Bayesian Adaptive Designs for Early Phase Clinical Trials

My dissertation focuses mainly on Bayesian adaptive designs for phase I and phase II clinical trials. It includes three specific topics: (1) proposing a novel two-dimensional dose-finding algorithm for biological agents, (2) developing Bayesian adaptive screening designs to provide more efficient and ethical clinical trials, and (3) incorporating missing late-onset responses to make an early stopping decision. Treating patients with novel biological agents is becoming a leading trend in oncology. Unlike cytotoxic agents, for which toxicity and efficacy monotonically increase with dose, biological agents may exhibit non-monotonic patterns in their dose-response relationships. Using a trial with two biological agents as an example, we propose a phase I/II trial design to identify the biologically optimal dose combination (BODC), which is defined as the dose combination of the two agents with the highest efficacy and tolerable toxicity. A change-point model is used to reflect the fact that the dose-toxicity surface of the combinational agents may plateau at higher dose levels, and a flexible logistic model is proposed to accommodate the possible non-monotonic pattern for the dose-efficacy relationship. During the trial, we continuously update the posterior estimates of toxicity and efficacy and assign patients to the most appropriate dose combination. We propose a novel dose-finding algorithm to encourage sufficient exploration of untried dose combinations in the two-dimensional space. Extensive simulation studies show that the proposed design has desirable operating characteristics in identifying the BODC under various patterns of dose-toxicity and dose-efficacy relationships. Trials of combination therapies for the treatment of cancer are playing an increasingly important role in the battle against this disease. To more efficiently handle the large number of combination therapies that must be tested, we propose a novel Bayesian phase II adaptive screening design to simultaneously select among possible treatment combinations involving multiple agents. Our design is based on formulating the selection procedure as a Bayesian hypothesis testing problem in which the superiority of each treatment combination is equated to a single hypothesis. During the trial conduct, we use the current values of the posterior probabilities of all hypotheses to adaptively allocate patients to treatment combinations. Simulation studies show that the proposed design substantially outperforms the conventional multi-arm balanced factorial trial design. The proposed design yields a significantly higher probability for selecting the best treatment while at the same time allocating substantially more patients to efficacious treatments. The proposed design is most appropriate for the trials combining multiple agents and screening out the efficacious combination to be further investigated. The proposed Bayesian adaptive phase II screening design substantially outperformed the conventional complete factorial design. Our design allocates more patients to better treatments while at the same time providing higher power to identify the best treatment at the end of the trial. Phase II trial studies usually are single-arm trials which are conducted to test the efficacy of experimental agents and decide whether agents are promising to be sent to phase III trials. Interim monitoring is employed to stop the trial early for futility to avoid assigning unacceptable number of patients to inferior treatments. We propose a Bayesian single-arm phase II design with continuous monitoring for estimating the response rate of the experimental drug. To address the issue of late-onset responses, we use a piece-wise exponential model to estimate the hazard function of time to response data and handle the missing responses using the multiple imputation approach. We evaluate the operating characteristics of the proposed method through extensive simulation studies. We show that the proposed method reduces the total length of the trial duration and yields desirable operating characteristics for different physician-specified lower bounds of response rate with different true response rates.

Screening for obstructive sleep apnea on the internet: randomized trial.

BACKGROUND: Obstructive sleep apnea is underdiagnosed. We conducted a pilot randomized controlled trial of an online intervention to promote obstructive sleep apnea screening among members of an Internet weight-loss community. METHODS: Members of an Internet weight-loss community who have never been diagnosed with obstructive sleep apnea or discussed the condition with their healthcare provider were randomized to intervention (online risk assessment+feedback) or control. The primary outcome was discussing obstructive sleep apnea with a healthcare provider at 12 weeks. RESULTS: Of 4700 members who were sent e-mail study announcements, 168 (97% were female, age 39.5 years [standard deviation 11.7], body mass index 30.3 [standard deviation 7.8]) were randomized to intervention (n=84) or control (n=84). Of 82 intervention subjects who completed the risk assessment, 50 (61%) were low risk and 32 (39%) were high risk for obstructive sleep apnea. Intervention subjects were more likely than control subjects to discuss obstructive sleep apnea with their healthcare provider within 12 weeks (11% [9/84] vs 2% [2/84]; P=.02; relative risk=4.50; 95% confidence interval, 1.002-20.21). The number needed to treat was 12. High-risk intervention subjects were more likely than control subjects to discuss obstructive sleep apnea with their healthcare provider (19% [6/32] vs 2% [2/84]; P=.004; relative risk=7.88; 95% confidence interval, 1.68-37.02). One high-risk intervention subject started treatment for obstructive sleep apnea. CONCLUSION: An online screening intervention is feasible and likely effective in encouraging members of an Internet weight-loss community to discuss obstructive sleep apnea with their healthcare provider.

The performance of human papillomavirus high-risk DNA testing in the screening and diagnostic settings.

OBJECTIVE: We sought to evaluate the performance of the human papillomavirus high-risk DNA test in patients 30 years and older. MATERIALS AND METHODS: Screening (n=835) and diagnosis (n=518) groups were defined based on prior Papanicolaou smear results as part of a clinical trial for cervical cancer detection. We compared the Hybrid Capture II (HCII) test result with the worst histologic report. We used cervical intraepithelial neoplasia (CIN) 2/3 or worse as the reference of disease. We calculated sensitivities, specificities, positive and negative likelihood ratios (LR+ and LR-), receiver operating characteristic (ROC) curves, and areas under the ROC curves for the HCII test. We also considered alternative strategies, including Papanicolaou smear, a combination of Papanicolaou smear and the HCII test, a sequence of Papanicolaou smear followed by the HCII test, and a sequence of the HCII test followed by Papanicolaou smear. RESULTS: For the screening group, the sensitivity was 0.69 and the specificity was 0.93; the area under the ROC curve was 0.81. The LR+ and LR- were 10.24 and 0.34, respectively. For the diagnosis group, the sensitivity was 0.88 and the specificity was 0.78; the area under the ROC curve was 0.83. The LR+ and LR- were 4.06 and 0.14, respectively. Sequential testing showed little or no improvement over the combination testing. CONCLUSIONS: The HCII test in the screening group had a greater LR+ for the detection of CIN 2/3 or worse. HCII testing may be an additional screening tool for cervical cancer in women 30 years and older.

Development and evaluation of a cervical cancer screening program for Chinese women

Despite of the proven efficacy of the Pap test, Asian populations still have low Pap screening compliance. The purpose of this dissertation was to investigate factors that influencing women's decision to obtain a Pap test, and to describe the development and evaluation of a cervical cancer educational program promoting the Pap screening behavior among women in Taiwan. ^ The first study examined factors associated with Pap screening compliance. Psychometric properties of measurement instruments were also assessed. The scale reliabilities were as the follows: Cronbach alpha 0.70 for knowledge scale, 0.88 for pros scale, 0.68 for cons scale, and 0.72 for perceived norms scale. Results from multiple logistic regression analysis, after adjusted for marital status, showed women who compliant to Pap screening guidelines had significantly higher knowledge, higher perceived benefits (pros), lower perceived barriers (cons), and higher perceived norms to receive a Pap test. ^ The second study described the development of a program called “Love yourself before you take care of your family”, designed to increase Pap screening behavior among women in Taiwan. The development of this program was guided by Intervention Mapping (IM), an innovative process of intervention design. The program used methods such as information transmission, modeling, persuasion, and facilitation. Strategies included direct mail campaigns, role model stories with women's testimonials, and phone intervention. ^ The third study examined the effectiveness of a randomized trial of the carefully-designed intervention (N = 424). Participants were female family members of inpatients admitted to one of the major teaching hospitals in Taiwan during August and September 1999. Women in the intervention group reported a higher rate of receiving a Pap test than women in the control group (50% versus 32%) after a three-month intervention (p = 0.002). Women in the intervention group showed increased knowledge (p = .016), perceived pros (p = 0.008), and susceptibility (p = .011) between baseline and follow-up. They also showed higher perceived pros of Pap tests than women in control group at follow-up (p = .031). This result suggested that program development based on theories and evidences could maximize the intervention impact for a specific target population. ^

Patient perceptions of balance in prostate cancer screening decision aids

Purpose. Recent reports reveals that studies of decision aids reported concern about the balance and accuracy of information included in decision aids. This study explores measures of balance in patient decision aids through a review of prostate cancer screening decision aid studies and analysis of patients’ rating of a patient decision aid for prostate cancer screening. ^ Methods. A data-abstraction form was used to collect the key characteristics, pertaining to balance, of studies included in the review. The key characteristics included (1) sample characteristics (age, race, family history of prostate cancer, and education), (2) description of the decision aid and how it was implemented, and (3) if a measure of balance was used for process evaluation and the rating. A summary table was used to report the findings. Deidentified data was received from a decision aid control trial and logistic regression analysis was used to test the association between the dependent variable (balance) and the independent variables (age, family history, race, screening preference at baseline, education, health insurance status). ^ Conclusion. Three sociodemographic variables remained significant in the final regression model: African American race, education and PSA history. Further research is needed to determine if these variables can predict a man’s perception of balance in prostate cancer screening decision aids. If a patient’s perceptions of balance can be predicted based on specific characteristics, patient report may not be the most objective method of evaluating the acceptability of a decision.^

Affect and cognition measures in preference-based decisions: Validity testing of the Ottawa Decisional Conflict Scale and a decision-specific anxiety measure with men eligible for prostate cancer screening

Background. At present, prostate cancer screening (PCS) guidelines require a discussion of risks, benefits, alternatives, and personal values, making decision aids an important tool to help convey information and to help clarify values. Objective: The overall goal of this study is to provide evidence of the reliability and validity of a PCS anxiety measure and the Decisional Conflict Scale (DCS). Methods. Using data from a randomized, controlled PCS decision aid trial that measured PCS anxiety at baseline and DCS at baseline (T0) and at two-weeks (T2), four psychometric properties were assessed: (1) internal consistency reliability, indicated by factor analysis intraclass correlations and Cronbach's α; (2) construct validity, indicated by patterns of Pearson correlations among subscales; (3) discriminant validity, indicated by the measure's ability to discriminate between undecided men and those with a definite screening intention; and (4) factor validity and invariance using confirmatory factor analyses (CFA). Results. The PCS anxiety measure had adequate internal consistency reliability and good construct and discriminant validity. CFAs indicated that the 3-factor model did not have adequate fit. CFAs for a general PCS anxiety measure and a PSA anxiety measure indicated adequate fit. The general PCS anxiety measure was invariant across clinics. The DCS had adequate internal consistency reliability except for the support subscale and had adequate discriminate validity. Good construct validity was found at the private clinic, but was only found for the feeling informed subscale at the public clinic. The traditional DCS did not have adequate fit at T0 or at T2. The alternative DCS had adequate fit at T0 but was not identified at T2. Factor loadings indicated that two subscales, feeling informed and feeling clear about values, were not distinct factors. Conclusions. Our general PCS anxiety measure can be used in PCS decision aid studies. The alternative DCS may be appropriate for men eligible for PCS. Implications: More emphasis needs to be placed on the development of PCS anxiety items relating to testing procedures. We recommend that the two DCS versions be validated in other samples of men eligible for PCS and in other health care decisions that involve uncertainty. ^

"I understood what she meant when she said '21st century medicine'": A qualitative analysis of patient responses for plans for colorectal cancer screening following a recommendation from their physician

Although physician recommendation has been significantly associated with colorectal cancer screening (CRCS), it still does not motivate all patients to get CRCS. Although improved physician recommendation for CRCS has been shown to increase patient CRCS screening, questions remain about what elements of that discussion may lead to screening. The objective of this study is to describe patients' perceptions and interpretations about their physician's recommendation for CRCS during their annual wellness exam. A subset of patients (n=51) participating in a supplement study of a behavioral intervention trial designed to increase CRCS completed a follow-up, open-ended interview two to four weeks after their annual wellness visit. Using qualitative methods, transcripts of these interviews were analyzed. Findings suggest that most patients would follow their physician's recommendation for CRCS despite not engaging in much discussion. Patients may refrain from CRCS discussion because of a commitment to CRCS, awareness of screening guidelines, and trust in physician's honesty and beneficence. Yet many patients left their wellness exams with questions, refraining because of future plans to consult with their physicians, perceived time constraints or a lack of a patient-physician relationship. If patients are leaving their wellness exams with unanswered questions, interventions should prepare physicians for patient reticence, teaching physicians how to assure patients that CRCS is a primary care activity where all questions and concerns, including cost and scheduling, may be resolved.^