Browsing by Author "Devaraj, Srikant"

Now showing 1 - 6 of 6
  • Thumbnail Image
    Item
    Casual analysis using two-part models : a general framework for specification, estimation and inference
    (2018-06-22) Hao, Zhuang; Terza, Joseph V.; Devaraj, Srikant; Liu, Ziyue; Mak, Henry; Ottoni-Wilhelm, Mark
    The two-part model (2PM) is the most widely applied modeling and estimation framework in empirical health economics. By design, the two-part model allows the process governing observation at zero to systematically differ from that which determines non-zero observations. The former is commonly referred to as the extensive margin (EM) and the latter is called the intensive margin (IM). The analytic focus of my dissertation is on the development of a general framework for specifying, estimating and drawing inference regarding causally interpretable (CI) effect parameters in the 2PM context. Our proposed fully parametric 2PM (FP2PM) framework comprises very flexible versions of the EM and IM for both continuous and count-valued outcome models and encompasses all implementations of the 2PM found in the literature. Because our modeling approach is potential outcomes (PO) based, it provides a context for clear definition of targeted counterfactual CI parameters of interest. This PO basis also provides a context for identifying the conditions under which such parameters can be consistently estimated using the observable data (via the appropriately specified data generating process). These conditions also ensure that the estimation results are CI. There is substantial literature on statistical testing for model selection in the 2PM context, yet there has been virtually no attention paid to testing the “one-part” null hypothesis. Within our general modeling and estimation framework, we devise a relatively simple test of that null for both continuous and count-valued outcomes. We illustrate our proposed model, method and testing protocol in the context of estimating price effects on the demand for alcohol.
  • Thumbnail Image
    Item
    Effect of Carvedilol vs Metoprolol Succinate on Mortality in Heart Failure with Reduced Ejection Fraction
    (Elsevier, 2018-05) Ajam, Tarek; Ajam, Samer; Devaraj, Srikant; Mohammed, Kahee; Sawada, Stephen; Medicine, School of Medicine
    Background Beta blocker therapy is indicated in all patients with heart failure with reduced ejection fraction (HFrEF) as per current guidelines. The relative benefit of carvedilol to metoprolol succinate remains unknown. This study aimed to compare survival benefit of carvedilol to metoprolol succinate. Methods The VA’s databases were queried to identify 114,745 patients diagnosed with HFrEF from 2007 to 2015 who were prescribed carvedilol and metoprolol succinate. The study estimated the survival probability and hazard ratio by comparing the carvedilol and metoprolol patients using propensity score matching with replacement techniques on observed covariates. Sub-group analyses were performed separately for men, women, elderly, duration of therapy of more than 3 months, and diabetic patients. Results A total of 43,941 metoprolol patients were matched with as many carvedilol patients. The adjusted hazard ratio of mortality for metoprolol succinate compared to carvedilol was 1.069 (95% CI: 1.046-1.092, P value: < .001). At six years, the survival probability was higher in the carvedilol group compared to the metoprolol succinate group (55.6% vs 49.2%, P value < .001). The sub-group analyses show that the results hold true separately for male, over or under 65 years old, therapy duration more than three months and non-diabetic patients. Conclusion Patients with HFrEF taking carvedilol had improved survival as compared to metoprolol succinate. The data supports the need for furthering testing to determine optimal choice of beta blockers in patients with heart failure with reduced ejection fraction.
  • Thumbnail Image
    Item
    Effect on Mortality of Higher Versus Lower β-Blocker (Metoprolol Succinate or Carvedilol) Dose in Patients With Heart Failure
    (Elsevier, 2018) Ajam, Tarek; Ajam, Samer; Devaraj, Srikant; Fudim, Marat; Kamalesh, Masoor; Medicine, School of Medicine
    This study aimed to compare the effect of β-blocker dose and heart rate (HR) on mortality in patients with heart failure with reduced ejection fraction (HFrEF). The Veteran Affairs databases were queried to identify all patients diagnosed with HFrEF based on International Classification of Diseases Ninth Revision codes from 2007 to 2015 and β-blocker (carvedilol or metoprolol succinate) use. 36,168 patients on low dose β blocker were then matched with 36,168 patients on high dose β-blocker using propensity score matching. The impact of β-blocker dose and HR was assessed on overall mortality using Cox proportional hazard model. After dividing average HR into separate quartiles and adjusting for patient characteristics, high β-blocker dose was associated with lower overall mortality as compared with a low dose of β blocker (hazard ratio 0.75, 95% confidence interval 0.73 to 0.77, p <0.01) independent of the HR achieved. The results held for all 4 quartiles of average HR. A higher β-blocker dose or a lower HR were independently and jointly associated with lower mortality for all quartiles of HR. In conclusion, higher dose of β-blocker therapy and a lower achieved HR were independently associated with a reduction in mortality in HFrEF patients.
  • Thumbnail Image
    Item
    Inflammation-associated microRNA changes in circulating exosomes of heart failure patients
    (BMC, 2017-12-19) Beg, Faheemullah; Wang, Ruizhong; Saeed, Zeb; Devaraj, Srikant; Masoor, Kamalesh; Nakshatri, Harikrishna; Surgery, School of Medicine
    Objective MiR-486 and miR-146a are cardiomyocyte-enriched microRNAs that control cell survival and self-regulation of inflammation. These microRNAs are released into circulation and are detected in plasma or in circulating exosomes. Little is known whether heart failure affects their release into circulation, which this study investigated. Results Total and exosome-specific microRNAs in plasma of 40 heart failure patients and 20 controls were prepared using the miRVana Kit. We measured exosomal and total plasma microRNAs separately because exosomes serve as cargos that transfer biological materials and alter signaling in distant organs, whereas microRNAs in plasma indicate the level of tissue damage and are mostly derived from dead cells. qRT-PCR was used to quantify miR-486, miR-146a, and miR-16. Heart failure did not significantly affect plasma miR-486/miR-16 and miR-146a/miR-16 ratio, although miR-146a/miR-16 showed a trend of elevated expression (2.3 ± 0.79, p = 0.27). By contrast, circulating exosomal miR-146a/miR-16 ratio was higher in heart failure patients (2.46 ± 0.51, p = 0.05). miR-146a is induced in response to inflammation as a part of inflammation attenuation circuitry. Indeed, Tnfα and Gm-csf increased miR-146a but not miR-486 in the cardiomyocyte cell line H9C2. These results, if confirmed in a larger study, may help to develop circulating exosomal miR-146a as a biomarker of heart failure.
  • Thumbnail Image
    Item
    Lower Post Myocardial Infarction Mortality among Women Treated at Veterans Affairs Hospitals Compared to Men
    (Elsevier, 2020-11) Ajam, Tarek; Devaraj, Srikant; Fudim, Marat; Ajam, Samer; Soleimani, Tahereh; Kamalesh, Masoor; Medicine, School of Medicine
    Background: There is conflicting evidence about whether mortality after myocardial infarction is higher among women than among men. This study aimed to compare sex differences in post myocardial infarction mortality in the Veterans Affairs system, a setting where the predominant subjects are men. Materials and methods: The Veterans Affairs Corporate Data Warehouse inpatient and laboratory chemistry databases were used to identify patients diagnosed with acute myocardial infarction from inpatient records from January 1st, 2005 to April 25th, 2015. Mortality data was obtained through the Veterans Affairs death registry. Results: A total of 130,241 patients were identified; 127,711 men (98%) and 2,530 women (2%). Men typically had more comorbidities including congestive heart failure (54% vs. 46%, P value < 0.001), diabetes mellitus (54% vs. 48%, P value < 0.001), and chronic kidney disease (39% vs. 28%, P value < 0.001). The peak troponin-I was significantly higher among men (16.0 vs. 10.7 ng/mL, P value = 0.03). The mean follow-up time was 1490.67 ± 8 days. After adjusting for differences in demographics and comorbidities, women had a significantly lower risk of mortality (hazard ration [HR]: 0.747, P value < 0.0001) as compared to men. Conclusions: In a health care system where the predominant subjects are men, women had better short- and long-term survival than men after an acute myocardial infarction. Further investigation is warranted to determine the reasons behind the improved outcomes in women post myocardial infarction in the veteran population.
  • Thumbnail Image
    Item
    Specification and estimation of the price responsiveness of alcohol demand: a policy analytic perspective
    (2016-01-13) Devaraj, Srikant; Tezra, Joseph V.; Antwi, Yaa Akosa; Jones, Josette; Wu, Jisong
    Accurate estimation of alcohol price elasticity is important for policy analysis – e.g.., determining optimal taxes and projecting revenues generated from proposed tax changes. Several approaches to specifying and estimating the price elasticity of demand for alcohol can be found in the literature. There are two keys to policy-relevant specification and estimation of alcohol price elasticity. First, the underlying demand model should take account of alcohol consumption decisions at the extensive margin – i.e., individuals' decisions to drink or not – because the price of alcohol may impact the drinking initiation decision and one's decision to drink is likely to be structurally different from how much they drink if they decide to do so (the intensive margin). Secondly, the modeling of alcohol demand elasticity should yield both theoretical and empirical results that are causally interpretable. The elasticity estimates obtained from the existing two-part model takes into account the extensive margin, but are not causally interpretable. The elasticity estimates obtained using aggregate-level models, however, are causally interpretable, but do not explicitly take into account the extensive margin. There currently exists no specification and estimation method for alcohol price elasticity that both accommodates the extensive margin and is causally interpretable. I explore additional sources of bias in the extant approaches to elasticity specification and estimation: 1) the use of logged (vs. nominal) alcohol prices; and 2) implementation of unnecessarily restrictive assumptions underlying the conventional two-part model. I propose a new approach to elasticity specification and estimation that covers the two key requirements for policy relevance and remedies all such biases. I find evidence of substantial divergence between the new and extant methods using both simulated and the real data. Such differences are profound when placed in the context of alcohol tax revenue generation.