HLHemin Lee

Hemin Lee

*This is an algorithmically generated author record based on publicly available information.

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


Hemin Lee has a total of 124 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Validation of serostatus of rheumatoid arthritis using ICD‐10 codes in administrative claims data

    by Hemin Lee, Jeffrey A. Sparks, Su Been Lee, Kazuki Yoshida, Joan E. Landon, Seoyoung C. Kim

    Purpose: To determine the accuracy of International Classification of Diseases‐ Tenth Revision (ICD‐10) diagnosis codes for rheumatoid arthritis (RA) serostatus using a U.S. claims database (Optum Clinformatics Data Mart, Optum) and to compare the results to a previous validation study performed in IBM Marketscan Research Database (sensitivity 73%, positive predictive value, PPV, 84%).Methods: In Optum (01/01/2016–03/31/2020) linked with laboratory results, we selected RA patients based on ≥2 ICD‐10 diagnosis codes for RA (M05 or M06) and at least one dispensing of RA treatments. We included individuals with at least one laboratory result for rheumatoid factor (RF) or anti‐cyclic citrullinated peptide (CCP) performed 365 days prior to and including the cohort entry date. An individual was “seropositive” if at least one of the 2 diagnosis codes used to define RA status was M05. “Seronegative” patients were required to have only M06. Secondary analyses were performed using subsets of M05 and M06 diagnosis codes. We calculated the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and kappa of M05 and M06 against the prespecified reference standard laboratory data.Results: We identified 14 490 adult RA patients who had at least 1 RF or anti‐CCP result. The number of patients identified for each reference standard definition ranged from 3315 (reference standard definition: high + anti‐CCP) to 13 636 (any + RF). PPV for seropositive RA, M05, was 77.1%. The PPV of M06 for seronegative RA was 61.6%. When we applied more restricted definitions of M05 and M06, the PPV for seropositive RA increased to 79.2%. The PPV for seronegative RA also notably increased to 89.5%.Conclusion: ICD‐10 codes (M05 and M06) can help identify RA serostatus in claims data, but their limitations should be acknowledged. The PPVs for seropositive and seronegative RA found in the Optum database were lower than those found in MarketScan, perhaps related to database variability or differing patient characteristics and clinical practice. When more restricted definitions of M05 and M06 were used, the PPVs for seropositive and seronegative RA improved to 79.2% and 89.5%, respectively.


  • Tofacitinib and Risk of Malignancy: Results From the Safety of Tofacitinib in Routine Care Patients With Rheumatoid Arthritis (STAR‐RA) Study

    by Farzin Khosrow‐Khavar, Rishi J. Desai, Hemin Lee, Su Been Lee, Seoyoung C. Kim

    Objectives: Results of the ORAL Surveillance safety trial have indicated that there is an increased risk for the development of malignancies with tofacitinib therapy when compared to treatment with tumor necrosis factor inhibitors (TNFi). This study was undertaken to further examine this safety concern in rheumatoid arthritis (RA) patients in a real‐world setting.Methods: Using US insurance claims data from Optum Clinformatics (2012–2020), IBM MarketScan Research Databases (2012–2018), and Medicare (parts A, B, and D, 2012–2017), we created 2 cohorts of RA patients who had initiated treatment with tofacitinib or TNFi. The first cohort, designated the real‐world evidence (RWE) cohort, included RA patients from routine care. For the second cohort, designated the randomized controlled trial (RCT)–duplicate cohort, we emulated the inclusion and exclusion criteria that were applied in the ORAL Surveillance trial of tofacitinib, which allowed us to assess the comparability of our results with the results of that trial. Cox proportional hazards models with propensity score fine‐stratification weighting were used to estimate hazard ratios (HRs) and 95% confidence intervals (95% CIs) for the risk of any malignancy (excluding nonmelanoma skin cancer). Database‐specific estimates were meta‐analyzed using fixed‐effects models with inverse‐variance weighting.Results: The RWE cohort consisted of 83,295 patients, including 10,504 patients (12.6%) who received treatment with tofacitinib. The pooled weighted HR for the primary outcome of any malignancy associated with tofacitinib treatment compared to any malignancy associated with TNFi therapy was 1.01 (95% CI 0.83, 1.22) in the RWE cohort and 1.17 (95% CI 0.85, 1.62) in the RCT‐duplicate cohort (compared to the ORAL Surveillance trial HR of 1.48 [95% CI 1.04, 2.09]).Conclusion: We did not find evidence of an increased risk of malignancy development with tofacitinib therapy, in comparison with TNFi therapy, in RA patients treated in a real‐world setting. However, our results cannot rule out the possibility of an increase in risk that may accrue with a longer duration of treatment with tofacitinib.


  • Top Cited Papers

    Tofacitinib and risk of cardiovascular outcomes: results from the Safety of TofAcitinib in Routine care patients with Rheumatoid Arthritis (STAR-RA) study

    by Farzin Khosrow-Khavar, Seoyoung C Kim, Hemin Lee, Su Been Lee, Rishi J Desai

    Objectives Recent results from ‘ORAL Surveillance’ trial have raised concerns regarding the cardiovascular safety of tofacitinib in patients with rheumatoid arthritis (RA). We further examined this safety concern in the real-world setting. Methods We created two cohorts of patients with RA initiating treatment with tofacitinib or tumour necrosis factor inhibitors (TNFI) using deidentified data from Optum Clinformatics (2012–2020), IBM MarketScan (2012–2018) and Medicare (parts A, B and D, 2012–2017) claims databases: (1) A ‘real-world evidence (RWE) cohort’ consisting of routine care patients and (2) A ‘randomised controlled trial (RCT)-duplicate cohort’ mimicking inclusion and exclusion criteria of the ORAL surveillance trial to calibrate results against the trial findings. Cox proportional hazards models with propensity score fine stratification weighting were used to estimate HR and 95% CIs for composite outcome of myocardial infarction and stroke and accounting for 76 potential confounders. Database-specific effect estimates were pooled using fixed effects models with inverse-variance weighting. Results In the RWE cohort, 102 263 patients were identified of whom 12 852 (12.6%) initiated tofacitinib. The pooled weighted HR (95% CI) comparing tofacitinib with TNFI was 1.01 (0.83 to 1.23) in RWE cohort and 1.24 (0.90 to 1.69) in RCT-duplicate cohort which aligned closely with ORAL-surveillance results (HR: 1.33, 95% CI 0.91 to 1.94). Conclusions We did not find evidence for an increased risk of cardiovascular outcomes with tofacitinib in patients with RA treated in the real-world setting; however, tofacitinib was associated with an increased risk of cardiovascular outcomes, although statistically non-significant, in patients with RA with cardiovascular risk factors. Trial registration number NCT04772248.


  • Open Access

    Patient characteristics associated with use of TNF vs interleukin inhibitors as first-line biologic treatment for psoriatic arthritis

    by Yinzhu Jin, Sarah K Chen, Hemin Lee, Joan E Landon, Joseph F Merola, Seoyoung C Kim

    BACKGROUND: Previous studies have examined treatment patterns among patients who use tumor necrosis factor (TNF) inhibitors for psoriatic arthritis (PsA). However, little data exist for a comparison between the TNF inhibitor treatment pattern and that of newly available biologics such as interleukin (IL)-12/23 or 17 inhibitors in the United States. OBJECTIVES: To (a) examine patient characteristics and their association with initiation of TNF inhibitors vs IL-12/23 or 17 inhibitors among PsA patients and (2) compare treatment persistence of PsA patients who initiated TNF inhibitors vs IL-12/23 or 17 inhibitors as first-line biologic treatment in a real-world setting in the United States. METHODS: Using claims data from MarketScan (2013-2017), we identified a cohort of PsA patients who initiated TNF inhibitors or IL-12/23 or 17 inhibitors. The primary outcome was treatment persistence, defined as continuous use of the index drug at 1 year, regardless of refill gaps. The secondary outcome was treatment persistence with high adherence at 1 year (ie, refill gaps ≤ 30 days). Multivariable logistic regression was used to assess the association between patient characteristics and treatment initiation and persistent use of TNF inhibitors vs IL-12/23 or 17 inhibitors. RESULTS: We identified 3,180 TNF inhibitor initiators and 214 IL-12/23 or 17 inhibitor initiators. Initiators of IL-12/23 or 17 inhibitors had more comorbidities than TNF inhibitor initiators. The proportion of patients with treatment persistence was 53.0% in TNF inhibitor initiators and 53.7% in IL-12/23 or 17 inhibitor initiators; 37.1% of TNF inhibitor users and 24.8% of IL-12/23 or 17 inhibitor users were treatment persistent with high adherence. There was no difference in 1-year treatment persistence between the 2 groups after adjusting for baseline characteristics (adjusted odds ratio [aOR] for TNF inhibitors vs IL-12/23 or 17 inhibitors: 0.86, 95% CI = 0.63-1.15). However, use of TNF inhibitors was associated with a greater treatment persistence with high adherence compared with use of IL-12/23 or 17 inhibitors (aOR = 1.61, 95% CI = 1.15-2.26). CONCLUSIONS: PsA patients who initiated an IL 12/23 or 17 inhibitor had a greater comorbidity burden compared with those who initiated TNF inhibitors. Treatment persistence was similar between the 2 groups, whereas medication adherence was higher with TNF inhibitors than with IL 12/23 or 17 inhibitors during the first year of treatment. DISCLOSURES: This study was funded by an investigator-initiated research grant from Pfizer, Inc (grant number: WI235988). The content is solely the responsibility of the authors. The sponsor was given the opportunity to make nonbinding comments on a draft of the manuscript. Publication of the manuscript was not contingent on approval by the sponsor. Kim has received research grants to the Brigham and Women’s Hospital from Roche, AbbVie, and Bristol-Myers Squibb for unrelated topics. Merola is a consultant and/or investigator for BMS, AbbVie, Dermavant, Lilly, Novartis, Janssen, UCB, Sun Pharma, and Pfizer. Jin, Chen, Lee, and Landon have nothing to disclose.


  • Risk of Hospitalization for Serious Infection After Initiation of Ustekinumab or Other Biologics in Patients With Psoriasis or Psoriatic Arthritis

    by Yinzhu Jin, Hemin Lee, Moa P. Lee, Joan E. Landon, Joseph F. Merola, Rishi J. Desai, Seoyoung C. Kim

    Objective: To compare the risk of serious infections requiring hospitalization in patients with psoriasis (PsO) or psoriatic arthritis (PsA) initiating ustekinumab versus other biologics or apremilast.Methods: In this multi‐database cohort study, we identified patients with PsO/PsA who initiated therapy with adalimumab, apremilast, certolizumab, etanercept, golimumab, ixekizumab, secukinumab, or ustekinumab between 2009 and 2018. The primary outcome measure was hospitalizations due to serious infections, which included bacterial, viral, or opportunistic infections. We estimated hazard ratios (HRs) comparing each study drug to ustekinumab after applying propensity score fine stratification weights for confounding control in each database. Database‐specific weighted HRs were combined by meta‐analysis.Results: We identified 123,383 patients with PsO/PsA who initiated one of the study drugs. During a total of 117,744 person‐years of follow‐up, 1,514 serious infections occurred with a crude incidence of 1.29 per 100 person‐years. After propensity score fine stratification and weighting, the incidence rates of serious infection among ustekinumab initiators ranged from 0.59 to 0.95 per 100 person‐years. Compared with ustekinumab, the combined weighted HRs (95% confidence interval [95% CI]) for serious infections were 1.66 (95% CI 1.34–2.06) for adalimumab, 1.42 (95% CI 1.02–1.96) for apremilast, 1.09 (95% CI 0.68–1.75) for certolizumab, 1.39 (95% CI 1.01–1.90) for etanercept, 1.74 (95% CI 1.00–3.03) for golimumab, 2.92 (95% CI 1.80–4.72) for infliximab, 2.98 (95% CI 1.20–7.41) for ixekizumab, and 1.84 (95% CI 1.24–2.72) for secukinumab.Conclusion: Other biologics and apremilast were associated with a 1.4‐ to 3‐times higher risk of hospitalization for serious infections in PsO/PsA patients when compared to ustekinumab; this finding should be considered in the safety profile of these therapies when selecting appropriate treatment regimens in patients with PsO/PsA.


  • Validation of claims‐based algorithms to identify patients with psoriasis

    by Hemin Lee, Mengdong He, Soo‐Kyung Cho, Lily Bessette, Angela Y. Tong, Joseph F. Merola, Lani R. Wegrzyn, Ryan D. Kilpatrick, Seoyoung C. Kim

    Purpose: Accurately identifying patients with psoriasis (PsO) is crucial for generating real‐world evidence on PsO disease course and treatment utilization.Methods: We developed nine claims‐based algorithms for PsO using a combination of the International Classification of Diseases (ICD)‐9 codes, specialist visit, and medication dispensing using Medicare linked to electronic health records data (2013–2014) in two healthcare provider networks in Boston, Massachusetts. We calculated positive predictive value (PPV) and 95% confidence interval (CI) for each algorithm using the treating physician's diagnosis of PsO via chart review as the gold standard. Among the confirmed PsO cases, we assessed their PsO disease activity.Results: The nine claims‐based algorithms identified 990 unique patient records. Of those, 918 (92.7%) with adequate information were reviewed. The PPV of the algorithms ranged from 65.1 to 82.9%. An algorithm defined as ≥1 ICD‐9 diagnosis code for PsO and ≥1 prescription claim for topical vitamin D agents showed the highest PPV (82.9%). The PPV of the algorithm requiring ≥2 ICD‐9 diagnosis codes and ≥1 prescription claim for PsO treatment excluding topical steroids was 81.1% but higher (82.5%) when ≥1 diagnosis was from a dermatologist. Among 411 PsO patients with adequate information on PsO disease activity in EHRs, 1.5–5.8% had no disease activity, 31.3–36.8% mild, and 26.9–35.1% moderate‐to‐severe across the algorithms.Conclusions: Claims‐based algorithms based on a combination of PsO diagnosis codes and dispensing for PsO‐specific treatments had a moderate‐to‐high PPV. These algorithms can serve as a useful tool to identify patients with PsO in future real‐world data pharmacoepidemiologic studies.


  • Open Access

    Identification of Acute Giant Cell Arteritis in Real‐World Data Using Administrative Claims‐Based Algorithms

    by Hemin Lee, Sara K. Tedeschi, Sarah K. Chen, Paul A. Monach, Erin Kim, Jun Liu, Attila Pethoe‐Schramm, Vincent Yau, Seoyoung C. Kim

    Objective: The objective of this study was to validate claims‐based algorithms for identifying acute giant cell arteritis (GCA) that will help generate real‐world evidence on comparative effectiveness research and epidemiologic studies. Among patients identified by the GCA algorithm, we further investigated whether GCA flares could be detected by using claims data.Methods: We developed five claims‐based algorithms based on a combination of International Classification of Diseases, Ninth Revision (ICD‐9) diagnosis codes, specialist visits, and dispensed medications using Medicare Parts A, B, and D linked to electronic medical records (2006‐2014). Acute cases of GCA were determined by chart review using the treating physician’s diagnosis of GCA as the gold standard. Among the patients identified with acute GCA, we assessed if a GCA flare occurred during the year after initial diagnosis.Results: The number of patients identified by each algorithm ranged from 220 to 896. Positive predictive values (PPVs) of the algorithms ranged from 60.7% to 84.8%. Requirement for disease‐specific workups, multiple diagnosis codes, or specialist visits improved the PPVs. The highest PPV (84.8%) was noted in an algorithm that required two or more diagnosis codes of GCA from inpatient, emergency department, or outpatient rheumatology visits plus a prednisone‐equivalent dose greater than or equal to 40 mg/day occurring 14 days before or after the second ICD‐9 diagnosis date, with the cumulative days’ supply greater than or equal to 14 days. Among patients identified as having GCA, 18.2% of patients had definite evidence of a flare and 25% had a potential flare.Conclusion: A claims‐based algorithm requiring two or more ICD‐9 diagnosis codes from inpatient, emergency department, or outpatient rheumatology visits and high‐dose glucocorticoid dispensing can be a useful tool to identify acute GCA cases in large administrative claims databases.


  • Validation of claims‐based algorithms for psoriatic arthritis

    by Hemin Lee, Julia A. Ford, Yinzhu Jin, Soo‐Kyung Cho, Adrian J. Santiago Ortiz, Angela Y. Tong, Seoyoung C. Kim

    Purpose: An increasing number of new medications are being developed and approved for psoriatic arthritis (PsA). To generate real‐world evidence on comparative safety and effectiveness of these drugs, a claims‐based algorithm that can accurately identify PsA is greatly needed.Methods: To identify patients with PsA, we developed seven claims‐based algorithms based on a combination of diagnosis codes and medication dispensing using the claims data from Medicare parts A/B/D linked to electronic medical records (2012‐2014). Two physicians independently conducted a chart review using the treating physician's diagnosis of PsA as the gold standard. We calculated the positive predictive value (PPV) and 95% confidence intervals of each algorithm.Results: Of the total 2157 records identified by the seven algorithms, 45% of the records had relevant clinical data to determine the presence of PsA. The PPV of the algorithms ranged from 75.2% (algorithm 1: ≥2 diagnosis codes for PsA and ≥1 diagnosis code for psoriasis) to 88.6% (algorithm 7: ≥2 diagnosis codes for PsA with ≥1 code by rheumatologist and ≥1 dispensing for PsA medication). Having ≥2 diagnosis codes and ≥1 dispensing for PsA medications (algorithm 6) also had PPV of 82.4%.Conclusions: All seven claims‐based algorithms demonstrated a moderately high PPV of 75% to 89% in identifying PsA. The use of ≥2 diagnosis codes plus ≥1 prescription claim for PsA appears to be a valid and efficient tool in identifying PsA patients in the claims data, while broader algorithms based on diagnoses without a prescription claim also have reasonably good PPVs.


  • Risk of Diabetes Mellitus in Patients with Juvenile Idiopathic Arthritis

    by Hemin Lee, Yinzhu Jin, Jun Liu, Ezra M. Cohen, Sarah K. Chen, Seoyoung C. Kim

    Objective.: To examine the risk of type 1 diabetes (T1D) associated with juvenile idiopathic arthritis (JIA).Methods.: Using the IBM MarketScan database, we compared JIA patients with asthma patients and healthy individuals for the risk of incident T1D.Results.: We included patients with 15,210 JIA, 76,050 patients with asthma, and 76,050 healthy individuals matched 1:5 on age, sex, and index date. After adjustment for confounders, the multivariable HR of T1D associated with JIA was 1.48 (95% CI 0.86–2.56) versus asthma and 1.81 (95% CI 1.03–3.17) versus healthy individuals.Conclusion.: JIA appears to be associated with an increased risk of T1D compared to patients with asthma and healthy children.


  • Open Access

    Front & Back Matter

    by Sven-Erik Dahlén, Dae Suk Kim, Dong Hyun Kim, Hemin Lee, Hyunjoong Jee, Young Lee, Min-Youl Chang, Taek-Jong Kwak, Chul-Hong Kim, Young-Ah Shin, et al.