Siona Prasad
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Siona Prasad has a total of 50 co-authors. Here are the top co-authors by number of shared publications:
- Yulin Hswen
Department of Epidemiology and Biostatistics and Medicine, University of California, San Francisco, USA
3 shared publications Alex Rankine
Department of Mechanical Engineering Tufts University Medford MA 02155 USA
1 shared publicationAlice Kim
Department of Medicine Brigham & Women's Hospital Boston Massachusetts USA
1 shared publication- Alyssa J. Moran
Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland
1 shared publication Ania Syrowatka
Harvard Medical School Boston Massachusetts USA
1 shared publication- Anna Li
The University of Hong Kong
1 shared publication
Latest Publications
- 2 October 2025
- Published by Wiley in American Journal of Hematology
- Vol. 100 (12) , 2238-2247
Development of an Algorithm for Estimating the Likelihood of Venous Thromboembolism in Primary Care Using Structured and Unstructured Electronic Health Record Data
Venous thromboembolism (VTE) is a major public health concern. It is often clinically difficult to diagnose and affects up to 900 000 individuals annually in the United States. Delayed or missed VTE diagnosis can impact treatment and increase morbidity and mortality. This retrospective study utilized structured and unstructured electronic health record (EHR) data from a large integrated care network in the northeastern US, focusing on 4678 adult patients presenting with at least one VTE‐associated sign or symptom at primary care visits during 2019–2020. Feature selection incorporated expert‐guided and data‐driven approaches, resulting in a final set of demographic, clinical history, and sign/symptom risk factors. The primary analysis developed seven machine learning models to predict VTE incidence. Secondary analyses included the prediction of timely and delayed VTE diagnoses. All models showed predictive ability with area under the curve (AUC) of 0.83–0.88. The logistic regression model demonstrated robust performance in predicting incident VTE cases, achieving an AUC of 0.88 (95% CI: 0.86–0.90). Multiple risk factors were identified, including cancer history, smoking history, and spinal cord trauma. Variations in the top risk factors between timely and delayed prediction models highlighted how certain patients were more likely to have a delayed or missed diagnosis. This study highlights the potential for data‐driven tools to facilitate timely, point‐of‐care VTE detection by leveraging structured and unstructured EHR data. The prediction model accurately estimated the likelihood of incident VTEs, especially in cases diagnosed late, showing potential to reduce costly diagnostic delays.
- 2 October 2025
- Published by Wiley in American Journal of Hematology
- Vol. 100 (12) , 2238-2247
Development of an Algorithm for Estimating the Likelihood of Venous Thromboembolism in Primary Care Using Structured and Unstructured Electronic Health Record Data
Venous thromboembolism (VTE) is a major public health concern. It is often clinically difficult to diagnose and affects up to 900 000 individuals annually in the United States. Delayed or missed VTE diagnosis can impact treatment and increase morbidity and mortality. This retrospective study utilized structured and unstructured electronic health record (EHR) data from a large integrated care network in the northeastern US, focusing on 4678 adult patients presenting with at least one VTE‐associated sign or symptom at primary care visits during 2019–2020. Feature selection incorporated expert‐guided and data‐driven approaches, resulting in a final set of demographic, clinical history, and sign/symptom risk factors. The primary analysis developed seven machine learning models to predict VTE incidence. Secondary analyses included the prediction of timely and delayed VTE diagnoses. All models showed predictive ability with area under the curve (AUC) of 0.83–0.88. The logistic regression model demonstrated robust performance in predicting incident VTE cases, achieving an AUC of 0.88 (95% CI: 0.86–0.90). Multiple risk factors were identified, including cancer history, smoking history, and spinal cord trauma. Variations in the top risk factors between timely and delayed prediction models highlighted how certain patients were more likely to have a delayed or missed diagnosis. This study highlights the potential for data‐driven tools to facilitate timely, point‐of‐care VTE detection by leveraging structured and unstructured EHR data. The prediction model accurately estimated the likelihood of incident VTEs, especially in cases diagnosed late, showing potential to reduce costly diagnostic delays.
- Open Access
- 9 June 2025
- Published by Elsevier in Mayo Clinic Proceedings: Digital Health
- Vol. 3 (3) , 100241
Development and Evaluation of an Artificial Intelligence–Powered Surgical Oral Examination Simulator: A Pilot Study
- 9 June 2025
- Published by Elsevier in Mayo Clinic Proceedings: Digital Health
- Vol. 3 (3) , 100241
Open AccessDevelopment and Evaluation of an Artificial Intelligence–Powered Surgical Oral Examination Simulator: A Pilot Study
- Open Access
- 13 February 2023
- Published by Public Library of Science (PLoS) in PLOS Global Public Health
- Vol. 3 (2) , e0001584
Image-based surgical site infection algorithms to support home-based post-cesarean monitoring: Lessons from Rwanda
- 13 February 2023
- Published by Public Library of Science (PLoS) in PLOS Global Public Health
- Vol. 3 (2) , e0001584
Open AccessImage-based surgical site infection algorithms to support home-based post-cesarean monitoring: Lessons from Rwanda
- Open Access
- 27 January 2022
- Published by MDPI AG in International Journal of Environmental Research and Public Health
- Vol. 19 (3) , 1410
Covidseeker: A Geospatial Temporal Surveillance Tool
Introduction: Geospatial temporal data derived from smartphones traditionally used for purposes of navigation may offer valuable information for public health surveillance and locational hot spotting. Our objective was to develop a web-based application, called Covidseeker, that captures continuous fine-grained geospatial temporal data from smartphones and leverages these data to study transmission patterns of COVID-19. Methods: This report describes the development of Covidseeker and the process by which it utilizes geospatial temporal data from smartphones and processes it into a usable format to study geospatial temporal patterns of COVID-19. We provide an overview of the design process, the principles, the software architecture, and the dashboard of the Covidseeker application and consider key challenges and strategic uses of capturing geospatial temporal data and the potential for future applications in outbreak surveillance. Results: A resource such as Covidseeker can support situational awareness by providing information about the location and timing of transmission of diseases such as COVID-19. Geospatial temporal data housed in smartphones hold tremendous potential to capture more depth about where and when transmission occurs and the patterns of human mobility that lead to increases in risk of COVID-19. Conclusion: An enormous and highly rich source of geospatial temporal information about human mobility can be used to provide highly localized discrete information that is difficult to capture by traditional sources. The architecture of Covidseeker can be applied to help track COVID-19 and should be integrated with traditional disease surveillance practices.
- 27 January 2022
- Published by MDPI AG in International Journal of Environmental Research and Public Health
- Vol. 19 (3) , 1410
Open AccessCovidseeker: A Geospatial Temporal Surveillance Tool
Introduction: Geospatial temporal data derived from smartphones traditionally used for purposes of navigation may offer valuable information for public health surveillance and locational hot spotting. Our objective was to develop a web-based application, called Covidseeker, that captures continuous fine-grained geospatial temporal data from smartphones and leverages these data to study transmission patterns of COVID-19. Methods: This report describes the development of Covidseeker and the process by which it utilizes geospatial temporal data from smartphones and processes it into a usable format to study geospatial temporal patterns of COVID-19. We provide an overview of the design process, the principles, the software architecture, and the dashboard of the Covidseeker application and consider key challenges and strategic uses of capturing geospatial temporal data and the potential for future applications in outbreak surveillance. Results: A resource such as Covidseeker can support situational awareness by providing information about the location and timing of transmission of diseases such as COVID-19. Geospatial temporal data housed in smartphones hold tremendous potential to capture more depth about where and when transmission occurs and the patterns of human mobility that lead to increases in risk of COVID-19. Conclusion: An enormous and highly rich source of geospatial temporal information about human mobility can be used to provide highly localized discrete information that is difficult to capture by traditional sources. The architecture of Covidseeker can be applied to help track COVID-19 and should be integrated with traditional disease surveillance practices.
- Open Access
- 14 October 2021
- Published by MDPI AG in International Journal of Environmental Research and Public Health
- Vol. 18 (20) , 10794
The Federal Menu Labeling Law and Twitter Discussions about Calories in the United States: An Interrupted Time-Series Analysis
Public awareness of calories in food sold in retail establishments is a primary objective of the menu labeling law. This study explores the extent to which we can use social media and internet search queries to understand whether the federal calorie labeling law increased awareness of calories. To evaluate the association of the federal menu labeling law with tweeting about calories we retrieved tweets that contained the term “calorie(s)” from the CompEpi Geo Twitter Database from 1 January through 31 December in 2016 and 2018. Within the same time period, we also retrieved time-series data for search queries related to calories via Google Trends (GT). Interrupted time-series analysis was used to test whether the federal menu labeling law was associated with a change in mentions of “calorie(s)” on Twitter and relative search queries to calories on GT. Before the implementation of the federal calorie labeling law on 7 May 2018, there was a significant decrease in the baseline trend of 4.37 × 10−8 (SE = 1.25 × 10−8, p < 0.001) mean daily ratio of calorie(s) tweets. A significant increase in post-implementation slope of 3.19 × 10−8 (SE = 1.34 × 10−8 , p < 0.018) mean daily ratio of calorie(s) tweets was seen compared to the pre-implementation slope. An interrupted time-series (ITS) analysis showed a small, statistically significant upward trend of 0.0043 (SE = 0.036, p < 0.001) weekly search queries for calories pre-implementation, with no significant level change post-implementation. There was a decrease in trend of 1.22 (SE = 0.27, p < 0.001) in search queries for calories post-implementation. The federal calorie labeling law was associated with a 173% relative increase in the trend of mean daily ratio of tweets and a -28381% relative change in trend for search queries for calories. Twitter results demonstrate an increase in awareness of calories because of the addition of menu labels. Google Trends results imply that fewer people are searching for the calorie content of their meal, which may no longer be needed since calorie information is provided at point of purchase. Given our findings, discussions online about calories may provide a signal of an increased awareness in the implementation of calorie labels.
- 14 October 2021
- Published by MDPI AG in International Journal of Environmental Research and Public Health
- Vol. 18 (20) , 10794
Open AccessThe Federal Menu Labeling Law and Twitter Discussions about Calories in the United States: An Interrupted Time-Series Analysis
Public awareness of calories in food sold in retail establishments is a primary objective of the menu labeling law. This study explores the extent to which we can use social media and internet search queries to understand whether the federal calorie labeling law increased awareness of calories. To evaluate the association of the federal menu labeling law with tweeting about calories we retrieved tweets that contained the term “calorie(s)” from the CompEpi Geo Twitter Database from 1 January through 31 December in 2016 and 2018. Within the same time period, we also retrieved time-series data for search queries related to calories via Google Trends (GT). Interrupted time-series analysis was used to test whether the federal menu labeling law was associated with a change in mentions of “calorie(s)” on Twitter and relative search queries to calories on GT. Before the implementation of the federal calorie labeling law on 7 May 2018, there was a significant decrease in the baseline trend of 4.37 × 10−8 (SE = 1.25 × 10−8, p < 0.001) mean daily ratio of calorie(s) tweets. A significant increase in post-implementation slope of 3.19 × 10−8 (SE = 1.34 × 10−8 , p < 0.018) mean daily ratio of calorie(s) tweets was seen compared to the pre-implementation slope. An interrupted time-series (ITS) analysis showed a small, statistically significant upward trend of 0.0043 (SE = 0.036, p < 0.001) weekly search queries for calories pre-implementation, with no significant level change post-implementation. There was a decrease in trend of 1.22 (SE = 0.27, p < 0.001) in search queries for calories post-implementation. The federal calorie labeling law was associated with a 173% relative increase in the trend of mean daily ratio of tweets and a -28381% relative change in trend for search queries for calories. Twitter results demonstrate an increase in awareness of calories because of the addition of menu labels. Google Trends results imply that fewer people are searching for the calorie content of their meal, which may no longer be needed since calorie information is provided at point of purchase. Given our findings, discussions online about calories may provide a signal of an increased awareness in the implementation of calorie labels.
- Open Access
- 4 May 2021
- Published by Wiley in Advanced NanoBiomed Research
- Vol. 1 (8) , 2000116
Atomic Force Microscopy Detects the Difference in Cancer Cells of Different Neoplastic Aggressiveness via Machine Learning
A novel method based on atomic force microscopy (AFM) working in Ringing mode (RM) to distinguish between two similar human colon epithelial cancer cell lines that exhibit different degrees of neoplastic aggressiveness is reported on. The classification accuracy in identifying the cell line based on the images of a single cell can be as high as 94% (the area under the receiver operating characteristic [ROC] curve is 0.99). Comparing the accuracy using the RM and the regular imaging channels, it is seen that the RM channels are responsible for the high accuracy. The cells are also studied with a traditional AFM indentation method, which gives information about cell mechanics and the pericellular coat. Although a statistically significant difference between the two cell lines is also seen in the indentation method, it provides the accuracy of identifying the cell line at the single‐cell level less than 68% (the area under the ROC curve is 0.73). Thus, AFM cell imaging is substantially more accurate in identifying the cell phenotype than the traditional AFM indentation method. All the obtained cell data are collected on fixed cells and analyzed using machine learning methods. The biophysical reasons for the observed classification are discussed.
- 4 May 2021
- Published by Wiley in Advanced NanoBiomed Research
- Vol. 1 (8) , 2000116
Open AccessAtomic Force Microscopy Detects the Difference in Cancer Cells of Different Neoplastic Aggressiveness via Machine Learning
A novel method based on atomic force microscopy (AFM) working in Ringing mode (RM) to distinguish between two similar human colon epithelial cancer cell lines that exhibit different degrees of neoplastic aggressiveness is reported on. The classification accuracy in identifying the cell line based on the images of a single cell can be as high as 94% (the area under the receiver operating characteristic [ROC] curve is 0.99). Comparing the accuracy using the RM and the regular imaging channels, it is seen that the RM channels are responsible for the high accuracy. The cells are also studied with a traditional AFM indentation method, which gives information about cell mechanics and the pericellular coat. Although a statistically significant difference between the two cell lines is also seen in the indentation method, it provides the accuracy of identifying the cell line at the single‐cell level less than 68% (the area under the ROC curve is 0.73). Thus, AFM cell imaging is substantially more accurate in identifying the cell phenotype than the traditional AFM indentation method. All the obtained cell data are collected on fixed cells and analyzed using machine learning methods. The biophysical reasons for the observed classification are discussed.
- 1 January 2021
- Published by Boston Congress of Public Health
- No. 26
Political Affiliation and Human Mobility Under Stay-at-Home Orders: A Difference-in-Difference Analysis with County and Time Fixed Effects
Background In late March 2020, state and local governments across the country issued stay-at-home directives to slow the spread of COVID-19. However, divergent messages from political parties on the severity of COVID-19 and differing levels of support of these social distancing measures have potentially prompted differential behaviors across political groups. This study examines state-level partisan differences in changes in human mobility during stay-at-home orders. Methods Aggregated and de-identified large-scale human mobility data was collected from Cuebiq, a mobility insights platform, to measure the fraction of users that staying at home. A difference-in-difference analysis was performed to evaluate the changes in human mobility before and after stay-at-home orders were implemented by state political afflictions. Results Before March 19th, 2020, there was a 0.82% difference (SE=0.003, p<0.001) between the percent of users staying at home in republican states versus democratic states. Difference-in-difference analysis revealed that on average, democratic states experienced a 4.11% (SE = 0.006, p<0.001) greater increase in the percent of users that stayed at home compared to republican states pre-post the implementation of stay-at-home orders. States that were not issued state-wide stay-at-home orders were excluded from the study. Conclusion Evidence supports the differential changes in adherence to stay-at-home orders by state political affiliation. These results suggest that political messaging may be a strong factor in influencing social distancing behaviors.
- 1 January 2021
- Published by Boston Congress of Public Health
- No. 26
Political Affiliation and Human Mobility Under Stay-at-Home Orders: A Difference-in-Difference Analysis with County and Time Fixed Effects
Background In late March 2020, state and local governments across the country issued stay-at-home directives to slow the spread of COVID-19. However, divergent messages from political parties on the severity of COVID-19 and differing levels of support of these social distancing measures have potentially prompted differential behaviors across political groups. This study examines state-level partisan differences in changes in human mobility during stay-at-home orders. Methods Aggregated and de-identified large-scale human mobility data was collected from Cuebiq, a mobility insights platform, to measure the fraction of users that staying at home. A difference-in-difference analysis was performed to evaluate the changes in human mobility before and after stay-at-home orders were implemented by state political afflictions. Results Before March 19th, 2020, there was a 0.82% difference (SE=0.003, p<0.001) between the percent of users staying at home in republican states versus democratic states. Difference-in-difference analysis revealed that on average, democratic states experienced a 4.11% (SE = 0.006, p<0.001) greater increase in the percent of users that stayed at home compared to republican states pre-post the implementation of stay-at-home orders. States that were not issued state-wide stay-at-home orders were excluded from the study. Conclusion Evidence supports the differential changes in adherence to stay-at-home orders by state political affiliation. These results suggest that political messaging may be a strong factor in influencing social distancing behaviors.