Jen-Pei Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Jen-Pei Liu has a total of 147 co-authors. Here are the top co-authors by number of shared publications:
Shein-Chung Chow
Department of Biostatistics & Bioinformatics, School of Medicine, Duke University, Durham, North Carolina, USA
62 shared publications- Chin-Fu Hsiao
Institute of Population Health Sciences, National Health Research Institutes, Zhunan, Taiwan
17 shared publications - Wei-Chu Chie
Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, Taipei City, Taiwan
8 shared publications - Hsiao-Hui Tsou
National Health Research Institutes
7 shared publications - Huey-Miin Hsueh
Department of Statistics, National Chengchi University, Taipei 11605, Taiwan
6 shared publications Shein‐Chung Chow
Department of Biostatistics and Bioinformatics Duke University School of Medicine Durham NC U.S.A.
6 shared publications
Latest Publications
- Open Access
- 27 April 2016
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 11 (4) , e0153525
Addressing Loss of Efficiency Due to Misclassification Error in Enriched Clinical Trials for the Evaluation of Targeted Therapies Based on the Cox Proportional Hazards Model
A key feature of precision medicine is that it takes individual variability at the genetic or molecular level into account in determining the best treatment for patients diagnosed with diseases detected by recently developed novel biotechnologies. The enrichment design is an efficient design that enrolls only the patients testing positive for specific molecular targets and randomly assigns them for the targeted treatment or the concurrent control. However there is no diagnostic device with perfect accuracy and precision for detecting molecular targets. In particular, the positive predictive value (PPV) can be quite low for rare diseases with low prevalence. Under the enrichment design, some patients testing positive for specific molecular targets may not have the molecular targets. The efficacy of the targeted therapy may be underestimated in the patients that actually do have the molecular targets. To address the loss of efficiency due to misclassification error, we apply the discrete mixture modeling for time-to-event data proposed by Eng and Hanlon [8] to develop an inferential procedure, based on the Cox proportional hazard model, for treatment effects of the targeted treatment effect for the true-positive patients with the molecular targets. Our proposed procedure incorporates both inaccuracy of diagnostic devices and uncertainty of estimated accuracy measures. We employed the expectation-maximization algorithm in conjunction with the bootstrap technique for estimation of the hazard ratio and its estimated variance. We report the results of simulation studies which empirically investigated the performance of the proposed method. Our proposed method is illustrated by a numerical example.
- 27 April 2016
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 11 (4) , e0153525
Open AccessAddressing Loss of Efficiency Due to Misclassification Error in Enriched Clinical Trials for the Evaluation of Targeted Therapies Based on the Cox Proportional Hazards Model
A key feature of precision medicine is that it takes individual variability at the genetic or molecular level into account in determining the best treatment for patients diagnosed with diseases detected by recently developed novel biotechnologies. The enrichment design is an efficient design that enrolls only the patients testing positive for specific molecular targets and randomly assigns them for the targeted treatment or the concurrent control. However there is no diagnostic device with perfect accuracy and precision for detecting molecular targets. In particular, the positive predictive value (PPV) can be quite low for rare diseases with low prevalence. Under the enrichment design, some patients testing positive for specific molecular targets may not have the molecular targets. The efficacy of the targeted therapy may be underestimated in the patients that actually do have the molecular targets. To address the loss of efficiency due to misclassification error, we apply the discrete mixture modeling for time-to-event data proposed by Eng and Hanlon [8] to develop an inferential procedure, based on the Cox proportional hazard model, for treatment effects of the targeted treatment effect for the true-positive patients with the molecular targets. Our proposed procedure incorporates both inaccuracy of diagnostic devices and uncertainty of estimated accuracy measures. We employed the expectation-maximization algorithm in conjunction with the bootstrap technique for estimation of the hazard ratio and its estimated variance. We report the results of simulation studies which empirically investigated the performance of the proposed method. Our proposed method is illustrated by a numerical example.
- Open Access
- 13 October 2014
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 9 (10) , e109746
Sample Size Determination for Individual Bioequivalence Inference
Statistical criterion for evaluation of individual bioequivalence (IBE) between generic and innovative products often involves a function of the second moments of normal distributions. Under replicated crossover designs, the aggregate criterion for IBE proposed by the guidance of the U.S. Food and Drug Administration (FDA) contains the squared mean difference, variance of subject-by-formulation interaction, and the difference in within-subject variances between the generic and innovative products. The upper confidence bound for the linearized form of the criterion derived by the modified large sample (MLS) method is proposed in the 2001 U.S. FDA guidance as a testing procedure for evaluation of IBE. Due to the complexity of the power function for the criterion based on the second moments, literature on sample size determination for the inference of IBE is scarce. Under the two-sequence and four-period crossover design, we derive the asymptotic distribution of the upper confidence bound of the linearized criterion. Hence the asymptotic power can be derived for sample size determination for evaluation of IBE. Results of numerical studies are reported. Discussion of sample size determination for evaluation of IBE based on the aggregate criterion of the second moments in practical applications is provided.
- 13 October 2014
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 9 (10) , e109746
Open AccessSample Size Determination for Individual Bioequivalence Inference
Statistical criterion for evaluation of individual bioequivalence (IBE) between generic and innovative products often involves a function of the second moments of normal distributions. Under replicated crossover designs, the aggregate criterion for IBE proposed by the guidance of the U.S. Food and Drug Administration (FDA) contains the squared mean difference, variance of subject-by-formulation interaction, and the difference in within-subject variances between the generic and innovative products. The upper confidence bound for the linearized form of the criterion derived by the modified large sample (MLS) method is proposed in the 2001 U.S. FDA guidance as a testing procedure for evaluation of IBE. Due to the complexity of the power function for the criterion based on the second moments, literature on sample size determination for the inference of IBE is scarce. Under the two-sequence and four-period crossover design, we derive the asymptotic distribution of the upper confidence bound of the linearized criterion. Hence the asymptotic power can be derived for sample size determination for evaluation of IBE. Results of numerical studies are reported. Discussion of sample size determination for evaluation of IBE based on the aggregate criterion of the second moments in practical applications is provided.
- 28 August 2014
- Published by Taylor & Francis in Aging & Mental Health
- Vol. 19 (6) , 548-559
Association of quality of life with laboratory measurements and lifestyle factors in community dwelling older people in Taiwan
Objectives: Little is known about the influence of routine laboratory measurements and lifestyle factors on generic quality of life (QOL) at older ages. We aimed to study the relationship between generic QOL and laboratory measurements and lifestyle factors in community dwelling older Chinese people. Methods: We conducted a cross-sectional analysis. Six hundred and ninety nine elders were randomly selected from the examinees of the annual health examination in Taipei City, Taiwan. Blood, urine and stool of the participants were examined and lifestyle data were collected. Participants completed the CASP-19 (control, autonomy, self-realization, pleasure) questionnaire, a 19-item QOL scale. The relationship between QOL and laboratory results and lifestyle factors was explored, using multiple linear regression and profile analysis. Results: The mean age of the participants was 75.5 years (SD = 6.5), and 49.5% were female. Male gender standardized β coefficients (β = 0.122) and exercise habit (β = 0.170) were associated with a better QOL, whereas advanced age (β = −0.242), blurred vision (β = −0.143), depression (β = −0.125), central obesity (β = −0.093), anemia (β = −0.095), rheumatoid arthritis (β = −0.073), Parkinsonism (β = −0.079), malignancy (β = −0.086) and motorcycle riding (β = −0.086) were associated with a lower QOL. Profile analysis revealed that young–old males, social drinkers, regular exercisers and car drivers had the best QOL (all p < 0.001). Conclusion: Of the many laboratory measurements, only anemia was associated with the lower QOL. By contrast, several lifestyle factors, such as social drinking, exercise habit and car driving, were associated with better QOL, whereas abdominal obesity and motorcycle riding were associated with lower QOL.
- 28 August 2014
- Published by Taylor & Francis in Aging & Mental Health
- Vol. 19 (6) , 548-559
Association of quality of life with laboratory measurements and lifestyle factors in community dwelling older people in Taiwan
Objectives: Little is known about the influence of routine laboratory measurements and lifestyle factors on generic quality of life (QOL) at older ages. We aimed to study the relationship between generic QOL and laboratory measurements and lifestyle factors in community dwelling older Chinese people. Methods: We conducted a cross-sectional analysis. Six hundred and ninety nine elders were randomly selected from the examinees of the annual health examination in Taipei City, Taiwan. Blood, urine and stool of the participants were examined and lifestyle data were collected. Participants completed the CASP-19 (control, autonomy, self-realization, pleasure) questionnaire, a 19-item QOL scale. The relationship between QOL and laboratory results and lifestyle factors was explored, using multiple linear regression and profile analysis. Results: The mean age of the participants was 75.5 years (SD = 6.5), and 49.5% were female. Male gender standardized β coefficients (β = 0.122) and exercise habit (β = 0.170) were associated with a better QOL, whereas advanced age (β = −0.242), blurred vision (β = −0.143), depression (β = −0.125), central obesity (β = −0.093), anemia (β = −0.095), rheumatoid arthritis (β = −0.073), Parkinsonism (β = −0.079), malignancy (β = −0.086) and motorcycle riding (β = −0.086) were associated with a lower QOL. Profile analysis revealed that young–old males, social drinkers, regular exercisers and car drivers had the best QOL (all p < 0.001). Conclusion: Of the many laboratory measurements, only anemia was associated with the lower QOL. By contrast, several lifestyle factors, such as social drinking, exercise habit and car driving, were associated with better QOL, whereas abdominal obesity and motorcycle riding were associated with lower QOL.
- 17 July 2014
- Published by Taylor & Francis in Journal of Biopharmaceutical Statistics
- Vol. 24 (6) , 1254-1263
Applications of the Bayesian Prior Information to Evaluation of Equivalence of Similar Biological Medicinal Products
In 2014, there are a number of patents of best-selling biotech drugs around the world about to expire. Large commercial interests will inevitably set off international competition among major pharmaceutical companies. The European Union, the United States, and other countries with advanced medical techniques have paid a high degree of attention to biological similarity (biosimilarity) for drug development and market management and have started to take corresponding measures. Because of the diverse definitions of biosimilarity between countries, which can determine whether the medicines on sale are in the review of the base grant, the competent authority will also encounter varying degrees of standards. Governments should review the corresponding guidelines as soon as possible because many countries around the world have actively amended the law in response to management need of biological similarity. The similarity of biological drugs should be assessed by clinical trial under current regulations. Pharmaceutical companies try to lower the cost of generic drugs to increase the competitiveness of their products. To reduce the number of subjects in the clinical trials for development of generic drugs, we refer to the clinical trial for evaluation of bridging studies. Hsiao et al. (2007) has proposed the use of Bayes method in evaluation of bridging studies. Prior information on the original region is used to reduce the number of subjects to lower the cost of biosimilar drug development. Take the results of the approval biologics as a priori information, and the information is appropriate to be embedded into the model evaluating the similar products in order to reduce the sample size required for the assessment of the test sample.
- 17 July 2014
- Published by Taylor & Francis in Journal of Biopharmaceutical Statistics
- Vol. 24 (6) , 1254-1263
Applications of the Bayesian Prior Information to Evaluation of Equivalence of Similar Biological Medicinal Products
In 2014, there are a number of patents of best-selling biotech drugs around the world about to expire. Large commercial interests will inevitably set off international competition among major pharmaceutical companies. The European Union, the United States, and other countries with advanced medical techniques have paid a high degree of attention to biological similarity (biosimilarity) for drug development and market management and have started to take corresponding measures. Because of the diverse definitions of biosimilarity between countries, which can determine whether the medicines on sale are in the review of the base grant, the competent authority will also encounter varying degrees of standards. Governments should review the corresponding guidelines as soon as possible because many countries around the world have actively amended the law in response to management need of biological similarity. The similarity of biological drugs should be assessed by clinical trial under current regulations. Pharmaceutical companies try to lower the cost of generic drugs to increase the competitiveness of their products. To reduce the number of subjects in the clinical trials for development of generic drugs, we refer to the clinical trial for evaluation of bridging studies. Hsiao et al. (2007) has proposed the use of Bayes method in evaluation of bridging studies. Prior information on the original region is used to reduce the number of subjects to lower the cost of biosimilar drug development. Take the results of the approval biologics as a priori information, and the information is appropriate to be embedded into the model evaluating the similar products in order to reduce the sample size required for the assessment of the test sample.
- 28 April 2014
- Published by Wiley in Statistics in Medicine
- Vol. 33 (19) , 3300-3317
An approximate approach to sample size determination in bioequivalence testing with multiple pharmacokinetic responses
The approval of generic drugs requires the evidence of average bioequivalence (ABE) on both the area under the concentration–time curve and the peak concentration Cmax. The bioequivalence (BE) hypothesis can be decomposed into the non‐inferiority (NI) and non‐superiority (NS) hypothesis. Most of regulatory agencies employ the two one‐sided tests (TOST) procedure to test ABE between two formulations. As it is based on the intersection–union principle, the TOST procedure is conservative in terms of the type I error rate. However, the type II error rate is the sum of the type II error rates with respect to each null hypothesis of NI and NS hypotheses. When the difference in population means between two treatments is not 0, no close‐form solution for the sample size for the BE hypothesis is available. Current methods provide the sample sizes with either insufficient power or unnecessarily excessive power. We suggest an approximate method for sample size determination, which can also provide the type II rate for each of NI and NS hypotheses. In addition, the proposed method is flexible to allow extension from one pharmacokinetic (PK) response to determination of the sample size required for multiple PK responses. We report the results of a numerical study. An R code is provided to calculate the sample size for BE testing based on the proposed methods. Copyright © 2014 John Wiley & Sons, Ltd.
- 28 April 2014
- Published by Wiley in Statistics in Medicine
- Vol. 33 (19) , 3300-3317
An approximate approach to sample size determination in bioequivalence testing with multiple pharmacokinetic responses
The approval of generic drugs requires the evidence of average bioequivalence (ABE) on both the area under the concentration–time curve and the peak concentration Cmax. The bioequivalence (BE) hypothesis can be decomposed into the non‐inferiority (NI) and non‐superiority (NS) hypothesis. Most of regulatory agencies employ the two one‐sided tests (TOST) procedure to test ABE between two formulations. As it is based on the intersection–union principle, the TOST procedure is conservative in terms of the type I error rate. However, the type II error rate is the sum of the type II error rates with respect to each null hypothesis of NI and NS hypotheses. When the difference in population means between two treatments is not 0, no close‐form solution for the sample size for the BE hypothesis is available. Current methods provide the sample sizes with either insufficient power or unnecessarily excessive power. We suggest an approximate method for sample size determination, which can also provide the type II rate for each of NI and NS hypotheses. In addition, the proposed method is flexible to allow extension from one pharmacokinetic (PK) response to determination of the sample size required for multiple PK responses. We report the results of a numerical study. An R code is provided to calculate the sample size for BE testing based on the proposed methods. Copyright © 2014 John Wiley & Sons, Ltd.
- 13 February 2014
- Published by Wiley in Journal of Chemometrics
- Vol. 28 (3) , 202-211
Sample size determination for within-device precision
- 13 February 2014
- Published by Wiley in Journal of Chemometrics
- Vol. 28 (3) , 202-211
Sample size determination for within-device precision
- 30 October 2013
- Published by Wiley
- p. 303-342
Analysis of Continuous Data
- 30 October 2013
- Published by Wiley
- p. 237-302
Classification of Clinical Trials