Important Baseline Covariates Research Articles

Case-control matching-guided exposure-efficacy relationship for avelumab in patients with urothelial carcinoma.

Exposure-response (E-R) analyses are an integral component of understanding the benefit/risk profile of novel oncology therapeutics. These analyses are typically conducted using data from the treatment arm to characterize the relationship between drug exposure (low vs. high) and efficacy or safety outcomes. For example, outcomes of patients with lower exposure in the treatment arm (e.g., Q1) might be compared to outcomes of those with higher drug exposure (Q2, Q3, and Q4). Outcomes from the lowest exposure quartile may be also compared to the control arm to evaluate whether the Q1 subgroup derived clinical benefit. However, the sample size and the distribution of patient baseline characteristics and disease risk factors are not balanced in such a comparison (Q1 vs. control), which may bias the analysis and causal interpretation of clinical benefit in the Q1 subgroup. Herein, we report the use of case-control matching to account for this bias and better understand the E-R relationship for avelumab in urothelial carcinoma, a PD-L1 inhibitor approved for the treatment of several cancers. Data from JAVELIN-100 was utilized which is a phase III study of avelumab in first-line maintenance treatment in patients with urothelial carcinoma; this clinical study demonstrated superiority of avelumab versus best-supportive care leading to approval in the United States, Europe, and other countries. A post hoc case-control matching method was implemented to compare the efficacy outcome between Q1 avelumab subgroup and matched patients extracted from the control arm with similar baseline characteristics, which showed a clinically relevant difference in overall survival in favor of the Q1 avelumab subgroup. This analysis demonstrates the importance of accounting for imbalance in important baseline covariates when comparing efficacy outcomes between subgroups within the treatment arm versus the control arm.

Assessing the Impacts of Cluster Effects and Covariate Imbalance in Cluster Randomized Equivalence Trials

Equivalence tests establish whether treatments are similar in their intended outcomes. This is in contrast to superiority tests, which establish whether a new treatment is better than a standard treatment or placebo. Few equivalence trials have employed a cluster randomized design, but they are subject to some of the same analysis pitfalls that are common to superiority trials—namely, a failure to adjust for either cluster effects or covariate imbalances resulting from randomization. Using real and simulated data from a cluster randomized trial comparing exercise protocols among U.S. Army soldiers, this study empirically demonstrates the consequences for power and Type I error rates when either or both of these effects have been ignored in the analysis. Analysis of real trial data showed that equivalence test outcomes can change depending on whether appropriate adjustments are applied. Simulations demonstrated that failing to adjust for important baseline covariates severely reduces statistical power, and failing to adjust for cluster effects increases the risk of false declarations of equivalence. As cluster randomized designs are increasingly employed for equivalence trials, analysts must be aware of the importance of adjusting for cluster effects and covariate imbalances to avoid false conclusions.

Assessing race and ethnicity differences in outcomes based on GDMT and target NT-proBNP in patients with heart failure with reduced ejection fraction: An analysis of the GUIDE-IT study

BackgroundThe GUIDE-IT trial was, a multicenter, randomized, parallel group, unblinded study that randomized patients to having heart failure therapy titrated to achieve an NT-proBNP <1000 pg/mL or to usual clinical care. Methods and resultsWe performed pre-specified subgroup analysis to look for the race and ethnicity-based differences in clinical outcomes of patients who were able to achieve GDMT or target NT-proBNP concentration of ≤1000 pg/mL at 90 days of follow-up. There were 894 patients enrolled in GUIDE-IT study. Of these, 733 participants had available data on 90-day guideline directed triple therapy and 616 on NT-proBNP. 35% of the patients were Black and 6% were Hispanic. Black patients were younger, had more comorbidities, lower EF, and higher NYHA class compared with non-Black.Adjusting for 90-day NT-proBNP and important baseline covariates, Black patients were at a higher risk than non-Black patients for HF hospitalization [HR, 2.19; 95% CI, 1.51–3.17; p < 0.0001], but at a similar risk for mortality [HR, 0.85.; 95% CI, 0.44–1.66; p = 0.64]. Similar results were seen adjusting for 90-day GDMT [HF hospitalization: Black vs non-Black, HR: 1.97; 1.41–2.77, P < 0.0001; mortality: HR: 0.70; 0.39–1.26, p = 0.23]. There were no significant differences between Hispanic and non-Hispanic patients with respect to heart failure hospitalization, cardiovascular or all-cause mortality. Over the study period, Black and Hispanic patients experienced smaller changes in physical function and quality of life as measured by the Kansas City Cardiomyopathy Questionnaire overall score. ConclusionCompared to non-Black patients, Black patients in GUIDE-IT study had a higher risk of heart failure hospitalization, but a comparable risk of mortality, despite improved use of GDMT and achievement of similar biomarker targets.

Evaluating performance of covariate-constrained randomization (CCR) techniques under misspecification of cluster-level variables in cluster-randomized trials

Covariate constrained randomization (CCR) is a method of controlling imbalance in important baseline covariates in cluster-randomized trials (CRT). We use simulated CRTs to investigate the performance (control of imbalance) of CCR relative to simple randomization (SR) under conditions of misspecification of the cluster-level variable used in the CCR algorithm.We use data from a Patient-Centered Outcomes Research Institute (PCORI)-funded CRT evaluating the Mothers and Babies (MB) intervention (AD-1507-31,473). CCR methodology was used in the MB study to control imbalance in, among other baseline variables, the percent minority (i.e., non-White) participants at each study site. Simulation schemes explored variation in degree of misspecification in the baseline covariate of interest, and include correct report, observed misspecification, and a range of simulated misspecification for intervals within and beyond that observed in the MB study. We also consider three within-site sample size scenarios: that observed in the MB study, small (mean 10) and large (mean 50). Simulations at every level of baseline covariate misspecification suggest that use of the CCR strategy provides between-arm imbalance that is simultaneously lower and less variable, on average, than that produced from the SR strategy. We find that the gains to using CCR over SR are nearly twice as high with accurate reporting (Δ = −5.33) compared to the observed study-level misspecification (Δ = −3.03). Although CCR still outperforms SR as the level of misspecification increases, the gains to using CCR over SR decrease; thus, every effort should still be made to obtain high-quality baseline data.

Digital marketing of smokeless tobacco: A longitudinal analysis of exposure and initiation among young adults

ObjectivesThis study examines the relationships between recall of exposure to digital marketing of smokeless tobacco, via the internet and social media, and subsequent initiation of smokeless tobacco use at one-year follow-up, among young adult never users of smokeless tobacco in Texas. MethodsData were from waves 6 (Spring 2017) and 7 (Spring 2018) of the Marketing and Promotions Across Colleges in Texas Study (Project M-PACT); a longitudinal study of two- and four-year Texas college students. Participants were 2731 young adult never smokeless tobacco users (ages 20–32) with complete data at both assessment periods. A multi-level, multiple logistic regression model was applied, accounting for school clustering, to examine the relationship between recall of exposure to digital marketing of smokeless tobacco at baseline (wave 6) and smokeless tobacco use initiation at one-year follow-up (wave 7). Analyses controlled for important baseline covariates (socio-demographic factors, other marketing exposure, other tobacco product use). ResultsOverall, 14.6% of never smokeless tobacco users reported exposure to smokeless tobacco marketing via digital media. Exposure to digital marketing of smokeless tobacco at baseline was associated with greater odds of smokeless tobacco use initiation among young adult never users (AOR: 2.14; 95%CI: 1.12 – 4.06) at one-year follow-up. ConclusionsFindings reveal exposure to smokeless tobacco marketing via digital media was common. Further, there appears to be a strong relationship between this exposure and subsequent smokeless tobacco use among young adult never smokeless tobacco users.

Data extraction for epidemiological research (DExtER): a novel tool for automated clinical epidemiology studies

The use of primary care electronic health records for research is abundant. The benefits gained from utilising such records lies in their size, longitudinal data collection and data quality. However, the use of such data to undertake high quality epidemiological studies, can lead to significant challenges particularly in dealing with misclassification, variation in coding and the significant effort required to pre-process the data in a meaningful format for statistical analysis. In this paper, we describe a methodology to aid with the extraction and processing of such databases, delivered by a novel software programme; the “Data extraction for epidemiological research” (DExtER). The basis of DExtER relies on principles of extract, transform and load processes. The tool initially provides the ability for the healthcare dataset to be extracted, then transformed in a format whereby data is normalised, converted and reformatted. DExtER has a user interface designed to obtain data extracts specific to each research question and observational study design. There are facilities to input the requirements for; eligible study period, definition of exposed and unexposed groups, outcome measures and important baseline covariates. To date the tool has been utilised and validated in a multitude of settings. There have been over 35 peer-reviewed publications using the tool, and DExtER has been implemented as a validated public health surveillance tool for obtaining accurate statistics on epidemiology of key morbidities. Future direction of this work will be the application of the framework to linked as well as international datasets and the development of standardised methods for conducting electronic pre-processing and extraction from datasets for research purposes.

Graphing and reporting heterogeneous treatment effects through reference classes

BackgroundExploration and modelling of heterogeneous treatment effects as a function of baseline covariates is an important aspect of precision medicine in randomised controlled trials (RCTs). Randomisation generally guarantees the internal validity of an RCT, but heterogeneity in treatment effect can reduce external validity. Estimation of heterogeneous treatment effects is usually done via a predictive model for individual outcomes, where one searches for interactions between treatment allocation and important patient baseline covariates. However, such models are prone to overfitting and multiple testing and typically demand a transformation of the outcome measurement, for example, from the absolute risk in the original RCT to log-odds of risk in the predictive model.MethodsWe show how reference classes derived from baseline covariates can be used to explore heterogeneous treatment effects via a two-stage approach. We first estimate a risk score which captures on a single dimension some of the heterogeneity in outcomes of the trial population. Heterogeneity in the treatment effect can then be explored via reweighting schemes along this axis of variation. This two-stage approach bypasses the search for interactions with multiple covariates, thus protecting against multiple testing. It also allows for exploration of heterogeneous treatment effects on the original outcome scale of the RCT. This approach would typically be applied to multivariable models of baseline risk to assess the stability of average treatment effects with respect to the distribution of risk in the population studied.Case studyWe illustrate this approach using the single largest randomised treatment trial in severe falciparum malaria and demonstrate how the estimated treatment effect in terms of absolute mortality risk reduction increases considerably in higher risk strata.Conclusions‘Local’ and ‘tilting’ reweighting schemes based on ranking patients by baseline risk can be used as a general approach for exploring, graphing and reporting heterogeneity of treatment effect in RCTs.Trial registrationISRCTN clinical trials registry: ISRCTN50258054. Prospectively registered on 22 July 2005.

2488. Virologic Failure in ART Naïve Patients Initiating on a Dolutegravir or Elvitegravir-Based Regimen

BackgroundRobust pharmacoeconomic modeling is dependent on high quality inputs, preferably from randomized clinical trials (RCT), but not all needed head to head comparisons occur in RCTs. We compared virologic outcomes in an antiretroviral (ART) naïve population initiating a dolutegravir (DTG) or elvitegravir (EVG)-based regimen using clinical trial-like criteria.MethodsART-naïve adults, initiating a DTG- or EVG-based regimen and meeting all study eligibility criteria (Figure 1) were identified in the OPERA® Observational Database, a collaboration of HIV caregivers following 100,000+ people living with HIV (PLWH) through electronic medical records. PLWH were followed from the date of first prescription until DTG- or EVG discontinuation, death, or study end (July 31, 2018). The primary outcome was verified (2 consecutive viral load (VL) ≥200 copies/mL or 1 VL ≥200 copies + discontinuation) virologic failure (VF), defined as either failure to achieve suppression (<50 copies/mL) prior to 36 weeks or failure to maintain suppression once achieved. Survival analyses were conducted with Kaplan–Meier methods and multivariate Cox Proportional Hazards modeling.ResultsA total of 1,688 (DTG) and 2,537 (EVG) met all eligibility criteria. Median (IQR) length of follow-up in the DTG users was 21 months (14–30), in the EVG users was 20 (14–32) months. Figure 2 characterizes baseline demographic/clinical characteristics. Figures 3 and 4 depict Kaplan–Meier curves and Cox model results, respectively. VF was experienced by 8.2% DTG and 10.9% EVG initiators at a rate (95% CI) per 1,000 person-years of 40.2 (33.8, 47.8) and 51.3 (45.3, 58.1), respectively. Younger age (18–25), being African American, having a baseline CD4 count ≤ 200, or having a government-based payer (ADAP, Ryan White, Medicaid, or Medicare) at baseline were associated with a significant (P < 0.05), increased hazard of VF. Initiating on DTG or initiating therapy with a lower baseline VL was associated with a significant, reduced hazard of VF. Compared with DTG, the adjusted hazard ratio for VF was 1.29 (95% CI: 1.02, 1.63) for EVG.ConclusionAmong ART-naïve patients, DTG users were significantly less likely to experience virologic failure than EVG users after adjustment for important baseline covariates. Disclosures All authors: No reported disclosures.

Relative risk of all-cause mortality in patients with nontuberculous mycobacterial lung disease in a US managed care population.

The risk of all-cause mortality of nontuberculous mycobacterial lung disease (NTMLD) in the United States (US) population is not well established. This study aims to assess the public health burden of NTMLD in the US by comparing the relative risk of all-cause mortality in the NTMLD population with an age- and sex-matched cohort from the general population. Patients with physician claims for NTMLD (ICD-9 0.031; ICD-10 A31.0) were identified between 2007 and 2016 from a large US national managed care insurance plan covering approximately 15-18 million members annually. A control group with no NTMLD ICD-9 or 10 codes was randomly selected from the general population and matched 3:1 to the NTMLD sample according to birth year, gender, and insurance benefit coverage. The date of first NTMLD diagnosis of each patient was assigned to the matched controls as the index date. The Cox proportional hazard method compared survival between cohorts, adjusting for demographic factors and baseline comorbidities. A total of 2005 patients with NTMLD and 6014 controls were identified, with a mean follow-up duration of 3.4 years and 3.7 years, respectively. The NTMLD group had substantially higher proportions of patients with asthma (23.3% versus 3.5%), bronchiectasis (36.5% versus 0.1%), COPD (52.0% versus 5.9%), arrhythmia (22.6% versus 6.5%), coronary artery disease (18.5% versus 6.6%), heart failure (11.9% versus 4.1%), and cancer (18.5% versus 5.0%). The unadjusted rate of all-cause mortality from the index date was 20.7 per 1000 person-years in the NTMLD group vs 5.6 per 1000 person-years in the control group (rate ratio = 3.73; 95% CI: 2.93-4.75). Multivariable Cox regression, adjusted for the above variables as well as all other important baseline covariates, showed a doubling risk of all-cause mortality (hazard ratio [HR] = 2.06; CI: 1.52-2.79; P < 0.001) in the NTMLD vs control group. All-cause mortality, adjusted for other factors, more than doubled with NTMLD compared with an age-sex-matched control group in a large US national managed care insurance plan.

Associations between preoperative OCT parameters and visual outcome 3months postoperatively in patients undergoing vitrectomy for idiopathic epiretinal membrane.

To evaluate the effect of preoperative optical coherence tomography (OCT) features on postoperative visual acuity and change in acuity after epiretinal membrane (ERM) surgery. This was a retrospective consecutive case series of patients with symptomatic idiopathic ERMs who underwent vitrectomy and membrane peel between 2008 and 2013. The main outcomes were postoperative letter score and change in letter score. The association between visual acuity and OCT features was estimated using linear regression, correcting for important baseline covariates. OCT features included central foveal thickness, maximum retinal thickness, retinal cross-sectional area, central photoreceptor thickness, central epiretinal membrane, intraretinal cyst presence and location, retinal contraction, foveal ellipsoid zone and external limiting membrane integrity, and pseudoholes. A total of 79 patients were included in the study. Mean visual acuity was 65.0 letters preoperatively and 68.9 letters postoperatively, with average follow-up of 92days. Better postoperative visual acuity was associated with the presence of preoperative retinal contraction (6.4 letters P = 0.027) and lower preoperative maximum retinal thickness values (0.34 letters per 10μm reduction, P = 0.003). Change in visual acuity was associated with preoperative foveal ellipsoid zone attenuation (6.6 letters more improvement when present, P = 0.013). Preoperative pseudoholes were associated with a smaller change in visual acuity (7.4 letters less improvement, P = 0.034). No other OCT features were significant, including intraretinal cyst presence and location. Better preoperative visual acuity was associated with better postoperative acuity, but less change in visual acuity postoperatively. Preoperative maximum retinal thickness and retinal contraction were predictive of better postoperative visual acuity, while patients with a preoperative attenuated foveal ellipsoid zone were most likely to improve. Patients with preoperative pseudoholes had slightly worse visual acuity postoperatively. Although better preoperative vision was associated with better postoperative vision, it was associated with less change in visual acuity, emphasizing the importance of careful consideration of the desired surgical goals before advising surgery for ERM.

Factors affecting the impact of covariate adjustment for binary outcomes in randomised clinical trials (RCTS): a simulation study

There is lack of consensus on whether analysis of RCT data should adjust for important baseline covariates. For binary outcomes, the estimated treatment effect can differ when logistic regression is carried out, even when the covariates are balanced between treatment groups. Simulation studies for RCTs with binary outcomes have shown that adjusting for a single baseline prognostic covariate increases the power to detect treatment effects. The impact on the size of the treatment effect was only notable when a strong prognostic covariate was included. The impact of adjustment for two covariates with different prognostic strengths and directions of the prognostic relationship to the outcome is not known. We present a detailed examination of factors affecting the impact of adjusted analysis. In particular, an adjusted logistic regression model with two covariates (1 continuous and 1 binary) are considered simultaneously. The simulation study addresses different levels of event rate and treatment effect, different distributions for the binary and continuous variables, variable prognostic strength of the covariates, and correlation between covariates. These simulations suggest that adjustment will only have a notable effect in extreme scenarios such as, treatment effect is very large and covariates are highly prognostic. We found that the relative difference between the unadjusted and adjusted odds ratios could be larger than 50%, though only under these scenarios. Otherwise the impact of adjustment is small.

Up, not down: The age curve in happiness from early adulthood to midlife in two longitudinal studies.

Happiness is an important indicator of well-being, and little is known about how it changes in the early adult years. We examined trajectories of happiness from early adulthood to midlife in 2 Canadian longitudinal samples: high school seniors followed from ages 18-43 and university seniors followed from ages 23-37. Happiness increased into the 30s in both samples, with a slight downturn by age 43 in the high school sample. The rise in happiness after high school and university remained after controlling for important baseline covariates (gender, parents' education, grades, self-esteem), time-varying covariates known to be associated with happiness (marital status, unemployment, self-rated physical health), and number of waves of participation. The upward trend in happiness runs counter to some previous cross-sectional research claiming a high point in happiness in the late teens, decreasing into midlife. As cross-sectional designs do not assess within-person change, longitudinal studies are necessary for drawing accurate conclusions about patterns of change in happiness across the life span.

Inconsistent treatment estimates from mis-specified logistic regression analyses of randomized trials.

When the difference between treatments in a clinical trial is estimated by a difference in means, then it is well known that randomization ensures unbiassed estimation, even if no account is taken of important baseline covariates. However, when the treatment effect is assessed by other summaries, for example by an odds ratio if the outcome is binary, then bias can arise if some covariates are omitted, regardless of the use of randomization for treatment allocation or the size of the trial. We present accurate closed-form approximations for this asymptotic bias when important normally distributed covariates are omitted from a logistic regression. We compare this approximation with ones in the literature and derive more convenient forms for some of these existing results. The expressions give insight into the form of the bias, which simulations show is usable for distributions other than the normal. The key result applies even when there are additional binary covariates in the model.

Venous Thromboembolism and Cardiovascular Risk: Results from the NAVIGATOR Trial

BackgroundContemporary studies suggest an association between venous thromboembolism and a higher incidence of major cardiovascular events, mostly attributed to arterial atherothrombosis. Using data from the Nateglinide and Valsartan in Impaired Glucose Tolerance Outcomes Research (NAVIGATOR) trial, we assessed the association of venous thromboembolism with major cardiovascular events. MethodsIn NAVIGATOR, patients with impaired glucose tolerance were randomly allocated to receive valsartan or placebo and nateglinide or placebo in addition to lifestyle modification. Baseline characteristics and prior history of venous thromboembolism were assessed. After adjusting for important baseline covariates, Cox proportional hazards regression models were used to assess the association between venous thromboembolism and major cardiovascular outcomes. ResultsOf the 9306 patients enrolled, 129 (1.4%) had a history of venous thromboembolism. Patients with venous thromboembolism were older, more frequently white and female, and had a higher body mass index. Patients with venous thromboembolism had higher 5-year event rates for the composite of death, myocardial infarction, and stroke, as compared with patients without venous thromboembolism (10.7% vs 5.9%; P < .001; adjusted hazard ratio 2.12; 95% confidence interval, 1.36-3.31; P = .001). ConclusionIn patients with impaired glucose tolerance at high risk for cardiovascular events, the prevalence of venous thromboembolism was rare but associated with worse long-term cardiovascular outcomes, including arterial events. Venous thromboembolism is a marker of risk, and attention should be paid to this high-risk group of patients.

Estimation of group means when adjusting for covariates in generalized linear models.

Generalized linear models are commonly used to analyze categorical data such as binary, count, and ordinal outcomes. Adjusting for important prognostic factors or baseline covariates in generalized linear models may improve the estimation efficiency. The model-based mean for a treatment group produced by most software packages estimates the response at the mean covariate, not the mean response for this treatment group for the studied population. Although this is not an issue for linear models, the model-based group mean estimates in generalized linear models could be seriously biased for the true group means. We propose a new method to estimate the group mean consistently with the corresponding variance estimation. Simulation showed the proposed method produces an unbiased estimator for the group means and provided the correct coverage probability. The proposed method was applied to analyze hypoglycemia data from clinical trials in diabetes.

Clinical characteristics and outcomes associated with the natural history of early repolarization in a young, biracial cohort followed to middle age: the Coronary Artery Risk Development in Young Adults (CARDIA) study.

Early repolarization (ER), a common electrocardiographic phenotype, has been associated with increased mortality risk in middle-aged adults. Data are sparse on long-term follow-up and outcomes associated with ER in younger adults. We prospectively examined 5039 participants (mean age, 25 years at baseline, 40% black) from the Coronary Artery Disease Risk in Adults (CARDIA) cohort for 23 years. Twelve-lead ECGs were recorded and analyzed at years 0, 7, and 20 and coded as definite or probable ER using a standardized algorithm. Cox regression was used, and models were adjusted for important baseline and clinical covariates. Kaplan-Meier curves were created for presence of ER and total mortality and cardiovascular mortality. Participants with ER were more likely to be black, male, smoke, have higher systolic blood pressure, lower heart rate and body mass index, higher exercise duration, and longer PR, QRS, and QT intervals. ER was associated with total mortality (hazard ratio, 1.77; confidence interval, 1.38-2.28; P<0.01) and cardiovascular mortality (hazard ratio, 1.59; confidence interval, 1.01-2.50; P=0.04) in unadjusted analyses, but adjustment for age, sex, and race attenuated associations almost completely. Sex-race stratified analyses showed no significant associations between ER and outcome for any of the subgroups except blacks. The presence of ER at any time point during 23 years of follow-up was not associated with adverse outcomes. Black race and male sex confound the unadjusted association of ER and outcomes, with no race-sex interactions noted. Additional studies are necessary to understand the factors associated with heightened risk of death in those who maintain ER into and beyond middle age.

Balancing continuous covariates based on Kernel densities

The balance of important baseline covariates is essential for convincing treatment comparisons. Stratified permuted block design and minimization are the two most commonly used balancing strategies, both of which require the covariates to be discrete. Continuous covariates are typically discretized in order to be included in the randomization scheme. But breakdown of continuous covariates into subcategories often changes the nature of the covariates and makes distributional balance unattainable. In this article, we propose to balance continuous covariates based on Kernel density estimations, which keeps the continuity of the covariates. Simulation studies show that the proposed Kernel-Minimization can achieve distributional balance of both continuous and categorical covariates, while also keeping the group size well balanced. It is also shown that the Kernel-Minimization is less predictable than stratified permuted block design and minimization. Finally, we apply the proposed method to redesign the NINDS trial, which has been a source of controversy due to imbalance of continuous baseline covariates. Simulation shows that imbalances such as those observed in the NINDS trial can be generally avoided through the implementation of the new method.

Minimal sufficient balance-a new strategy to balance baseline covariates and preserve randomness of treatment allocation.

In many clinical trials, baseline covariates could affect the primary outcome. Commonly used strategies to balance baseline covariates include stratified constrained randomization and minimization. Stratification is limited to few categorical covariates. Minimization lacks the randomness of treatment allocation. Both apply only to categorical covariates. As a result, serious imbalances could occur in important baseline covariates not included in the randomization algorithm. Furthermore, randomness of treatment allocation could be significantly compromised because of the high proportion of deterministic assignments associated with stratified block randomization and minimization, potentially resulting in selection bias. Serious baseline covariate imbalances and selection biases often contribute to controversial interpretation of the trial results. The National Institute of Neurological Disorders and Stroke recombinant tissue plasminogen activator Stroke Trial and the Captopril Prevention Project are two examples. In this article, we propose a new randomization strategy, termed the minimal sufficient balance randomization, which will dually prevent serious imbalances in all important baseline covariates, including both categorical and continuous types, and preserve the randomness of treatment allocation. Computer simulations are conducted using the data from the National Institute of Neurological Disorders and Stroke recombinant tissue plasminogen activator Stroke Trial. Serious imbalances in four continuous and one categorical covariate are prevented with a small cost in treatment allocation randomness. A scenario of simultaneously balancing 11 baseline covariates is explored with similar promising results. The proposed minimal sufficient balance randomization algorithm can be easily implemented in computerized central randomization systems for large multicenter trials.

Verteporfin Photodynamic Therapy Cohort Study: Report 1: Effectiveness and Factors Influencing Outcomes

To compare the visual outcomes after verteporfin photodynamic therapy (VPDT) administered in routine clinical practice with those observed in the Treatment of Age-related macular degeneration with Photodynamic therapy (TAP) trials and to quantify the effects of clinically important baseline covariates on outcome. A prospective longitudinal study of patients treated with VPDT in 45 ophthalmology departments in the United Kingdom with expertise in the management of neovascular age-related macular degeneration (nAMD). Patients with wholly or predominantly classic choroidal neovascularization (CNV) of any cause with a visual acuity >or=20/200 in the eye to be treated. Refracted best-corrected visual acuity (BCVA) and contrast sensitivity were measured in VPDT-treated eyes at baseline and subsequent visits. Eyes were retreated at 3 months if CNV was judged to be active. Baseline angiograms were graded to quantify the percentages of classic and occult CNV. Treated eyes were categorized as eligible or ineligible for TAP, or unclassifiable. Best-corrected visual acuity and contrast sensitivity during 1 year of follow-up after initial treatment. A total of 7748 treated patients were recruited. Data from 4043 patients with a diagnosis of nAMD were used in the present analysis. Reading center determination of lesion type showed that 87% were predominantly classic CNV. Eyes received 2.4 treatments in year 1 and 0.4 treatments in year 2. Deterioration of BCVA over 1 year was similar to that observed in the VPDT arms of the TAP trials and was not influenced by TAP eligibility classification. Best-corrected visual acuity deteriorated more quickly in current smokers; with increasing proportion of classic CNV, increasing age, and better baseline BCVA; and when the fellow eye was the better eye. Patients in the cohort who would have been eligible for the TAP trials demonstrated deterioration in BCVA similar to VPDT-treated TAP participants but with fewer treatments. Clinical covariates with a significant impact on BCVA outcomes were identified.

A substantial and confusing variation exists in handling of baseline covariates in randomized controlled trials: a review of trials published in leading medical journals

Objective Statisticians have criticized the use of significance testing to compare the distribution of baseline covariates between treatment groups in randomized controlled trials (RCTs). Furthermore, some have advocated for the use of regression adjustment to estimate the effect of treatment after adjusting for potential imbalances in prognostically important baseline covariates between treatment groups. Study Design and Setting We examined 114 RCTs published in the New England Journal of Medicine, the Journal of the American Medical Association, The Lancet, and the British Medical Journal between January 1, 2007 and June 30, 2007. Results Significance testing was used to compare baseline characteristics between treatment arms in 38% of the studies. The practice was very rare in British journals and more common in the U.S. journals. In 29% of the studies, the primary outcome was continuous, whereas in 65% of the studies, the primary outcome was either dichotomous or time-to-event in nature. Adjustment for baseline covariates was reported when estimating the treatment effect in 34% of the studies. Conclusions Our findings suggest the need for greater editorial consistency across journals in the reporting of RCTs. Furthermore, there is a need for greater debate about the relative merits of unadjusted vs. adjusted estimates of treatment effect.