Seshadri C. Mudumbai
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Seshadri C. Mudumbai has a total of 158 co-authors. Here are the top co-authors by number of shared publications:
- Edward Rivera Mariano
Stanford University School of Medicine Stanford CA USA
16 shared publications - Randall Scott Stafford
Affiliation Stanford University School of Medicine, Stanford Prevention Research Center, Stanford, California, United States of America
6 shared publications - Todd H. Wagner
Health Economics Resource Center, VA Palo Alto Health Care System, Menlo Park, California
6 shared publications M. Frances Davies
4 shared publications- Paul A. Heidenreich
Medical Service, VA Palo Alto Healthcare System, CA (R.L.T., N.M.K., A.V., N.D., A.S., P.H.).
4 shared publications Randall S Stafford
Department of Medicine Primary, Stanford Prevention Research Center, Stanford University School of Medicine , Stanford, CA 94305
4 shared publications
Latest Publications
- Open Access
- 8 October 2026
- Published by Elsevier in Drug and Alcohol Dependence
- Vol. 288
Incidence of opioid use disorder among patients prescribed long-term opioid therapy: An observational study
- 8 October 2026
- Published by Elsevier in Drug and Alcohol Dependence
- Vol. 288
Open AccessIncidence of opioid use disorder among patients prescribed long-term opioid therapy: An observational study
- 23 May 2026
- Published by Springer Nature in Journal of Medical Systems
- Vol. 50 (1) , 85
A Standardized Framework for Evaluating State Opioid Surveillance Dashboards: A Health Informatics Assessment
Background. State public health agencies across the U.S. depend on opioid surveillance dashboards to monitor overdose trends and guide interventions. Despite their central role in decision-making, no standardized framework exists to evaluate how well these dashboards actually function. Objective. We developed a health informatics framework to systematically assess state opioid surveillance dashboards and applied it across all 50 U.S. states. Methods. Our framework evaluates dashboards across nine domains (including temporal analysis, geographic granularity, data disaggregation, visualization features, narrative support, simplification techniques, and interactive functionality) using 28 specific sub-variables. Sub-variables were selected based on public health informatics literature on effective dashboard design, including features identified in COVID-19 dashboard evaluations and usability frameworks. We scored each feature as either present or absent, creating composite scores with a theoretical maximum of 28 points. All 50 U.S. state-level dashboards were reviewed twice by the study team, with an external consistency check on a subset of 10 dashboards. We calculated the correlation coefficient between dashboard scores and state population estimates. Results. Dashboard scores varied widely, ranging from 12 to 22 points (mean:19.18, standard deviation:2.20). Three states (Delaware, Indiana, and New York) achieved the highest score of 22 by offering comprehensive geographic detail, extensive data breakdowns, and full interactive capabilities. Interactive filtering was present in 92% of dashboards, but only 80% enabled switching between visualization modes. While nearly all dashboards showed time trends, only 6% provided zip code-level detail, and over a third lacked race and ethnicity data. Single external reviewer demonstrated strong correlation (r = 0.84, p = 0.002). Dashboard scores were not correlated with state population size (r = 0.187, p = 0.193). Conclusions. State opioid surveillance dashboards varied substantially in observable capabilities. This framework provides a standardized method to compare dashboard features and identify gaps that may inform future refinement.
- 23 May 2026
- Published by Springer Nature in Journal of Medical Systems
- Vol. 50 (1) , 85
A Standardized Framework for Evaluating State Opioid Surveillance Dashboards: A Health Informatics Assessment
Background. State public health agencies across the U.S. depend on opioid surveillance dashboards to monitor overdose trends and guide interventions. Despite their central role in decision-making, no standardized framework exists to evaluate how well these dashboards actually function. Objective. We developed a health informatics framework to systematically assess state opioid surveillance dashboards and applied it across all 50 U.S. states. Methods. Our framework evaluates dashboards across nine domains (including temporal analysis, geographic granularity, data disaggregation, visualization features, narrative support, simplification techniques, and interactive functionality) using 28 specific sub-variables. Sub-variables were selected based on public health informatics literature on effective dashboard design, including features identified in COVID-19 dashboard evaluations and usability frameworks. We scored each feature as either present or absent, creating composite scores with a theoretical maximum of 28 points. All 50 U.S. state-level dashboards were reviewed twice by the study team, with an external consistency check on a subset of 10 dashboards. We calculated the correlation coefficient between dashboard scores and state population estimates. Results. Dashboard scores varied widely, ranging from 12 to 22 points (mean:19.18, standard deviation:2.20). Three states (Delaware, Indiana, and New York) achieved the highest score of 22 by offering comprehensive geographic detail, extensive data breakdowns, and full interactive capabilities. Interactive filtering was present in 92% of dashboards, but only 80% enabled switching between visualization modes. While nearly all dashboards showed time trends, only 6% provided zip code-level detail, and over a third lacked race and ethnicity data. Single external reviewer demonstrated strong correlation (r = 0.84, p = 0.002). Dashboard scores were not correlated with state population size (r = 0.187, p = 0.193). Conclusions. State opioid surveillance dashboards varied substantially in observable capabilities. This framework provides a standardized method to compare dashboard features and identify gaps that may inform future refinement.
- 14 May 2026
- Published by Springer Nature in Journal of Medical Systems
- Vol. 50 (1) , 81
Large Language Models for Semantic Interoperability in Value-Based Perioperative Care
Proven protocols remain isolated inside institutions because EHR systems cannot understand each other’s data semantically. Current data models each advance semantic interoperability, but none fully solves the problem of cross-institutional protocol portability. The authors suggest that large language models can translate data across systems using existing data model frameworks as “clinical constitutions” to achieve interoperability without universal standards.
- 14 May 2026
- Published by Springer Nature in Journal of Medical Systems
- Vol. 50 (1) , 81
Large Language Models for Semantic Interoperability in Value-Based Perioperative Care
Proven protocols remain isolated inside institutions because EHR systems cannot understand each other’s data semantically. Current data models each advance semantic interoperability, but none fully solves the problem of cross-institutional protocol portability. The authors suggest that large language models can translate data across systems using existing data model frameworks as “clinical constitutions” to achieve interoperability without universal standards.
- 23 March 2026
- Published by Wolters Kluwer Health in Anesthesia & Analgesia
- Vol. 143 (3) , 472-479
Study Design and Statistical Methods for Health Equity Research in Perioperative Medicine
- 23 March 2026
- Published by Wolters Kluwer Health in Anesthesia & Analgesia
- Vol. 143 (3) , 472-479
Study Design and Statistical Methods for Health Equity Research in Perioperative Medicine
- 24 December 2025
- Published by BMJ in Regional Anesthesia & Pain Medicine
Utilization and opioid outcomes of a transitional pain service in high-risk surgical veterans: a cohort study
Background: The immediate postoperative period is high-risk for patients with psychiatric comorbidities or substance use disorders, who are prone to prolonged opioid use. Transitional Pain Services (TPSs) support pain control, tapering, and recovery, but predictors of high utilization are unclear. Methods: We conducted a retrospective cohort study of surgical patients managed by TPS at a Veterans Affairs medical center from February 2024 to February 2025. Outcomes included total postoperative opioid use (morphine milligram equivalents, MME), TPS duration (days), and provider contacts, analyzed using multivariable linear regression (one-sided winsorization at 90th percentile). Predictors included demographics, pain catastrophizing, psychiatric/substance use history, and preoperative opioid use. Results: Among 345 patients, 94% had chronic pain, 26% used opioids preoperatively, 69% had psychiatric diagnosis, and 79% had substance use history. Of preoperative opioid users (n=88), 68% achieved dose reduction; only 3.2% of the full cohort were discharged on higher doses. Median postoperative use was 218 MME (IQR 165–428), TPS duration 31 days (IQR 12–36), provider contacts 5 (IQR 3–6). Pain catastrophizing was the strongest predictor: each point increase was associated with 31 MME higher opioid use (p=0.023), 3.9 additional TPS days (p<0.001), and increased contacts (p=0.015). Active nicotine use (β=109 MME, p=0.005) and preoperative opioid ≥50 MME/day (β=169 MME, p=0.006) also predicted higher opioid consumption. Conclusions: Pain catastrophizing and substance use markers predicted higher TPS utilization and opioid use, highlighting the value of preoperative psychosocial screening. Limiting dose escalation to 3.2%, despite increased baseline risks, demonstrates the effectiveness of well-resourced TPS programs.
- 24 December 2025
- Published by BMJ in Regional Anesthesia & Pain Medicine
Utilization and opioid outcomes of a transitional pain service in high-risk surgical veterans: a cohort study
Background: The immediate postoperative period is high-risk for patients with psychiatric comorbidities or substance use disorders, who are prone to prolonged opioid use. Transitional Pain Services (TPSs) support pain control, tapering, and recovery, but predictors of high utilization are unclear. Methods: We conducted a retrospective cohort study of surgical patients managed by TPS at a Veterans Affairs medical center from February 2024 to February 2025. Outcomes included total postoperative opioid use (morphine milligram equivalents, MME), TPS duration (days), and provider contacts, analyzed using multivariable linear regression (one-sided winsorization at 90th percentile). Predictors included demographics, pain catastrophizing, psychiatric/substance use history, and preoperative opioid use. Results: Among 345 patients, 94% had chronic pain, 26% used opioids preoperatively, 69% had psychiatric diagnosis, and 79% had substance use history. Of preoperative opioid users (n=88), 68% achieved dose reduction; only 3.2% of the full cohort were discharged on higher doses. Median postoperative use was 218 MME (IQR 165–428), TPS duration 31 days (IQR 12–36), provider contacts 5 (IQR 3–6). Pain catastrophizing was the strongest predictor: each point increase was associated with 31 MME higher opioid use (p=0.023), 3.9 additional TPS days (p<0.001), and increased contacts (p=0.015). Active nicotine use (β=109 MME, p=0.005) and preoperative opioid ≥50 MME/day (β=169 MME, p=0.006) also predicted higher opioid consumption. Conclusions: Pain catastrophizing and substance use markers predicted higher TPS utilization and opioid use, highlighting the value of preoperative psychosocial screening. Limiting dose escalation to 3.2%, despite increased baseline risks, demonstrates the effectiveness of well-resourced TPS programs.
- Open Access
- 13 July 2025
- Published by BMJ in Regional Anesthesia & Pain Medicine
Optimal multimodal analgesia combinations to reduce pain and opioid use following non-cardiac surgery: an instrumental variable analysis
Background Multimodal analgesia (MMA) is a perioperative pain management strategy that targets various pain pathways, resulting in reduced postoperative pain and opioid use. Unfortunately, the optimal combinations of pain medications to use with perioperative MMA remain uncertain. Our goal was to estimate the treatment effect of MMA on postoperative pain and opioid use and identify optimal non-opioid medication combinations to enhance MMA benefits. Methods The study population includes all patients undergoing elective non-cardiac surgery with general anesthesia between 1 January 2017 and 31 December 2022 at six geographically similar Veterans Health Administration hospitals. An instrumental variable (IV) analysis was conducted using the anesthesiologist as the instrument to emulate randomization to receiving specific pain medication combinations. Outcomes were self-reported pain and opioid use after surgery. Results Of the 23 238 procedures included in the study, 46.1% received MMA. MMA was more common in younger patients, females and those with a lower probability of mortality. With IV analysis, inpatients with MMA required 6.8 fewer oral morphine equivalents (OMEs, 95% CI −10.2, to −3.4) in the postoperative period, and outpatients with MMA reported postoperative pain scores that were, on average, 1.0 unit lower than patients who did not receive MMA (95% CI −1.6 to −0.4). Combinations of non-steroidal anti-inflammatory drugs (NSAIDs) plus dexamethasone or regional anesthesia resulted in the greatest reductions in postoperative opioid use (mean reduction −29.5 OMEs, 95% CI −36.9 to −19.5 and mean reduction −28.4 OMEs, 95% CI −40.1 to −16.8, respectively). Conclusion Our findings further support existing evidence on the effectiveness of MMA in reducing postoperative pain and opioid use following non-cardiac surgery. Importantly, our study highlights that dexamethasone and NSAIDs, not acetaminophen, which is almost universally used in MMA regimens, resulted in the greatest reduction of postoperative pain and postoperative opioid use. This has significant implications for the continued use of NSAIDs and dexamethasone in MMA protocols and underscores the need for future studies exploring the independent effect of intravenous acetaminophen on postoperative pain.
- 13 July 2025
- Published by BMJ in Regional Anesthesia & Pain Medicine
Open AccessOptimal multimodal analgesia combinations to reduce pain and opioid use following non-cardiac surgery: an instrumental variable analysis
Background Multimodal analgesia (MMA) is a perioperative pain management strategy that targets various pain pathways, resulting in reduced postoperative pain and opioid use. Unfortunately, the optimal combinations of pain medications to use with perioperative MMA remain uncertain. Our goal was to estimate the treatment effect of MMA on postoperative pain and opioid use and identify optimal non-opioid medication combinations to enhance MMA benefits. Methods The study population includes all patients undergoing elective non-cardiac surgery with general anesthesia between 1 January 2017 and 31 December 2022 at six geographically similar Veterans Health Administration hospitals. An instrumental variable (IV) analysis was conducted using the anesthesiologist as the instrument to emulate randomization to receiving specific pain medication combinations. Outcomes were self-reported pain and opioid use after surgery. Results Of the 23 238 procedures included in the study, 46.1% received MMA. MMA was more common in younger patients, females and those with a lower probability of mortality. With IV analysis, inpatients with MMA required 6.8 fewer oral morphine equivalents (OMEs, 95% CI −10.2, to −3.4) in the postoperative period, and outpatients with MMA reported postoperative pain scores that were, on average, 1.0 unit lower than patients who did not receive MMA (95% CI −1.6 to −0.4). Combinations of non-steroidal anti-inflammatory drugs (NSAIDs) plus dexamethasone or regional anesthesia resulted in the greatest reductions in postoperative opioid use (mean reduction −29.5 OMEs, 95% CI −36.9 to −19.5 and mean reduction −28.4 OMEs, 95% CI −40.1 to −16.8, respectively). Conclusion Our findings further support existing evidence on the effectiveness of MMA in reducing postoperative pain and opioid use following non-cardiac surgery. Importantly, our study highlights that dexamethasone and NSAIDs, not acetaminophen, which is almost universally used in MMA regimens, resulted in the greatest reduction of postoperative pain and postoperative opioid use. This has significant implications for the continued use of NSAIDs and dexamethasone in MMA protocols and underscores the need for future studies exploring the independent effect of intravenous acetaminophen on postoperative pain.
- 6 June 2025
- Published by Oxford University Press (OUP) in Military Medicine
- Vol. 190 (11-12) , e2550-e2557
Influence of Social Support and Personality Traits on Digital Health Technology Adoption Among U.S. Veterans: A Mixed-Methods Study
Introduction: Digital health technologies (DHT), including telemedicine and world wide web-based tools, have transformed healthcare delivery by enhancing patient self-management and clinical monitoring capabilities. Despite widespread adoption in many populations, U.S. veterans show notably lower rates of technology use compared to other demographic groups. The objective of this study was to investigate factors influencing the adoption of DHT among veterans at a tertiary referral hospital. Materials and Methods: Following Institutional Review Board (IRB) approval, we conducted structured telephone interviews to assess technology use, knowledge, and preferences using the Technology Acceptance Model (TAM) framework. We supplemented interview data with information, including sociodemographics from electronic health records. Participants were classified as either non-users or adopters (i.e., those with some-or-more use) of DHT. Analysis included descriptive statistics; logistic regression to identify adoption predictors; and consensus clustering to characterize user subgroups based on demographic, psychosocial, and personality factors. Results: The study included 141 veterans, predominantly older white males from suburban areas. DHT adopters (n = 50, 35.4%) were younger than non-users (median age 61 vs. 70 years, P = .012). The presence of technology-using family or friends strongly predicted adoption (odds ratios [OR] = 5.59, 95% confidence interval [CI], 1.91-18.12, P = .002). Personality traits associated with adoption included openness to new experiences (OR = 7.09, 95% CI, 1.89-32.20, P = .006) and anxiety (OR = 3.11, 95% CI, 1.12-9.27, P = .034). No significant associations were found between DHT adoption and age, gender, race, income, or education. Among adopters, consensus clustering identified 2 robust subgroups: Cluster 1 with a median age of 62 years and high extroversion (93.3%), and Cluster 2 with a median age of 53 years and lower extroversion (6.7%). Conclusions: This study demonstrates that social support networks and personality characteristics, rather than demographic factors, primarily influence veterans’ adoption of DHT. Our findings suggest a need for personalized implementation strategies that consider psychological and social factors. Future initiatives should focus on developing targeted educational programs that leverage social support systems while accounting for individual personality traits.
- 6 June 2025
- Published by Oxford University Press (OUP) in Military Medicine
- Vol. 190 (11-12) , e2550-e2557
Influence of Social Support and Personality Traits on Digital Health Technology Adoption Among U.S. Veterans: A Mixed-Methods Study
Introduction: Digital health technologies (DHT), including telemedicine and world wide web-based tools, have transformed healthcare delivery by enhancing patient self-management and clinical monitoring capabilities. Despite widespread adoption in many populations, U.S. veterans show notably lower rates of technology use compared to other demographic groups. The objective of this study was to investigate factors influencing the adoption of DHT among veterans at a tertiary referral hospital. Materials and Methods: Following Institutional Review Board (IRB) approval, we conducted structured telephone interviews to assess technology use, knowledge, and preferences using the Technology Acceptance Model (TAM) framework. We supplemented interview data with information, including sociodemographics from electronic health records. Participants were classified as either non-users or adopters (i.e., those with some-or-more use) of DHT. Analysis included descriptive statistics; logistic regression to identify adoption predictors; and consensus clustering to characterize user subgroups based on demographic, psychosocial, and personality factors. Results: The study included 141 veterans, predominantly older white males from suburban areas. DHT adopters (n = 50, 35.4%) were younger than non-users (median age 61 vs. 70 years, P = .012). The presence of technology-using family or friends strongly predicted adoption (odds ratios [OR] = 5.59, 95% confidence interval [CI], 1.91-18.12, P = .002). Personality traits associated with adoption included openness to new experiences (OR = 7.09, 95% CI, 1.89-32.20, P = .006) and anxiety (OR = 3.11, 95% CI, 1.12-9.27, P = .034). No significant associations were found between DHT adoption and age, gender, race, income, or education. Among adopters, consensus clustering identified 2 robust subgroups: Cluster 1 with a median age of 62 years and high extroversion (93.3%), and Cluster 2 with a median age of 53 years and lower extroversion (6.7%). Conclusions: This study demonstrates that social support networks and personality characteristics, rather than demographic factors, primarily influence veterans’ adoption of DHT. Our findings suggest a need for personalized implementation strategies that consider psychological and social factors. Future initiatives should focus on developing targeted educational programs that leverage social support systems while accounting for individual personality traits.
- 5 February 2025
- Published by BMJ in Regional Anesthesia & Pain Medicine
- Vol. 50 (2) , 184-187
Lessons learned from establishing a transitional pain service in the USA
- 5 February 2025
- Published by BMJ in Regional Anesthesia & Pain Medicine
- Vol. 50 (2) , 184-187
Lessons learned from establishing a transitional pain service in the USA
- 2 February 2025
- Published by Wiley in Anaesthesia
- Vol. 80 (4) , 451-453
Beyond the ‘black box’: choosing interpretable machine learning models for predicting postoperative opioid trends
- 2 February 2025
- Published by Wiley in Anaesthesia
- Vol. 80 (4) , 451-453
Beyond the ‘black box’: choosing interpretable machine learning models for predicting postoperative opioid trends