About

Sid Banerjee is an Associate Professor in the School of Operations Research and Information Engineering at Cornell University. He joined the faculty in July 2015 and is affiliated with the graduate fields of Applied Mathematics, Computer Science, Data Science, and Electrical and Computer Engineering. His research focuses on stochastic modeling and the design of algorithms and incentives for large-scale systems, particularly in settings where a large number of agents interact through communication or social networks, or via algorithmic mechanisms. His work spans areas such as online marketplace design, control of information flows, large-scale social network computing, and learning and recommendation systems for the Internet. Banerjee's recent research includes incentive mechanisms in collaborative platforms, dynamic pricing in ridesharing platforms like Uber and Lyft, and scalable algorithms for web search personalization.

Research topics

• Medicine
• Psychology
• Psychiatry
• Internal medicine
• Clinical psychology

Selected publications

• Remote Symptom Assessment in Ambulatory Palliative Care: User-Centered Development of an mHealth App

  Journal of Palliative Medicine · 2026-04-10

  article

  BACKGROUND: Ambulatory palliative care (PC) focuses on managing complex symptoms, yet assessment relies on recall during infrequent visits to manage daily fluctuating symptoms. We aimed to design a PC-anchored remote symptom assessment (RSA) tool with patients and clinicians. METHODS: = 14) from an ambulatory PC clinic participated in staged prototype development and testing through focus groups and interviews. RESULTS: EMA-PAL features patient- and provider-facing interfaces that visualize daily Edmonton Symptom Assessment System (ESAS) scores, highlight severe symptoms, and support shared review during visits. Participants endorsed EMA-PAL's potential to enhance communication and workflow efficiency and identified priorities for future development (medication tracking and workflow integration). CONCLUSION: This exploratory pilot suggests that PC-anchored RSA tools such as EMA-PAL may help address gaps in symptom assessment and timelier and patient-centered symptom management in ambulatory PC.

  PublisherDOI

• Abstract TU248: Loneliness and Days at Home Before and After Cardiovascular Disease Hospitalization: the REasons for Geographic And Racial Differences in Stroke (REGARDS) Study

  Circulation · 2026-03-24

  article

  Background: Social factors may influence older adults’ ability to recover or “bounce back” after a hospitalization for cardiovascular disease (CVD). Loneliness is a potentially modifiable social factor that could be intervened upon post-hospitalization. Objective: To evaluate the association of self-reported loneliness prior to CVD hospitalization (myocardial infarction, heart failure, or stroke) with days at home, defined as being alive and without inpatient, emergency department, observation unit, or skilled nursing facility care. Methods: Loneliness was assessed longitudinally using a single item approximately every 2 years in the REasons for Geographic And Racial Differences in Stroke (REGARDS) study and dichotomized as <1 or ≥1 day in the prior week. The loneliness assessment before and closest to adjudicated CVD hospitalization was selected. Linked Medicare claims were used to detect healthcare utilization. We estimated the probability of being home on each day during the year before and after hospitalization among participants surviving at least that long using a generalized estimating equation with logit link, binomial distribution, exchangeable correlation structure, and restricted quadratic splines for time, allowing for discontinuity at hospitalization and interactions between time and loneliness. Results: Among 1,780 participants with a CVD hospitalization, mean age at hospitalization was 77.9 years (SD 7.2 years), 45.7% were women, 30% Black and 70% White, and 42.4% were hospitalized for myocardial infarction, 37.5% for heart failure, and 33.4% for stroke; 23.5% reported being lonely at least 1 day in the prior week during the most recent assessment prior to hospitalization. Predicted probabilities and 95% confidence intervals for being at home in the year before and after CVD hospitalization are depicted in the Figure . Conclusions: Older adults who felt lonely had similar probability of being at home as those who were not lonely prior to CVD hospitalization, but they had lower probability of being home following CVD hospitalization. Findings suggest that loneliness may identify individuals particularly susceptible to poor recovery after CVD events.

  PublisherDOI

• Dynamic Association Between Passive Sensing Activity Levels and Behavioral Activation in Late-Life Depression (Preprint)

  2026-01-28

  articleOpen accessSenior author

  <sec> <title>BACKGROUND</title> A key symptom of depression is reduced behavioral activation, namely, low activity levels and meaningful engagement with the external environment. Thus, objective and timely measures of activity levels are useful tools to precisely track individuals’ activity levels during treatment. Prior adult depression studies have shown that activity levels measured using passive sensing (e.g., step count, time spent away from home) predict depression relapse, persistence and poor response to psychosocial interventions. Yet there is scarce research on how passive sensing measures relate to behavioral activation, and especially in late-life depression. </sec> <sec> <title>OBJECTIVE</title> We examined the association between passive sensing activity levels and self-reported behavioral activation during psychosocial interventions in community-dwelling older adults with depression. </sec> <sec> <title>METHODS</title> Sample was comprised of depressed older adults from three clinical trials at the Weill Cornell ALACRITY Center (N=75; Mean Age: 71 (IQR: [65.0, 77.5]); 91% Female; 61% White). Participants were randomized to 9 weeks of either behavioral interventions or comparison conditions. Activity levels were measured by smartphone-recorded daily step count and time away from home. Self-reported behavioral activation was measured using the BADS. We applied a functional regression (scalar-on-function) to test the association between activity levels and pre-post intervention changes in behavioral activation. </sec> <sec> <title>RESULTS</title> In the behavioral intervention group, higher daily activity – particularly step count – was initially associated with lower behavioral activation but became positively associated later in treatment (β(43) = 0.54, 95% CI: [0.05, 1.03]). In contrast, in the comparison group, greater time spent away from home showed a consistent negative association with behavioral activation throughout the intervention (minimum: β(32) = -0.80, 95% CI: [-1.11, -0.48]; maximum: β(2) = -0.04, 95% CI: [-0.04, -0.04]). </sec> <sec> <title>CONCLUSIONS</title> Our results suggest that passive sensing in older adults can measure changes in activity levels during interventions for late-life depression. It is a promising alternative to self-reports that can guide future development and personalization of interventions for older adults with depression. </sec>

  PublisherDOI

• Care Coordination and Hospitalization in Older Adults With or at Risk for Cardiovascular Disease

  JAMA Network Open · 2026-04-28

  articleOpen access

  Importance: Patients with or at risk for cardiovascular disease (CVD) often see many ambulatory physicians who may not communicate with each other. Care coordinators can bridge gaps in communication among physicians, but there are too few of them for all patients who might benefit. Objective: To compare the effectiveness of 2 strategies for allocating patients with or at risk for CVD to care coordination. Design, Setting, and Participants: This randomized clinical trial performed randomization and outreach from May 15 to November 30, 2023, and completed follow-up on May 31, 2024, in an accountable care organization in New York, New York. Participants included patients who were 65 years or older, had CVD or at least 1 CVD risk factor, had highly fragmented ambulatory care the previous year (fragmentation score ≥0.85), and had been attributed by Medicare to the accountable care organization. Intervention: Usual care assigned patients to care coordinators after any hospitalization. The intervention moved the time of care coordination earlier, offering care coordination proactively (without respect to hospitalization) to those who reported problems with care coordination on a telephone survey. Main Outcomes and Measures: The primary outcome was emergency department (ED) visits or hospitalizations during follow-up. The main secondary outcome was acceptability of the intervention, followed by appropriateness, fidelity, and efficiency. Results: A total of 400 participants (202 in the intervention group and 198 in the control group) were included in the analysis. The mean (SD) age of participants was 75.8 (7.0) years; 287 (71.8%) were female. Participants had a median of 14 (IQR, 9-22) visits to 8 (IQR, 6-11) physicians during the previous year. In the intervention group, 13 participants of 49 eligible (26.5%) accepted care coordination, compared with 17 of 17 (100%) in the control group. The most common reason for declining care coordination was that participants were coordinating care themselves. There was no difference in ED visits or hospitalizations (0.25 [95% CI, 0.21-0.31] events per 100 person-days alive in the intervention group vs 0.21 [95% CI, 0.17-0.27] events per 100 person-days alive in the control group; P = .29). Conclusions and Relevance: In this randomized clinical trial, proactive outreach for offering care coordination in advance of hospitalization did not result in better outcomes compared with usual care offering posthospitalization coordination. Many participants declined the proactive outreach offer. Trial Registration: ClinicalTrials.gov Identifier: NCT05820295.

  PublisherDOI

• Neural network models for predicting readmission among patients undergoing peripheral vascular intervention using electronic health record data and clinical registry data

  BMJ Surgery Interventions & Health Technologies · 2025-06-01

  articleOpen access

  Objectives: To determine whether neural network models based on electronic health record (EHR) data can match and augment the performance of models based on clinical registry data in predicting readmission after peripheral vascular intervention (PVI). Design: Observational cohort study. Setting: Vascular Quality Initiative registry and INSIGHT Clinical Research Network EHR data from multiple academic institutions in New York City. Participants: Patients undergoing PVI during January 1, 2013 to September 30, 2021. Main outcome measures: Our outcome variable was 90-day readmission. We developed logistic regression (LR), multilevel perceptron (MLP), and recurrent neural network (RNN) models using registry alone, EHR data alone, and combined registry-EHR data. EHR data were evaluated using derived variables to match registry variables (EHR-derived data) and clinically meaningful code aggregation (EHR-direct data). Models were evaluated using area under the curve (AUC) for discrimination, Spiegelhalter z score for calibration, and Brier score for overall performance. Results: The analytical cohort included 2348 patients undergoing PVI (mean age: 69.9±11.5 years). 832 (35%) patients were readmitted within 90 days. LR to predict 90-day readmission based on registry data alone had an AUC of 0.710, Spiegelhalter z score of 1.021, and Brier score of 0.211. MLP based on registry data alone had similar performance. MLP and RNN based on EHR-direct data (MLP: AUC=0.742, Spiegelhalter z=0.933, Brier=0.204; RNN: AUC=0.737, Spiegelhalter z=1.026, Brier=0.206) and registry+EHR-direct data (MLP: AUC=0.756, Spiegelhalter z=0.794, Brier=0.199; RNN: AUC=0.751, Spiegelhalter z=1.057, Brier=0.200) had improved performances. LR based on EHR-direct data and combined registry+EHR-direct data had worse performances. Conclusions: EHR data, when used with neural network models, can be useful to establish readmission predictive models or augment clinical registry data. EHR-based models can be potentially embedded in the clinical workflow, but model performance may be constrained by the absence of certain information in clinical encounters, such as social determinants of health.

  PublisherOA PDFDOI

• Precision Assignment to Psychosocial Interventions for Late-Life Depression: An Automated Treatment Decision Rule

  2025-03-07 · 2 citations

  preprintOpen access

  Importance: Most older adults with depression do not have access to efficacious psychotherapies, due to critical clinician shortage. Even when treated, response rates are limited to about 50%. A Treatment Decision Rule (TDR) can maximize treatment efficacy and resources by assigning patients to their optimal intervention. This is the first study to develop TDR for late-life depression designed for community settings. Objectives: To develop a scalable TDR for assignment to simple psychotherapies or care-as-usual interventions for late-life depression that can be delivered easily in community settings. Participants: The sample included 427 older adults with major depression.Design: In this prognostic study, older adults aged 60 or older with major depression participated in one of four randomized controlled trials comparing psychotherapy to care-as-usual. Setting: Participants were recruited outpatient and community settings of Weill Cornell Medicine and the University of California San Francisco (UCSF) between 2002-2011. Data were analyzed from May 2023 to January 2025.Interventions: Participants received 8 to 14 sessions of (1) simple psychotherapies (problem-solving therapy, psychotherapy for late-life depression and medical burden) or (2) care-as-usual conditions (supportive therapy, treatment as usual, or case management).Main Outcomes and Measures: Our primary outcome was mean reduction in depression severity (Hamilton Depression Rating Scale; HAM-D). We applied a data-driven Generated Effect Modifier TDR to identify the optimal intervention for each patient based on baseline characteristics (demographics, depression severity, social support, cognitive impairment, and disability). The selected TDR model maximized depression reduction and proportion of patients treated with care-as-usual interventions.Results: The TDR-based assignment improved expected reduction in depression severity by 34% compared to care-as-usual assignment and was superior to assignment of all patients to psychotherapy. Older adults with higher depression severity, stronger social support, and lower cognitive functioning should be assigned to psychotherapy. Older adults with lower depression severity, higher cognitive functioning, and low social support would benefit from care-as-usual interventions.Conclusions and Relevance: This automatic TDR can be applied in community settings to inform assignment based on baseline characteristics to increase precision, cost-effectiveness and response rates among older adults with depression.Trial Registration: NCT00601055, NCT00151372, NCT00052091, NCT00540865

  PublisherDOI

• Comparative Effectiveness of Two Methods for Assigning Care Coordinators to People Living With Dementia

  Journal of the American Geriatrics Society · 2025-06-14 · 4 citations

  articleOpen access

  BACKGROUND: We sought to determine the comparative effectiveness of two strategies for assigning care coordinators to people living with dementia (PLWD) and their caregivers. METHODS: We conducted a pragmatic randomized clinical trial embedded in a Medicare accountable care organization (ACO) in New York, NY in 2022-2024. We included community-dwelling PLWD ≥ 65 years who were attributed to the ACO and had highly fragmented ambulatory care in the previous year (reversed Bice-Boxerman Index ≥ 0.86). The trial compared usual care (assigning care coordinators to PLWD after hospital discharge) to usual care plus proactive outreach, which assigned care coordinators to PLWD if they or their caregivers reported difficulty with care coordination on a telephone survey. Participants were followed for the combined outcome of emergency department (ED) visit or hospitalization. RESULTS: Among the 385 PLWD in the trial, the mean age was 82.6 years (SD 6.9), and 56.4% were female. Overall, participants had had a mean of 14.9 ambulatory visits to 8.9 different providers the previous year. The acceptance rate of care management was higher in the control group (73.7%) than in the intervention group (38.0%). Care coordinators were ultimately assigned to 14 of 192 PLWD in the control group (7.3%) and 19 of 193 PLWD in the intervention group (9.8%). The intention-to-treat analysis (N = 385) found a trend toward fewer ED visits in the intervention group (0.14 ED visits per 100 person-days alive vs. 0.18 ED visits per 100 person-days alive, p = 0.07) but no difference in the combined outcome of ED visit or hospitalization (p = 0.71). CONCLUSION: Although the particular intervention we tested was not more effective than usual care, this trial is novel in that it used highly fragmented care as an inclusion criterion and shows that more work is needed to address fragmented care among PLWD.

  PublisherOA PDFDOI

• Precision Assignment to Psychosocial Interventions for Late-Life Depression

  JAMA Psychiatry · 2025-09-17 · 2 citations

  articleOpen access

  Importance: Most older adults with depression lack access to efficacious psychotherapies due to a critical clinician shortage. Even when treated, response rates are limited to approximately 50%. A treatment decision rule (TDR) may maximize treatment efficacy and resources by assigning patients to their optimal intervention. This is the first study to propose a TDR for late-life depression designed for community settings. Objective: To develop a scalable TDR for assignment to a psychotherapy or usual care intervention for late-life depression that can be delivered easily in community settings. Design, Setting, and Participants: In this prognostic study, adults 60 years or older with major depression participated in randomized controlled trials comparing psychotherapy with usual care. Participants were recruited from outpatient and community settings of Weill Cornell Medicine and the University of California San Francisco between 2002 and 2011. Data were analyzed from May 2023 to May 2025. Interventions: Participants received either psychotherapy (problem-solving therapy, psychotherapy for late-life depression and medical burden) or usual care (supportive therapy, treatment as usual, or case management). Main Outcomes and Measures: The primary outcome was mean reduction in depression severity (measured by the Hamilton Depression Rating Scale [HAM-D]). A generated effect modifier TDR was applied to identify the optimal intervention for each patient based on baseline characteristics (demographics, depression severity, social support, cognition, and disability). The TDR maximized depression severity reduction and the proportion of patients treated with the usual care intervention. Results: In 427 older adults with late-life depression (mean [SD] age, 72.7 [8.7] years; 70% female), the predicted HAM-D score reduction with TDR-based intervention was a mean of 49.1% (95% CI, 47.4%-51.0%). The TDR improved expected depression severity reduction by 34% compared with usual care (HAM-D reduction, 36.6% [95% CI, 34.5%-38.7%]) and the TDR was somewhat superior to assigning all patients to receive psychotherapy (HAM-D reduction, 46.7% [95% CI, 44.2%-48.8%]). Older adults with higher depression severity, stronger social support, and lower cognitive functioning should receive psychotherapy; those with lower depression severity, higher cognitive functioning, and low social support would benefit from usual care. Conclusions and Relevance: In this study of older adults with depression, pending prospective testing, the automatic TDR may be used in community settings to inform treatment assignment. The TDR has the potential to increase precision, cost-effectiveness, and response rates among older adults with depression. Trial Registration: ClinicalTrials.gov Identifiers: NCT00601055, NCT00151372, NCT00052091, NCT00540865.

  PublisherOA PDFDOI

• Geography and risk of suicidal ideation and attempts post outpatient psychiatric visit in commercially insured US adults

  Journal of Psychiatric Research · 2025-01-30 · 1 citations

  articleOpen access
  PublisherOA PDFDOI

• Longitudinal Trajectories of Symptom Change During Antidepressant Treatment Among Managed Care Patients With Depression and Anxiety

  Biological Psychiatry · 2025-04-09 · 1 citations

  article
  PublisherDOI

Recent grants

• ALACRITY for Late- and Mid-Life Mood Disorders

  NIH · $11.5M · 2017–2024

Frequent coauthors

• George S. Alexopoulos

  Cornell University

  145 shared

• Aparna Vasanthakumar

  AbbVie (United States)

  82 shared

• Fabrizio Tabbò

  82 shared

• Lucy A. Godley

  Robert H. Lurie Comprehensive Cancer Center of Northwestern University

  82 shared

• Biljana Čuljković

  82 shared

• ShaoNing Yang

  82 shared

• Micheal Leser

  Vanderbilt University

  82 shared

• Amy Chadburn

  82 shared

Education

• PhD, Biostatistics

  University of Alabama at Birmingham

  2008

• Master of Statistics (M. Stat)

  Indian Statistical Institute

  2003

• Bachelor of Statistics (B. Stat)

  Indian Statistical Institute

  2001

Awards & honors

• WNCG Student Leadership Award (2013)
• Institute Silver Medal (2007)
• Governor's Gold Medal (2007)