A hybrid method in combining treatment effects from matched and unmatched studies

In the biomedical studies, the general data structures have been the matched (paired) and unmatched designs. Recently, many researchers are interested in Meta-Analysis to obtain a better understanding from several clinical data of a medical treatment. The hybrid design, which is combined two data structures, may create the fundamental question for statistical methods and the challenges for statistical inferences. The applied methods are depending on the underlying distribution. If the outcomes are normally distributed, we would use the classic paired and two independent sample T-tests on the matched and unmatched cases. If not, we can apply Wilcoxon signed rank and rank sum test on each case. ^ To assess an overall treatment effect on a hybrid design, we can apply the inverse variance weight method used in Meta-Analysis. On the nonparametric case, we can use a test statistic which is combined on two Wilcoxon test statistics. However, these two test statistics are not in same scale. We propose the Hybrid Test Statistic based on the Hodges-Lehmann estimates of the treatment effects, which are medians in the same scale.^ To compare the proposed method, we use the classic meta-analysis T-test statistic on the combined the estimates of the treatment effects from two T-test statistics. Theoretically, the efficiency of two unbiased estimators of a parameter is the ratio of their variances. With the concept of Asymptotic Relative Efficiency (ARE) developed by Pitman, we show ARE of the hybrid test statistic relative to classic meta-analysis T-test statistic using the Hodges-Lemann estimators associated with two test statistics.^ From several simulation studies, we calculate the empirical type I error rate and power of the test statistics. The proposed statistic would provide effective tool to evaluate and understand the treatment effect in various public health studies as well as clinical trials.^

A nonparametric method of comparing the partial areas under two correlated receiver operating characteristic curves

A non-parametric method was developed and tested to compare the partial areas under two correlated Receiver Operating Characteristic curves. Based on the theory of generalized U-statistics the mathematical formulas have been derived for computing ROC area, and the variance and covariance between the portions of two ROC curves. A practical SAS application also has been developed to facilitate the calculations. The accuracy of the non-parametric method was evaluated by comparing it to other methods. By applying our method to the data from a published ROC analysis of CT image, our results are very close to theirs. A hypothetical example was used to demonstrate the effects of two crossed ROC curves. The two ROC areas are the same. However each portion of the area between two ROC curves were found to be significantly different by the partial ROC curve analysis. For computation of ROC curves with large scales, such as a logistic regression model, we applied our method to the breast cancer study with Medicare claims data. It yielded the same ROC area computation as the SAS Logistic procedure. Our method also provides an alternative to the global summary of ROC area comparison by directly comparing the true-positive rates for two regression models and by determining the range of false-positive values where the models differ. ^

A novel method for evaluating an interaction effect of correlated continuous covariates in a linear model

Interaction effect is an important scientific interest for many areas of research. Common approach for investigating the interaction effect of two continuous covariates on a response variable is through a cross-product term in multiple linear regression. In epidemiological studies, the two-way analysis of variance (ANOVA) type of method has also been utilized to examine the interaction effect by replacing the continuous covariates with their discretized levels. However, the implications of model assumptions of either approach have not been examined and the statistical validation has only focused on the general method, not specifically for the interaction effect.^ In this dissertation, we investigated the validity of both approaches based on the mathematical assumptions for non-skewed data. We showed that linear regression may not be an appropriate model when the interaction effect exists because it implies a highly skewed distribution for the response variable. We also showed that the normality and constant variance assumptions required by ANOVA are not satisfied in the model where the continuous covariates are replaced with their discretized levels. Therefore, naïve application of ANOVA method may lead to an incorrect conclusion. ^ Given the problems identified above, we proposed a novel method modifying from the traditional ANOVA approach to rigorously evaluate the interaction effect. The analytical expression of the interaction effect was derived based on the conditional distribution of the response variable given the discretized continuous covariates. A testing procedure that combines the p-values from each level of the discretized covariates was developed to test the overall significance of the interaction effect. According to the simulation study, the proposed method is more powerful then the least squares regression and the ANOVA method in detecting the interaction effect when data comes from a trivariate normal distribution. The proposed method was applied to a dataset from the National Institute of Neurological Disorders and Stroke (NINDS) tissue plasminogen activator (t-PA) stroke trial, and baseline age-by-weight interaction effect was found significant in predicting the change from baseline in NIHSS at Month-3 among patients received t-PA therapy.^

Estimates of hospital and physician costs of epilepsy in the United States for 1995 comparing the provider -based survey method and a patient -based medical charts and billing method

This investigation compares two different methodologies for calculating the national cost of epilepsy: provider-based survey method (PBSM) and the patient-based medical charts and billing method (PBMC&BM). The PBSM uses the National Hospital Discharge Survey (NHDS), the National Hospital Ambulatory Medical Care Survey (NHAMCS) and the National Ambulatory Medical Care Survey (NAMCS) as the sources of utilization. The PBMC&BM uses patient data, charts and billings, to determine utilization rates for specific components of hospital, physician and drug prescriptions. ^ The 1995 hospital and physician cost of epilepsy is estimated to be $722 million using the PBSM and $1,058 million using the PBMC&BM. The difference of $336 million results from $136 million difference in utilization and $200 million difference in unit cost. ^ Utilization. The utilization difference of $136 million is composed of an inpatient variation of $129 million, $100 million hospital and $29 million physician, and an ambulatory variation of $7 million. The $100 million hospital variance is attributed to inclusion of febrile seizures in the PBSM, $−79 million, and the exclusion of admissions attributed to epilepsy, $179 million. The former suggests that the diagnostic codes used in the NHDS may not properly match the current definition of epilepsy as used in the PBMC&BM. The latter suggests NHDS errors in the attribution of an admission to the principal diagnosis. ^ The $29 million variance in inpatient physician utilization is the result of different per-day-of-care physician visit rates, 1.3 for the PBMC&BM versus 1.0 for the PBSM. The absence of visit frequency measures in the NHDS affects the internal validity of the PBSM estimate and requires the investigator to make conservative assumptions. ^ The remaining ambulatory resource utilization variance is $7 million. Of this amount, $22 million is the result of an underestimate of ancillaries in the NHAMCS and NAMCS extrapolations using the patient visit weight. ^ Unit cost. The resource cost variation is $200 million, inpatient is $22 million and ambulatory is $178 million. The inpatient variation of $22 million is composed of $19 million in hospital per day rates, due to a higher cost per day in the PBMC&BM, and $3 million in physician visit rates, due to a higher cost per visit in the PBMC&BM. ^ The ambulatory cost variance is $178 million, composed of higher per-physician-visit costs of $97 million and higher per-ancillary costs of $81 million. Both are attributed to the PBMC&BM's precise identification of resource utilization that permits accurate valuation. ^ Conclusion. Both methods have specific limitations. The PBSM strengths are its sample designs that lead to nationally representative estimates and permit statistical point and confidence interval estimation for the nation for certain variables under investigation. However, the findings of this investigation suggest the internal validity of the estimates derived is questionable and important additional information required to precisely estimate the cost of an illness is absent. ^ The PBMC&BM is a superior method in identifying resources utilized in the physician encounter with the patient permitting more accurate valuation. However, the PBMC&BM does not have the statistical reliability of the PBSM; it relies on synthesized national prevalence estimates to extrapolate a national cost estimate. While precision is important, the ability to generalize to the nation may be limited due to the small number of patients that are followed. ^