comparison cohort (1.3 [2.0] vs 5.3 [10.4] days; P < .001), totaling 3505 bed-days avoided (mean [SD], 4.0 [10.6] bed-days saved per patient), with no significant difference in all-cause mortality at last follow-up (2.6% [23 of 876] vs 4.0% [64 of 1590]; P = .07). Safer@Home patients and control patients also had no significant difference in the proportion experiencing 30-day hospital readmission (19.9% [174 of 876] vs 16.7% [266 of 1590]; P = .06). As intended, more Safer@Home than control patients had at least one 30-day return urgent care visit (37.3% [327 of 876] vs 5.2% [82 of 1590]; P < .001). In contrast, the Safer@Home and control cohorts did not significantly differ in experiencing at least one 30-day return ED visit (15.2% [133 of 876] vs 12.5% [199 of 1590]; P = .06). Safer@Home patients had significantly fewer mean (SD) total 30-day return ED visits per patient than control patients (0.19 [0.50] vs 0.21 [0.85]; P < .001). Conclusions and Relevance: In this cohort study, patients receiving acute, virtual, home care with remote monitoring and as-needed return urgent care visits had markedly shorter hospital stays than patients receiving standard inpatient hospital care, with no significant increase in mortality, ED revisits, or return hospitalizations. This new care model is promising for systems that cannot staff Medicare-compliant hospital-at-home visits.

  PublisherOA PDFDOI

• Findings from a public engagement programme for school children aged 13–15 years about endocrinology

  Endocrine Connections · 2024-10-05

  articleOpen access

  Objectives: To investigate the utility and effectiveness of a school outreach programme in areas of lower socioeconomic status to improve understanding of common endocrine topics and the medical profession. Methods: Two secondary school outreach sessions were conducted in July 2022. Students were invited to attend lectures delivered by medical professionals and engage in poster-making sessions using the knowledge they had gained throughout the day. Participants completed anonymised pre- and post-session surveys. Outcomes were identified using Kirkpatrick's training evaluation model. Self-reported perceptions and beliefs (Kirkpatrick's level 2a) were compared using chi-square tests. Thematic analysis of team-led poster presentations was performed. Results: Of the 254 participants included, the response rates of pre- and post-session questionnaires were 75.6% and 56.2%, respectively. The outreach day increased students' understanding of obesity and diabetes, polycystic ovary syndrome, and Health Technology. The most well-received activities from the outreach day were voted to be the poster challenge (43.4%) and poster presentation (14.7%). Following the session, there was a trend towards an increased understanding of medical careers and interest in pursuing a medical career, although these did not reach statistical significance. Conclusions: Outreach programmes could be a practical and effective approach to engaging prospective medical applicants from areas of lower socioeconomic status. Further studies are required to expand outreach programmes and investigate the efficacy of school engagement programmes.

  PublisherDOI

• CoMICs (Concise Medical Information Cines) videos on diabetes mellitus and polycystic ovary syndrome have better quality, content, and reliability compared to videos from other sources

  Endocrine Abstracts · 2023-05-02 · 1 citations

  article

  Searchable abstracts of presentations at key conferences in endocrinology ISSN 1470-3947 (print) | ISSN 1479-6848 (online)

  PublisherDOI

• Real-World Application of Oral Therapy for Infective Endocarditis: A Multicenter, Retrospective, Cohort Study

  Clinical Infectious Diseases · 2023 · 53 citations

  • Medicine
  • Internal medicine
  • Surgery

  BACKGROUND: We sought to compare the outcomes of patients treated with intravenous (IV)-only vs oral transitional antimicrobial therapy for infective endocarditis (IE) after implementing a new expected practice within the Los Angeles County Department of Health Services (LAC DHS). METHODS: We conducted a multicentered, retrospective cohort study of adults with definite or possible IE treated with IV-only vs oral therapy at the 3 acute care public hospitals in the LAC DHS system between December 2018 and June 2022. The primary outcome was clinical success at 90 days, defined as being alive and without recurrence of bacteremia or treatment-emergent infectious complications. RESULTS: We identified 257 patients with IE treated with IV-only (n = 211) or oral transitional (n = 46) therapy who met study inclusion criteria. Study arms were similar for many demographics; however, the IV cohort was older, had more aortic valve involvement, were hemodialysis patients, and had central venous catheters present. In contrast, the oral cohort had a higher percentage of IE caused by methicillin-resistant Staphylococcus aureus. There was no significant difference between the groups in clinical success at 90 days or last follow-up. There was no difference in recurrence of bacteremia or readmission rates. However, patients treated with oral therapy had significantly fewer adverse events. Multivariable regression adjustments did not find significant associations between any selected variables and clinical success across treatment groups. CONCLUSIONS: These results demonstrate similar outcomes of real-world use of oral vs IV-only therapy for IE, in accord with prior randomized, controlled trials and meta-analyses.

  PublisherOA PDFDOI

• Global Research Highlights

  Canadian Journal of Emergency Medicine · 2023-01-01

  articleOpen access
  PublisherOA PDFDOI

Frequent coauthors

• Brad Spellberg

  Los Angeles Medical Center

  21 shared

• Paul Holtom

  Los Angeles Medical Center

  13 shared

• Noah Wald‐Dickler

  12 shared

• Allison B Chambliss

  University of California, Los Angeles

  12 shared

• Catherine P. Canamar

  9 shared

• C. Edward Coffey

  Los Angeles Medical Center

  9 shared

• Soodtida Tangpraphaphorn

  Los Angeles Medical Center

  9 shared

• Jan Shoenberger

  Los Angeles Medical Center

  5 shared