Optum Database Research Articles

Impact of Hospitalization and Medication Switching on Post-discharge Adherence to Oral Anticoagulants in Patients With Atrial Fibrillation.

Adherence to chronic medications remains poor in practice. There is limited evidence on how hospitalization affects post-discharge adherence to oral anticoagulants (OACs) in individuals with atrial fibrillation. The aim of this study was to examine the impact of hospitalization and medication switching on post-discharge adherence to OACs in the population with atrial fibrillation. A quasi-experimental pre-post observational study was conducted using United States commercial insurance health care claims from the 2009 to 2016 Optum database. Adults with atrial fibrillation taking OACs who had a random hospitalization occurring after the first observed OAC prescription fill and no other admission in the preceding and following 6 months were identified. OAC adherence was estimated by the proportion of days covered within 6 and 12 months before and after hospitalization. Difference-in-difference analysis was employed to compare the pre-hospitalization and post-hospitalization proportion of days covered, stratified by reasons for hospitalization (i.e., bleeding vs non-bleeding-related reasons) and adjusting for imbalanced baseline characteristics between groups. Change in adherence when the OAC was switched at discharge was also examined. The 22,429 individuals who met study criteria were predominantly male (52.4%), white (77.2%), and older age (median 74years). A clinically significant hemorrhage was the reason for 1029 (4.5%) of qualifying hospitalizations. After covariate adjustment, there was a reduction in the proportion of days covered after discharge, regardless of admission diagnosis (p<0.0001). The 6-month difference-in-difference analyses revealed that adherence was incrementally reduced by 3.2% (p=0.0003) in the bleeding group compared with the nonbleeding group, whereas switching from warfarin to a direct oral anticoagulant after hospitalization was associated with a smaller reduction by 3.4% in adherence (p=0.0342) compared with other switchers, regardless of the reason for hospitalization. The 12-month difference-in-difference analyses revealed similar results. Hospitalization is temporally associated with a reduction in adherence to OACs, regardless of reason for hospitalization. More effective strategies are needed to improve OAC adherence, particularly during transition of care.

Association of Pain on Hospital Discharge with the Risk of 30-Day Readmission in Patients with Total Hip and Knee Replacement

BackgroundIt is not clear if there is a risk of 30-day readmissions following total hip and knee arthroplasty in patients reporting high levels of pain at hospital discharge. We examined the relationship between post-surgical pain on the day of discharge and 30-day readmission in patients who received total knee and hip arthroplasty. MethodsRetrospective cohort study was conducted of patients who received total knee (n = 155,284) or hip arthroplasty (n = 89,283) from 2011 to 2018 using electronic health records from the Optum database. Four categories of pain at discharge were created, from none to severe. Multivariate logistic regression models to predict 30-day all-cause readmission were adjusted for patient and clinical characteristics and built separately for knee and hip arthroplasty patients. ResultsMean ages for hip and knee patients were 64.4 (standard deviation 11.3) and 65.7 (standard deviation 9.7) years, respectively. The majority of patients were female (hip: 54.4%; knee: 61.5%). The unadjusted rate of 30-day readmission was 3.54% for hip replacement and 3.66% for knee replacement. In models adjusted for patient and clinical characteristics, for patients with total hip replacement, the odds of 30-day readmission for those with severe pain score at discharge vs those with no pain at discharge were 1.60 (95% confidence interval 1.33-1.92). Similarly, readmission likelihood increased as pain at discharge increased (severe pain vs no pain) for patients with total knee arthroplasty (odds ratio 1.38, 95% confidence interval 1.19-1.59). ConclusionOur findings demonstrated that the pain scores on the day of discharge are associated with 30-day hospital readmission.

PMD34 DEEP JOINT INFECTION COMPARISON BETWEEN SIMULTANEOUS AND STAGED BILATERAL TOTAL KNEE ARTHROPLASTY

Patients with bilateral knee arthritis are candidates for total knee arthroplasty and may decide between simultaneous (SIMTKA) or staged bilateral TKA (STATKA). Inconsistent outcomes have been reported for SIMTKA and STATKA. This study aims to compare the deep joint infection following SIMTKA and STATKA surgery. The electronic medical records (EHR) from the Optum database (2007-2019) were used to identify TKA patients. The EHR data were available for a geographically diverse population of 5.5 million health plan members who have been successfully linked to the Optum EHR database. Short interval STATKA was defined as having staged surgery within 2 weeks after the first unilateral TKA. The postoperative outcome assessed was deep joint infection six months after surgery. Multivariable logistic regression was developed to investigate association between surgery type and deep joint infection with age, Elixhauser comorbidity score and provider volume adjustment. A total of 6,135 SMITKA and 8,203 two-week STATKA were included in the analysis. SIMTKA were younger (mean 63±8 vs 66±10, p<0.001), with more comorbidities (Elixhauser score ≥4: 21% vs 18.6%, p<0.001), suffering more pain (back pain: 12.8% vs 6.5%, p<0.001; knee pain diagnosis: 44.6% vs 16.9%, p<0.001), and more likely to be treated in facilities with higher volume of TKA surgeries (annual volume: 120±95 vs 114±149). In the multivariable logistic regression, patients who had simultaneous TKA were more likely to suffer from deep joint infection six month after surgery compared to patients having 2-week staged TKA (aOR: 3.4, 95% CI: 2.2, 5.5, incidence: 1.14% vs 0.33%). This study demonstrates that simultaneous TKA is associated with increase risk of deep joint infection compared to 2-week staged TKA. Studies are warranted to investigate the best interval of staged TKA and to compare other surgical outcomes among SIMTKA and STATKA.

Hepatocellular carcinoma surveillance, early detection and survival in a privately insured US cohort.

Semiannual hepatocellular carcinoma (HCC) surveillance is recommended in patients with cirrhosis; however, recent studies have raised questions over its utility. We investigated the impact of surveillance on early detection and survival in a nationally representative database. We included patients with cirrhosis and HCC from the Optum database (2001-2015) with >6months of follow-up between cirrhosis and HCC diagnoses. Surveillance adherence was defined as proportion of time covered (PTC), with each 6-month period after abdominal imaging defined as 'covered'. To determine the association between surveillance and mortality, we compared PTC between fatal and non-fatal HCC. Of 1001 patients with cirrhosis and HCC, 256 died with median follow-up 30months. Median PTC by any imaging was greater in early-stage vs late-stage HCC (43.6% vs 37.4%, P=.003) and non-fatal vs fatal HCC (40.8% vs 34.3%, P=.001). In multivariable analyses, each 10% increase in PTC was associated with increased early HCC detection (OR 1.07, 95% CI 1.01-1.12) and decreased mortality (HR 0.95; 95% CI 0.90-1.00). On subgroup analysis, PTC by CT/MRI was associated with early tumour detection and decreased mortality; however, PTC by ultrasound was only associated with early detection but not decreased mortality. These findings were robust across sensitivity analyses. In a US cohort of privately insured HCC patients, PTC by any imaging modality was associated with increased early detection and decreased mortality. Continued evaluation of HCC surveillance strategies and effectiveness is warranted.

Inferring disease severity in rheumatoid arthritis using predictive modeling in administrative claims databases.

BackgroundConfounding by disease severity is an issue in pharmacoepidemiology studies of rheumatoid arthritis (RA), due to channeling of sicker patients to certain therapies. To address the issue of limited clinical data for confounder adjustment, a patient-level prediction model to differentiate between patients prescribed and not prescribed advanced therapies was developed as a surrogate for disease severity, using all available data from a US claims database.MethodsData from adult RA patients were used to build regularized logistic regression models to predict current and future disease severity using a biologic or tofacitinib prescription claim as a surrogate for moderate-to-severe disease. Model discrimination was assessed using the area under the receiver (AUC) operating characteristic curve, tested and trained in Optum Clinformatics® Extended DataMart (Optum) and additionally validated in three external IBM MarketScan® databases. The model was further validated in the Optum database across a range of patient cohorts.ResultsIn the Optum database (n = 68,608), the AUC for discriminating RA patients with a prescription claim for a biologic or tofacitinib versus those without in the 90 days following index diagnosis was 0.80. Model AUCs were 0.77 in IBM CCAE (n = 75,579) and IBM MDCD (n = 7,537) and 0.75 in IBM MDCR (n = 36,090). There was little change in the prediction model assessing discrimination 730 days following index diagnosis (prediction model AUC in Optum was 0.79).ConclusionsA prediction model demonstrated good discrimination across multiple claims databases to identify RA patients with a prescription claim for advanced therapies during different time-at-risk periods as proxy for current and future moderate-to-severe disease. This work provides a robust model-derived risk score that can be used as a potential covariate and proxy measure to adjust for confounding by severity in multivariable models in the RA population. An R package to develop the prediction model and risk score are available in an open source platform for researchers.

US Optum Database Study in Polycythemia Vera Patients: Thromboembolic Events (TEs) with Hydroxyurea (HU) Vs Ruxolitinib Switch Therapy and Machine-Learning Model to Predict Incidence of Tes and HU Failure

US Optum Database Study in Polycythemia Vera Patients: Thromboembolic Events (TEs) with Hydroxyurea (HU) Vs Ruxolitinib Switch Therapy and Machine-Learning Model to Predict Incidence of Tes and HU Failure

1144-P: Distribution of Prescribed Dose During the First Year of Metformin Monotherapy in the U.S.

International guidelines recommend metformin as initial pharmacotherapy for patients with type 2 diabetes (T2D). Guidelines recommend metformin is up-titrated to a maximally effective (~2000 mg/day) dose based on glycemic control and tolerability; however, the extent to which this dose titration occurs in the real-world setting is unknown. To address this question in a U.S. population, we conducted a retrospective cohort study using the Optum database to evaluate the dose distribution of metformin monotherapy (MM) over 12 months in adult (≥21 years) T2D patients initiating MM during 2013-2016. Of the 30,352 eligible patients, 49% were female and median age was 65 years. The Table describes the distribution of metformin dose over the 12-month follow-up period. The majority (88%) of patients were initiated on MM at ≤1000 mg/day. At 3, 6 and 12 months, 84%, 82% and 80% of MM patients, respectively, were on doses ≤1000 mg/day. At 12 months, only 16% of patients were on a metformin dose &amp;gt;1500 mg/day. Among those who were up-titrated during follow-up (n=5003), the median highest dose of metformin was 1500 mg/d and median time to highest dose was 165 days. In conclusion, the majority of MM patients were not up-titrated and remained on a sub-maximally effective dose of metformin at 12 months despite international guidelines. Further research will be conducted to examine the relationship between MM dose and glycemic control. Disclosure K. Iglay: Employee; Self; Merck &amp; Co., Inc. Stock/Shareholder; Self; Merck &amp; Co., Inc. B. Sawhney: Consultant; Self; Merck &amp; Co., Inc. S. Rajpathak: None. M. Crutchlow: Consultant; Spouse/Partner; Crinetics, Novartis Pharmaceuticals Corporation, ProSciento, Soleno Therapeutics, Zealand Pharma A/S. Employee; Self; Merck &amp; Co., Inc. Research Support; Spouse/Partner; Dexcom, Inc. G. Fernandes: None. K. Khunti: Advisory Panel; Self; Amgen Inc., AstraZeneca, Eli Lilly and Company, Merck Sharp &amp; Dohme Corp., Novo Nordisk A/S, Sanofi. Consultant; Self; AstraZeneca, Boehringer Ingelheim International GmbH, Eli Lilly and Company, Janssen Pharmaceuticals, Inc., Merck Sharp &amp; Dohme Corp., Novartis AG, Novo Nordisk A/S, Pfizer Inc., Sanofi-Aventis, Servier, Takeda Pharmaceutical Company Limited. Research Support; Self; AstraZeneca, Boehringer Ingelheim International GmbH, Eli Lilly and Company, Merck Sharp &amp; Dohme Corp., Novartis AG, Novo Nordisk A/S, Pfizer Inc., Sanofi-Aventis. Speaker's Bureau; Self; AstraZeneca, Berlin-Chemie AG, Boehringer Ingelheim International GmbH, Eli Lilly and Company, Janssen Pharmaceuticals, Inc., Menarini Group, Merck Sharp &amp; Dohme Corp., Novartis AG, Novo Nordisk A/S, Roche Pharma, Sanofi, Servier, Takeda Pharmaceutical Company Limited.

Abstract 23: Patient Level Prediction Models Developed With Large Observational Databases Outperformed Existing Clinical Prediction Scores For Stratifying Bleeding Risk In Patients With Atrial Fibrillation

Background: Vitamin K antagonists (e.g., warfarin) and direct oral anticoagulants (DOACs; e.g., rivaroxaban, apixaban, dabigatran, edoxaban) are effective therapies for lowering the risk of thrombotic outcomes including stroke in patients with atrial fibrillation (AF). Like all anticoagulants, they are associated with an increased risk of bleeding. Several models (e.g., ATRIA, ORBIT, HAS-BLED, CHADS 2 and CHA 2 DS 2 -VASc) are available to predict the risk of major bleeding but their performance has not been widely evaluated in observational databases. Objective: To develop patient-level bleeding risk prediction models and evaluate their performance compared to existing clinical models of bleeding. Methods: Based on tools available through the Observational Health Data Science and Informatics Collaborative and Observed Medical Outcomes Partnership Common Data Model framework, we developed patient-level prediction (PLP) models to predict the risk of major bleeding events in patients with non-valvular AF who were new users of warfarin or DOACs. A LASSO regularized logistic regression technique including over 100,000 baseline covariates was performed across each of 4 US databases: IBM MarketScan ® (Commercial (CCAE), Medicaid (MDCD), and Medicare Supplemental (MDCR)) and Optum Clinformatics ® Extended Data Mart (Optum). We evaluated the performance of the PLP models compared with the existing models using the Area Under the Curve (AUC), with 75% training and 25% testing sets. The models were then externally validated by applying them to the other three databases. Results: In each database, all the PLP models achieved higher internal validity (AUCs 0.69 - 0.83) compared with the existing clinical models. The highest AUC among the existing clinical models was 0.76 for CHA 2 DS 2 -VASc run on the DOACs new user population in the CCAE database; the comparative AUC for the PLP model was 0.79. The external validation of the PLP models was somewhat lower, the lowest being the new user DOACs models learned on the MDCD database and applied to the other three databases with AUCs between 0.56 and 0.58. This is possibly in part due to differences between patients in MDCD versus other databases (e.g., age, disability, socioeconomic status). The highest performing was the new user warfarin model learned on the Optum database validated on the CCAE database with an AUC of 0.75. Conclusion: A patient level prediction model outperforms many existing clinical bleeding risk models currently in use.

A new model to predict ischemic stroke in patients with atrial fibrillation using warfarin or direct oral anticoagulants

Stroke risk stratification scores (eg, CHA2DS2-VASc) are used to tailor therapeutic recommendations for patients with atrial fibrillation (AF) in different risk groups. The purpose of this study was to develop a tool to estimate stroke risk in patients receiving oral anticoagulants (OACs) and to identify patients who remain at high risk for stroke despite anticoagulation therapy. Patients with nonvalvular AF initiating OACs were identified in the MarketScan data from 2007 to 2015. Using bootstrapping methods and backward selection of 44 candidate variables, we developed a model that selected variables predicting stroke. The final model was validated in patients with nonvalvular AF in the Optum database in the period 2009-2015. In both databases, the discrimination of existing stroke scores were individually evaluated and compared with our new model termed the AntiCoagulaTion-specific Stroke (ACTS) score. Among 135,523 patients with AF initiating OACs in the MarketScan dataset, 2028 experienced an ischemic stroke after anticoagulant initiation. The stepwise model identified 11 variables (including type of OAC) associated with ischemic stroke. The discrimination (C statistic) of the model was adequate (0.68; 95% confidence interval [CI] 0.66-0.70), showing excellent calibration (χ2 = 6.1; P = .73). ACTS was then applied to 84,549 AF patients in the Optum dataset (1408 stroke events) and showed similar discrimination (C statistic 0.67; 95% CI 65-0.69). However, previously developed predictive models had similar discriminative ability (CHA2DS2-VASc 0.67; 95% CI 0.65-0.68). A novel model to identify AF patients at higher risk of ischemic stroke, using extensive administrative health care data including type of anticoagulant, did not perform better than established simpler models.

Economic Burden of Patients Treated for Higher-Risk Myelodysplastic Syndromes (HR-MDS) in Routine Clinical Care in the United States.

Background and ObjectiveSignificant clinical burden is associated with higher-risk myelodysplastic syndromes (HR-MDS); however, the economic burden has not been fully examined. We examined cost of care and healthcare utilization (HCU) in HR-MDS patients engaged in routine care in the United States (US).MethodsAdult US patients diagnosed with HR-MDS from 1/1/2008 to 10/31/2015 were identified from the Optum database. Patients were followed until death, progression to acute myeloid leukemia (AML), end of enrollment, or end of study (12/31/2015). Myelodysplastic syndrome (MDS)-related costs/HCU (including medical/pharmacy claims with a primary diagnosis of MDS, MDS-related treatment, or supportive care) and non-MDS-related costs/HCU were evaluated. Costs were calculated as per-patient per-month (PPPM) costs adjusted to 2015 US dollars.ResultsOf the 209 HR-MDS patients included, median follow-up was 9.9 months (interquartile range 4.6–17.9), and 69.4% had at least one inpatient admission, 56.9% had at least one emergency department visit, and nearly all patients had at least one outpatient visit. Average PPPM costs over follow-up were $17,361; year 1 versus year 2 costs were higher ($17,337 vs $12,976) following HR-MDS diagnosis. The majority of costs were for MDS-related medical services ($10,327 PPPM). MDS-related medical PPPM costs decreased from $10,557 (year 1) to $6530 (year 2). The main drivers of MDS-related medical costs and the decrease in year 2 were chemotherapy and supportive care costs.ConclusionsThe economic burden of HR-MDS is considerable, particularly within the first year of diagnosis. Treatment/supportive care costs accounted for a significant portion of MDS-related costs. As HR-MDS treatment evolves, the economic impact and HCU need to be further investigated.

PSS23 - SWITCH PATTERNS AND TOTAL COSTS IN BIOLOGIC-NAIVE PATIENTS INITIATING APREMILAST OR A BIOLOGIC FOR THE TREATMENT OF PSORIASIS

Psoriasis affects ≈7.4 million US adults. As psoriasis progresses, patients may switch therapies. This study compared treatment switching and total cost of care among biologic-naive psoriasis patients who initiated treatment with apremilast or biologics. Adults with psoriasis were selected if they had initiated apremilast or biologics from January 1, 2014‒September 30, 2016, and had a minimum 12 months of pre- and post-index continuous enrollment in the Optum Clinformatics™ claims database. Patients initiating apremilast were 1:2 propensity score matched (PSM) to patients initiating biologics. Switch was defined as a claim for a new psoriasis therapy after initiation of the index medication. Days to switch was defined as the number of days between the index date and date of switch to a new psoriasis therapy. Total direct healthcare costs were measured during follow-up by type of service and based on standard prices in the Optum database. T-test, chi-square test, and Wilcoxon rank-sum test were used to evaluate differences between the 2 cohorts for continuous and categorical variables as appropriate. 533 biologic-naive psoriasis patients initiating apremilast were matched to 955 biologic-naive psoriasis patients initiating biologics. Patient characteristics were similar between cohorts after PSM. Twelve-month switch rates were 14.1% (n=75) for patients initiating apremilast vs 17.8% (n=170) for those initiating biologics (P=0.06). Among those who switched, apremilast patients had a mean days to switch to a biologic of 187 days. Regardless of whether patients switched or did not switch treatments, mean 12-month total healthcare costs were significantly lower among those initiating apremilast vs those initiating biologics (all: $24,452 vs $46,290, P<0.0001; switched: $37,462 vs $53,724, P<0.0001; did not switch: $22,322 vs $44,680, P<0.0001). Biologic-naive patients treated with apremilast vs biologics had lower switch rates and significantly lower healthcare costs, even if they switched to biologics during the 12-month post-index period.

PMS36 - SWITCH PATTERNS AND TOTAL HEALTHCARE COSTS IN BIOLOGIC-NAIVE PATIENTS INITIATING APREMILAST OR A BIOLOGIC FOR THE TREATMENT OF PSORIATIC ARTHRITIS: RESULTS FROM A US CLAIMS ANALYSIS

Psoriatic arthritis (PsA), a chronic systemic inflammatory disease, is prevalent in an estimated 0.3% to 1.0% of adults. As PsA progresses, patients may switch therapies. This study compared treatment switching and total cost of care among biologic-naive PsA patients who initiated treatment with apremilast or a biologic. Adults with PsA were selected if they had initiated apremilast or biologic therapy from January 1, 2014‒September 30, 2016, and had a minimum 12 months of pre- and post-index continuous enrollment in the Optum Clinformatics™ claims database. Patients initiating apremilast were 1:2 propensity score matched (PSM) to patients initiating biologics. Switch was defined as a claim for a new PsA therapy after initiation of the index medication. Days to switch was defined as the number of days between the index and switch dates. Total direct costs were measured during follow-up by type of service and based on standard prices in the Optum database. T-test, chi-square test, and Wilcoxon rank-sum test were used to evaluate differences. 170 biologic-naive PsA patients initiating apremilast were matched to 327 biologic-naive PsA patients initiating biologics. Patient characteristics were similar between cohorts after PSM. Twelve-month switch rates were 17.7% (n=30) for apremilast patients vs 25.1% (n=82) for biologic patients (P=0.06), with mean days to switch of 201 vs 181 days, respectively. Regardless of whether patients switched treatments, mean 12-month total healthcare costs were significantly lower among those initiating apremilast vs those initiating biologics (all: $28,423 vs 41,178, P<0.0001; switched: $39,803 vs $51,517, P<0.004; did not switch: $25,984 vs $37,717, P<0.0001). Biologic-naive patients treated with apremilast vs biologics trended towards lower switch rates and significantly lower healthcare costs, even if they switched to a biologic during the 12-month post-index period. Future studies with larger sample sizes are needed to increase our understanding of switch rates.

Association between provider specialty and healthcare costs and glycemic control for patients with diabetes

Aims: To analyze the association between provider, healthcare costs, and glycemic control for patients with diabetes mellitus (DM).Materials and methods: This cross-sectional study identified adults with type 1 or 2 DM (T1D, T2D) in the Optum database. The main independent variable was provider (endocrinologist or primary care). Regression analysis compared total medical and pharmacy costs, adjusting for health status and other patient differences, by provider.Results: For all patients, HbA1C improvement was greater, and medical costs significantly lower with an endocrinologist rather than a primary care provider. The largest HbA1C improvement (4%) occurred for insulin-dependent patients seen by endocrinologists. Significant medical savings with endocrinologist management occurred within the Medicare Advantage population in every sub-group of patients, with 14% lower costs ($4,767) for patients with T1D, 11% lower costs ($3,160) for patients with macro- and microvascular complications, and 10% lower costs ($2,237) for insulin-dependent patients. Within the commercial insurance population, medical costs were reduced by ≥9% in every sub-group of patients, with a 20% reduction ($8,450) for patients with micro- and macrovascular complications. Overall total costs (medical and pharmacy) were 8% ($1,541) higher for patients receiving endocrinologist rather than primary care, although endocrinologist care resulted in a 9% reduction (–$3,710) in costs for Medicare Advantage patients with T1D. Total medical costs (excluding pharmacy costs) may be a more accurate indicator of costs associated with patients in various stages of DM.Limitations: There was insufficient data to develop risk-adjustment payments for pharmacy costs based on disease severity. The cross-sectional design identifies associations and not cause–effect relationships.Conclusion: DM management by an endocrinologist was associated with greater HbA1C improvement and significantly lower medical costs. Total costs were higher with an endocrinologist, but for patients with T1D lower costs were seen, ranging from 2–9% regardless of insurance type.

Abstract P021: Oral Anticoagulant Use and Risk of Incident Dementia in Persons With Atrial Fibrillation

Background: Current guidelines recommend oral anticoagulation for the prevention of stroke in most patients with atrial fibrillation (AF). Direct oral anticoagulants (DOACs) are associated with lower risk of ischemic stroke and intracranial bleeding than warfarin and, therefore, may reduce risk of dementia among patients with AF. Hypothesis: AF patients initiating DOACs for stroke prevention will have lower risk of dementia than patients initiating warfarin. Methods: We used data from two US healthcare claim databases, MarketScan (2007-2015) and Optum Clinformatics (2009-2015). Using validated algorithms, we identified patients with AF who initiated oral anticoagulation. Dementia was defined as having an International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code of 290.xx, 294.xx, or 331.0. Comorbidities and use of other medications were defined based on inpatient, outpatient, and pharmacy claims. We performed a series of head-to-head comparisons of different oral anticoagulants (warfarin, dabigatran, rivaroxaban, and apixaban) in propensity score-matched cohorts. In each database, multivariable-adjusted Cox regression models were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for incident dementia for each comparison of propensity score-matched cohorts, adjusting for age, sex, comorbidities, use of other medications, and markers of socioeconomic status (in the Optum database only). We meta-analyzed database-specific results using a random-effects model. Results: The analysis included 307,099 AF patients from the MarketScan database and 161,346 from the Optum database, of which 6,572 and 4,391 respectively had a diagnosis of incident dementia. Mean follow up ranged between 0.7 to 2.3 years and incident of diagnosed dementia between 7 to 14 cases per 1000 person-years across different oral anticoagulant initiator groups. Patients initiating DOACs had a lower risk of incident dementia than those initiating warfarin (dabigatran: HR 0.85, 95%CI 0.71, 1.01; rivaroxaban: HR 0.85, 95%CI 0.76, 0.94; apixaban: HR 0.80, 95%CI 0.65, 0.97). There were no differences in risk of incident dementia comparing DOAC user groups (dabigatran vs. rivaroxaban: HR 1.02, 95%CI 0.79, 1.32; dabigatran vs. apixaban: HR 0.92, 95%CI 0.63, 1.36; apixaban vs. rivaroxaban: HR 1.01, 95%CI 0.86, 1.19). Conclusions: Patients with AF initiating DOACs experienced lower rates of incident dementia than warfarin users. No obvious benefit was observed for any particular DOAC in relation to dementia rates.

Transfusion Burden and Outcomes in Patients with Low- or Unspecified-Grade Myelodysplastic Syndromes Treated with Erythropoiesis-Stimulating Agents: Analysis of a United States Commercial Claims Database

Introduction: Myelodysplastic syndromes (MDS) are a heterogeneous group of hematopoietic disorders characterized by 1 or more peripheral blood cytopenias.Anemia is the most common cytopenia associated with low-grade MDS (based on the International Prognostic Scoring System). Erythropoiesis-stimulating agents (ESA) are commonly used as frontline treatment of anemia in patients (pts) with low-grade MDS; red blood cell transfusions are also important in supportive care. The present database analysis was conducted to understand the transfusion burden and outcomes among pts in the United States who were diagnosed with low- or unspecified-grade MDS and treated with ESA.Methods: A retrospective analysis of theOptum integrated claims and Electronic Medical Record database ("Optum Database") identified pts aged ≥ 18 years with an index diagnosis of MDS (low-grade, ICD 9 code: 238.72; high-grade, ICD 9 code: 283.73; del(5q), ICD 9 code: 238.74; or unspecified-grade, ICD 9 code: 238.75) between 2006 and 2014 who had 365 days of retrospective data preceding index diagnosis. Pts with baseline acute myeloid leukemia were excluded. Demographics and clinical outcomes were analyzed by descriptive summary. Kaplan-Meier survival analysis was performed to define overall survival (OS) and time to transfusion dependence or transfusion independence.Results: Among the 8,493 pts in the Optum Database who met the inclusion criteria, MDS was categorized as low-grade in 2,136 (25.2%), high-grade in 367 (4.3%), del(5q) in 198 (2.3%), and unspecified-grade in 5,792 (68.2%); median follow-up was 2.3 years. In each MDS category, roughly half of pts were male, and the majority were white and non-Hispanic. Median ages at diagnosis were 75, 74, 74, and 76 years, and 977 (46%), 224 (61%), 83 (42%), and 2,071 (36%) received ≥ 1 treatment in the low-grade, high-grade, del(5q), and unspecified-grade MDS categories, respectively. Frontline therapy with ESA was the mainstay of treatment for pts with low-grade (n = 732; 75%) and unspecified-grade (n = 1,284; 62%) MDS. Average durations of frontline treatment with ESA were 228.1 and 204.5 days in pts with low-grade and unspecified-grade MDS, respectively. Median OS among low- and unspecified-grade MDS pts receiving frontline ESA was 5.4 and 4.6 years, respectively. Of the pts who received frontline ESA, 53 (7.2%) with low-grade and 107 (8.3%) with unspecified-grade MDS were transfusion dependent (≥ 2 transfusions in the 16 weeks prior to treatment initiation). Transfusion dependence in pts receiving frontline ESA was associated with shorter median OS vs transfusion independence (1.7 years vs 5.7 years in low-grade MDS, P < 0.0001; 1.7 years vs 5.2 years in unspecified-grade MDS, P < 0.0001;Figure 1). Among the 160 frontline ESA pts with low-/unspecified-grade MDS who were transfusion dependent, 108 (68%) became transfusion independent (defined as ≥ 60 days without transfusion; 37 of 53 with low-grade MDS; 71 of 107 with unspecified-grade MDS); median time to transfusion independence was 137 and 139 days in pts with low-grade and unspecified-grade MDS, respectively. Most pts with low- or unspecified-grade MDS who received frontline ESA (52% and 51%, respectively) went on to receive second-line treatment, and treatment with ESA was the most common second-line therapy (42% and 40%, respectively). 163 (8.7%) of post-front-line ESA pts were transfusion dependent. Among pts with low-/unspecified-grade MDS who received second-line therapy, median OS was shorter for those who were transfusion dependent after frontline treatment with ESA vs those who were transfusion independent (1.97 years vs 4.5 years, respectively, P < 0.001). 92 (56%) of these transfusion-dependent pts became transfusion independent in a median of 181 days.Conclusions: ESA therapy was the most common treatment approach, both frontline and second-line, for pts with low- or unspecified-grade MDS; however, treatment with ESA did not always eliminate the need for transfusion. Similar to previous reports in pts with low-grade MDS, survival outcomes were substantially worse if pts were transfusion dependent prior to frontline or second-line treatment or became transfusion dependent following frontline treatment. These data indicate the critical importance for treatments to reduce transfusion burden among pts with MDS, in both the frontline and relapsed settings. [Display omitted] DisclosuresMehra:Janssen: Employment, Equity Ownership. Potluri:Janssen Research & Development, LLC: Other: Contracted to perform research; SmartAnalyst, Inc.: Employment. He:Janssen Global Services, LLC: Employment, Equity Ownership. Rizo:Janssen Research & Development, LLC: Employment, Equity Ownership. Mundle:Janssen Research & Development, LLC: Employment, Equity Ownership. Bussolari:Janssen Research & Development, LLC: Employment, Equity Ownership.

Real-world Canagliflozin Utilization: Glycemic Control Among Patients With Type 2 Diabetes Mellitus—A Multi-Database Synthesis

PurposeRandomized controlled trials have found that treatment of type 2 diabetes mellitus with canagliflozin, a sodium glucose co-transporter 2 inhibitor, is associated with significant reductions in glycosylated hemoglobin (HbA1c) levels. However, very few studies have evaluated the effectiveness of sodium glucose co-transporter 2 inhibitors in a real-world context. This data synthesis aims to examine the demographic characteristics and glycemic control among patients treated with canagliflozin in clinical practice, using results obtained from 2 US-specific retrospective administrative claims databases. MethodsData included in the synthesis were derived from 2 large claims databases (the Optum Research Database and the Inovalon MORE2 Registry, Research Edition) and were obtained from 3 recently published retrospective observational studies of adult patients with type 2 diabetes mellitus who were treated with canagliflozin. Two of the studies used the Optum database (3-month and 6-month follow-up) and 1 study used the Inovalon database (mean follow-up of 4 months). Patient demographic characteristics, clinical characteristics, treatment utilization, and achievement of glycemic goals at baseline and after canagliflozin treatment were evaluated across the 3 studies. Results were assessed using univariate descriptive statistics. FindingsBaseline demographic characteristics were generally similar between the Optum and Inovalon cohorts. Mean baseline HbA1c was 8.7% in the Optum and 8.3% in the Inovalon cohort. Seventy-five percent of the Optum (3-month study) cohort and 74% of the Inovalon cohort used 2 or more antihyperglycemic agents. During follow-up, in both cohorts, the proportion of patients who achieved tight glycemic control (HbA1c <7.0%) more than doubled, while the proportion who had poor control (HbA1c ≥9.0%) decreased by approximately 50%. Among patients who had baseline HbA1c ≥7.0%, 21% of the Optum cohort and 24% of the Inovalon cohort achieved tight glycemic control (HbA1c <7.0%), and the proportion of patients achieving HbA1c <8.0% more than doubled in both cohorts (from 30% to 61% in the Optum cohort, and from 33% to 69% in the Inovalon cohort). ImplicationsThis synthesis of real-world data from 2 large patient databases suggests that treatment of type 2 diabetes mellitus with canagliflozin is associated with significant and consistent improvements in glycemic control. Patients with varying HbA1c control and multiple antihyperglycemic agent use were able to lower their HbA1c levels with canagliflozin treatment. Additional studies with longer follow-up would be beneficial to evaluate the durability of the real-world effectiveness of canagliflozin.

Analyses of Real World Data on Overall Survival in Multiple Myeloma Patients with at Least 3 Prior Lines of Therapy Including a PI and an IMiD, or Double Refractory to a PI and an IMiD

Background: To fully evaluate the potential benefit of novel agents for the treatment of patients with multiple myeloma (MM) who are heavily pretreated and refractory, it is important to understand the outcomes of this patient population based on current real-world experience. An International Myeloma Working Group study determined that the median overall survival (OS) of patients refractory to bortezomib (proteasome inhibitor, PI) and at least 1 immunomodulatory drug (IMiD) was 9 months (Kumar S et al. Leukemia 2012; 26: 149). Since then, other therapies have been approved for relapsed and refractory MM in the United States (US), including pomalidomide (IMiD) and carfilzomib (PI). In this analysis, real-world data were used to define the treatment landscape and outcomes of patients with MM refractory to PIs and IMiDs or who had received ³3 prior lines of therapy (LOT; including a PI and an IMiD) and provide context to results from the single-agent daratumumab phase 2 study MMY2002 (Sirius) recently presented at ASCO 2015 (Lonial S. J Clin Oncol 33, 2015 suppl; abstr LBA8512).Methods: Two independent databases were analyzed.TheIMS LifeLink: IMS Oncology Electronic Medical Records (EMR) Database (IMS Health Incorporated, Danbury, CT) and the OPTUM Database (OPTUM, Inc., Eden Prairie, MN) both comprised US patients only. For the IMS LifeLink database, patient records from the index period of 2000-2011 were screened. For the OPTUM database, the indexing period was 2007-2014. Median OS was assessed for cohorts that met the criteria of disease that was double refractory to a PI and IMiD (Criteria 1) or had been treated with ³3 LOT including a PI and IMiD and showed disease progression within 60 days on completion of last regimen (Criteria 2). Patients who met Criteria 1 could have received ³3 prior LOT, however those who met Criteria 2 only did not meet the double refractory criteria. Subgroup analyses of the eligible population were conducted on those who were only double refractory and triple/quadruple refractory.Results: For the IMS LifeLink database, 4,030 patients with MM were screened, approximately 90% of patients were diagnosed with MM in 2006 or later, and 500 met the criteria for the target population. Of the 500 patients, 323 patients met Criteria 1 and 177 patients only met Criteria 2. For the OPTUM database, 3,837 patients with MM were screened, approximately 90% of patients were diagnosed after 2009, and 162 met the criteria for the target population, 120 of whom met Criteria 1 and 42 of whom only met Criteria 2. In the total eligible populations, median OS was 239 days in the IMS LifeLink dataset compared with 240 days in the OPTUM dataset (P = 0.5358). Among patients that were only double refractory (triple/quadruple refractory patients excluded), median OS was 228 days (n = 253) in the IMS LifeLink dataset compared with 259 days (n = 97) in the OPTUM dataset (P = 0.8052). In triple/quadruple refractory patients, median OS was 154 days (n = 70) in the IMS LifeLink dataset and 95 days (n = 23) in the OPTUM dataset (P = 0.6675). The results from both databases were consistent, hence the data were pooled for further analyses; the pooled analyses indicated that the median OS was 240 days for the eligible population (n = 662), 237 days for patients who were only double refractory (n = 350), and 154 days for patients who were triple/quadruple refractory (n = 93). A naïve comparison of the OS curves from the MMY2002 study and the pooled analysis suggests a survival benefit with daratumumab versus the real-world historical control (Figure).Conclusions: Analyses of real-world data from two independent US patient databases indicated that outcomes remain poor among patients with MM who are heavily pretreated and/or highly refractory despite the availability and use of newer PIs and IMiDs, such as carfilzomib and pomalidomide. Median OS of approximately 8 months was observed in patients with ≥3 LOT (including a PI and IMiD) or refractory to a PI and IMiD. These data not only highlight the critical need for new MM treatments for patients with advanced MM, but also provide a point of reference against which novel agents such as daratumumab could be evaluated. [Display omitted] DisclosuresUsmani:Onyx: Consultancy, Honoraria, Research Funding; Janssen: Research Funding; Celgene Corporation: Consultancy, Honoraria. Ahmadi:Janssen: Employment. Ng:Janssen: Employment. Lam:Janssen: Employment. Potluri:Smart Analyst: Employment. Mehra:Janssen: Employment.