Jenny W. Sun
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
Research Subjects
Jenny W. Sun has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
Brian T. Bateman
Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine Brigham and Women's Hospital and Harvard Medical School Boston Massachusetts USA
2 shared publicationsKrista F. Huybrechts
Department of Epidemiology Harvard T. H. Chan School of Public Health Boston Massachusetts USA
2 shared publicationsSebastien Haneuse
Department of Biostatistics, Harvard T. H. Chan School of Public Health , Boston, MA, USA
2 shared publicationsSengwee Toh
Department of Population Medicine, Harvard Medical School and Harvard Pilgrim Health Care Institute, Boston, Massachusetts
2 shared publicationsSonia Hernández-Díaz
Department of Epidemiology, Harvard T. H. Chan School of Public Health , Boston, MA, USA
2 shared publicationsAaron L. Sarvet
Department of Population Medicine, Harvard Medical School and Harvard Pilgrim Health Care Institute, Boston, Massachusetts
1 shared publication
Latest Publications
- 13 December 2025
- Published by Elsevier in The Journal of Pediatrics
- Vol. 291, 114941
Application of the Target Trial Emulation Framework to Studies in the Pediatric Population
- 13 December 2025
- Published by Elsevier in The Journal of Pediatrics
- Vol. 291, 114941
Application of the Target Trial Emulation Framework to Studies in the Pediatric Population
- Open Access
- 5 April 2022
- Published by American Medical Association (AMA) in JAMA Network Open
- Vol. 5 (4) , e226484
Comparison of Rates of Type 2 Diabetes in Adults and Children Treated With Anticonvulsant Mood Stabilizers
Over the past several decades, the prevalence of type 2 diabetes (T2D) has increased in the US and globally.1-4 Approximately 1 in 10 people in the US were living with T2D from 2013 to 2016,1,5 and youth-onset T2D has increased in the US by approximately 5% annually from 2002 to 2015.5-7 There are several known risk factors for T2D, including lifestyle, environmental, and psychosocial factors that are key focus areas for prevention.8 Patients and clinicians would also benefit from guidance on choosing between similarly effective medications that may have different metabolic adverse effects.
- 5 April 2022
- Published by American Medical Association (AMA) in JAMA Network Open
- Vol. 5 (4) , e226484
Open AccessComparison of Rates of Type 2 Diabetes in Adults and Children Treated With Anticonvulsant Mood Stabilizers
Over the past several decades, the prevalence of type 2 diabetes (T2D) has increased in the US and globally.1-4 Approximately 1 in 10 people in the US were living with T2D from 2013 to 2016,1,5 and youth-onset T2D has increased in the US by approximately 5% annually from 2002 to 2015.5-7 There are several known risk factors for T2D, including lifestyle, environmental, and psychosocial factors that are key focus areas for prevention.8 Patients and clinicians would also benefit from guidance on choosing between similarly effective medications that may have different metabolic adverse effects.
- 6 January 2022
- Published by Oxford University Press (OUP) in American Journal of Epidemiology
- Vol. 191 (4) , 711-723
Use of Linked Databases for Improved Confounding Control: Considerations for Potential Selection Bias
Pharmacoepidemiologic studies are increasingly conducted within linked databases, often to obtain richer confounder data. However, the potential for selection bias is frequently overlooked when linked data is available only for a subset of patients. We highlight the importance of accounting for potential selection bias by evaluating the association between antipsychotics and type 2 diabetes in youths within a claims database linked to a smaller laboratory database. We used inverse probability of treatment weights (IPTW) to control for confounding. In analyses restricted to the linked cohorts, we applied inverse probability of selection weights (IPSW) to create a population representative of the full cohort. We used pooled logistic regression weighted by IPTW only or IPTW and IPSW to estimate treatment effects. Metabolic conditions were more prevalent in linked cohorts compared with the full cohort. Within the full cohort, the confounding-adjusted hazard ratio was 2.26 (95% CI: 2.07, 2.49) comparing initiation of antipsychotics with initiation of control medications. Within the linked cohorts, a different magnitude of association was obtained without adjustment for selection, whereas applying IPSW resulted in point estimates similar to the full cohort’s (e.g., an adjusted hazard ratio of 1.63 became 2.12). Linked database studies may generate biased estimates without proper adjustment for potential selection bias.
- 6 January 2022
- Published by Oxford University Press (OUP) in American Journal of Epidemiology
- Vol. 191 (4) , 711-723
Use of Linked Databases for Improved Confounding Control: Considerations for Potential Selection Bias
Pharmacoepidemiologic studies are increasingly conducted within linked databases, often to obtain richer confounder data. However, the potential for selection bias is frequently overlooked when linked data is available only for a subset of patients. We highlight the importance of accounting for potential selection bias by evaluating the association between antipsychotics and type 2 diabetes in youths within a claims database linked to a smaller laboratory database. We used inverse probability of treatment weights (IPTW) to control for confounding. In analyses restricted to the linked cohorts, we applied inverse probability of selection weights (IPSW) to create a population representative of the full cohort. We used pooled logistic regression weighted by IPTW only or IPTW and IPSW to estimate treatment effects. Metabolic conditions were more prevalent in linked cohorts compared with the full cohort. Within the full cohort, the confounding-adjusted hazard ratio was 2.26 (95% CI: 2.07, 2.49) comparing initiation of antipsychotics with initiation of control medications. Within the linked cohorts, a different magnitude of association was obtained without adjustment for selection, whereas applying IPSW resulted in point estimates similar to the full cohort’s (e.g., an adjusted hazard ratio of 1.63 became 2.12). Linked database studies may generate biased estimates without proper adjustment for potential selection bias.
- 30 October 2020
- Published by Oxford University Press (OUP) in American Journal of Epidemiology
- Vol. 190 (5) , 918-927
Development and Validation of a Pediatric Comorbidity Index
Comorbidity scores are widely used to help address confounding bias in nonrandomized studies conducted within health-care databases, but existing scores were developed to predict all-cause mortality in adults and might not be appropriate for use in pediatric studies. We developed and validated a pediatric comorbidity index, using health-care utilization data from the tenth revision of the International Classification of Diseases. Within the MarketScan database of US commercial claims data, pediatric patients (aged ≤18 years) continuously enrolled between October 1, 2015, and September 30, 2017, were identified. Logistic regression was used to predict the 1-year risk of hospitalization based on 27 predefined conditions and empirically identified conditions derived from the most prevalent diagnoses among patients with the outcome. A single numerical index was created by assigning weights to each condition based on its β coefficient. We conducted internal validation of the index and compared its performance with existing adult scores. The pediatric comorbidity index consisted of 24 conditions and achieved a C statistic of 0.718 (95% confidence interval (CI): 0.714, 0.723). The index outperformed existing adult scores in a pediatric population (C statistics ranging from 0.522 to 0.640). The pediatric comorbidity index provides a summary measure of disease burden and can be used for risk adjustment in epidemiologic studies of pediatric patients.
- 30 October 2020
- Published by Oxford University Press (OUP) in American Journal of Epidemiology
- Vol. 190 (5) , 918-927
Development and Validation of a Pediatric Comorbidity Index
Comorbidity scores are widely used to help address confounding bias in nonrandomized studies conducted within health-care databases, but existing scores were developed to predict all-cause mortality in adults and might not be appropriate for use in pediatric studies. We developed and validated a pediatric comorbidity index, using health-care utilization data from the tenth revision of the International Classification of Diseases. Within the MarketScan database of US commercial claims data, pediatric patients (aged ≤18 years) continuously enrolled between October 1, 2015, and September 30, 2017, were identified. Logistic regression was used to predict the 1-year risk of hospitalization based on 27 predefined conditions and empirically identified conditions derived from the most prevalent diagnoses among patients with the outcome. A single numerical index was created by assigning weights to each condition based on its β coefficient. We conducted internal validation of the index and compared its performance with existing adult scores. The pediatric comorbidity index consisted of 24 conditions and achieved a C statistic of 0.718 (95% confidence interval (CI): 0.714, 0.723). The index outperformed existing adult scores in a pediatric population (C statistics ranging from 0.522 to 0.640). The pediatric comorbidity index provides a summary measure of disease burden and can be used for risk adjustment in epidemiologic studies of pediatric patients.
- 29 October 2020
- Published by Wiley in Diabetes, Obesity and Metabolism
- Vol. 23 (2) , 444-454
Antidiabetic medication use in commercially insured children and adolescents in the United States from 2004 to 2019
Aim: To describe the patterns of non‐insulin antidiabetic medication use, initiation and adherence in the paediatric population. Methods: We conducted a descriptive study of non‐insulin antidiabetic medication use in children and adolescents (aged 10‐18 years) using real‐world data from a nationwide US commercial claims database (January 2004–September 2019). Trends in the prevalence of non‐insulin antidiabetic medication use overall and by class were evaluated. Among new users of non‐insulin antidiabetic agents, medication adherence was examined using group‐based trajectory models. Results: In a cohort of more than 1 million paediatric patients, the prevalence of any non‐insulin antidiabetic medication use was 75.7 per 100 000 patients in 2004 and more than doubled to 162.0 per 100 000 in 2019. Biguanides (metformin) was by far the most widely used medication class. The use of newer classes was low (<10 per 100 000), but there was an uptake in the use of glucagon‐like peptide‐1 receptor agonists after liraglutide received paediatric approval in 2019. Medication adherence was poor during the 18 months after treatment initiation: 79.6% of initiators experienced an early treatment interruption (median time to interruption: 90 days among metformin monotherapy initiators) and 21% of initiators did not return for a prescription refill after the first month. Conclusions: There was a substantial increase in non‐insulin antidiabetic medication use among commercially insured paediatric patients from 2004 to 2019. Nearly all patients were treated with metformin, while the use of newer agents remained low. Despite the increase in medication use, short treatment episodes were observed, even among patients with a diagnosis of type 2 diabetes, raising concern over poor adherence.
- 29 October 2020
- Published by Wiley in Diabetes, Obesity and Metabolism
- Vol. 23 (2) , 444-454
Antidiabetic medication use in commercially insured children and adolescents in the United States from 2004 to 2019
Aim: To describe the patterns of non‐insulin antidiabetic medication use, initiation and adherence in the paediatric population. Methods: We conducted a descriptive study of non‐insulin antidiabetic medication use in children and adolescents (aged 10‐18 years) using real‐world data from a nationwide US commercial claims database (January 2004–September 2019). Trends in the prevalence of non‐insulin antidiabetic medication use overall and by class were evaluated. Among new users of non‐insulin antidiabetic agents, medication adherence was examined using group‐based trajectory models. Results: In a cohort of more than 1 million paediatric patients, the prevalence of any non‐insulin antidiabetic medication use was 75.7 per 100 000 patients in 2004 and more than doubled to 162.0 per 100 000 in 2019. Biguanides (metformin) was by far the most widely used medication class. The use of newer classes was low (<10 per 100 000), but there was an uptake in the use of glucagon‐like peptide‐1 receptor agonists after liraglutide received paediatric approval in 2019. Medication adherence was poor during the 18 months after treatment initiation: 79.6% of initiators experienced an early treatment interruption (median time to interruption: 90 days among metformin monotherapy initiators) and 21% of initiators did not return for a prescription refill after the first month. Conclusions: There was a substantial increase in non‐insulin antidiabetic medication use among commercially insured paediatric patients from 2004 to 2019. Nearly all patients were treated with metformin, while the use of newer agents remained low. Despite the increase in medication use, short treatment episodes were observed, even among patients with a diagnosis of type 2 diabetes, raising concern over poor adherence.
- Open Access
- 20 October 2020
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 15 (10) , e0241083
Predicting overdose among individuals prescribed opioids using routinely collected healthcare utilization data
With increasing rates of opioid overdoses in the US, a surveillance tool to identify high-risk patients may help facilitate early intervention. To develop an algorithm to predict overdose using routinely-collected healthcare databases. Within a US commercial claims database (2011–2015), patients with ≥1 opioid prescription were identified. Patients were randomly allocated into the training (50%), validation (25%), or test set (25%). For each month of follow-up, pooled logistic regression was used to predict the odds of incident overdose in the next month based on patient history from the preceding 3–6 months (time-updated), using elastic net for variable selection. As secondary analyses, we explored whether using simpler models (few predictors, baseline only) or different analytic methods (random forest, traditional regression) influenced performance. We identified 5,293,880 individuals prescribed opioids; 2,682 patients (0.05%) had an overdose during follow-up (mean: 17.1 months). On average, patients who overdosed were younger and had more diagnoses and prescriptions. The elastic net model achieved good performance (c-statistic 0.887, 95% CI 0.872–0.902; sensitivity 80.2, specificity 80.1, PPV 0.21, NPV 99.9 at optimal cutpoint). It outperformed simpler models based on few predictors (c-statistic 0.825, 95% CI 0.808–0.843) and baseline predictors only (c-statistic 0.806, 95% CI 0.787–0.26). Different analytic techniques did not substantially influence performance. In the final algorithm based on elastic net, the strongest predictors were age 18–25 years (OR: 2.21), prior suicide attempt (OR: 3.68), opioid dependence (OR: 3.14). We demonstrate that sophisticated algorithms using healthcare databases can be predictive of overdose, creating opportunities for active monitoring and early intervention.
- 20 October 2020
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 15 (10) , e0241083
Open AccessPredicting overdose among individuals prescribed opioids using routinely collected healthcare utilization data
With increasing rates of opioid overdoses in the US, a surveillance tool to identify high-risk patients may help facilitate early intervention. To develop an algorithm to predict overdose using routinely-collected healthcare databases. Within a US commercial claims database (2011–2015), patients with ≥1 opioid prescription were identified. Patients were randomly allocated into the training (50%), validation (25%), or test set (25%). For each month of follow-up, pooled logistic regression was used to predict the odds of incident overdose in the next month based on patient history from the preceding 3–6 months (time-updated), using elastic net for variable selection. As secondary analyses, we explored whether using simpler models (few predictors, baseline only) or different analytic methods (random forest, traditional regression) influenced performance. We identified 5,293,880 individuals prescribed opioids; 2,682 patients (0.05%) had an overdose during follow-up (mean: 17.1 months). On average, patients who overdosed were younger and had more diagnoses and prescriptions. The elastic net model achieved good performance (c-statistic 0.887, 95% CI 0.872–0.902; sensitivity 80.2, specificity 80.1, PPV 0.21, NPV 99.9 at optimal cutpoint). It outperformed simpler models based on few predictors (c-statistic 0.825, 95% CI 0.808–0.843) and baseline predictors only (c-statistic 0.806, 95% CI 0.787–0.26). Different analytic techniques did not substantially influence performance. In the final algorithm based on elastic net, the strongest predictors were age 18–25 years (OR: 2.21), prior suicide attempt (OR: 3.68), opioid dependence (OR: 3.14). We demonstrate that sophisticated algorithms using healthcare databases can be predictive of overdose, creating opportunities for active monitoring and early intervention.