Selection of the optimal personalized treatment from multiple treatments with multivariate outcome measures.

In this work, we propose a novel method for individualized treatment selection when the treatment response is multivariate. For the K treatment (K ≥2) scenario we compare quantities that are suitable indexes based on outcome variables for each treatment conditional on patient-specific scores constructed from collected covariate measurements. Our method covers any number of treatments and outcome variables, and it can be applied for a broad set of models. The proposed method uses a rank aggregation technique to estimate an ordering of treatments based on ranked lists of treatment performance measures such as smooth conditional means and conditional probability of a response for one treatment dominating others. The method has the flexibility to incorporate patient and clinician preferences to the optimal treatment decision on an individual case basis. A simulation study demonstrates the performance of the proposed method in finite samples. We also present data analyses using HIV and Diabetes clinical trials data to show the applicability of the proposed procedure for real data.

Personalized treatment selection using data from crossover designs with carry-over effects.

In this work, we propose a semiparametric method for estimating the optimal treatment for a given patient based on individual covariate information for that patient when data from a crossover design are available. Here, we assume there are carry-over effects for patients switching from one treatment to another. For the K treatment (K ≥ 2) scenario, we show that nonparametric estimation of carry-over effects can have the undesirable property that comparison of treatment means can only be done using independent outcome measurements from different groups of patients rather than using available joint measurements for each patient. To overcome this barrier, we compare probabilities of outcome variable of each treatment dominating outcome variables for all other treatments conditional on patient-specific scores constructed from patient covariates. We suggest single-index models as appropriate models connecting outcome variables to covariates and our empirical investigations show that frequencies of correct treatment assignments are highly accurate. The proposed method is also rather robust against departures from a single-index model structure. We also conduct a real data analysis to show the applicability of the proposed procedure.

Estimation of average treatment effects among multiple treatment groups by using an ensemble approach.

In observational studies, generalized propensity score (GPS)-based statistical methods, such as inverse probability weighting (IPW) and doubly robust (DR) method, have been proposed to estimate the average treatment effect (ATE) among multiple treatment groups. In this article, we investigate the GPS-based statistical methods to estimate treatment effects from two aspects. The first aspect of our investigation is to obtain an optimal GPS estimation method among four competing GPS estimation methods by using a rank aggregation approach. We further examine whether the optimal GPS-based IPW and DR methods would improve the performance for estimating ATE. It is well known that the DR method is consistent if either the GPS or the outcome models are correctly specified. The second aspect of our investigation is to examine whether the DR method could be improved if we ensemble outcome models. To that end, bootstrap method and rank aggregation method are used to obtain the ensemble optimal outcome model from several competing outcome models, and the resulting outcome model is incorporated into the DR method, resulting in an ensemble DR (enDR) method. Extensive simulation results indicate that the enDR method provides the best performance in estimating the ATE regardless of the method used for estimating GPS. We illustrate our methods using the MarketScan healthcare insurance claims database to examine the treatment effects among three different bones and substitutes used for spinal fusion surgeries. We draw conclusions based on the estimates from the enDR method coupled with the optimal GPS estimation method.

Flexible semi-parametric regression of state occupational probabilities in a multistate model with right-censored data.

Inference for the state occupation probabilities, given a set of baseline covariates, is an important problem in survival analysis and time to event multistate data. We introduce an inverse censoring probability re-weighted semi-parametric single index model based approach to estimate conditional state occupation probabilities of a given individual in a multistate model under right-censoring. Besides obtaining a temporal regression function, we also test the potential time varying effect of a baseline covariate on future state occupation. We show that the proposed technique has desirable finite sample performances and its performance is competitive when compared with three other existing approaches. We illustrate the proposed methodology using two different data sets. First, we re-examine a well-known data set dealing with leukemia patients undergoing bone marrow transplant with various state transitions. Our second illustration is based on data from a study involving functional status of a set of spinal cord injured patients undergoing a rehabilitation program.

Propensity scores based methods for estimating average treatment effect and average treatment effect among treated: A comparative study.

Propensity score based statistical methods, such as matching, regression, stratification, inverse probability weighting (IPW), and doubly robust (DR) estimating equations, have become popular in estimating average treatment effect (ATE) and average treatment effect among treated (ATT) in observational studies. Propensity score is the conditional probability receiving a treatment assignment with given covariates, and propensity score is usually estimated by logistic regression. However, a misspecification of the propensity score model may result in biased estimates for ATT and ATE. As an alternative, the generalized boosting method (GBM) has been proposed to estimate the propensity score. GBM uses regression trees as weak predictors and captures nonlinear and interactive effects of the covariate. For GBM-based propensity score, only IPW methods have been investigated in the literature. In this article, we provide a comparative study of the commonly used propensity score based methods for estimating ATT and ATE, and examine their performances when propensity score is estimated by logistic regression and GBM, respectively. Extensive simulation results indicate that the estimators for ATE and ATT may vary greatly due to different methods. We concluded that (i) regression may not be suitable for estimating ATE and ATT regardless of the estimation method of propensity score; (ii) IPW and stratification usually provide reliable estimates of ATT when propensity score model is correctly specified; (iii) the estimators of ATE based on stratification, IPW, and DR are close to the underlying true value of ATE when propensity score is correctly specified by logistic regression or estimated using GBM.

Selection of the optimal personalized treatment from multiple treatments with right-censored multivariate outcome measures.

We propose a novel personalized concept for the optimal treatment selection for a situation where the response is a multivariate vector that could contain right-censored variables such as survival time. The proposed method can be applied with any number of treatments and outcome variables, under a broad set of models. Following a working semiparametric Single Index Model that relates covariates and responses, we first define a patient-specific composite score, constructed from individual covariates. We then estimate conditional means of each response, given the patient score, correspond to each treatment, using a nonparametric smooth estimator. Next, a rank aggregation technique is applied to estimate an ordering of treatments based on ranked lists of treatment performance measures given by conditional means. We handle the right-censored data by incorporating the inverse probability of censoring weighting to the corresponding estimators. An empirical study illustrates the performance of the proposed method in finite sample problems. To show the applicability of the proposed procedure for real data, we also present a data analysis using HIV clinical trial data, that contained a right-censored survival event as one of the endpoints.

Optimal Personalized Treatment Selection with Multivariate Outcome Measures in a Multiple Treatment Case.

In this work we propose a novel method for individualized treatment selection when there are correlated multiple treatment responses. For the K treatment (K ≥ 2) scenario, we compare quantities that are suitable indexes based on outcome variables for each treatment conditional on patient-specific scores constructed from collected covariate measurements. Our method covers any number of treatments and outcome variables, and it can be applied for a broad set of models. The proposed method uses a rank aggregation technique that takes into account possible correlations among ranked lists to estimate an ordering of treatments based on treatment performance measures such as the smooth conditional mean. The method has the flexibility to incorporate patient and clinician preferences into the optimal treatment decision on an individual case basis. A simulation study demonstrates the performance of the proposed method in finite samples. We also present data analyses using HIV clinical trial data to show the applicability of the proposed procedure for real data.

Personalized treatment selection using observational data.

Estimating the optimal treatment regime based on individual patient characteristics has been a topic of discussion in many forums. Advanced computational power has added momentum to this discussion over the last two decades and practitioners have been advocating the use of new methods in determining the best treatment. Treatments that are geared toward the 'best' outcome for a patient based on his/her genetic markers and characteristics are of high importance. In this article, we develop an approach to predict the optimal personalized treatment based on observational data. We have used inverse probability of treatment weighted machine learning methods to obtain score functions to predict the optimal treatment. Extensive simulation studies showed that our proposed method has desirable performance in selecting the optimal treatment. We provided a case study to examine the Statin use on cognitive function to illustrate the use of our proposed method.

Multi-Response Based Personalized Treatment Selection with Data from Crossover Designs for Multiple Treatments.

In this work we propose a novel method for treatment selection based on individual covariate information when the treatment response is multivariate and data are available from a crossover design. Our method covers any number of treatments and it can be applied for a broad set of models. The proposed method uses a rank aggregation technique to estimate an ordering of treatments based on ranked lists of treatment performance measures such as smooth conditional means and conditional probability of a response for one treatment dominating others. An empirical study demonstrates the performance of the proposed method in finite samples.

Prevalence and Treatment Utilization of Patients Diagnosed with Depression and Anxiety Disorders Based on Kentucky Medicaid 2012-2019 Datasets.

Objectives: To examine the prevalence and treatment utilization of patients diagnosed with Depression and Anxiety Disorders (DAD) based on Kentucky Medicaid 2012-2019 datasets. Methods: The study was based on Kentucky Medicaid claims data from 2012 through 2019 for patients 14 years and older. We constructed yearly patient-level databases using ICD_9 CM and ICD_10 CM codes to identify the patients with DAD, using the Current Procedure Terminology (CPT) codes to identify individual psychotherapy and group psychotherapy and using the National drug codes to categorize pharmacotherapy. Based on these data, we constructed summary tables that reflected the trends in prevalence of DAD across eight Kentucky Medicaid regions and for different demographic subgroups. Next, we implemented logistic regression on the constructed yearly patient-level data to formally assess the impact of risk factors and treatments on the prevalence of DAD. The potential risk factors included age, gender, race/ethnicity, geographic characteristics, comorbidities such as alcohol use disorder and tobacco use. Results: The prevalence of DAD increased from 30.84% in 2012 to 36.04% in 2019. The prevalence of DAD was significantly higher in patients with the following characteristics: non-Hispanic white, females, aged between 45 and 54 years old, living in rural areas, having alcohol use disorder, and using tobaccos. Other than 2013, the utilization of pharmacotherapy maintained at about 62%. The utilization of psychotherapy increased over years from 24.4% in 2012 to 36.5% in 2019. Overall, the utilization of any treatment slightly increased from 70.9% in 2012 to 73.3% in 2019 except a drastic decline in 2013 due to the reduction of benzodiazepine prescription. Patients being whites, females, and living in rural areas were more likely to use pharmacotherapy, and patients living in rural areas were less likely to use psychotherapy than those residing in urban areas. Conclusion: The prevalence of DAD has increased over time from 2012 to 2019. The utilization of pharmacotherapy maintained at 62% over eight years except 2013, and the utilization of psychotherapy has steadily increased over time.

Prevalence and Treatment for Alcohol Use Disorders Based on Kentucky Medicaid 2012-2019 Datasets.

Alcohol use is the leading substance use in the United States. Persons with alcohol use disorder (AUD) face enormous health consequences and family problems. Analysis of Medicaid enrollee data is critical to understand different aspects of AUD and the treatment utilization for patients with AUD. Yearly patient-level data were constructed from the Kentucky 2012-2019 Medicaid claims data. ICD-9-CM and ICD-10-CM codes were used to identify patients with AUD and their comorbid conditions, the 11-digit National Drug Codes were used to identify medication treatments, and procedure codes were used to identify psychosocial and behavioral therapies. Logistic regression models were used to examine factors that were associated with AUD prevalence and AUD treatments. The prevalence of AUD trended up over time. Patients living in metro areas, between ages 45-54, having mental disorders, tobacco use, and with a family history of alcoholism had significantly higher rates of AUD. About 60% of patients diagnosed with AUD had major depressive disorder or anxiety. The treatment utilization for AUD also trended up from 2012 to 2019; however, it was still lower than 25% in 2019. Pharmacological treatments were used in only 2.89% of AUD cases in 2012, which increased to 8.13% in 2019. Psychosocial treatments were used in only 1.59% of AUD cases in 2012 that increased to 18.95% in 2019. The prevalence of AUD trended up over years. However, the treatment utilization for AUD was lower than 25%, even as of 2019. There is an urgent need for comprehensive, evidence-based, personalized AUD treatments.

A probability based method for selecting the optimal personalized treatment from multiple treatments.

In this work we propose a method for optimal treatment assignment based on individual covariate information for a patient. For the K treatment ( K≥2 ) scenario, we compare quantities that are suitable surrogates to true conditional probabilities of outcome variable of each treatment dominating outcome variables for all other treatments conditional on patient specific scores constructed from patient-specific covariates. As opposed to methods based on conditional means, our method can be applied for a broad set of models and error structures. Furthermore, the proposed method has very desirable large sample properties. We suggest Single Index Models as appropriate models connecting outcome variables to covariates and our empirical investigations show that correct treatment assignments are highly accurate. The proposed method is also rather robust against departures from a Single Index Model structure. Furthermore, selection of a treatment using the proposed metric appears to incur no losses in terms of the average reward for cases when two treatments are close in terms of this metric. We also conduct a real data analysis to show the applicability of the proposed procedure. This analysis highlights possible gains both in terms of average response and survival time if one were to use the proposed method.

Characterization and classification of lupus patients based on plasma thermograms.

OBJECTIVE: Plasma thermograms (thermal stability profiles of blood plasma) are being utilized as a new diagnostic approach for clinical assessment. In this study, we investigated the ability of plasma thermograms to classify systemic lupus erythematosus (SLE) patients versus non SLE controls using a sample of 300 SLE and 300 control subjects from the Lupus Family Registry and Repository. Additionally, we evaluated the heterogeneity of thermograms along age, sex, ethnicity, concurrent health conditions and SLE diagnostic criteria. METHODS: Thermograms were visualized graphically for important differences between covariates and summarized using various measures. A modified linear discriminant analysis was used to segregate SLE versus control subjects on the basis of the thermograms. Classification accuracy was measured based on multiple training/test splits of the data and compared to classification based on SLE serological markers. RESULTS: Median sensitivity, specificity, and overall accuracy based on classification using plasma thermograms was 86%, 83%, and 84% compared to 78%, 95%, and 86% based on a combination of five antibody tests. Combining thermogram and serology information together improved sensitivity from 78% to 86% and overall accuracy from 86% to 89% relative to serology alone. Predictive accuracy of thermograms for distinguishing SLE and osteoarthritis / rheumatoid arthritis patients was comparable. Both gender and anemia significantly interacted with disease status for plasma thermograms (p<0.001), with greater separation between SLE and control thermograms for females relative to males and for patients with anemia relative to patients without anemia. CONCLUSION: Plasma thermograms constitute an additional biomarker which may help improve diagnosis of SLE patients, particularly when coupled with standard diagnostic testing. Differences in thermograms according to patient sex, ethnicity, clinical and environmental factors are important considerations for application of thermograms in a clinical setting.

Smooth estimation of the reliability function.

Problems with censored data arise quite frequently in reliability applications. Estimation of the reliability function is usually of concern. Reliability function estimators proposed by Kaplan and Meier (1958), Breslow (1972), are generally used when dealing with censored data. These estimators have the known properties of being asymptotically unbiased, uniformly strongly consistent, and weakly convergent to the same Gaussian process, when properly normalized. We study the properties of the smoothed Kaplan-Meier estimator with a suitable kernel function in this paper. The smooth estimator is compared with the Kaplan-Meier and Breslow estimators for large sample sizes giving an exact expression for an appropriately normalized difference of the mean square error (MSE) of the two estimators. This quantifies the deficiency of the Kaplan-Meier estimator in comparison to the smoothed version. We also obtain a non-asymptotic bound on an expected L1-type error under weak conditions. Some simulations are carried out to examine the performance of the suggested method.