Journal: Frontiers in Digital Health

Article Title: Guideline-based strategies to identify severe cytokine release syndrome in COVID-19 and cancer immunotherapy using large-scale electronic health records

doi: 10.3389/fdgth.2025.1625889

Figure Lengend Snippet: Definition of the cohorts of interests, inclusion and exclusion criteria. From the Optum® COVID-19 data, subjects with active COVID-19 or who received TCE treatment (blinatumomab) were isolated. The first COVID-19 infection or TCE treatment was considered a triggering event. Subjects with missing information on gender or age were excluded. Subjects with pre-existing comorbidities that share traits with CRS at least 7 days before the triggering event were excluded. A significant number of COVID-19 patients had no reported data 30 days around the triggering event and were excluded. Patients diagnosed with Sepsis within 30 days after onset were also excluded. Three cohorts of interest were used for subsequent CRS case identification: ‘COVID-19 adult cohort’, ‘TCE adult cohort’ and ‘TCE pediatric cohort’.

Article Snippet: Using the Optum® de-identified COVID-19 EHR dataset, we isolated 2.5 million patients with active COVID-19 and 171 individuals treated with the T-cell Engager (TCE) blinatumomab.

Techniques: Isolation, Infection

Journal: Frontiers in Digital Health

Article Title: Guideline-based strategies to identify severe cytokine release syndrome in COVID-19 and cancer immunotherapy using large-scale electronic health records

doi: 10.3389/fdgth.2025.1625889

Figure Lengend Snippet: CRS grading algorithm (decision tree) on EHR datasets following the ASTCT grading guideline. (A) The latest and commonly used ASTCT CRS grading , keeping identical wording as in the original publication. (B) From the list of reported patient features within a time window of 30 days after the triggering event (TCE administration or COVID-19 diagnosis), patients are first graded into ‘grade N+’ and then separated into definite grades: grade 1+ includes patients with fever ≥38 °C (‘strict’ definition) or those with potentially mitigated fever by corticosteroid or cytokine blocker (anti-IL1 or anti-IL6) therapy (‘mitigated’ definition). Grade 2+ to 4+ are defined based on the grade-defining interventions: grade 4+: CPAP or invasive ventilation or use of multiple vasopressors; grade 3+ (one vasopressor or non-CPAP ventilation); and grade 2+ (evidence for hypoxia or hypotension). Notably, we assumed that the use of vasopressors or ventilation indicated hypoxia or hypotension, even if the reported cardiovascular or respiratory parameters were within the reference range. Patients without grade 2+ were classified as “definite” grades if lab values were in range, “probable” grades if hypoxia or hypotension were not measured, or as non-classifiable. We proposed a definition for CRS grade 2+ (or grade 3+) positive and negative (i.e., control) cohorts. SaO2 = arterial oxygen saturation, SBP = systolic blood pressure SpO2 = peripheral oxygen saturation, PaO2 = partial arterial oxygen pressure, PvO2 = venous oxygen tension, and DBP = diastolic blood pressure.

Article Snippet: Using the Optum® de-identified COVID-19 EHR dataset, we isolated 2.5 million patients with active COVID-19 and 171 individuals treated with the T-cell Engager (TCE) blinatumomab.

Techniques: Biomarker Discovery, Control

Journal: Frontiers in Digital Health

Article Title: Guideline-based strategies to identify severe cytokine release syndrome in COVID-19 and cancer immunotherapy using large-scale electronic health records

doi: 10.3389/fdgth.2025.1625889

Figure Lengend Snippet: Identified patients following different implementations of the consensus ASTCT CRS grading on the COVID-19 and TCE cohorts. (A) Grading into N + groups, from grade 1+ to grade 4+, depending on the implementations: strict, extended, extended+mitigations. (B) Breakdown of the cohorts by grades based on the ‘extended+mitigations’ implementation, including details of definite and probable grading, as well as ambiguous patients who show symptoms of a higher grade but do not qualify for grade 1, and deceased patients with (probable grade 5) or without (not gradable) grade 2+ features. Notably, sepsis patients have been excluded, and the cohort includes patients who could be CRS positive or negative. Therefore, the percentages are calculated within the “usable cohort” of patients who did not experience sepsis, but these numbers should be adjusted to include the entire cohort, including sepsis cases, if prevalence needs to be determined.

Article Snippet: Using the Optum® de-identified COVID-19 EHR dataset, we isolated 2.5 million patients with active COVID-19 and 171 individuals treated with the T-cell Engager (TCE) blinatumomab.

Techniques:

Journal: Alzheimer's & Dementia

Article Title: Real‐world observations of GLP‐1 receptor agonists and SGLT‐2 inhibitors as potential treatments for Alzheimer's disease

doi: 10.1002/alz.70639

Figure Lengend Snippet: Associations with Alzheimer's disease: glucagon‐like peptide‐1 receptor agonists compared to dipeptidyl peptidase‐4 inhibitors (referent); (A) Optum Clinformatics ® data analyses; (B) Northwestern University Electronic Health Record data analyses (adjustment variables: age, sex, race, weight, comorbidity, and antidiabetic drug exposure). CI, confidence interval; EHR, electronic health record.

Article Snippet: We used Optum's de‐identified Clinformatics ® Data Mart database (Optum Clinformatics ® [2007–2021]) and electronic health record (EHR) data from the Northwestern Medicine Enterprise Data Warehouse (NMEDW [2005–2023]).

Techniques:

Journal: Alzheimer's & Dementia

Article Title: Real‐world observations of GLP‐1 receptor agonists and SGLT‐2 inhibitors as potential treatments for Alzheimer's disease

doi: 10.1002/alz.70639

Figure Lengend Snippet: Associations with Alzheimer's disease in individuals with age between 60 and 74 years, and PS‐matched analyses: glucagon‐like peptide‐1 receptor agonists compared to dipeptidyl peptidase‐4 inhibitors (referent); (A) Optum Clinformatics ® data analyses; (B) Northwestern University Electronic Health Record data analyses (adjustment variables: age, sex, race, weight, comorbidity, and antidiabetic drug exposure). CI, confidence interval; EHR, electronic health record; PS, propensity score.

Article Snippet: We used Optum's de‐identified Clinformatics ® Data Mart database (Optum Clinformatics ® [2007–2021]) and electronic health record (EHR) data from the Northwestern Medicine Enterprise Data Warehouse (NMEDW [2005–2023]).

Techniques:

Journal: Alzheimer's & Dementia

Article Title: Real‐world observations of GLP‐1 receptor agonists and SGLT‐2 inhibitors as potential treatments for Alzheimer's disease

doi: 10.1002/alz.70639

Figure Lengend Snippet: Associations with Alzheimer's disease: sodium‐glucose cotransporter‐2 inhibitors compared to dipeptidyl peptidase‐4 inhibitors (referent); (A) Optum Clinformatics ® data analyses; (B) Northwestern University Electronic Health Record data analyses (adjustment variables: age, sex, race, weight, comorbidity, and antidiabetic drug exposure). CI, confidence interval; EHR, electronic health record.

Article Snippet: We used Optum's de‐identified Clinformatics ® Data Mart database (Optum Clinformatics ® [2007–2021]) and electronic health record (EHR) data from the Northwestern Medicine Enterprise Data Warehouse (NMEDW [2005–2023]).

Techniques:

Journal: Alzheimer's & Dementia

Article Title: Real‐world observations of GLP‐1 receptor agonists and SGLT‐2 inhibitors as potential treatments for Alzheimer's disease

doi: 10.1002/alz.70639

Figure Lengend Snippet: Associations with Alzheimer's disease in individuals with age between 60 and 74 years, and PS‐matched analyses: sodium‐glucose cotransporter‐2 inhibitors compared to dipeptidyl peptidase‐4 inhibitors (referent); (A) Optum Clinformatics ® data analyses; (B) Northwestern University Electronic Health Record data analyses (adjustment variables: age, sex, race, weight, comorbidity, and antidiabetic drug exposure). CI, confidence interval; EHR, electronic health record; PS, propensity score.

Article Snippet: We used Optum's de‐identified Clinformatics ® Data Mart database (Optum Clinformatics ® [2007–2021]) and electronic health record (EHR) data from the Northwestern Medicine Enterprise Data Warehouse (NMEDW [2005–2023]).

Techniques:

Journal: BMC Medical Research Methodology

Article Title: The necessity of validity diagnostics when drawing causal inferences from observational data: lessons from a multi-database evaluation of the risk of non-infectious uveitis among patients exposed to Remicade ®

doi: 10.1186/s12874-024-02428-7

Figure Lengend Snippet: Empirical equipoise, covariate balance, empirical calibration validity diagnostics and representativeness for IBD primary analysis Key: Amb EMR = IQVIA Ambulatory Electronic Medical Records; ASMD = absolute standardized mean difference; CCAE = Merative™ MarketScan ® Commercial Database; CI = Confidence Interval; Clinformatics ® = Optum ® De-Identified Clinformatics ® Data Mart Database; EASE = expected absolute systematic error; HR = Hazard ratio; IBD = irritable bowel diseases (Crohn’s disease or ulcerative colitis); IBD comparator = golimumab, certolizumab pegol, ustekinumab, or vedolizumab; Optum ® EHR = Optum ® De-Identified Electronic Health Record; Pharmetrics = IQVIA Adjudicated Health Plan Claims Data; Remicade ® (m) = Remicade ® exposure; Target covariate prevalence = prevalence of baseline covariates in the initial Remicade ® exposure cohort before study design restrictions were applied; Analytic covariate prevalence = prevalence of baseline covariates in Remicade ® exposure cohort after study design restrictions were applied (i.e., PS matching)

Article Snippet: The EHR databases included Optum ® de-identified Electronic Health Record Dataset (Optum ® EHR) and IQVIA Ambulatory EMR (Amb EMR).

Techniques:

Journal: BMC Medical Research Methodology

Article Title: The necessity of validity diagnostics when drawing causal inferences from observational data: lessons from a multi-database evaluation of the risk of non-infectious uveitis among patients exposed to Remicade ®

doi: 10.1186/s12874-024-02428-7

Figure Lengend Snippet: Attrition diagrams for inflammatory bowel diseases (IBD); patient attrition counts and proportions after sequential design choices applied Key: Amb EMR = IQVIA Ambulatory Electronic Medical Records; ASMD = absolute standardized mean difference; CCAE = Merative™ MarketScan ® Commercial Database; CI = Confidence Interval; Clinformatics ® = Optum ® De-Identified Clinformatics ® Data Mart Database; EASE = expected absolute systematic error; HR = Hazard ratio; Optum ® EHR = Optum ® De-Identified Electronic Health Record; Pharmetrics = IQVIA Adjudicated Health Plan Claims Data; RA = rheumatoid arthritis; RA comparator = certolizumab pegol or tocilizumab; Remicade ® (m) = Remicade ® exposure with concurrent methotrexate; Target covariate prevalence = prevalence of baseline covariates in the initial Remicade ® exposure cohort before study design restrictions were applied; Analytic covariate prevalence = prevalence of baseline covariates in Remicade ® exposure cohort after study design restrictions were applied (i.e., PS matching)

Article Snippet: The EHR databases included Optum ® de-identified Electronic Health Record Dataset (Optum ® EHR) and IQVIA Ambulatory EMR (Amb EMR).

Techniques:

Journal: BMC Medical Research Methodology

Article Title: The necessity of validity diagnostics when drawing causal inferences from observational data: lessons from a multi-database evaluation of the risk of non-infectious uveitis among patients exposed to Remicade ®

doi: 10.1186/s12874-024-02428-7

Figure Lengend Snippet: Empirical equipoise, covariate balance, empirical calibration validity diagnostics and representativeness for RA primary analysis Key – Target: patients with inflammatory bowel diseases newly exposed to Remicade ® , Comparator: patients with inflammatory bowel diseases newly exposed to [golimumab, certolizumab pegol, ustekinumab, or vedolizumab], CCAE: Merative™ MarketScan ® Commercial Database, Optum ® EHR = Optum ® De-Identified Electronic Health Record; Pharmetrics = IQVIA Adjudicated Health Plan Claims Data

Article Snippet: The EHR databases included Optum ® de-identified Electronic Health Record Dataset (Optum ® EHR) and IQVIA Ambulatory EMR (Amb EMR).

Techniques:

Journal: BMC Medical Research Methodology

Article Title: The necessity of validity diagnostics when drawing causal inferences from observational data: lessons from a multi-database evaluation of the risk of non-infectious uveitis among patients exposed to Remicade ®

doi: 10.1186/s12874-024-02428-7

Figure Lengend Snippet: Risk of non-infectious uveitis (NIU) among patients with inflammatory bowel diseases (IBD) Key – PS: propensity score, OT: on-treatment, ITT: intention-to-treat, T: Remicade ® new users with IBD, C: golimumab, certolizumab pegol, ustekinumab, or vedolizumab new users with IBD, IR: incidence rate, PYs: person-years, CCAE: Merative™ MarketScan ® Commercial Database, Optum ® EHR: Optum ® De-Identified Electronic Health Record, Pharmetrics: IQVIA Adjudicated Health Plan Claims Data

Article Snippet: The EHR databases included Optum ® de-identified Electronic Health Record Dataset (Optum ® EHR) and IQVIA Ambulatory EMR (Amb EMR).

Techniques: