SEARCH · PubMed Health

Results for “bootstrap aggregation”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 37 records · Page 2Linked to original sources

PubMed2006

Total and component health care costs in a non-Medicare HMO population of patients with and without type 2 diabetes and with and without macrovascular disease.

BACKGROUND: Type 2 diabetes (T2DM) is one of the most prevalent and costly chronic conditions in the United States. Macrovascular disease (MVD) remains a common and costly comorbidity in T2DM. Understanding the impact of MVD on total health care costs in patients with T2DM is of great importance to managed care organizations (MCOs). OBJECTIVE: To examine from the perspective of an MCO the impact of MVD on health care costs in patients with T2DM and in a matched comparison group of patients without diabetes. METHODS: This study involved retrospective analysis of administrative claims (eligibility, pharmacy, and medical) using data from a commercial health maintenance organization population of approximately 700,000 members in an East Coast health plan. Patients were included in this study if they (a) had 2 or more claims for T2DM ( International Classification of Diseases, Ninth Revision, Clinical Modification[ICD-9-CM] codes 250.X0 or 250.X2), or (b) had a prescription drug claim for insulin and a diagnosis of T2DM, or (c) had at least 1 pharmacy claim for an oral glycemic-modifying agent during the 12-month period from January 1, 2003, through December 31, 2003. Patients with 2 or more medical claims for type 1 diabetes (ICD-9-CM codes 250.X1 or 250.X3) were excluded from the study. A random group of comparison patients without diabetes (ICD-9 code 250.xx) were matched on age group and sex. Study patients in these 2 groups were subdivided into 4 groups based on the presence of medical claims with diagnosis codes for MVD (acute myocardial infarction, other ischemic heart disease, coronary artery bypass surgery, percutaneous transluminal angioplasty, congestive heart failure, cerebrovascular accident, peripheral vascular disease, cerebrovascular disease, and peripheral vascular disease). Direct medical costs were aggregated for 12 months after the index date for patients in all 4 groups. Bootstrapping technique was used to compare the health care costs between patients with T2DM and those without diabetes, stratified by MVD status. RESULTS: A total of 9,059 patients with T2DM were identified and were matched by age group and sex to a random group of patients without diabetes. MVD was present in 26.9% (n=2,441) of patients with T2DM versus 11.3% (n=1,027) of patients without diabetes. Patients with MVD and T2DM were, on average, a year younger than patients with MVD but without diabetes (54.55 vs. 55.55 years, P <0.001). Patients with T2DM but without MVD were nearly the same age as patients with neither diabetes nor MVD (50.44 vs. 50.59 years, P=0.092). The T2DM patients with MVD had average 12-month costs more than 3 times the costs for patients with T2DM but without medical claims with diagnosis codes for MVD--10,450 dollars versus 3,385 dollars, respectively. Pharmacy costs accounted for 29.0% and inpatient hospital costs accounted for 43.9% of total medical costs in T2DM patients with MVD versus 55.0% and 17.3%, respectively, in T2DM patients without MVD. Patients with MVD diagnoses and T2DM had total average medical costs that were 1.7 times the total medical costs for MVD patients without T2DM--10,450 dollars versus 6,090 dollars, respectively. CONCLUSIONS: The results of this analysis suggest that MVD may triple the total medical care costs in patients with T2DM. These economic consequences would appear to support the importance of interventions intended to prevent macrovascular events in patients with T2DM.

Adult↗

PubMed2003

On the calculation of intestinal schistosome infection intensity.

The effects of using different methods to calculate individual infection intensities on the age-infection distribution of Schistosoma mansoni field data are demonstrated. Methods are tested on a maximum of three stool samples per person collected on three consecutive days; the methods considered for the calculation of individual infection intensities are the geometric mean (GM), arithmetic mean (AM) and pseudo geometric mean (GM of stool samples instead of replicates). In addition, the effects of calculating the infection intensity for each age group using either AMs or GMs are compared. Differences occur in the shape of the age-infection profiles obtained by using either the arithmetic or geometric group mean. When using the AM, peak infection intensity occurs in a younger age group compared to using the GM, and all three methods of calculating individual infection intensity give the first peak of infection in the same age group. However, differences occur in the position of the second peak which occurs earlier with the two GMs than with the AM. Bootstrapping procedures show that the individual AM, gives a different age group for the first peak of infection at least 25% of the time when compared to either of the GMs, and 31% of the time for the second peak, while the two GMs give the same peak age groups around 90-92% of the time for both peaks. When using the GM, to calculate infection intensity for each age group, there are no differences between the three methods used to calculate individual infection intensity. This is confirmed by bootstrapping procedures. The results are discussed in relation to the distribution of parasites and levels of parasite aggregation. The implications of the results for field studies are also discussed.

Adolescent↗

PubMedDate not supplied

Random forest: a classification and regression tool for compound classification and QSAR modeling.

A new classification and regression tool, Random Forest, is introduced and investigated for predicting a compound's quantitative or categorical biological activity based on a quantitative description of the compound's molecular structure. Random Forest is an ensemble of unpruned classification or regression trees created by using bootstrap samples of the training data and random feature selection in tree induction. Prediction is made by aggregating (majority vote or averaging) the predictions of the ensemble. We built predictive models for six cheminformatics data sets. Our analysis demonstrates that Random Forest is a powerful tool capable of delivering performance that is among the most accurate methods to date. We also present three additional features of Random Forest: built-in performance assessment, a measure of relative importance of descriptors, and a measure of compound similarity that is weighted by the relative importance of descriptors. It is the combination of relatively high prediction accuracy and its collection of desired features that makes Random Forest uniquely suited for modeling in cheminformatics.

Journal Article↗

PubMed1997

Additive effects of older brothers and homosexual brothers in the prediction of marriage and cohabitation.

Research has shown that male homosexuality tends to cluster in families and that homosexual males have, on average, a greater number of older brothers than do heterosexual males. This study investigated whether the former, between-families effect and the latter, within-families effect are additive. The subjects were 717 full siblings over age 40 reported by 343 heterosexual and homosexual male probands examined in Southern Ontario in 1994-1995. The sibling's history of legal marriage or cohabitation in a heterosexual relationship was taken as a proxy variable for sexual orientation. There were no significant findings for the female siblings. As expected, the never-married male siblings were more likely to come from the sibships of the homosexual probands, and they had a greater average number of older brothers. A bootstrapped logistic regression analysis showed that an additive model best explained the male siblings' data. The results suggest that the familial aggregation of male homosexuality cannot be explained by the birth order effect and that older brothers and family membership reflect separate influences on sexual orientation or sexual orientation-correlated behavior.

Adult↗

PubMed2005

Prediction of penicillin resistance in Staphylococcus aureus isolates from dairy cows with mastitis, based on prior test results.

AIM: To gauge how well prior laboratory test results predict in vitro penicillin resistance of Staphylococcus aureus isolates from dairy cows with mastitis. METHODS: Population-based data on the farm of origin (n=79), genotype based on pulsed-field gel electrophoresis (PFGE) results, and the penicillin-resistance status of Staph. aureus isolates (n=115) from milk samples collected from dairy cows with mastitis submitted to two diagnostic laboratories over a 6-month period were used. Data were mined stochastically using the all-possible-pairs method, binomial modelling and bootstrap simulation, to test whether prior test results enhance the accuracy of prediction of penicillin resistance on farms. RESULTS: Of all Staph. aureus isolates tested, 38% were penicillin resistant. A significant aggregation of penicillin-resistance status was evident within farms. The probability of random pairs of isolates from the same farm having the same penicillin-resistance status was 76%, compared with 53% for random pairings of samples across all farms. Thus, the resistance status of randomly selected isolates was 1.43 times more likely to correctly predict the status of other isolates from the same farm than the random population pairwise concordance probability (p=0.011). This effect was likely due to the clonal relationship of isolates within farms, as the predictive fraction attributable to prior test results was close to nil when the effect of within-farm clonal infections was withdrawn from the model. CONCLUSIONS: Knowledge of the penicillin-resistance status of a prior Staph. aureus isolate significantly enhanced the predictive capability of other isolates from the same farm. In the time and space frame of this study, clinicians using previous information from a farm would have more accurately predicted the penicillin-resistance status of an isolate than they would by chance alone on farms infected with clonal Staph. aureus isolates, but not on farms infected with highly genetically heterogeneous bacterial strains.

Animals↗

PubMed2004

Methods for testing familial aggregation of diseases in population-based samples: application to Hodgkin lymphoma in Swedish registry data.

We use data on lymphoma in families of Hodgkin lymphoma (HL) cases from the Swedish Family Cancer Database (Hemminki et al. 2001) to illustrate survival methods for detecting familial aggregation in first degree relatives of case probands compared to first degree relatives of control probands, from registries that permit sampling of all cases. Because more than one case may occur in a given family, the first degree relatives of case probands are not necessarily independent, and we present procedures that allow for such dependence. A bootstrap procedure also accommodates matching of case and control probands by resampling the matching clusters, defined as the combined set of all first degree relatives of the matched case and control probands. Regarding families as independent sampling units leads to inferences based on "sandwich variance estimators" and accounts for dependencies from having more than one proband in a family, but not for matching. We compare these methods in analysis of familial aggregation of HL and also present simulations to compare survival analyses with analyses of binary outcome data.

Adult↗

PubMed2006

[Decisions in case of "problematic" cost-effectiveness ratios based on the example of a clinical trial in rehabilitation care].

PURPOSE: Economic assessment of an additional psychological intervention in the rehabilitation of patients with chronic low-back pain and evaluation of results by decision makers. METHODS: Piggy-back cost-utility analysis of a randomised clinical trial, including a bootstrap analysis. Costs were measured by using the cost accounting systems of the rehabilitation clinics and by surveying patients. Health-related quality of life was measured using the EQ-5D. Implications of different representations of the decision problem and corresponding decision rules concerning the cost-effectiveness plane are discussed. RESULTS: As compared with the 126 patients of the control arm, the 98 patients in the intervention arm gained 3.5 days in perfect health on average as well as 1219 euro cost saving. However, because of the uncertainty involved, the results of a bootstrap analysis cover all quadrants of the cost-effectiveness plane. Using maximum willingness-to-pay per effect unit gained, decision rules can be defined for parts of the cost-effectiveness plane. These have to be aggregated in a further valuation step. CONCLUSIONS: Study results show that decisions on stochastic economic evaluation results may require an additional valuation step aggregating the various parts of the cost-effectiveness plane.

Cost-Benefit Analysis↗

PubMed2006

Assessing performance of prediction rules in machine learning.

INTRODUCTION: An important goal in machine learning is to assess the degree to which prediction rules are robust and replicable, since these rules are used for decision making and for planning follow-up studies. This requires an estimate of a prediction rule's true error rate, a statistic that can be estimated by resampling data. However, there are many possible approaches depending upon whether we draw observations with or without replacement, or sample once, repeatedly, or not at all, and the pros and cons of each are often unclear. This study illustrates and compares different methods for estimating true error with the aim of providing practical guidance to users of machine learning techniques. METHODS: We conducted Monte Carlo simulation studies using four different error estimators: bootstrap, split sample, resubstitution and a direct estimate of true error. Here, 'split sample' refers to a single random partition of the data into a pair of training and test samples, a popular scheme. We used stochastic gradient boosting as a learning algorithm, and considered data from two studies for which the underlying data mechanism was known to be complex: a library of 6000 tripeptide substrates collected for analysis of proteasome inhibition as part of anticancer drug design, and a cardiovascular study involving 600 subjects receiving antiplatelet treatment for acute coronary syndrome. RESULTS: There were important differences in the performance of the various error estimators examined. Error estimators for split sample and resubstitution, while being the most transparent in action and the simplest to apply, did not quantify the performance of prediction rules as accurately as the bootstrap. This was true for both types of study data, despite their highly different nature. CONCLUSIONS: The robustness and reliability of decisions based on analysis of genomics data could, in many cases, be improved by following best practices for prediction error estimation. For this, techniques such as bootstrap should be considered.

Algorithms↗

PubMed2003

Bagging tree classifiers for laser scanning images: a data- and simulation-based strategy.

Diagnosis based on medical image data is common in medical decision making and clinical routine. We discuss a strategy to derive a classifier with good performance on clinical image data and to justify the properties of the classifier by an adapted simulation model of image data. We focus on the problem of classifying eyes as normal or glaucomatous based on 62 routine explanatory variables derived from laser scanning images of the optic nerve head. As learning sample we use a case-control study of 98 normal and 98 glaucomatous subjects matched by age and sex. Aggregating multiple unstable classifiers allows substantial reduction of misclassification error in many applications and bench mark problems. We investigate the performance of various classifiers for the clinical learning sample as well as for a simulation model of eye morphologies. Bagged classification trees (bagged-CTREE) are compared to single classification trees and linear discriminant analysis (LDA). We additionally compare three estimators of misclassification error: 10-fold cross-validation, the 0.632+ bootstrap and the out-of-bag estimate. In summary, the application of our strategy of a knowledge-based decision support shows that bagged classification trees perform best for glaucoma classification.

Case-Control Studies↗

PubMed2004

Bootstrap confidence intervals for medical costs with censored observations.

Medical costs data with administratively censored observations often arise in cost-effectiveness studies of treatments for life-threatening diseases. Mean of medical costs incurred from the start of a treatment until death or a certain time point after the implementation of treatment is frequently of interest. In many situations, due to the skewed nature of the cost distribution and non-uniform rate of cost accumulation over time, the currently available normal approximation confidence interval has poor coverage accuracy. In this paper, we propose a bootstrap confidence interval for the mean of medical costs with censored observations. In simulation studies, we show that the proposed bootstrap confidence interval had much better coverage accuracy than the normal approximation one when medical costs had a skewed distribution. When there is light censoring on medical costs (< or =25 per cent), we found that the bootstrap confidence interval based on the simple weighted estimator is preferred due to its simplicity and good coverage accuracy. For heavily censored cost data (censoring rate > or =30 per cent) with larger sample sizes (n > or =200), the bootstrap confidence intervals based on the partitioned estimator has superior performance in terms of both efficiency and coverage accuracy. We also illustrate the use of our methods in a real example.

Computer Simulation↗

PubMed2026

Phylogenomic subsampling and upsampling for efficient evolutionary analyses of big data.

Long runtimes, high memory demands, and reliance on high-performance computing impede phylogenomic analyses. We review a scalable phylogenomic subsampling with upsampling (PSU) framework to address this challenge, which reduces runtime and memory requirements by orders of magnitude. In PSU, small subsamples of sites from a concatenated alignment are analyzed, which are expanded by upsampling before inference, and the resulting inferences are aggregated to obtain evolutionary estimates. PSU harnesses the fact that the computational cost of maximum likelihood analysis is strongly influenced by the number of distinct site patterns in the concatenated alignment, whereas statistical power depends primarily on the amount of evolutionary information represented by the total number of sites and substitutions. By reducing the former while restoring the latter through upsampling, PSU can approximate many full-alignment analyses at substantially lower computational cost. Analysis of simulated and empirical datasets shows that PSU can accurately estimate bootstrap support values, select the optimal substitution model, test evolutionary hypotheses, and infer branch lengths, divergence times, and associated uncertainty measures. PSU also provides distributions of inferred clade support across independent subsamples, enabling detection of conflicting phylogenetic signals that may remain hidden in conventional bootstrap analysis of concatenated alignments. Automated tuning of subsample size, the number of subsamples, and the number of upsampling replicates make PSU practical. We suggest that PSU is a general approach for scalable phylogenomic inference using a broad range of statistical methods. By enabling analyses of genome-scale alignments on commodity hardware, PSU broadens research access and reduces environmental and infrastructural costs of big-data phylogenomics.

Phylogeny↗

PubMed2026

Phylogenomic subsampling and upsampling for efficient evolutionary analyses of big data.

Long runtimes, high memory demands, and reliance on high-performance computing impede phylogenomic analyses. We review a scalable phylogenomic subsampling with upsampling (PSU) framework, in which small subsamples of sites from a concatenated alignment are expanded by upsampling before inference, and the resulting analyses are then aggregated to obtain evolutionary estimates. PSU harnesses the fact that the computational cost of maximum likelihood analysis is strongly influenced by the number of distinct site patterns in the concatenated alignment, whereas statistical power depends primarily on the amount of evolutionary information represented by the total number of sites and substitutions. By reducing the former while restoring the latter through upsampling, PSU can approximate many full-data analyses at substantially lower computational cost. Analysis of simulated and empirical datasets shows that PSU can accurately estimate bootstrap support values, select the optimal substitution model, test evolutionary hypotheses, and infer branch lengths, divergence times, and associated uncertainty measures, while reducing runtime and memory requirements by orders of magnitude. PSU also provides distributions of inferred clade support across independent subsamples, enabling detection of conflicting phylogenetic signals that may remain hidden in conventional bootstrap analysis. Automated tuning of subsample size, the number of subsamples, and the number of upsampling replicates make PSU practical across diverse datasets. We suggest that PSU is a general strategy for scalable phylogenomic inference using a broad range of statistical methods. By enabling analyses of genome-scale alignments on commodity hardware, PSU broadens research access and reduces environmental and infrastructural costs of big-data phylogenomics.

confidence limits↗

PubMed2004

Ageing and health-care expenditure: the red herring argument revisited.

Zweifel and colleagues have previously proposed that proximity to death is a more important influence on health-care costs than age, suggesting that demographic change per se will not have a large impact on future aggregate health expenditure. However, issues of econometric methodology have led to challenges of the robustness of these findings. This paper revisits the analysis. Using a longitudinal hospital data set from Oxfordshire, England, the two-step Heckman model from the Zweifel study is first replicated, to find that neither age nor proximity to death have a significant effect on hospital costs. Econometric problems with the model are demonstrated, and instead a two-part model shows both age and proximity to death to have significant effects on quarterly hospital costs. Cost predictions, calculated with bootstrapped 95% confidence intervals, further demonstrate that while age may significantly affect quarterly costs, these cost changes are small compared to the tripling of quarterly costs that occurs with approaching death in the last year of life. The analyses show the importance of model selection to properly assess the determinants of health-care expenditures.

Aged↗

PubMed2003

Roles of land use resolution and unit-area load rates in assessment of diffuse nutrient emissions.

In contrast to its counterparts in Europe and North America, the Australian National Pollutant Inventory (NPI) includes estimates of aggregated emissions of nutrients (total nitrogen and total phosphorus) from catchments and facilities. Sparse or inadequate data limit the extent to which nutrient exports may be estimated from direct observations. The paucity of data for calibration and simulation limits the use of sophisticated models in most Australian catchments. Therefore, a simple unit-area load model-Catchment Management Support System (CMSS)-was selected to estimate aggregated catchment emissions for the NPI. Estimates from models like CMSS are sensitive to spatial and categorical resolution of land uses identified within the catchment and to nutrient generation rates selected for each land use category. Using three Hawkesbury-Nepean subcatchments, we show that while high spatial resolution of land use mapping is useful, only four or five major land use categories with carefully selected generation rates were required to estimate potential nutrient exports sufficiently well and to determine subcatchments contributing most. Nutrient emission estimates proved to be highly dependent on selection of generation rates so a bootstrap technique was adopted to reduce subjectivity and to improve estimates of confidence limits. This led to a specification of new generation rates for Natural, Unimproved pasture, Rural and Urban land uses and to establishment of uncertainty limits.

Agriculture↗

PubMed2002

p values for pathogens: statistical inference from infectious-disease data.

Certain features of infectious-disease data-including the aggregated nature of the data, confounding variables, correlated variables, and non-linear relations-complicate the use of standard statistical procedures. Using data on a helminth infection, we review the use of three parametric tests (analysis of variance, linear regression, and logistic regression), address the complications arising from violation of the assumptions for these tests, and suggest methods of correction. We also compare the relative merits of parametric methods with equivalent non-parametric approaches, and illustrate the differences produced with results from a Kruskal-Wallis test and a t test. The value of using a resampling method-bootstrapping-is also shown. Finally, we discuss problems arising from use of a study design that requires data on the same attribute to be collected from the same individual over a period of time, and present three methods for overcoming this complication, showing that, in the example used, the mixed effect model and generalised estimating equation give similar results.

Analysis of Variance↗

PubMed2003

AFLP analysis of genetic relationships among aromatic grapevines (Vitis vinifera).

Genotypic diversity has been detected among aromatic grapevines (Vitis vinifera) by molecular markers (AFLPs). The 22 primer-pairs generated a total of 1,331 bands of which 564 (40%) were polymorphic over all the genotypes. The bootstrap analysis pointed out that a large number of polymorphic bands (200-400) has to be used for a better estimation of the genetic distances among genotypes; 383 polymorphic AFLP bands were used for the cluster and the principal coordinate analyses because they did not present missing data across all the genotypes. The cluster analysis (UPGMA), based on polymorphic AFLP markers, revealed no relationship between the Moscato and Malvasia grapevines. The Malvasias, unlike the Moscatos distinguished by their distinct muscat aroma, have to be considered a more complex group because it includes muscat and non-muscat grapevines. The principal coordinate analysis (PCO) confirmed the pattern of the cluster analysis only for those varieties which presented a low coefficient of dissimilarity, while for the other varieties there was no correspondence between the two analyses. The pattern of aggregation among aromatic grapevines in the cluster and principal coordinate analyses does not support any classification that might include an aromatic grapevine group in V. vinifera. Even though some synonyms and homonyms are present among aromatic grapevines (V. vinifera), genetic diversity exists among genotypes in AFLP markers.

Cluster Analysis↗

PubMed2005

Long-term cost effectiveness of early and sustained dual oral antiplatelet therapy with clopidogrel given for up to one year after percutaneous coronary intervention results: from the Clopidogrel for the Reduction of Events During Observation (CREDO) trial.

OBJECTIVES: This study sought to evaluate the long-term cost effectiveness of a clopidogrel loading strategy before percutaneous coronary intervention (PCI) followed by continued treatment for one year. BACKGROUND: The Clopidogrel for the Reduction of Events During Observation (CREDO) trial, a randomized trial of 2,116 patients, showed the effectiveness of antiplatelet therapy with clopidogrel 300 mg before PCI and 75 mg daily for one year afterward compared with placebo load and placebo days 29 to 365 in reducing the combined risk of death, myocardial infarction, and stroke. All patients received clopidogrel on days 1 to 28 and aspirin on days 1 to 365. METHODS: All hospitalizations were assigned a diagnosis-related group. Associated costs were estimated three ways (including professional costs): 1) Medicare costs, 2) MEDSTAT costs, and 3) blend with Medicare for those age > or = 65 years and MEDSTAT for those age <65 years. Clopidogrel 75 mg cost 3.22 dollars. Life expectancy in trial survivors was estimated using external data. Confidence intervals were assessed by bootstrap. RESULTS: The primary composite end point occurred in 89 (8.45%) clopidogrel patients and in 122 (11.48%) placebo patients (relative risk reduction [RRR] 26.9%; 95% confidence interval [CI] 3.9% to 44.4%). The number of life-years gained (LYG) with clopidogrel was 0.1526 (95% CI 0.0263 to 0.2838) using Framingham data and 0.1920 (95% CI 0.054 to 0.337) using Saskatchewan data. Average total costs were 664 dollars higher for the clopidogrel arm (95% CI -461 dollars to 1,784 dollars). The incremental cost-effectiveness ratios (ICERs) based on Framingham data ranged from 3,685 dollars/LYG to 4,353 dollars/LYG, with over 97% of bootstrap-derived ICER estimates below 50,000 dollars/LYG. The ICERs based on Saskatchewan data were 2,929 dollars/LYG to 3,460 dollars/LYG, with over 98% of estimates below 50,000 dollars/LYG. CONCLUSIONS: Platelet inhibition with clopidogrel loading before PCI followed by therapy for one year is highly cost effective.

Aspirin↗

PubMed2005

Long-term cost-effectiveness of clopidogrel given for up to one year in patients with acute coronary syndromes without ST-segment elevation.

OBJECTIVES: We sought to evaluate the long-term cost-effectiveness of clopidogrel for up to one year after an acute coronary syndrome (ACS) without ST-segment elevation. BACKGROUND: The efficacy of platelet inhibition with clopidogrel for up to one year after ACS was demonstrated in the Clopidogrel in Unstable angina to prevent Recurrent Events (CURE) trial, a randomized trial of 12,562 patients in 28 countries that was conducted between 1998 and 2000. Patients were given clopidogrel (300-mg load followed by 75 mg/day) versus placebo, both in addition to aspirin, for a mean of nine months. METHODS: We used patient-level clinical outcomes and resource use from the CURE trial and estimates of life expectancy gains as a result of the prevention of the clinical events of death, stroke, and myocardial infarction on the basis of data from external sources. RESULTS: Excluding clopidogrel costs, average costs of hospitalizations alone were 325 dollars less for the clopidogrel arm (95% confidence interval -722 dollars to 45 dollars) using diagnosis-related group-based Medicare reimbursement rates. When including clopidogrel costs (766 dollars greater for the clopidogrel arm), average total costs were 442 dollars higher for the clopidogrel arm (95% confidence interval 62 dollars to 820 dollars). The incremental cost-effectiveness ratio (ICER) on the basis of the Framingham Heart Study was 6,318 dollars per life-year gained (LYG) with clopidogrel, with 94% of bootstrap-derived ICER estimates <50,000 dollars/LYG; based on Saskatchewan, the ICER was 6,475 dollars/LYG with 98% of estimates <50,000 dollars. CONCLUSIONS: Platelet inhibition with clopidogrel in patients for up to one year after presentation with an acute coronary syndrome is both effective and cost-effective.

Acute Disease↗