One problem with assessing effects of smoking cessation interventions on withdrawal symptoms is that symptoms are affected by whether participants abstain from smoking during trials. Those who enter a randomized trial but do not change smoking behavior might not experience withdrawal related symptoms.
We present a tutorial of how one can use a principal stratification sensitivity analysis to account for abstinence in the estimation of smoking cessation intervention effects. The paper is intended to introduce researchers to principal stratification and describe how they might implement the methods.
We provide a hypothetical example that demonstrates why estimating effects within observed abstention groups is problematic. We demonstrate how estimation of effects within groups defined by potential abstention that an individual would have in either arm of a study can provide meaningful inferences. We describe a sensitivity analysis method to estimate such effects, and use it to investigate effects of a combined behavioral and nicotine replacement therapy intervention on withdrawal symptoms in a female prisoner population.
Overall, the intervention was found to reduce withdrawal symptoms but the effect was not statistically significant in the group that was observed to abstain. More importantly, the intervention was found to be highly effective in the group that would abstain regardless of intervention assignment. The effectiveness of the intervention in other potential abstinence strata depends on the sensitivity analysis assumptions.
We make assumptions to narrow the range of our sensitivity parameter estimates. While appropriate in this situation, such assumptions might not be plausible in all situations.
A principal stratification sensitivity analysis provides a meaningful method of accounting for abstinence effects in the evaluation of smoking cessation interventions on withdrawal symptoms. Smoking researchers have previously recommended analyses in subgroups defined by observed abstention status in the evaluation of smoking cessation interventions. We believe that principal stratification analyses should replace such analyses as the preferred means of accounting for post-randomization abstinence effects in the evaluation of smoking cessation programs.
The paired availability design for historical controls postulated four classes corresponding to the treatment (old or new) a participant would receive if arrival occurred during either of two time periods associated with different availabilities of treatment. These classes were later extended to other settings and called principal strata. Judea Pearl asks if principal stratification is a goal or a tool and lists four interpretations of principal stratification. In the case of the paired availability design, principal stratification is a tool that falls squarely into Pearl's interpretation of principal stratification as “an approximation to research questions concerning population averages.” We describe the paired availability design and the important role played by principal stratification in estimating the effect of receipt of treatment in a population using data on changes in availability of treatment. We discuss the assumptions and their plausibility. We also introduce the extrapolated estimate to make the generalizability assumption more plausible. By showing why the assumptions are plausible we show why the paired availability design, which includes principal stratification as a key component, is useful for estimating the effect of receipt of treatment in a population. Thus, for our application, we answer Pearl's challenge to clearly demonstrate the value of principal stratification.
principal stratification; causal inference; paired availability design
Pearl’s article provides a useful springboard for discussing further the benefits and drawbacks of principal stratification and the associated discomfort with attributing effects to post-treatment variables. The basic insights of the approach are important: pay close attention to modification of treatment effects by variables not observable before treatment decisions are made, and be careful in attributing effects to variables when counterfactuals are ill-defined. These insights have often been taken too far in many areas of application of the approach, including instrumental variables, censoring by death, and surrogate outcomes. A novel finding is that the usual principal stratification estimand in the setting of censoring by death is by itself of little practical value in estimating intervention effects.
principal stratification; causal inference
Pearl (2011) asked for the causal inference community to clarify the role of the principal stratification framework in the analysis of causal effects. Here, I argue that the notion of principal stratification has shed light on problems of non-compliance, censoring-by-death, and the analysis of post-infection outcomes; that it may be of use in considering problems of surrogacy but further development is needed; that it is of some use in assessing “direct effects”; but that it is not the appropriate tool for assessing “mediation.” There is nothing within the principal stratification framework that corresponds to a measure of an “indirect” or “mediated” effect.
causal inference; mediation; non-compliance; potential outcomes; principal stratification; surrogates
This commentary takes up Pearl's welcome challenge to clearly articulate the scientific value of principal stratification estimands that we and colleagues have investigated, in the area of randomized placebo-controlled preventive vaccine efficacy trials, especially trials of HIV vaccines. After briefly arguing that certain principal stratification estimands for studying vaccine effects on post-infection outcomes are of genuine scientific interest, the bulk of our commentary argues that the “causal effect predictiveness” (CEP) principal stratification estimand for evaluating immune biomarkers as surrogate endpoints is not of ultimate scientific interest, because it evaluates surrogacy restricted to the setting of a particular vaccine efficacy trial, but is nevertheless useful for guiding the selection of primary immune biomarker endpoints in Phase I/II vaccine trials and for facilitating assessment of transportability/bridging surrogacy.
principal stratification; causal inference; vaccine trial
Intermediate outcomes are common and typically on the causal pathway to the final outcome. Some examples include noncompliance, missing data, and truncation by death like pregnancy (e.g. when the trial intervention is given to non-pregnant women and the final outcome is preeclampsia, defined only on pregnant women). The intention-to-treat approach does not account properly for them, and more appropriate alternative approaches like principal stratification are not yet widely known. The purposes of this study are to inform researchers that the intention-to-treat approach unfortunately does not fit all problems we face in experimental research, to introduce the principal stratification approach for dealing with intermediate outcomes, and to illustrate its application to a trial of long term calcium supplementation in women at high risk of preeclampsia.
Principal stratification and related concepts are introduced. Two ways for estimating causal effects are discussed and their application is illustrated using the calcium trial, where noncompliance and pregnancy are considered as intermediate outcomes, and preeclampsia is the main final outcome.
The limitations of traditional approaches and methods for dealing with intermediate outcomes are demonstrated. The steps, assumptions and required calculations involved in the application of the principal stratification approach are discussed in detail in the case of our calcium trial.
The intention-to-treat approach is a very sound one but unfortunately it does not fit all problems we find in randomized clinical trials; this is particularly the case for intermediate outcomes, where alternative approaches like principal stratification should be considered.
Intermediate outcomes; Intention-to-treat approach; Principal stratification; Causal effects
The papers of this special issue have the dual focus of reviewing research, especially clinical trials, testing self-determination theory (SDT) and of discussing the relations between SDT and motivational interviewing (MI). Notably, trials are reviewed that examined interventions either for behaviors such as physical activity and smoking cessation, or for outcomes such as weight loss. Although interventions were based on and intended to test the SDT health-behavior-change model, authors also pointed out that they drew techniques from MI in developing the interventions. The current paper refers to these studies and also clarifies the meaning of autonomy, which is central to SDT and has been shown to be important for effective change. We clarify that the dimension of autonomy versus control is conceptually orthogonal to the dimension of independence versus dependence, and we emphasize that autonomy or volition, not independence, is the important antecedent of effective change. Finally, we point out that SDT and MI have had much in common for each has emphasized autonomy. However, a recent MI article seems to have changed MI's emphasis from autonomy to change talk as the key ingredient for change. We suggest that change talk is likely to be an element of effective change only to the degree that the change talk is autonomously enacted and that practitioners facilitate change talk in an autonomy supportive way.
If not handled appropriately, missing data can result in biased estimates and, quite possibly, incorrect conclusions about treatment efficacy. This article aimed to demonstrate how ordinary use of generalized estimating equations (GEE) can be problematic if the assumption of missing completely at random (MCAR) is not met.
We tested whether results differed for different analytic methods depending on whether the MCAR assumption was violated. This example used data from a published randomized controlled trial examining whether varying the timing of a weight management intervention, in concert with smoking cessation, improved cessation rates for adult female smokers. Participants were 284 women with at least one report of smoking status during Visits 4–16. Smoking status was assessed at each visit via self-report and biologically verified using expired carbon monoxide.
Results showed that while the GEE analysis found differences in smoking status between conditions, tests of the MCAR assumption demonstrated that it was not valid for this dataset. Additional analyses using tests that do not require the MCAR assumption found no differences between conditions. Thus, GEE is not an appropriate choice for this analysis.
While GEE is an appropriate technique for analyzing dichotomous data when the MCAR assumption is not violated, weighted GEE or mixed-effects logistic regression are more appropriate when the missing data mechanism is not MCAR.
Clinical trial registration information: NCT00113711
The habitual "any other comments" general open question at the end of structured questionnaires has the potential to increase response rates, elaborate responses to closed questions, and allow respondents to identify new issues not captured in the closed questions. However, we believe that many researchers have collected such data and failed to analyse or present it.
General open questions at the end of structured questionnaires can present a problem because of their uncomfortable status of being strictly neither qualitative nor quantitative data, the consequent lack of clarity around how to analyse and report them, and the time and expertise needed to do so. We suggest that the value of these questions can be optimised if researchers start with a clear understanding of the type of data they wish to generate from such a question, and employ an appropriate strategy when designing the study. The intention can be to generate depth data or 'stories' from purposively defined groups of respondents for qualitative analysis, or to produce quantifiable data, representative of the population sampled, as a 'safety net' to identify issues which might complement the closed questions.
We encourage researchers to consider developing a more strategic use of general open questions at the end of structured questionnaires. This may optimise the quality of the data and the analysis, reduce dilemmas regarding whether and how to analyse such data, and result in a more ethical approach to making best use of the data which respondents kindly provide.
Meta-analysis of genome-wide association studies (GWAS) has become a useful tool to identify genetic variants that are associated with complex human diseases. To control spurious associations between genetic variants and disease that are caused by population stratification, double genomic control (GC) correction for population stratification in meta-analysis for GWAS has been implemented in the software METAL and GWAMA and is widely used by investigators. In this research, we conducted extensive simulation studies to evaluate the double GC correction method in meta-analysis and compared the performance of the double GC correction with that of a principal components analysis (PCA) correction method in meta-analysis. Results show that when the data consist of population stratification, using double GC correction method can have inflated type I error rates at a marker with significant allele frequency differentiation in the subpopulations (such as caused by recent strong selection). On the other hand, the PCA correction method can control type I error rates well and has much higher power in meta-analysis compared to the double GC correction method, even though in the situation that the casual marker does not have significant allele frequency difference between the subpopulations. We applied the double GC correction and PCA correction to meta-analysis of GWAS for two real datasets from the Atherosclerosis Risk in Communities (ARIC) project and the Multi-Ethnic Study of Atherosclerosis (MESA) project. The results also suggest that PCA correction is more effective than the double GC correction in meta-analysis.
genome-wide association studies; meta-analysis; double genomic control correction; principal components analysis; population stratification
STUDY QUESTION: Continuous quality improvement (CQI) has been implemented at least to some degree in many health care settings, yet randomized controlled trials (RCTs) of CQI are rare. We ask whether, when, and how RCTs of CQI might be designed. STUDY DESIGN: We consider two applications of CQI: as a general philosophy of management and (by analogy with the use of conceptual models from the behavioral sciences) as a conceptual model for developing specific interventions. The example of warfarin therapy for stroke prevention among patients with atrial fibrillation is used throughout. PRINCIPAL FINDINGS: While it is impractical to use RCTs to study CQI as a general management philosophy, RCT methodology is appropriate for studying CQI as a conceptual model for generating interventions. RCTs of CQI might be considered when the process change under consideration is very large, its implications (e.g., in terms of cost, outcomes of care, etc.) are very great, and the best approach is uncertain. When designing RCTs of CQI, critical decisions include (1) the unit of randomization; (2) whether the focus is on CQI as a method for generating interventions or, instead, is on specific interventions in and of themselves; and (3) the flexibility available to local personnel to modify the intervention's operational details. CONCLUSIONS: RCTs of CQI as a conceptual model for generating interventions are feasible.
Participants in longitudinal studies on the effects of drug treatment and criminal justice system interventions are at high risk for institutionalization (e.g., spending time in an environment where their freedom to use drugs, commit crimes, or engage in risky behavior may be circumscribed). Methods used for estimating treatment effects in the presence of institutionalization during follow-up can be highly sensitive to assumptions that are unlikely to be met in applications and thus likely to yield misleading inferences. In this paper, we consider the use of principal stratification to control for institutionalization at follow-up. Principal stratification has been suggested for similar problems where outcomes are unobservable for samples of study participants because of dropout, death, or other forms of censoring. The method identifies principal strata within which causal effects are well defined and potentially estimable. We extend the method of principal stratification to model institutionalization at follow-up and estimate the effect of residential substance abuse treatment versus outpatient services in a large scale study of adolescent substance abuse treatment programs. Additionally, we discuss practical issues in applying the principal stratification model to data. We show via simulation studies that the model can only recover true effects provided the data meet strenuous demands and that there must be caution taken when implementing principal stratification as a technique to control for post-treatment confounders such as institutionalization.
Principal Stratification; Post-Treatment Confounder; Institutionalization; Causal Inference
Although researchers develop evidence-based programs for public health practice, rates of adoption and implementation are often low. This qualitative study aimed to better understand implementation of the Program to Encourage Active, Rewarding Lives for Seniors (PEARLS), a depression care management program at a Seattle-King County area agency on aging.
We used stratified, purposive sampling in 2008 to identify 38 PEARLS clients and agency staff for participation. In 9 focus groups and 1 one-on-one interview, we asked participants to identify benefits and negative consequences of PEARLS, facilitators of and barriers to program implementation, and strategies for overcoming the barriers. Two independent researchers used thematic analysis to categorize data into key themes and subthemes.
PEARLS benefits clients by decreasing depression symptoms and addressing other concerns, such as health problems. For staff, PEARLS provides "another set of eyes" and is a comprehensive program to help them meet clients' mental health needs. Barriers included issues with implementation process (eg, lack of communication) and the perception that eligibility criteria were more rigid than those of other agency programs. Recommended solutions included changing eligibility criteria, providing additional staff training, increasing communication, and clarifying referral procedures, roles, and responsibilities.
Barriers to PEARLS delivery discourage referrals to what is generally viewed as a beneficial program. Implementing participants' strategies for overcoming these barriers can enhance delivery of PEARLS to a greater number of older adults and help them improve their depression symptoms.
This overview highlights some recent advances in the epidemiology, diagnosis, risk stratification and treatment of acute coronary syndromes. The
sheer volume of new studies reflects the robust state of global cardiovascular research but the focus here is on findings that are of most interest
to the practising cardiologist. Incidence and mortality rates for myocardial infarction are in decline, probably owing to a combination of lifestyle changes, particularly smoking cessation, and improved pharmacological and interventional treatment. Troponins remain central for diagnosis and new high-sensitivity assays are further lowering detection thresholds and improving outcomes. The incremental diagnostic value of other circulating biomarkers remains unclear and for risk stratification simple clinical algorithms such as the GRACE score have proved more useful. Primary PCI with minimal treatment delay is the most effective reperfusion strategy in ST elevation myocardial infarction (STEMI). Radial access is associated with less bleeding than with the femoral approach, but outcomes appear similar. Manual thrombectomy limits distal embolisation and infarct size while drug-eluting stents reduce the need for further revascularisation procedures. Non-culprit disease is best dealt with electively as a staged procedure after primary PCI has been completed. The development of antithrombotic and antiplatelet regimens for primary PCI continues to evolve, with new indications for fondaparinux and bivalirudin as well as small-molecule glycoprotein (GP)IIb/IIIa inhibitors. If timely primary PCI is unavailable, fibrinolytic treatment remains an option but a strategy of early angiographic assessment is recommended for all patients. Non-ST segment elevation myocardial infarction (NSTEMI) is now the dominant phenotype and outcomes after the acute phase are significantly worse than for STEMI. Many patients with NSTEMI remain undertreated and there is a large body of recent work seeking to define the most effective antithrombotic and antiplatelet regimens for this group of patients. The benefits of early invasive treatment for most patients are not in dispute but optimal timing remains unresolved. Cardiac rehabilitation is recommended for all patients with acute myocardial infarction but take-up rates are disappointing. Home-based programmes are effective and may be more acceptable for many patients. Evidence for the benefits of lifestyle modification and pharmacotherapy for secondary prevention continues to accumulate but the argument for omega-3 fatty acid supplements is now hard to sustain following recent negative trials. Implantable cardioverter-defibrillators for patients with severe myocardial infarction protect against sudden death but for primary prevention should be based on left ventricular ejection fraction measurements late (around 40 days) after presentation, earlier deployment showing no mortality benefit.
acute coronary syndromes; advances in clinical cardiology.
Given a randomized treatment Z, a clinical outcome Y, and a biomarker S measured some fixed time after Z is administered, we may be interested in addressing the surrogate endpoint problem by evaluating whether S can be used to reliably predict the effect of Z on Y. Several recent proposals for the statistical evaluation of surrogate value have been based on the framework of principal stratification. In this paper, we consider two principal stratification estimands: joint risks and marginal risks. Joint risks measure causal associations of treatment effects on S and Y, providing insight into the surrogate value of the biomarker, but are not statistically identifiable from vaccine trial data. While marginal risks do not measure causal associations of treatment effects, they nevertheless provide guidance for future research, and we describe a data collection scheme and assumptions under which the marginal risks are statistically identifiable. We show how different sets of assumptions affect the identifiability of these estimands; in particular, we depart from previous work by considering the consequences of relaxing the assumption of no individual treatment effects on Y before S is measured. Based on algebraic relationships between joint and marginal risks, we propose a sensitivity analysis approach for assessment of surrogate value, and show that in many cases the surrogate value of a biomarker may be hard to establish, even when the sample size is large.
Estimated likelihood; Identifiability; Principal stratification; Sensitivity analysis; Surrogate endpoint; Vaccine trials
Smoking is responsible for over 400,000 premature deaths in the United States every year, making it the leading cause of preventable death. In addition, smoking-related illness leads to billions of dollars in healthcare expenditures and lost productivity annually. The public is increasingly aware that successfully abstaining from smoking at any age can add years to one’s life and reduce many of the harmful effects of smoking. Although the majority of smokers desire to quit, only a small fraction of attempts to quit are actually successful. The symptoms associated with nicotine withdrawal are a primary deterrent to cessation and they need to be quelled to avoid early relapse. This review will focus on the neuroadaptations caused by chronic nicotine exposure and discuss how those changes lead to a withdrawal syndrome upon smoking cessation. Besides examining how nicotine usurps the endogenous reward system, we will discuss how the habenula is part of a circuit that plays a critical role in the aversive effects of high nicotine doses and nicotine withdrawal. We will also provide an updated summary of the role of various nicotinic receptor subtypes in the mechanisms of withdrawal. This growing knowledge provides mechanistic insights into current and future smoking cessation therapies.
Reward; negative motivation; dopamine; withdrawal; VTA; nucleus accumbens; habenula; addiction; drug abuse; nicotinic receptors; nicotinic knockout mice
The health risks associated with smoking justify efforts at cessation. Of the 50 million smokers in the United States, about 20 million attempt to quit each year. Approximately 6% are successful. Nicotine, the addictive agent within tobacco smoke, acts to enhance the release of neurotransmitters in the pleasure reinforcing area of the brain. Nicotine replacement therapy has been successfully used to relieve patients' withdrawal symptoms when cessation has been attempted. Nicotine replacement is available as a gum, patch, inhaler, and nasal spray. Bupropion, an antidepressant, is the first non-nicotine drug approved for smoking cessation. It blocks the neuronal uptake of serotonin and norepinephrine. Bupropion, like nicotine replacement therapy, is twice as effective as a placebo in smoking cessation.
Population stratification can cause spurious associations in population–based association studies. Several statistical methods have been proposed to reduce the impact of population stratification on population–based association studies. We simulated a set of stratified populations based on the real haplotype data from the HapMap ENCODE project, and compared the relative power, type I error rates, accuracy and positive prediction value of four prevailing population–based association study methods: traditional case-control tests, structured association (SA), genomic control (GC) and principal components analysis (PCA) under various population stratification levels. Additionally, we evaluated the effects of sample sizes and frequencies of disease susceptible allele on the performance of the four analytical methods in the presence of population stratification. We found that the performance of PCA was very stable under various scenarios. Our comparison results suggest that SA and PCA have comparable performance, if sufficient ancestral informative markers are used in SA analysis. GC appeared to be strongly conservative in significantly stratified populations. It may be better to apply GC in the stratified populations with low stratification level. Our study intends to provide a practical guideline for researchers to select proper study methods and make appropriate inference of the results in population-based association studies.
This study investigated the effects of a peer-mediated intervention on the social interaction of five triads comprised of preschoolers with autism and their typical peers. Strategies thought to facilitate interaction were selected based on analyses of a descriptive data base. Peers were taught to attend to, comment on, and acknowledge the behavior of their classmates with disabilities. These are behaviors preschoolers typically exhibit frequently, but that do not obligate responses to the same extent as questions and requests do. The ABCB reversal designs revealed that improved rates of social interaction during play were clearly associated with the peer intervention for 4 of the 5 children with autism. This intervention offers an alternative peer-intervention package for increasing interaction between children with and without disabilities.
Journal editorials are an important medium for communicating information about medical innovations. Evaluative statements contained in editorials pertain to the innovation's technical merits, as well as its probable economic, social and political, and ethical consequences. This information will either promote or impede the subsequent diffusion of innovations. This paper analyzes the evaluative information contained in thirty editorials that pertain to the topic of computer-assisted decision making (CDM). Most editorials agree that CDM technology is effective and economical in performing routine clinical tasks; controversy surrounds the use of more sophisticated CDM systems for complex problem solving. A few editorials argue that the innovation should play an integral role in transforming the established health care system. Most, however, maintain that it can or should be accommodated within the existing health care framework. Finally, while few editorials discuss the ethical ramifications of CDM technology, those that do suggest that it will contribute to more humane health care. The editorial analysis suggests that CDM technology aimed at routine clinical task will experience rapid diffusion. In contrast, the diffusion of more sophisticated CDM systems will, in the foreseeable future, likely be sporadic at best.
Delay discounting (DD) describes how the value of a reinforcer decreases as delay to its delivery increases. Relationships between DD and various aspects of drug abuse have been demonstrated reliably. A potential barrier to wider adoption of DD techniques is that results are often expressed in terms that may be too abstract or unfamiliar to a broader audience, particularly when describing or comparing hyperbolic DD functions or values of k. In an effort to potentially make DD results more accessible, the current report explores use of an ED50 value in characterizing DD functions, similar to that used in pharmacology research for characterizing dose-effect functions. The ED50 proposed with regard to DD is the delay that is effective in discounting the subjective value of the delayed reinforcer by 50%. Additionally, a convenient method for calculating ED50 values for DD is discussed.
Delay Discounting; Impulsivity; Drug Abuse
We examined subjective responses to smoking the first cigarette of the day and investigated how these responses related to smoking cessation treatment outcome. Data from participants (N = 207) in a clinical trial of message framing for smoking cessation with bupropion, obtained prior to the targeted quit day, were used to examine indices of craving, withdrawal, and affect before and after smoking the first cigarette of the day. After smoking the initial cigarette, craving, withdrawal symptoms, and negative affect were lessened, and positive affect increased. Greater decreases in craving as measured by the Questionnaire on Smoking Urges-Brief (QSU-Brief) predicted relapse at the end of treatment (6 weeks) and at the 3 month follow-up time point. These associations do not appear to be mediated by established measures of dependence. Thus, this preliminary study provides evidence that there are significant changes in craving, withdrawal, and affect related to smoking the first cigarette of the day, with the largest of these changes observed for craving. Moreover, changes in tobacco craving in response to the first cigarette of the day may be a novel predictor of smoking relapse that should be tested in future studies.
First cigarette; craving; withdrawal; affect; smoking cessation
This study examined the interaction of gender and lifetime psychiatric status on the experience of nicotine withdrawal using retrospective data from the National Comorbidity Survey (NCS; N = 816). Multiple regression analyses were performed to examine the main and interactive effects of gender and major depression, alcohol abuse/dependence, panic disorder, and PTSD on indices of withdrawal. Major depression and alcohol abuse/dependence were associated with longer duration of withdrawal symptoms in women. Women also showed stronger associations between major depression and recurrent withdrawal symptoms and PTSD and smoking relapse to alleviate withdrawal. Men showed a stronger association between alcohol abuse/dependence and smoking relapse to alleviate withdrawal. When developing and providing smoking cessation interventions, it is important to consider the gender-specific effects of lifetime psychiatric status on withdrawal.
The ‘birthweight paradox’ describes the phenomenon whereby birthweight-specific mortality curves cross when stratified on other exposures, most notably cigarette smoking. The paradox has been noted widely in the literature and numerous explanations and corrections have been suggested. Recently, causal diagrams have been used to illustrate the possibility for collider-stratification bias in models adjusting for birthweight. When two variables share a common effect, stratification on the variable representing that effect induces a statistical relation between otherwise independent factors. This bias has been proposed to explain the birthweight paradox.
Causal diagrams may illustrate sources of bias, but are limited to describing qualitative effects. In this paper, we provide causal diagrams that illustrate the birthweight paradox and use a simulation study to quantify the collider-stratification bias under a range of circumstances. Considered circumstances include exposures with and without direct effects on neonatal mortality, as well as with and without indirect effects acting through birthweight on neonatal mortality. The results of these simulations illustrate that when the birthweight-mortality relation is subject to substantial uncontrolled confounding, the bias on estimates of effect adjusted for birthweight may be sufficient to yield opposite causal conclusions, i.e. a factor that poses increased risk appears protective. Effects on stratum-specific birthweight-mortality curves were considered to illustrate the connection between collider-stratification bias and the crossing of the curves. The simulations demonstrate the conditions necessary to give rise to empirical evidence of the paradox.
collider-stratification bias; birthweight; directed acyclic graphs; neonatal nortality
A tightly reasoned justification is presented for the procedures used in our test of the linear-no threshold theory of radiation carcinogenesis by comparing lung cancer rates with average radon exposure in U.S. counties. A key point is its dependence on ecological variables rather than on characteristics of individuals and the principal problems involve treatment of potential confounding factors (CF). The method of stratification is introduced and shown to be preferable to multiple regression for evaluating effects of confounding. Utilizing numerous available CF reduces the problem of representing a complex confounding relationship by the average value of a single CF. The requirements on a CF for affecting the results are quantified in terms of its correlations with lung cancer rates and radon levels and it is shown that the existence of an unknown confounder satisfying these requirements is highly implausible. Effects of combinations of confounding factors are treated and shown not to be important. The problem of confounding factors on the level of individuals is resolved. Consideration of plausibility of correlations is used in several applications, including treatments of uncertainty in smoking prevalence, within county differences in radon exposure between smokers and non-smokers, variations in intensity of smoking, differences between measured radon levels and actual exposures, etc. Examples are presented for all applications. The differences between our study and case-control studies, and the advantages of each for testing the linear-no threshold theory, are discussed.
linear-no threshold; radiation carcinogenesis; confounding; stratification; plausibility of correlation; dose-response