In phase I clinical trials, the ‘3+3’ algorithmic design has been unparalleled in its popularity. The statistical properties of the ‘3+3’ design have been studied in the literature either in comparison with other methods or by deriving exact formulae of statistical quantities. However, there is still much to be known about its capabilities of describing and accounting for uncertainties in the observed data.
The objective of this study is to provide a probabilistic support for analyzing the heuristic performance of the ‘3+3’ design. The operating characteristics of the algorithm are computed under different hypotheses, levels of evidence, and true (or best guessed) toxicity rates. The dose-finding rules are further compared with those generated by the modified toxicity probability interval (mTPI) design, and generalized for implementation in all ‘A+B’ designs.
Our likelihood method is based on the evidential paradigm. Two hypotheses are chosen to correspond to two hypothesized dose limiting toxicity (DLT) rates, e.g., H1 - unsafe vs. H2 - acceptable. Given observed toxicities per dose, the likelihood-ratio is calculated and compared to a certain k threshold (level of evidence). Under various true toxicities, the probabilities of weak evidence, favoring H1, and favoring H2 were computed under four sets of hypotheses and several k thresholds.
For scenarios where the midpoint of the two hypothesized DLT rates is around 0.30, and for a threshold of k = 2, the ‘3+3’ design has a reduced probability (≈0.50) of identifying unsafe doses, but high chances of identifying acceptable doses. For more extreme scenarios targeting a relatively high or relatively low DLT rate, the ‘3+3’ design has no probabilistic support, and therefore it should not be used. In our comparisons, the likelihood method is in agreement with the mTPI design for the majority of hypothesized scenarios. Even so, based on the evidential paradigm, a ‘3+3’ design is often incapable of providing sufficient levels of evidence of acceptability for doses under reasonable scenarios.
Given the small sample size per dose, the levels of evidence are limited in their ability to provide strong evidence favoring either of the hypotheses.
In many situations, the ‘3+3’ design does not treat enough patients per dose to have confidence in correct dose selection, and the safety of the selected/unselected doses. This likelihood method allows consistent inferences to be made at each dose level, and evidence to be quantified regardless of cohort size. The approach can be used in phase I studies for identifying acceptably safe doses, but also for defining stopping rules in other types of dose-finding designs.
phase I clinical trials; ‘3+3’ algorithm; likelihood method; evidential paradigm
Informed consent is intended to ensure that individuals understand the purpose, risks, and benefits of research studies, and then can decide, voluntarily, whether to enroll. However, research suggests that consent procedures do not always lead to adequate participant understanding and may be longer and more complex than necessary. Studies also suggest some consent interventions, including enhanced consent forms and extended discussions with patients, increase understanding, yet methodologic challenges have been raised in studying consent in actual trial settings. This study aimed to examine the feasibility of testing two consent interventions in actual studies and also to measure effectiveness of interventions in improving understanding of trials.
Participants enrolling in any of eight ongoing clinical trials (“collaborating studies”) were, for the purposes of this study, sequentially assigned to one of three study arms involving different informed consent procedures (one control and two intervention). Control participants received standard consent form and processes. Participants in the 1st intervention arm received a bulleted fact-sheet providing simple summaries of all study components in addition to the standard consent form. Participants in the 2nd intervention arm received the bulleted fact-sheet and standard consent materials and then also engaged with a member of the collaborating study staff in a feedback Q&A session. Following consent procedures, we administered closed and open ended questions to assess patient understanding and we assessed literacy level. Descriptive statistics, Wilcoxon-Mann-Whitney and Kruskal-Wallis tests were generated to assess correlations; regression analysis determined predictors of patient understanding.
144 participants enrolled. Using regression analysis participants receiving the 2nd intervention, which included a standard consent form, bulleted fact sheet and structured question and answer session with a study staff member, had open-ended question scores that were 7.6 percentage points higher (p=.02) than participants who received the control arm (standard consent only), although unadjusted comparisons did not reach statistical significance. Eleven clinical trial investigators agreed to participate and 8 trials provided sufficient data to be included, thereby demonstrating feasibility of consent research in actual settings.
Our study supports the hypothesis that patients receiving both bulleted fact sheets and a question and answer session have higher understanding compared to patients receiving standard consent form and procedures alone. Fact sheets and short structured dialog are quick to administer and easy to replicate across studies and should be tested in larger samples for effectiveness.
research ethics; informed consent; understanding; bioethics; health literacy
The U.S. federal regulation “Exception from Informed Consent (EFIC) for Emergency Research,” 21 Code of Federal Regulations 50.24, permits emergency research without informed consent under limited conditions. Additional safeguards to protect human subjects include requirements for community consultation and public disclosure prior to starting the research. Because the regulations are vague about these requirements, Institutional Review Boards (IRBs) determine the adequacy of these activities at a local level. Thus there is potential for broad interpretation and practice variation.
To describe the variation of community consultation and public disclosure activities approved by IRBs, and the effectiveness of this process for a multi-center, EFIC, pediatric status epilepticus clinical research trial. Methods: Community consultation and public disclosure activities were analyzed for each of 15 participating sites. Surveys were conducted with participants enrolled in the status epilepticus trial to assess the effectiveness of public disclosure dissemination prior to study enrollment.
Every IRB, among the 15 participating sites, had a varied interpretation of EFIC regulations for community consultation and public disclosure activities. IRBs required various combinations of focus groups, interviews, surveys, and meetings for community consultation; news releases, mailings, and public service announcements for public disclosure. At least 4,335 patients received information about the study from these efforts. 158 chose to be included in the “Opt Out” list. Of the 304 participants who were enrolled under EFIC, 12 (5%) had heard about the study through community consultation or public disclosure activities. The activities reaching the highest number of participants were surveys and focus groups associated with existing meetings. Public disclosure activities were more efficient and cost-effective if they were part of an in-hospital resource for patients and families.
There is substantial variation in IRBs' interpretations of the federal regulations for community consultation and public disclosure. One of the goals of community consultation and public disclosure efforts for emergency research is to provide community members an opportunity to opt-out of EFIC research; however, rarely do patients or their legally authorized representatives report having learned about a study prior to enrollment.
Emergency Research; Exception from informed consent; Community Consultation; Public Disclosure; Pediatrics; Multi-centered randomized double blinded controlled study
While rarely used for supplementation trials in the U.S., schools present a practical alternative to a clinical setting.
We describe the successful recruitment and retention of urban schoolchildren into a 6-month randomized, double-blind vitamin D3 supplementation trial.
Boston-area urban schoolchildren, aged 8-15 years, were recruited in 2011-2012 through classroom and auditorium presentations. Informed consent forms in five languages were sent home to parents. Retention methods included regular telephone calls and gift cards for completed study visits.
In total, 691 schoolchildren enrolled. Their mean (SD) age was 11.7 (1.4) years; 59% were from racial/ethnic minorities and 68% qualified for free or reduced-price school meals. Multi-level, culturally-sensitive, creative approaches contributed to success in recruitment and retention. Of 691 participants, 81% completed the 6-month intervention period. Reasons for attrition included missed appointments and fear of a blood draw. More children from households with higher incomes were retained than those from households with lower incomes (85% vs. 79%, respectively, P=0.04).
The need for three fasting blood draws over the 6-month supplementation period was a limiting factor in the recruitment and retention of children in this study.
Recruitment of urban children into a school-based randomized controlled trial represents a feasible approach for a supplementation study. Particular attention to children of lower socioeconomic status may enhance participation and retention when conducting intervention studies among diverse populations.
recruitment; retention; school-based intervention; vitamin D; randomized controlled trial; children
There is compelling evidence supporting the importance of maintaining confidentiality of interim data in clinical trials designed to reliably address the benefit-to-risk profile of interventions. While this is widely recognized, creative approaches are needed to achieve this in challenging settings where interim data are released for regulatory review and action, even though the trial would be continued to address its primary hypothesis. An illustration is the recently emerging setting of cardiovascular safety trials in type 2 diabetes mellitus. At the first stage of such trials, if large relative increases in cardiovascular major morbidity/mortality can be ruled out, data can be released solely for the purpose of allowing regulatory decision making about marketing approval. The trial then is continued in the post-marketing setting to address the primary hypothesis regarding whether smaller relative increases can be ruled out. Active rather than passive approaches are needed to protect the integrity of cardiovascular safety trials. Given the importance to trial integrity of maintaining confidentiality of interim data such as the estimated relative effect on cardiovascular risk, a Data Access Plan should be in place in these trials to ensure such data are not revealed to study participants and their caregivers, investigators involved in trial conduct, the sponsor's management team, and the public, until trial completion. A Performance Standards Document also should be developed to pre-specify targeted and minimally acceptable levels for recruitment rate, best real world achievable adherence, avoidance of cross-ins, and retention rate. This document should specify creative approaches for achieving these targets, oversight procedures during trial conduct to monitor performance levels, and actions to be taken if emerging data indicate minimally acceptable levels are not being reached. In settings where meaningful breaches in confidentiality have occurred, such oversight allows adverse effects on trial integrity to be detected earlier and more effectively addressed.
Data Monitoring Committee; Data Access Plan; Performance Standards Document; Cardiovascular Safety Trials
Ethical evaluation of risk/benefit in clinical trials is premised on the achievability of resolving research questions motivating an investigation.
To determine the fraction and number of patients enrolled in trials that were at risk of not meaningfully addressing their primary research objective due to unsuccessful patient accrual.
We used the National Library of Medicine clinical trial registry to capture all initiated phase 2 and 3 intervention clinical trials that were registered as closed in 2011. We then determined the number that had been terminated due to unsuccessful accrual and the number that had closed after less than 85% of the target number of human subjects had been enrolled. Five factors were tested for association with unsuccessful accrual.
Of 2579 eligible trials, 481 (19%) either terminated for failed accrual or completed with less than 85% expected enrolment, seriously compromising their statistical power. Factors associated with unsuccessful accrual included greater number of eligibility criteria (p=0.013), non-industry funding (25% vs. 16%, p <0.0001), earlier trial phase (23% vs. 16%, p <0.0001), fewer number of research sites at trial completion (p <0.0001) and at registration (p<0.0001), and an active (non-placebo) comparator (23% vs. 16%, p <0.001).
48,027 patients had enrolled in trials closed in 2011 that were unable to answer the primary research question meaningfully. Ethics bodies, investigators, and data monitoring committees should carefully scrutinize trial design, recruitment plans, and feasibility of achieving accrual targets when designing and reviewing trials, monitor accrual once initiated, and take corrective action when accrual is lagging.
PMID: 25475878 CAMSID: cams4853
Although smoking prevalence remains strikingly high in homeless populations (~70% and three times the US national average), smoking cessation studies usually exclude homeless persons. Novel evidence-based interventions are needed for this high-risk subpopulation of smokers.
To describe the aims and design of a first-ever smoking cessation clinical trial in the homeless population. The study was a two-group randomized community-based trial that enrolled participants (n = 430) residing across eight homeless shelters and transitional housing units in Minnesota. The study objective was to test the efficacy of motivational interviewing (MI) for enhancing adherence to nicotine replacement therapy (NRT; nicotine patch) and smoking cessation outcomes.
Participants were randomized to one of the two groups: active (8 weeks of NRT + 6 sessions of MI) or control (NRT + standard care). Participants attended six in-person assessment sessions and eight retention visits at a location of their choice over 6 months. Nicotine patch in 2-week doses was administered at four visits over the first 8 weeks of the 26-week trial. The primary outcome was cotinine-verified 7-day point-prevalence abstinence at 6 months. Secondary outcomes included adherence to nicotine patch assessed through direct observation and patch counts. Other outcomes included the mediating and/or moderating effects of comorbid psychiatric and substance abuse disorders.
Lessons learned from the community-based cessation randomized trial for improving recruitment and retention in a mobile and vulnerable population included: (1) the importance of engaging the perspectives of shelter leadership by forming and convening a Community Advisory Board; (2) locating the study at the shelters for more visibility and easier access for participants; (3) minimizing exclusion criteria to allow enrollment of participants with stable psychiatric comorbid conditions; (4) delaying the baseline visit from the eligibility visit by a week to protect against attrition; and (5) regular and persistent calls to remind participants of upcoming appointments using cell phones and shelter-specific channels of communication.
The study’s limitations include generalizability due to the sample drawn from a single Midwestern city in the United States. Since inclusion criteria encompassed willingness to use NRT patch, all participants were motivated and were ready to quit smoking at the time of enrollment in the study. Findings from the self-select group will be generalizable only to those motivated and ready to quit smoking. High incentives may limit the degree to which the intervention is replicable.
Lessons learned reflect the need to engage communities in the design and implementation of community-based clinical trials with vulnerable populations.
Dynamic Treatment Regimes; Adaptive Trials; Statistical Issues in Clinical Trials; Sequential multiple assignment randomized trials (SMARTs)
Clinical Trials; Dynamic Treatment Regimes; Conference Proceedings
The classification system for categorizing the riskiness of a clinical trial is largely defined by the body of federal regulations known as the Common Rule (45 CFR 46, Subpart A) and by regulations governing the US Food and Drug Administration (FDA) codified in 21 CFR 50. This rule is applied according to the interpretation of institutional review boards (IRBs) charged with overseeing the research. If a clinical trial is determined by an IRB to constitute “minimal risk,” there are important practical implications: the IRB may allow waiver or alteration of the informed consent process; the study may be carried out in certain vulnerable populations; or the study may be reviewed by IRBs using an expedited process. However, it is unclear how the risk levels of pragmatic clinical trials (PCTs) should be assessed. Such trials typically compare existing, widely used medical therapies or interventions in the setting of routine clinical practice. Some of the therapies may be considered risky of themselves but the study comparing them may or may not add to that pre-existing level of risk. In this paper, we examine current research regulations and common interpretations of those regulations and suggest that current interpretation and application of regulations governing minimal-risk classification are marked by a high degree of variability and confusion, which in turn may ultimately harm patients by delaying or hindering potentially beneficial research. We advocate for a clear differentiation between the risks associated with a given therapy and the incremental risk incurred during research evaluating those therapies as a basic principle for evaluating the risk of a clinical study. We then examine two studies that incorporate aspects of PCTs and consider how various factors including patient perspectives, clinical equipoise, practice variation, and research methods such as cluster randomization contribute to current and evolving concepts of minimal risk, and how this understanding in turn affects the design and conduct of PCTs.
Common Rule; Pragmatic clinical trial; Minimal risk; Patient-centered outcomes research; Institutional review board; Ethics committees; research; Ethics; research
There are situations in which the requirement to obtain conventional written informed consent can impose significant or even insurmountable barriers to conducting pragmatic clinical research, including some comparative effectiveness studies and cluster-randomized trials. Although certain federal regulations governing research in the United States (45 CFR 46) define circumstances in which any of the required elements may be waived, the same standards apply regardless of whether any single element is to be waived or whether consent is to be waived in its entirety. Using the same threshold for a partial or complete waiver limits the options available to IRBs as they seek to optimize a consent process. In this article, we argue that new standards are necessary in order to enable important pragmatic clinical research while at the same time protecting patients’ rights and interests.
Clinical trials frequently spend considerable effort to collect data on patients who were assessed for eligibility but not enrolled. The Consolidated Standards of Reporting Trials (CONSORT) guidelines’ recommended flow diagram for randomized clinical trials reinforces the belief that the collection of screening data is a necessary and worthwhile endeavor. The rationale for collecting screening data includes scientific, trial management, and ethno-socio-cultural reasons.
We posit that the cost of collecting screening data is not justified, in part due to inability to centrally monitor and verify the screening data in the same manner as other clinical trial data.
To illustrate the effort and site-to-site variability, we analyzed the screening data from a multi-center, randomized clinical trial of patients with transient ischemic attack or minor ischemic stroke (POINT).
Data were collected on over 27,000 patients screened across 172 enrolling sites, 95% of whom were not enrolled. Although the rate of return of screen failure logs was high overall (95%), there were a considerable number of logs that were returned with “no data to report” (23%), often due to administrative reasons rather than no patients screened.
In spite of attempts to standardize the collection of screening data, due to differences in site processes, multi-center clinical trials face challenges in collecting those data completely and uniformly. The efforts required to centrally collect high-quality data on an extensive number of screened patients may outweigh the scientific value of the data. Moreover, the lack of a standardized definition of “screened” and the challenges of collecting meaningful characteristics for patients who have not signed consent limits the ability to compare across studies and to assess generalizability and selection bias as intended.
Consolidated Standards of Reporting Trials guidelines; Screening logs; Screen Failure; TIA; minor ischemic stroke
A dynamic treatment regime (DTR) comprises a sequence of decision rules, one per stage of intervention, that recommends how to individualize treatment to patients based on evolving treatment and covariate history. These regimes are useful for managing chronic disorders, and fit into the larger paradigm of personalized medicine. The Value of a DTR is the expected outcome when the DTR is used to assign treatments to a population of interest.
The Value of a data-driven DTR, estimated using data from a sequential multiple assignment randomized trial, is both a data-dependent parameter and a non-smooth function of the underlying generative distribution. These features introduce additional variability that is not accounted for by standard methods for conducting statistical inference, e.g., the bootstrap or normal approximations, if applied without adjustment. Our purpose is to provide a feasible method for constructing valid confidence intervals for this quantity of practical interest.
We propose a conceptually simple and computationally feasible method for constructing valid confidence intervals for the Value of an estimated DTR based on subsampling. The method is self-tuning by virtue of an approach called the double bootstrap. We demonstrate the proposed method using a series of simulated experiments.
The proposed method offers considerable improvement in terms of coverage rates of the confidence intervals over the standard bootstrap approach.
In this paper, we have restricted our attention to Q-learning for estimating the optimal DTR. However, other methods can be employed for this purpose; to keep the discussion focused, we have not explored these alternatives.
Subsampling-based confidence intervals provide much better performance compared to standard bootstrap for the Value of an estimated DTR.
Background and purpose
A behavioral intervention is a program aimed at modifying behavior for the purpose of treating or preventing disease, promoting health, and/or enhancing well-being. Many behavioral interventions are dynamic treatment regimens, that is, sequential, individualized multicomponent interventions in which the intensity and/or type of treatment is varied in response to the needs and progress of the individual participant. The multiphase optimization strategy (MOST) is a comprehensive framework for development, optimization, and evaluation of behavioral interventions, including dynamic treatment regimens. The objective of optimization is to make dynamic treatment regimens more effective, efficient, scalable, and sustainable. An important tool for optimization of dynamic treatment regimens is the sequential, multiple assignment, randomized trial (SMART). The purpose of this article is to discuss how to develop optimized dynamic treatment regimens within the MOST framework.
Methods and results
The article discusses the preparation, optimization, and evaluation phases of MOST. It is shown how MOST can be used to develop a dynamic treatment regimen to meet a prespecified optimization criterion. The SMART is an efficient experimental design for gathering the information needed to optimize a dynamic treatment regimen within MOST. One signature feature of the SMART is that randomization takes place at more than one point in time.
MOST and SMART can be used to develop optimized dynamic treatment regimens that will have a greater public health impact.
Multiphase optimization strategy; MOST; sequential; multiple assignment; randomized trial; SMART
Lack of clinical trial awareness is a known obstacle to clinical trial enrollment. We sought to define the prevalence of clinical trial awareness in the United States population, determine characteristics associated with increased trial awareness, and explore potential disparities in trial awareness.
We utilized data from the Health Information National Trends Survey (HINTS) from 2008 and 2012. Logistic regression was utilized to assess predictors of clinical trial awareness, particularly socio-demographic variables and information-seeking preferences. Trial awareness and information seeking preferences were compared in patient subgroups and between the two time periods.
Clinical trial awareness increased from 68% to 74% between 2008 and 2012. In the 2012 dataset, higher education level (odds ratio (OR) 3.52, 95% confidence interval (CI) 2.16–5.74), higher yearly income category (OR 1.84, 95% CI 1.17–2.89), and Internet-use (OR 2.13, 95% CI 1.52–3.00) were significantly associated with clinical trial awareness. Hispanic ethnicity (OR 0.41, 95% CI 0.25–0.68) was significantly associated with decreased awareness. Clinical trial awareness increased in African-American/blacks (Δ10.6%) and Hispanics (Δ10.7%) between 2008 and 2012, as did Internet-use in both subgroups (Δ14.2%, Δ18.1%, respectively).
Overall clinical trial awareness has increased between 2008 and 2012, though a large subset of the population still lacks general awareness of clinical trials. Racial and ethnic disparities in trial awareness exist, though disparities may be decreasing among the black population. These findings may help target educational efforts and inform approaches to increasing trial awareness.
Clinical trial awareness; Clinical trial barriers; Clinical trial disparities; Health communication; Health information; National cross-sectional survey; Internet
Investigators may elect to extend follow-up of participants enrolled in a randomized clinical trial after the trial comes to its planned end. The additional follow-up may be initiated to learn about longer term effects of treatments including adverse events, costs related to treatment, or for reasons unrelated to treatment such as to observe the natural course of the disease using the established cohort from the trial.
We examine transitioning from trials to extended follow-up studies when the goal of additional follow-up is to observe longer term treatment effects.
We conducted a literature search in selected journals from 2000–2012 to identify trials that extended follow-up for the purpose of studying longer term treatment effects and extracted information on the operational and logistical issues in the transition. We also draw experience from three trials coordinated by the Johns Hopkins Coordinating Centers that made transitions to extended followup: the Alzheimer’s Disease Anti-inflammatory Prevention Trial (ADAPT); Multicenter Uveitis Steroid Treatment (MUST) trial; and Childhood Asthma Management Program (CAMP).
Transitions are not uncommon in multicenter clinical trials, even in trials that continued to the planned end of the trial. Transitioning usually necessitates new participant consents. If study infrastructure is not maintained during the transition, participants will be lost and re-establishing the staff and facilities will be costly. Merging data from the trial and follow-up study can be complicated by changes in data collection measures and schedules.
Our discussion and recommendations are limited to issues that we have experienced in transitions from trials to follow-up studies.
We discuss issues such as maintaining funding, IRB and consent requirements, contacting participants, and combining data from the trial and follow-up phases. We conclude with a list of recommendations to facilitate transitions from a trial to an extended follow-up study.
linical trial; transitioning; morphing; extended follow-up
There is growing consensus that the US clinical trials system is broken, with trial costs and complexity increasing exponentially, and many trials failing to accrue. Yet concerns about the expense and failure rate of randomized trials are only the tip of the iceberg; perhaps what should worry us most is the number of trials that are not even considered because of projected costs and poor accrual. Several initiatives, including the Clinical Trials Transformation Initiative and the “Sensible Guidelines Group” seek to push back against current trends in clinical trials, arguing that all aspects of trials - including design, approval, conduct, monitoring, analysis and dissemination - should be based on evidence rather than contemporary norms. Proposed here are four methodologic fixes for current clinical trials. The first two aim to simplify trials, reducing costs and increasing patient acceptability by dramatically reducing eligibility criteria - often to the single criterion that the consenting physician is uncertain which of the two randomized arms is optimal - and by clinical integration, investment in data infrastructure to bring routinely collected data up to research grade to be used as endpoints in trials. The second two methodologic fixes aim to shed barriers to accrual, either by cluster randomization of clinicians (in the case of modifications to existing treatment) or by early consent, where patients are offered the chance of being randomly selected to be offered a novel intervention if disease progresses at a subsequent point. Such solutions may be partial, or result in a new set of problems of their own. Yet the current crisis in clinical trials mandates innovative approaches: randomized trials have resulted in enormous benefits for patients and we need to ensure that they continue to do so.
Bayesian predictive probabilities can be used for interim monitoring of clinical trials to estimate the probability of observing a statistically significant treatment effect if the trial were to continue to its predefined maximum sample size.
We explore settings in which Bayesian predictive probabilities are advantageous for interim monitoring compared to Bayesian posterior probabilities, p-values, conditional power, or group sequential methods.
For interim analyses that address prediction hypotheses, such as futility monitoring and efficacy monitoring with lagged outcomes, only predictive probabilities properly account for the amount of data remaining to be observed in a clinical trial and have the flexibility to incorporate additional information via auxiliary variables.
Computational burdens limit the feasibility of predictive probabilities in many clinical trial settings. The specification of prior distributions brings additional challenges for regulatory approval.
The use of Bayesian predictive probabilities enables the choice of logical interim stopping rules that closely align with the clinical decision making process.
predictive probability; predictive power; conditional power; posterior probability; interim monitoring; Bayesian; clinical trials; group sequential
Recent advances in medical research suggest that the optimal treatment rules should be adaptive to patients over time. This has led to an increasing interest in studying dynamic treatment regimes (DTRs), a sequence of individualized treatment rules, one per stage of clinical intervention, which map present patient information to a recommended treatment. There has been a recent surge of statistical work for estimating optimal DTRs from randomized and observational studies. The purpose of this paper is to review recent methodological progress and applied issues associated with estimating optimal DTRs.
We discuss Sequential Multiple Assignment Randomized Trials (SMARTs), a clinical trial design used to study treatment sequences. We use a common estimator of an optimal DTR that applies to SMART data as a platform to discuss several practical and methodological issues.
We provide a limited survey of practical issues associated with modeling SMART data. We review some existing estimators of optimal dynamic treatment regimes and discuss practical issues associated with these methods including: model building; missing data; statistical inference; and choosing an outcome when only non-responders are re-randomized. We mainly focus on the estimation and inference of DTRs using SMART data. DTRs can also be constructed from observational data, which may be easier to obtain in practice, however, care must be taken to account for potential confounding.
Adaptive treatment strategies; Dynamic treatment regimes; Missing data; Personalized treatment; Q-learning; Sequential Multiple Assignment Randomized Trials; Outcome weighted learning; Augmented value maximization; Structural mean models
Missing data are unavoidable in most randomized controlled clinical trials, especially when measurements are taken repeatedly. If strong assumptions about the missing data are not accurate, crude statistical analyses are biased and can lead to false inferences. Furthermore, if we fail to measure all predictors of missing data, we may not be able to model the missing data process sufficiently. In longitudinal randomized trials, measuring a patient's intent to attend future study visits may help to address both of these problems. Leon et al. developed and included the Intent to Attend assessment in the Lithium Treatment—Moderate dose Use Study (LiTMUS), aiming to remove bias due to missing data from the primary study hypothesis .
The purpose of this study is to assess the performance of the Intent to Attend assessment with regard to its use in a sensitivity analysis of missing data.
We fit marginal models to assess whether a patient's self-rated intent predicted actual study adherence. We applied inverse probability of attrition weighting (IPAW) coupled with patient intent to assess whether there existed treatment group differences in response over time. We compared the IPAW results to those obtained using other methods.
Patient-rated intent predicted missed study visits, even when adjusting for other predictors of missing data. On average, the hazard of retention increased by 19% for every one-point increase in intent. We also found that more severe mania, male gender, and a previously missed visit predicted subsequent absence. Although we found no difference in response between the randomized treatment groups, IPAW increased the estimated group difference over time.
LiTMUS was designed to limit missed study visits, which may have attenuated the effects of adjusting for missing data. Additionally, IPAW can be less efficient and less powerful than maximum likelihood or Bayesian estimators, given that the parametric model is well-specified.
In LiTMUS, the Intent to Attend assessment predicted missed study visits. This item was incorporated into our IPAW models and helped reduce bias due to informative missing data. This analysis should both encourage and facilitate future use of the Intent to Attend assessment along with IPAW to address missing data in a randomized trial.
intent to attend; inverse probability weighting; attrition; intermittent missing data; bipolar disorder; LiTMUS
Informed consent is the cornerstone for protection of human subjects in clinical trials. However, a growing body of evidence suggests that reform of the informed consent process in the United States is needed.
The Clinical Trials Transformation Initiative conducted interviews with 25 experienced observers of the informed consent process to identify limitations and actionable recommendations for change.
There was broad consensus that current practices often fail to meet the ethical obligation to inform potential research participants during the informed consent process. The most frequent single recommendation, which would affect all participants in federally regulated clinical research, was reform of the informed consent document. The interviews also identified the need for reform of clinical research review by institutional review boards, including transitioning to a single institutional review board for multi-site trials.
The consensus recommendations from the interviewees provide a framework for meaningful change in the informed consent process. Although some proposed changes are feasible for rapid implementation, others such as substantive reform of the informed consent document may require change in federal regulations.
Informed consent; institutional review board; research ethics; decision-making; clinical research; health policy
principal stratification; noncompliance; randomized trial
Traditionally, the purpose of a dose-finding design in cancer is to find the maximum tolerated dose based solely on toxicity. However, for molecularly targeted agents, little toxicity may arise within the therapeutic dose range and the dose-response curves may not be monotonic. This challenges the principle that more is better, which is widely accepted for conventional chemotherapy.
We propose three adaptive dose-finding designs for trials evaluating molecularly targeted agents, for which the dose-response curves are unimodal or plateaued. The goal of these designs is to find the optimal biological dose, which is defined as the lowest dose with the highest rate of efficacy while safe. The first proposed design is parametric and assumes a logistic dose-efficacy curve for dose finding; the second design is nonparametric and uses the isotonic regression to identify the optimal biological dose; and the third design has the spirit of a “semiparametric” approach by assuming a logistic model only locally around the current dose.
We conducted extensive simulation studies to investigate the operating characteristics of the proposed designs. Simulation studies show that the nonparametric and semiparametirc designs have good operating characteristics for finding the optimal biological dose.
The proposed designs assume a binary endpoint. Extension of the proposed designs to ordinal and time-to-event endpoints worth further investigation.
Among the three proposed designs, the nonparametric and semiparametirc designs yield consistently good operating characteristics and thus are recommended for practical use. The software to implement these two designs is available for free download at http://odin.mdacc.tmc.edu/~yyuan/.
dose-finding; optimal biological dose; Bayesian adaptive design; isotonic regression; molecularly targeted agent