Background An extensive literature has covered the statistical properties of the Continual Reassessment Method (CRM) and the modifications of this method. While there are some applications of CRM designs in recent Phase I trials, the standard method (SM) of escalating doses after three patients with an option for an additional three patients SM remains very popular, mainly due to its simplicity. From a practical perspective, clinicians are interested in designs that can estimate the MTD using fewer patients for a fixed number of doses, or can test more dose levels for a given sample size.

Purpose This article compares CRM-based methods with the SM in terms of the number of patients needed to reach the MTD, total sample size required, and trial duration.

Methods The comparisons are performed under two alternative schemes: a fixed or a varying sample approach with the implementation of a stopping rule. The stopping rule halts the trial if the confidence interval around the MTD is within a pre-specified bound. Our simulations evaluated several CRM-based methods under different scenarios by varying the number of dose levels from five to eight and the location of the true MTD.

Results CRM and SM are comparable in terms of how fast they reach the MTD and the total sample size required when testing a limited number of dose levels (≤5), but as the number of dose levels increases, CRM reaches the MTD in fewer patients when used with a fixed sample of 20 patients. However, a sample size of 20—25 patients is not sufficient to achieve a narrow precision around the estimated toxicity rate at the MTD.

Limitations We focused on methods with practical design features that are of interest to clinicians. However, there are several alternative CRM-based designs that are not investigated in this manuscript, and hence our results are not generalizable to other designs.

Conclusions We show that CRM-based methods are an improvement over the SM in terms of accuracy and optimal dose allocation in almost all cases, except when the true dose is among the lower levels. Clinical Trials 2008; 5: 465—477. http://ctj.sagepub.com

Background Qualitative checklists for phase III trials have been proposed, to improve the reporting of such trials and to assess the validity of their results.

Purpose Our objective was to develop such a scale for phase I cancer trials.

Methods From a review of existing guidelines and checklists for phase III clinical trials, a staff team was responsible for the first selection of items and the construction of the questionnaire. The proposed quality assessment measures were rated by the survey respondents comprised of phase I research clinicians and statisticians on a 4-point Likert scale. Selected items from the quantitative analysis of the questionnaires were reviewed by an expert team who was responsible for providing the final items list. This was then applied to 103 recently published cancer phase I trials.

Results Of the 48 initial items proposed by the staff team, 17 were selected from the quantitative analysis of the 99 participants' ratings. After qualitative analysis by the expert team, a 15-item checklist was derived, with 5 items related to trial objective, 5 to design, and 5 to analysis. The application to 103 recent journal articles on phase I cancer trials evaluating cytotoxic drugs showed on average the report of 10 items (range: 6—13) with 4 items reported in more than 95% of papers, while 2 were poorly reported.

Limitations The response rate of participants was 20.7%.

Conclusions A quality assessment checklist was developed for improved critical appraisal of the reporting of cytotoxic, dose-finding phase I oncology trials. This may be a first step toward a minimum standard of quality measures for all phase I clinical trial reports. Clinical Trials 2008; 5: 478—485. http://ctj.sagepub.com

Background The first applications of cluster randomized trials with three instead of two levels are beginning to appear in health research, for instance, in trials where different strategies to implement best-practice guidelines are compared. In such trials, the strategy is implemented in health care units (`clusters') and aims at changing the behavior of health care professionals working in this unit (`subjects'), while the effects are measured at patient level (`evaluations').

Purpose To guide the choice of number of clusters, number of subjects per cluster, and number of evaluations per subject.

Methods We derive a sample size formula and investigate the influence of sample allocation on power or number of clusters required.

Results The required sample size is the product of the sample size in absence of correlation and two variance inflation factors (VIFs) that describe the clustering of evaluations within subjects and of subjects within cluster, respectively. Because each VIF is expressed in terms of an interpretable Pearson correlation, subject matter knowledge can be incorporated. Moreover, these Pearson's correlations are related to intracluster correlations (ICCs) from comparable, but 2-level cluster randomized trials. Formulas are obtained to guide the sample allocation (number of clusters, subjects, and evaluations) for minimizing total sample size, minimizing the number of clusters, or maximizing power given a budget constraint.

Limitations Empirical estimates of variance components or ICCs from 3-level cluster trials are scarce which limits reliably powering.

Conclusions When parameterized in terms of Pearson correlations, the two variance inflation factors give quantitative insight into the impact of the number of clusters, subjects and evaluations on power. Moreover, subject matter knowledge as well as ICCs from 2-level cluster randomized trials can be incorporated in the sample size calculation, when empirical estimates of variance components or ICCs from a pilot or comparable 3-level study are lacking. Clinical Trials 2008; 5: 486—495. http://ctj.sagepub.com

Background Verification bias occurs when the percentage of subjects receiving disease verification differs according to the test result. Statistical adjustment yields unbiased sensitivity and specificity under a missing at random (MAR) assumption.

Purpose To use an example from an international study to show how the assumptions needed for unbiased statistical adjustment for verification bias can be undermined by conditions on the ground, and that accuracy of estimates is also compromised by too low a sampling fraction of subjects who test negative.

Methods A study in Zimbabwe assessed the accuracy of a screening test for cervical cancer screening, visual inspection with acetic acid (VIA). The study was conducted in two phases, Phase I, where only 10% of subjects with negative tests received verification, and Phase II, in which nearly all subjects were verified. Unadjusted, simple- and covariate-adjusted estimates were compared to investigate factors affecting differences. Bootstrap simulations were used to illustrate the effect of varying test negative sampling fractions.

Results Phase I unadjusted sensitivity and specificity were 0.66 (0.61—0.70) and 0.34 (0.31—0.36), respectively. Simple-weighted adjusted estimators accounting only for VIA status were 0.20 (0.17—0.23) and 0.80 (0.78—0.81), respectively, suggesting the test to be useless. It was found that verification (colposcopy) capacity in-country had been exceeded, and that random selection of test negative patients for colposcopy had been compromised. Phase II estimates of sensitivity and specificity were 0.77 and 0.64, respectively. With 9% disease prevalence, a VIA test-negative sampling fraction >50% was necessary for the confidence intervals for sensitivity to have more than a 90% probability of including the true value. Limitations Phase I statistical adjustment was not made for MAR deviations unexplained by the two auxiliary factors, Pap results and STD history. Adjustment was not possible for other unmeasured co-factors.

Conclusions While there are standard formulae for correcting for verification bias, these will be biased if the MAR assumption is not met, which can occur through the actions of study personnel or subjects. Design of such studies in low resource environments needs to either require 100% verification, or employ procedures ensuring that the sample of test negatives who receive verification is indeed random. In addition, required test negative sampling fractions need to incorporate information on both disease prevalence and overall sample size. Clinical Trials 2008; 5: 496—503. http://ctj.sagepub.com

Background Healthy People 2010 underscores the relevance of eliminating health disparities. Thus, it is paramount to create interventions that promote health for all individuals.

Purpose This study examined differences in rates of and reasons for ineligibility among non-Hispanic blacks and whites in a randomized controlled physical activity intervention study.

Methods Participants (1245 adults) responded to community advertising for the research study. Eligibility at the four pre-randomization assessment sessions was determined by self-reported medical information, resting EKG, 7-Day Physical Activity Recall, fitness test and Stage of Change. We used t-tests to examine the rates of eligibility among participant subgroups.

Results Blacks had higher rates of overall ineligibility (86.9%) than whites (75.1%; p < 0.01) and were more likely to be ineligible due to lack of interest or no-show at a pre-randomization appointment (35.4% vs. 24.3%; p < 0.01). Blacks were more likely to be ineligible for medical reasons after the telephone screen (16.3% vs. 7.8%; p = 0.01).

Limitations This study did not use a random sampling of potential participants from each of the racial/ethnic groups and thus, there is the potential for selection bias.

Conclusions Blacks were more likely to choose not to enroll in the study due to a lack of interest, but had similar rates of overall medical ineligibility to whites. This highlights the importance of strategies that enhance interest among blacks, who initially respond to recruitment advertising. Clinical Trials 2008; 5: 504—516. http://ctj.sagepub.com

Background Transplantation of cells derived through the manipulation of pluripotent stem cells may involve great uncertainty and the possibility of serious risks.

Purpose To develop guidelines for the ethical conduct of clinical trials using such stem cells.

Methods Review of literature on clinical trials ethics and clinical applications of stem cells; critical deliberation on potential guidelines.

Results Such transplantation should be allowed in clinical practice only after clinical trials demonstrate efficacy and safety. These clinical trials should follow ethical principles that guide all clinical research. Additional requirements to strengthen trial design, coordinate scientific and ethics review, verify that participants understand key features of the trial, and ensure publication of findings are also warranted because of the highly innovative nature of the intervention, limited experience in humans, and the high hopes of patients who have no alternative effective treatments.

Limitations These recommendations will need to be modified in light of actual experience with stem cell clinical trials.

Conclusions These recommendations will help guarantee that the efficacy and safety of innovative stem cell interventions will be rigorously established, while also protecting study participants. Clinical Trials 2008; 5: 517—522. http:// ctj.sagepub.com

Background Observational studies and small clinical trials suggested that hormone replacement therapy (HRT) decreases risk of cognitive loss and Alzheimer's disease (AD) in postmenopausal women and may have value in primary prevention. Purpose A clinical trial was designed to determine if HRT delays AD or memory loss. This report describes the rationale and original design of the trial and details extensive modifications that were required to respond to unanticipated findings that emerged from other studies during the course of the trial.

Methods The trial was designed as a multi-center, placebo-controlled primary prevention trial for women 65 years of age or older with a family history of dementia. Recruitment from local sites was supplemented by centralized efforts to use names of Medicare beneficiaries. Inclusion criteria included good general health and intact memory functioning. Participants were randomized to HRT or placebo in a 1:1 ratio. Assignment was stratified by hysterectomy status and site. The primary outcomes were incident AD and memory decline on neuropsychological testing.

Results Enrollment began in March 1998. In response to the Women's Health Initiative (WHI) May 2002 report of increased incidence of heart disease, stroke, pulmonary embolism, and breast cancer among women randomized to HRT, participants were re-consented with a revised consent form. Procedural modifications, including discontinuation of study medication and a modification of the planned primary outcome based on a final enrollment below the target enrollment (N = 477), were enacted in response to the subsequent WHI Memory Study report of increased risk of dementia and poorer cognitive function with HRT. The mean length of treatment exposure prior to discontinuation was 2.14 years. Participants' mean age at baseline was 72.8; mean education was 14.2 years. Minority participation was 19% and 34% had a hysterectomy. The study continues to follow these participants for a total of 5 years blind to the original medication assignment.

Limitations Results reported from the WHI during the course of this study mandated extensive procedural modifications, including discontinuing recruitment before completion and halting study medication. Alternative strategies for study redesign that were considered are discussed. Clinical Trials 2008; 5: 523—533. http://ctj.sagepub.com

Background Surgery is of increasing importance in the treatment and outcome of diseases of the pancreas worldwide. The incidence of pancreatic cancer (7—11/ 100,000 per year) has risen over the last years and surgical resection remains the only option for definite cure. Twenty-five percent of all resections are left of the superior mesenteric vein (distal pancreatectomy) and the appropriate closure technique for the pancreatic remnant remains unclear. Pancreatic fistulas are the most common (0—40%) and relevant postoperative complication. The optimal surgical strategy for pancreatic resection needs to be identified from the large number of surgical procedures available today.

Purpose To evaluate the effectiveness of the two most common surgical techniques for distal pancreatectomy: stapler versus hand-sewn closure of the pancreatic remnant.

Methods In order to account for the uncertainty and clinical heterogeneity in the management of the pancreatic remnant following distal pancreatectomy, a study protocol is developed on the basis of a retrospective survey of patients in a center of excellence for pancreatic surgery and a systematic review with meta-analysis.

Results The DISPACT trial is a multicentered, randomized, controlled and patient-and observer-blinded trial using a two-group parallel group-sequential superiority design to compare the two techniques mentioned above. It will include approximately 336 randomized patients at up to 20 centers of excellence in pancreatic surgery, who are undergoing elective distal pancreatectomy for resectable benign, malign, and neuroendocrine tumors, chronic pancreatitis and pseudocysts of the pancreatic body and tail. The combination of the rate of postoperative pancreatic fistula and mortality will be evaluated as the primary endpoint. In addition, a set of general and surgical parameters will be analyzed. Pre-specified treatment manuals and continuous intra-operative (photo-documentation of surgical procedures and blinded evaluation thereafter) and on-site monitoring will assure that the treatment of the study patients conforms to protocol and will minimize clinical heterogeneity. Due to uncertainties about the effect sizes of the primary endpoint, an a priori planned interim analysis of the primary endpoint will be conducted after 224 evaluable patients are selected in order to reassess the initially planned sample size. Limitations Since pre-existing evidence was limited our initial sample size calculation is based on uncertain assumptions and may need to be modified in a planned interim analysis. Moreover, since surgical experience remains a potential confounder in surgical trials, learning curve bias has to be taken into account when analyzing the results. Given the participating trial sites, standardization of peri-and postoperative treatment represents a major issue of trial conduct.

Conclusions A group-sequential study design accounts for the uncertainty of pre-existing evidence. Also, standardization of surgical and postoperative care and blinded outcome assessment as well as adjustment for varying surgical expertise will contribute to a high validity and generalizability of the results. Clinical Trials 2008; 5: 534—545. http://ctj.sagepub.com

Background The Weight Loss Maintenance Trial (WLM) is a multi-center, randomized, controlled trial that compares the effects of two 30-month maintenance interventions, i.e., Personal Contact (PC) and Interactive Technology (IT) to a self-directed usual care control group (SD), in overweight or obese individuals who are at high risk for cardiovascular disease.

Purpose This paper provides an overview of the design and methods, and design considerations and lessons learned from this trial.

Methods All participants received a 6-month behavioral weight loss program consisting of weekly group sessions. Participants who lost 4 kg were randomized to one of three conditions (PC, IT, or SD). The PC condition provided monthly contacts with an interventionist primarily via telephone and quarterly face-to-face visits. The IT condition provided frequent, individualized contact through a tailored, website system. Both the PC and IT maintenance programs encouraged the DASH dietary pattern and employed theory-based behavioral techniques to promote maintenance.

Results Design considerations included choice of study population, frequency and type of intervention visits, and choice of primary outcome. Overweight or obese persons with CVD risk factors were studied. The pros and cons of studying this population while excluding others are presented. We studied intervention contact strategies that made fewer demands on participant time and travel, while providing frequent opportunities for interaction. The primary outcome variable for the trial was change in weight from randomization to end of follow-up (30 months). Limitations Limits to generalizability are discussed. Individuals in need of weight loss strategies may have been excluded due to barriers associated with internet use. Other participants may have been excluded secondary to a comorbid condition.

Conclusions This paper highlights the design and methods of WLM and informs readers of discussions of critical issues and lessons learned from the trial. Clinical Trials 2008; 5: 546—556. http://ctj.sagepub.com